Keywords: Mechatronics in Manufacturing Processes, Sensors and Sensing Systems
Abstract: The sea lamprey, sometimes known as "vampire fish", is an invasive species in the Great Lakes region that threatens its ecosystems and billion-dollar fisheries. The parasitic sea lamprey uses suctorial mouth to prey on various host fish by attaching to the fish and draining its body fluids. In this talk we first describe our effort in developing a soft pressure sensor array as an electronic skin (e-skin), for detecting the suction by adult sea lampreys during their upstream migration for spawning. Such e-skins can be mounted at strategically chosen places, such as selective fishways, to facilitate the capture and population assessment of sea lampreys. We discuss regularized least-square algorithms for mitigating the crosstalk in the resistor network of the sensor array, to properly reconstruct the pressure profile under lamprey suction. Machine learning is further adopted to automate the lamprey detection process, as verified with data from animal experiments

In the second part of the talk we explore tracking the movement of fish, such as sea lampreys, with mobile acoustic telemetry, which provides key information about fish migration patterns and habitat uses and is thus critical to decision-making in fishery management. In mobile acoustic telemetry, acoustic tags are implanted in fish and emit pings periodically, which are picked up by acoustic receivers mounted on robots to infer the fish location. We discuss the use of gliding robotic fish and unmanned surface vehicles for tracking acoustic tags, and specifically, we show how distributed filtering by a group of robots can result in localization of a moving target based on the time-difference-of-arrivals (TDOAs) of the emitted signal.

Keywords: Modeling and Design of Mechatonic Systems, Actuators
Abstract: This study focuses on nonlinear series elastic actuators (SEAs) which have the potential to overcome the performance limit of fixed stiffness SEAs, design a spring unit, and experimentally validate the static characteristic. The result will provide insights into the design of nonlinear SEAs for human assist devices.

Keywords: Control Application in Mechatronics, Identification and Estimation in Mechatronics, Robot Dynamics and Control
Abstract: Soft wearable robots that leverage pneumatic-based actuators are envisioned as a comfortable, portable, and safe alternative to traditionally rigid designs. However, control of such devices is still an outstanding challenge. Most tasks in daily life require adjustments of joint impedance in order to physically interact with dynamic environments. In addition, under conditions of dynamic instability, increased impedance is often required to reduce the effects of neural motion noise. Therefore, it is critical to endow soft wearable robots with impedance modulation capabilities for successful deployment in assistive wearable robot scenarios aimed at augmenting user independence. Our model-based approach leverages the force-deflection characteristics of a single actuator to develop an inverse model for joint impedance of the antagonistic setup. The inverse model takes as input the desired equilibrium and stiffness, and outputs the required pressures of each actuator to achieve the desired impedance.

Keywords: Intelligent Process Automation, Mechatronics in Manufacturing Processes
Abstract: Tank manufacturing for LNG ships is currently mostly manual work due to the size and difficulty of the work. Research on the automation of LNG cargo production using a mobile welding robot is in progress. In this study, software development that determines the operation of a mobile welding robot is introduced.

Keywords: Mechatronics in Manufacturing Processes, Control Application in Mechatronics, Robot Dynamics and Control
Abstract: Tendon-sheath robots have several advantages for entering hazardous confined-space environments given their re- duced size and weight, in addition to being intrinsically-safe when there is a risk of explosion. When teleoperated due to limited sensing in confined-space environments, backlash and stick slip friction in tendon-sheath robots cause hunting-like oscillations around the goal position. The main contributions of this work are to (1) develop an automation method that converges each joint to target values while avoiding backlash and stick slip friction and (2) apply the automation method using traded control to de-hunt the teleoperation. Experimental results show the approach leads to an end effector precision of five-thousands of an inch, an order of magnitude better joint level precision than current tendon- sheath control schemes [1]. Trials of the automated method show a 57.1% reduction in completion time for a precision location task. Initial user studies (N=3) of the method shows a 38.5% reduction in user completion time compared to teleoperation with existing compensation methods.

Keywords: Novel Industry Applications of Mechatroinics, Service Robots, Modeling and Design of Mechatonic Systems
Abstract: This paper proposes a robot-based unmanned automatic charging system that can charge multiple vehicles with a single robot. The robot base is mounted to the ceiling and is designed with a beam and an upper rail to allow horizontal movement. The parking spaces are arranged symmetrically around the beam to minimize the robot's movement while enabling the attachment and detachment of the charging coupler to multiple vehicles.

Keywords: Actuators in Mechatronic Systems, Opto-Mechatronic Sensors, Sensor Integration, Data Fusion
Abstract: Detection of slip during object grasping and manipulation plays a vital role in object handling. Existing solutions largely depend on visual information to devise a strategy for grasping. Nonetheless, in order to achieve proficiency akin to humans and achieve consistent grasping and manipulation of unfamiliar objects, the incorporation of artificial tactile sensing has become a necessity in robotic systems. In this poster, we present a novel physics-informed, data-driven method to detect slip continuously in real time. The GelSight Mini, an optical tactile sensor, is mounted on custom grippers to acquire tactile readings. Our work leverages the inhomogeneity of tactile sensor readings during slip events to develop distinctive features and formulates slip detection as a classification problem. To evaluate our approach, we test multiple data-driven models on 10 common objects under different loading conditions, textures, and materials. Our results show that the best classification algorithm achieves an average accuracy of 99%. We demonstrate the application of this work in a dynamic robotic manipulation task in which real-time slip detection and prevention algorithm is implemented.

Keywords: Automotive Systems, Vehicle Technology
Abstract: This paper presents the design and control of a solar panel cleaning mobile robot carried by a drone. The robot has tracked wheels to stick to the slanted solar panels and move. Control between suction pad and wheel velocities has to be done with care in order not to slip down while moving. Experimental studies of moving on the solar panel were demonstrated to confirm the feasibility.

Keywords: Vehicle Control, Motion Vibration and Noise Control, Control Application in Mechatronics
Abstract: This study proposes an H-infinity control synthesis for solving the actuator saturation problem and conducts an experimental study of H-infinity control for the active suspension system of the quarter car. In the H-infinity control design procedure for the active suspension system, actuator saturation is directly handled by introducing a dynamic model with a saturation function. To improve ride comfort for passengers, the acceleration of the vehicle body is selected as a controlled output, and the H-infinity norm of the transfer function from disturbance to controlled output is optimized. Based on Lyapunov stability theory, the control synthesis problem is formulated as a non-convex bilinear matrix inequality. This design difficulty is overcome by the proposed single-objective distributed quantum-behaved particle swarm optimizer, which efficiently explores the optimal controller that provides the minimum upper limit of the H-infinity norm. The simulation and experimental tests are performed using the Quanser’s active suspension system platform and the road profile generated by trigonometric functions. The results demonstrate the effectiveness of the proposed H-infinity controller.

Keywords: Mobile Robots, Modeling and Design of Mechatonic Systems, Legged Robots
Abstract: This paper presents a mobile robot for amphibious surface locomotion called ARMoR. The locomotion system of ARMoR consists of two wheel-and-leg transformable mechanisms and a customizable balancing tail. A sphere body chassis containing electronic components assembles the wheels and the tail. A combination of chassis design and transformable wheels allows ARMoR to safely navigate various environments, including diverse terrains and water surfaces. The robot is controlled and operated using an embedded microprocessor interfacing with sensing, communicating, and powering modules, including the Global Positioning System (GPS), camera, Inertial Measurement Unit (IMU), wireless communication module, and batteries. ARMoR was tested for its locomotion capabilities on concrete, dirt, grass, rocky surface, low brush, stairs, and water. On concrete, dirt, and grass, ARMoR operated in the wheeled mode; on other surfaces, the wheels transformed into the legged configuration enabling the robot to traverse challenging surface conditions effectively. ARMoR successfully traversed all terrains, and the traversal speeds were measured.

Keywords: Underwater robotics, Actuators in Mechatronic Systems, Fuel Cells and Alternative Power Sources
Abstract: A Proton Exchange Membrane (PEM) fuel cell enabled buoyancy control device (BCD) is developed to address the challenge of depth control in underwater devices by compensating for buoyancy changes during underwater manipulations. The BCD splits distilled water into hydrogen and oxygen gases, increasing the volume of balloons attached to it, providing a permanent change in buoyancy of the device, and reducing the need for motors to run constantly. Results indicate that energy savings can reach up to 85% in comparison with experiments that use DC motors only. Additionally, the response is smoother and device stability is increased when DC motors are not running. Experimental results show the trajectory profiles when the BCD is active and passive, respectively. Furthermore, the PEM electrolyzers can be used as fuel cells to charge a battery or run another mechanical device after the operation, resulting in around 28% energy savings.

Keywords: Mobile Robots, Modeling and Design of Mechatonic Systems
Abstract: Marine transportation is an important means of transportation, in which ships are in constant use and accidents are unavoidable. In this study, we developed an amphibious robot to help rescue people in an accident environment that are too low or dangerous for rescuers to enter. Angled spoke paddling wheel(ASPW) makes it possible to drive on the water surface, while taking advantage of the Angled Spoke Wheel(ASW). The ASPW changes the paddle force by rotating the paddle when the wheel rotates so that the total force is generated in the direction of the robot’s movement. The paddle is optimized using the Taguchi method. The driving experiment was conducted in four environments: ground, water surface, transition between ground and water, and obstacle overcoming situation. The maximum driving speed was 0.47 m/s on the ground and 0.1 m/s on the water surface. The maximum average inclination degree in the transition situation was 25° from the ground to the water surface and 20° from the water surface to the ground. Additionally, it was possible to overcome an obstacle with a maximum height of 41.3 mm, demonstrating that the proposed robot excels in amphibious driving, while possessing the ability to overcome the ASW level obstacle.

Keywords: Control Application in Mechatronics, Biomechatronics, Underwater robotics
Abstract: An autonomous underwater vehicle (AUV) is a crewless robotic vehicle that dives into the water and performs without human assistance. This paper focuses on developing trajectory tracking control for bio-mimetic AUV system under uncertain environments. Therefore, a relatively new control technique called time delay-based estimation control is proposed for trajectory tracking under multiple uncertainties. This algorithm estimates the total disturbance in the system using immediate past information of input and output of feedback state and control variables. The benefit of this scheme is that it avoids assumptions about a priori upper bound information of disturbance. Further, the control structure is simple and does not require any high-frequency switching or high gain to nullify the effects of disturbance. The theoretical analysis of the proposed scheme guarantees the uniformly ultimate bounded stability of the closed-loop system. The numerical analysis is also carried out to validate the control performance of the given algorithm for lemniscate reference path tracking.

Keywords: Modeling and Design of Mechatonic Systems, Underwater robotics, Robot Dynamics and Control
Abstract: This paper presents a design and physics-based dynamic model of an underwater depth control device based on buoyancy change. Negative buoyancy is achieved by decreasing the volume while keeping the mass constant, resulting in a higher density. The nonlinearities are discussed in the context of the operation conditions and the practically fulfilled assumptions. With reasonable considerations, the dynamic model is linearized and state-space equations are developed for such a system. The stability of such systems, when subject to constrained control input, is studied. A Model-based Predictive Control (MPC), which optimizes the control energy and output error with disturbance rejection and considers constrained control input, is then developed. The controller was simulated using MATLAB with a discrete plant model. An experimental setup was also created to test the controller. The developed MPC was then tested experimentally on the hardware which shows the validity of the dynamic model as well.

Keywords: Mobile Robots, Modeling and Design of Mechatonic Systems
Abstract: This paper proposes a novel design of rigid-wing bi-directional sailboat and its method of sailing. The proposed design features a bi-directional hull, a rudder on both ends of the vessel and a freely rotating rigid wing sail whose angle of attack is controlled by a tail rudder. The proposed method of sailing involves reversing the vessel’s travel direction at every turn and is ideal for micro-scale autonomous robotic platforms as it overcomes the difficulty in turning the boat through the eye of the wind during an upwind sail. A simulation model using aerodynamic and hydrodynamic forces is developed to test the effectiveness of the proposed design and method of sailing and compared to the traditional method of sailing in station-keeping and sailing upwind scenarios. The effectiveness of the proposed method is demonstrated with a prototype model equipped with a rigid wing sail.

Keywords: Identification and Estimation in Mechatronics, Robot Dynamics and Control, Human -Machine Interfaces
Abstract: This paper investigates the design of optimal inputs for dynamic system identification. Specifically, this paper concerns the perturbation design for system identification experiments where target human systems are perturbed by mechanical inputs produced by an active device. Although conventional perturbation design criteria are generally applicable, including the scenario described above, problems arise due to the dynamics of the active device. A low-bandwidth active device may distort the input signal and thereby void the optimality of the input. To address this issue, the paper formulates an optimization problem for optimal input design that explicitly incorporates the active device dynamics. The cost function is the determinant of a modified covariance lower bound that takes the active device dynamics into consideration. The proposed method is demonstrated with an identification of a linear dynamics model simulating human arm impedance. Simulation results show that, compared with a standard optimal input and an input with a flat spectrum, the proposed optimal input with active device compensation achieved a smaller parameter covariance. Furthermore, the proposed optimization problem suggests that the optimal covariance lower bound can be achieved by active devices with different dynamics properties. This allows the control design of the active device to satisfy a wide variety of requirements without sacrificing its ability to perform system identification.

Keywords: Fault Detection and diagnosis in Manufacturing, Identification and Estimation in Mechatronics, Modeling and Design of Mechatonic Systems
Abstract: This article proposes a method with a serial manipulator calibration technique using the linearized finite screw deviation model for calibrating the manipulator to reduce error in the workspace. With the emergence of robot technology, applications of the multi-axis manipulator have significantly increased. In previous research, many studies calibrated the multi-axis manipulator using end-effector tracking devices such as vision-sensor-based apparatus. However, the appliance of a special device for calibration can be resource-consuming in practical applications for applying every damaged manipulator. In this research, the proposed method helps the keep manipulator works appropriately in the workspace when the situation cannot afford those tracking devices because of the lack of resources such as budget and time. By designing a linearized finite screw deviation model to calculate the error of each axis of the manipulator from the non-continuous points trajectory of the end-effector, the error model becomes an affine function applicable in the global optimization method which can identify the deviation of the manipulator and leads to reduction of the error in the workspace. To validate the performance of the proposed method, simulation and the experiment were carried out with seven degrees of freedom manipulator.

Keywords: Modeling and Design of Mechatonic Systems, Identification and Estimation in Mechatronics, Compuational Models and Methods
Abstract: Just as optimal control addresses the inexact science of selecting controller gains, optimal system identification balances the effects of linearization, estimation and order reduction, to obtain the “best fit” approximation of a target electrical, mechanical, or electro-mechanical system. Like any engineering design problem, it involves matching a set of free design parameters to a requirement specification that defines what “best” means. In this paper, closed-form metrics of a normalized second-order system are used to develop a clear and simple design process to identify a 2nd order approximation that exhibits the most relevant dynamic characteristics of the target system. The process identifies the optimal parameters of an under or over-damped system from its step-response, and refines the approximation using its impulse-response. The approach is formulaic, non-iterative, and may be used to fit a second-order approximation to a higher-order system response, without the need for a complex search algorithm.

Keywords: Neural Networks, Identification and Estimation in Mechatronics, Vehicle Control
Abstract: Normalizing flows have gained increasing attention in the area of probabilistic modeling. For solving inverse problems, BayesFlow is a state-of-the-art Bayesian inference method based on normalizing flows. However, BayesFlow suffers from overfitting in many real-world scenarios. Therefore, we put forward stochastic BayesFlow, enhancing BayesFlow through stochastic normalizing flows. Apart from being less prone to overfitting, stochastic BayesFlow performs more robustly in parameter identification from noisy observations. Moreover, we develop a stochastic BayesFlow algorithm to solve stochastic inverse problems and validate it using the inverse uncertainty quantification of a vehicle dynamics model.

Keywords: Identification and Estimation in Mechatronics, Modeling and Design of Mechatonic Systems, Neural Networks
Abstract: This paper proposed a new torque estimation method based on an equivalent efficiency model and back propagation (BP) neural network to obtain accurate torque. Firstly, the joint transmission efficiency model is obtained based on experiments to correct the torque observer (TOB). Then the relationship between joint position, velocity, current, and torque estimation is established by BP neural network, and the torque estimation error is compensated further. Finally, several comparative experiments are carried out. The results show that the proposed method can obtain more accurate torque compared with traditional TOB.

Keywords: Identification and Estimation in Mechatronics, Artificial Intelligence in Mechatronics, Machine Learning
Abstract: Accurate dynamical models form a main driver for high performance mechatronic applications. Conventional modeling of mechatronic systems is often limited in its ability to handle poorly understood phenomena and may not be adequate in instances where the underlying dynamics are not fully known nor fully captured by sensory data. To overcome these limitations, we propose a physics-based data-driven state-space modeling approach. We phrase the problem as a probabilistic representation learning problem. The hybrid model combines known physical relations with parametrized functions, represented as neural networks, to serve as substitutes for the previously unidentified substructures. The identification problem is solved using the Expectation-Maximization (EM) algorithm. In the Expectation step, Bayesian smoothers are utilized to provide complete state estimates from partial observations. In the M-step, the hybrid model is fitted onto the smoothed data. Although the physics based prior model comes at the loss of expressiveness, it serves as a strong model prior. The use of a physical model prior is beneficial both to improve the accuracy of the inference during the E-step as well as to reduce the complexity of the M-step. The proposed methodology is applied and validated for the identification of friction in both joints of an acrobat, with only measurements available in one joint. Numerical experiments demonstrate the methods capability of identifying comprehensive representations of the friction characteristics in both joints and possessing accurate predictive abilities.

Keywords: Control Application in Mechatronics, Robot Dynamics and Control
Abstract: In this paper, a grinding robot targeting large areas and an impedance-based force-control method applied to the robot are described. With the development of science and technology, the demand for various industrial robots has increased, and among them, the demand and importance of grinding robots that require high risk and precision has also grown. A robot consisting of a manipulator and grinding module with a 2-Dof parallel structure is proposed as the design of a new grinding robot. The control method is based on impedance force control, but to overcome the limitations of using only impedance control, the impedance control via model-based prediction optimization (MPO) is proposed as a control technique for grinding robots. Experiments were conducted to verify the force-following ability of the proposed control approach, resulting in a 28.1% improvement in settling time for the desired force. Even for disturbance, improved recovery performance over conventional controllers has been verified. As a result, the proposed impedance force control via MPO is presented as an appropriate control method for grinding robots targeting large areas.

Keywords: Modeling and Design of Mechatonic Systems, Actuators in Mechatronic Systems, Mechatronics in Manufacturing Processes
Abstract: In many separation processes, filtration performance degrades over time due to retained particles blocking the flow through the filter membrane. A novel bearingless spinfilter extends the long-term performance by self-cleaning effects. The filter rotor is magnetically levitated and actuated by two self-bearing motors inside a hermetically sealed housing, which eliminates the need for bearings and rotary sealings, that both lead to process fluid contamination. Both bearingless motors have integrated electronics and independently control the levitation of the spinfilter rotor. A first prototype is designed and the concept is validated by the separation of a yeast cell culture. Special focus is placed on the internal rotary seal between the feed and filtrate regions, that is inevitably created when the filter membrane is in motion. Any leakage flow through the seal leads to filtrate impurities, which is minimized in this paper with an embedded impeller as a pressure compensation method. A constant filtrate flux of SI{1750}{literperhourpersquaremeter} and a filtrate purity of SI{75}{percent} was achieved in a first series of tests.

Keywords: Mechatronics in Manufacturing Processes, Novel Industry Applications of Mechatroinics, Modeling and Design of Mechatonic Systems
Abstract: Wire straightening is essential for obtaining a piece of straight wire from a roll of wire stock. The objectives of this research were to develop a two-roll wire straightener and find the appropriate roll gap and roll angle for the straightener based on the finite element analysis results. A PLC was employed as the main controller to control motors that adjust the parameters. An image acquisition and processing unit were also built to obtain the deformation of the aluminum wire offline. The wire used in the study is a 1050 series aluminum wire with a length of 400 mm and a diameter of 2.98 mm. Three wires were analyzed with initial deformations of 5.0 mm, 7.5 mm, and 10.0 mm. Full factorial experiments of the straightening process were simulated using the finite element analysis. Then, a two-way analysis of variance was performed on experimental data. Subsequently, the straightening parameters with the smallest residual deformation were adopted as the optimal parameters for the wire straightener. Using the parameters found in this study can improve the residual deformation by over 90%.

Keywords: Identification and Estimation in Mechatronics, Control Application in Mechatronics, Modeling and Design of Mechatonic Systems
Abstract: Maintaining web tension in web processing line under changing web speed is a key factor in achieving good final product quality. Tension control systems with dancers play an integral role in such applications. A dancer roll system is essentially a preloaded idler arm with two machine tasks: imposing the web tension and acting as a controlling mechanism to avoid web damage caused by a range of mechanical issues like eccentric and non-circular web rolls or dynamic web speed trajectories. All the dancer control strategies described in literature have one thing in common: they do not adapt their control parameters to the type of web material. The characteristics of the web are not known in advance, and can vary from roll to roll and from time to time. Identification of the web parameters allows more advanced control. In this paper, a digital twin based feed forward controller is proposed to control the dancer in order to obtain variation-free even under varying web speed.

Keywords: Mechatronics in Manufacturing Processes, Identification and Estimation in Mechatronics, Sensors and Sensing Systems
Abstract: Normal force is one of the most important factors in the grinding and polishing process. This project focuses on system identification of the grinding tool Mini Robot, which can not only spin the grinding wheel but also actively control the force along with the shaft direction using a voice coil motor, achieving a fine grinding quality. This advantage can widely benefit many collaboration platforms such as an automation grinding platform with manipulator. System identification of this grinding tool was carried out to build a state observer feedback-control system. The force was estimated using the information on current and displacement of the voice coil motor. Feedback-control design can later be added into the system.

Keywords: Learning and Neural Control in Mechatronics, Control Application in Mechatronics, Mechatronics in Manufacturing Processes
Abstract: This work explores the feasibility of a geometry-agnostic feedback control strategy for laser powder bed fusion (L-PBF) using reinforcement learning. The controller is designed to be capable of anticipating and responding to geometry-induced inhomogeneities in the measurements, while accomplishing feedback control of the process. % Because the measurements from the process exhibit high noise levels, the training is done on a noiseless simulation model of the process. To train the reinforcement learning controller, first a reduced-order simulation model is fit to experimental data. Then, the optimal control strategy is found through reinforcement learning on this reduced-order model. After the training, we demonstrate that the learned control strategy can reduce up to 55% of the error 2-norm and 59% of the standard deviation with respect to a given reference value. Moreover, the learned control strategy is applicable to novel build geometries without any additional tuning, or modification of the controller, in which we find that the controller attenuated 2-norm error by 62% and variation levels by 60% when deployed on a new (test) geometry, presenting the efficacy of the proposed controller. Finally, the experimental validation of the algorithm in a `playback' setting resulted in a 24% reduction of both 2-norm error and variation levels, highlighting its potential in an industrial L-PPBF system.

Keywords: Design Optimization in Mechatronics
Abstract: Design optimization of mechanisms is a promising research area as it results in more energy-efficient machines without compromising performance. However, machine builders do not actually use the potential described in the literature, as these methods require too much theoretical analysis. Therefore, this paper proposes a novel industrial applicable approach that enables the design optimization of reciprocating mechanisms using CAD models.

The 3D multi-body software is used to perform motion simulations, from which the objective value samples can be extracted. In this paper, the considered objective value is the required torque, for a specific combination of design parameters, to fulfil the movement. Dedicated software can execute multiple motion simulations sequentially and interchange data between the different simulations, which automates the process of retrieving objective value samples. Therefore, without in-depth analytical design analysis, a machine designer can evaluate multiple designs at a low cost. Moreover, an optimal design that meets the objective can be found by implementing an optimization algorithm. In a case study of an emergency ventilator mechanism which considers three link lengths as design parameters (DP's), 39 CAD motion simulations allowed a reduction of the RMS torque of the mechanism by 57.2%.

Keywords: Design Optimization in Mechatronics, Modeling and Design of Mechatonic Systems, Control Application in Mechatronics
Abstract: Modular drivetrains have already been introduced in literature as an approach to deal with load variations and provide an easily adaptable machine design. Although some research regarding the performance of a modular drivetrain has already been performed, a method to evaluate and compare several modular drivetrain architectures on multiple performance criteria is not yet available. This paper presents a sensitivity analysis framework that can be used to evaluate the different architectures against each other and to make a comparison with the traditional benchmark alternative. A benchmark case of a single motor driven shaft with variable loads and a dynamic speed profile is used to illustrate the functionality of the framework. Two modular variants of the benchmark system are presented. From the evaluation, the modular architectures are found to outperform the benchmark case regarding energy consumption and tracking error. However, a cost increase is observed. The trade-off between the additional investment cost and the increased performance can be assessed using this sensitivity analysis framework.

Keywords: Design Optimization in Mechatronics, Rapid Prototyping, Modeling and Design of Mechatonic Systems
Abstract: The objective of this research is to demonstrate that conventional robot designs, typically characterised by orthogonal axis placement, can be improved, tailoring the design to a task or task ensemble. For that purpose we develop a co-optimization pipe-line specialized to the combined optimization of design and task execution. Our pipe-line pursues high realism by proposing a link deformation strategy that preserves manufacturability of the deformed links, specifically leaving the actuators unaffected. We demonstrate the approach on a proprietary industrial robot, tailoring the design to two time optimal task definitions. We demonstrate that task execution times can be reduced up to 15%.

Keywords: Modeling and Design of Mechatonic Systems, Robot Dynamics and Control, Compuational Models and Methods
Abstract: Single-axis motion profiles are typically generated by an engineer by adjusting acceleration, jerk, and velocity at the bounds of system specifications to find a trajectory that meets the system performance requirements. This is an often trite task that does not allow the engineer to consider many secondary system performance options, such as energy usage. In multi-degree-of-freedom systems, secondary considerations are typically brushed aside even more as the task's complexity requires more focus on the coordinated motion of multiple stages. This paper presents an optimization framework for generating optimal trajectories offline for single- and multi-axis systems based entirely on basic kinematic relationships. The framework, implemented in the Python package Pyomo, involves setting a minimization target and appropriate norm in conjunction of a set of operational constraints. Trajectory generation results for minimal power, jerk, acceleration, and velocity considerations using 1-, 2-, and infinity-norms are compared, and a nonlinear optimization scheme to minimize move time is also examined. Results show that trapezoidal/S-curve profiles embody minimization of velocity norms as well as move time, while other optimal profiles can be utilized in cases where system specifications require different performance options. Normalized peak power and total system energy cost functions are also developed. This paper provides practicing engineers with an easily implementable tool to facilitate system component selection and performance optimization.

Keywords: Modeling and Design of Mechatonic Systems, Design Optimization in Mechatronics, Novel Industry Applications of Mechatroinics
Abstract: The importance of dynamic wireless power transfer (DWPT) has increased significantly resulting from the development of electric mobility devices (EMDs), e.g. electric vehicles (EVs) and automated guided vehicles (AGVs). How to enlarge endurance with limited battery capacity has become a vital issue in the studies of EMDS, and the DWPT system has been foreseen as an emerging technology for this issue. In this paper, a continuous DWPT system has been investigated with the use of DD-shaped (bipolar) coils and Q-shaped (unipolar) coils in order to solve the problem of output power pulsation during the movement. To meet the practice of curved roadways and transfer efficiency, the design and analysis of transmitters and receivers for circular and athletic-field-like roadways as presented. The coupling coefficients and mutual inductance of the continuous DWPT system are addressed provided simulation validation and experimental test.

Keywords: Design Optimization in Mechatronics, Modeling and Design of Mechatonic Systems, Actuators in Mechatronic Systems
Abstract: Adaptive structures are equipped with sensors and actuators to counteract deformations caused by external loads. Previous work has shown that active control employed in high-rise buildings to compensate for static loads and dampen vibrations allows for reducing resource consumption for construction by half. In order to achieve this, proper placement of actuators is a delicate part in the design process of adaptive structures. In this paper, a general two-stage optimization procedure is proposed to place actuators for adaptive structures under static loads. Former studies focused on placing actuators to minimize displacements due to external loads. However, this leads to unnecessary large actuator forces, as displacements do not have to be compensated as much as possible, but only to reach certain service criteria. For this, different objectives are introduced using the proposed optimization procedure, focusing on either actuator force minimization while maintaining displacement constraints or displacement minimization. The optimization results are systematically compared for an example structure in terms of the resulting displacements, the required actuator forces and the respective optimal actuator configurations.

Keywords: Software Design for System Integration, Human -Machine Interfaces, Service Robots
Abstract: We present the Interactive Task Encoding System (ITES) for teaching robots to perform manipulative tasks. ITES is designed as an input system for the Learning-from-Observation (LfO) framework, which enables household robots to be programmed using few-shot human demonstrations without the need for coding. In contrast to previous LfO systems that rely solely on visual demonstrations, ITES leverages both verbal instructions and interaction to enhance recognition robustness, thus enabling multimodal LfO. ITES identifies tasks from verbal instructions and extracts parameters from visual demonstrations. Meanwhile, the recognition result was reviewed by the user for interactive correction. Evaluations conducted on a real robot demonstrate the successful teaching of multiple operations for several scenarios, suggesting the usefulness of ITES for multimodal LfO. The source code is available at https://github.com/microsoft/symbolic-robot-teaching-interface.

Keywords: Human -Machine Interfaces, Unmanned Aerial Vehicles, Vehicles and Space Exploration
Abstract: To deal with changes and uncertainties in controlling a Unmanned Aerial Vehicle (UAV) reliably, a new platform is proposed to synergize human and machine intelligence, and it is based on a new Brain Computer Interface (BCI) to (1) quantify human’s affections in arbitrating human and machine intelligence and alleviating adverse effects by human’s mistakes, (2) fuse human and machine’s control commands in different frequencies seamlessly and generate control commands at motor levels for real-time performance. In this paper, existing works on BCIs are discussed to identify the limitations of traditional Human-Machine Interactions (HMIs), a new framework of HMIs is proposed for a supervisory control of UAV; in particular, it is equipped with an arbitrating mechanism to optimize the shared control of UAVs based on quantified states of human’ affection. It is expected to improve adaptability, agility, and reliability, responsiveness, and resilience of UAVs. This is an ongoing project, and the development platform for the feasibility study of the proposed method is introduced as our plan for future work.

Keywords: Human -Machine Interfaces, Control Application in Mechatronics, Neural and Fuzzy Control in Mechatronics
Abstract: In this paper, a method is proposed to estimate the velocity of the hand cooperating with the robot as human intent based on human surface electromyography signals, and to make the robot move in a predictive assistive motion. Specifically, an index to measure the matching index between human hand velocity and robot motion is proposed, based on the estimation of human hand velocity from human surface electromyography signals by a recurrent neural network, and using this as the human intent. Furthermore, the strength of the robot's support is adjusted based on the matching index to achieve predictive assistive motion behavior of the robot during cooperative work. In the experiment, an index to measure the burden of human work was defined, and the effectiveness of the proposed method was verified. Specifically, the proposed method was implemented and evaluated for the task of manipulating an object in cooperation with an impedance-controlled robot. The results show that the proposed method can reduce the burden on humans.

Keywords: Human -Machine Interfaces, Flexible Manipulators and Structures, Control Application in Mechatronics
Abstract: Flexible systems are difficult to control because they deflect in response to any applied force and they tend to oscillate around the desired path or set point. Human operators driving such systems are challenged by the deflection and vibration that makes the system difficult to move and accurately position. Such systems can be augmented with an intelligent control system that aids the human operator. Numerous types of controllers can be used for such applications; however, it is challenging to balance the control authority of the human operator and the augmenting controller. Input shaping is a control technique that reduces unwanted flexible system responses by modifying the human-operator command in real-time. This paper investigates the use of input shaping as an augmenting controller to aid in the accurate positioning of highly-oscillatory systems. Results from operator testing verify some of the key advantages of this controller.

Keywords: Machine Learning, Human -Machine Interfaces, Machine Vision
Abstract: Instead of taking images with a camera, synthetic data is generated in a computer simulation. One advantage of this is that training data can be generated on-demand, e.g., to automatically retrain robots when a task changes. While this makes synthetic data a promising approach for autonomous productions, realizing such autonomous setups is difficult with current systems for generating synthetic data, which usually require a programmer for every dataset to be generated.

To overcome this problem, we present a novel framework for generating synthetic data. This framework restructures the generation process into asynchronous phases to increase the level of autonomy in two ways. First, by letting programmers write parameterized scripts, many different datasets can be autonomously generated. Secondly, by introducing a user interface, domain experts are enabled to influence the generation process on their own without a programmer. Furthermore, by being built as a new layer on top of existing systems for generating synthetic data, our framework shows a new way to maximize compatibility with other research on synthetic data generation.

To test our framework, we have developed a fully functional prototype based on it. Successfully using this prototype for an example experiment, we conclude that our ideas work. Future research can use our prototype for more elaborate experiments on autonomous productions and to further assess its usability.

Keywords: Identification and Estimation in Mechatronics, Motion Vibration and Noise Control, Control Application in Mechatronics
Abstract: Adaptive structures are equipped with sensors and actuators to counteract deformations caused by external loads. Concerning railway bridges, previous work has shown that active vibration damping allows to extend the service life. Trains as external loads represent the decisive influencing factor for bridge vibration and has to be taken into account when applying model-based control concepts. This paper proposes a state and disturbance estimator (SDE) for bridge structures based on a moving point load train model and estimating the average train axle mass. The model employed for state and disturbance estimation is linear time variant, which allows use of an augmented Kalman filter. Estimability is analyzed based on the Fisher information and the proposed SDE is systematically tested through simulations. A linear quadratic regulator is designed and combined with the proposed SDE to evaluate the closed-loop performance for damping the bridge vibrations during train crossing.

Keywords: Control Application in Mechatronics, Robot Dynamics and Control, Identification and Estimation in Mechatronics
Abstract: In conventional robust motion control systems, disturbance observer (DOB) nominal models are designed with same order as the actual plant such that the nominal model directly cancels with the actual plant dynamics. However, for multi-DOF systems such as 6-DOF industrial robots, identifying the higher-order model is laborious. Moreover, there is a high risk of obtaining a nominal model with large deviation from the actual plant due to severe parameter uncertainty. Thus, a reduced-order nominal model is derived from the actual plant model and compared with the one which same order as the actual plant in this paper. The designed model is simple, easy to identify and implement. From the analyses and experiment results, DOB with the proposed nominal model is not affected by severe robot model uncertainty and show significant improvement in motion control performance in terms of transient response and tracking accuracy.

Keywords: Identification and Estimation in Mechatronics, Motion Vibration and Noise Control, Control Application in Mechatronics
Abstract: In this paper, we present identification methods of normal-direction motor parameters and a force ripple reduction method in the normal direction for permanent-magnet linear synchronous motors (PMLSMs). This paper discusses the force generation mechanism of a PMLSM both in the normal and tangential directions by the D- and Q-axis currents, which is utilized for identifying and suppressing the motor force ripple in the normal direction. Using the proposed identification method on the experimental setup of a linear stage driven by an iron-cored PMLSM, we identify the normal-direction force constant with an error of only 7% as compared to a direct measured data from a dynamometer. The geometry-driven ripple in the normal direction is also identified by experimentally estimating the dominant 2nd- and 6th-order spatial harmonics, which also show significant fidelity with an NRMSE (peak-to-peak normalized root-mean-square error) of only 3.39% as compared to the dynamometer measurement. Using the identified motor parameters, we achieve the force ripple reduction in the normal direction by 84.6% and 87.8% in peak-to-peak and root-mean square (RMS) values, respectively.

Keywords: Motion Vibration and Noise Control, Identification and Estimation in Mechatronics, Control Application in Mechatronics
Abstract: Vibration suppression is critical in precision mechatronic systems for nanofabrication. For automated wafer handling in semiconductor plants, Overhead Hoist Transport (OHT) vehicles transport wafers carried in Front Opening Unified Pods (FOUPs); while the wafers are transported in a FOUP, semiconductor chips are at risk of damage by excited small particles due to mechanical vibration. Active suppression of the FOUP vibrations has been proposed to improve the production yield. However, there are two main challenges that make it a non-trivial problem. First, moving FOUPs carried by the OHT vehicles have no external anchoring point as a momentum source for control efforts. Second, no sensor attachments are permitted on mass-production FOUPs, which makes feedback control more challenging without measurement. Since the goal is to suppress the large FOUP acceleration peaks instead of eliminating all vibration, an inertia-based counterbalancing system is designed to address these challenges. To validate this idea, a custom testbed is designed for multi-axis vibration generation and suppression. A Disturbance Observer-Based Controller (DOBC) is developed and implemented on the hardware. During the experiment, 38 percent of the OHT hand unit vibration (and 42 percent of FOUP vibration) suppression is achieved in the OHT travel direction. Moreover, multi-axis FOUP-level acceleration-peak reduction is achieved to verify the effectiveness of the proposed method.

Keywords: Control Application in Mechatronics, Motion Vibration and Noise Control, Actuators
Abstract: This article proposes a method of feed-forward cancellation of disturbance effects and reports on the results of verifying its performance using the controller included with the dual-stage actuator and a newly designed controller for the magnetic disk device benchmark problem proposed by the Specialist Committee on Investigation of High Value-Added by Precision Servo Systems (PSS3). As a result of the verification, the controller with the proposed method was able to reduce the displacement of the PZT actuator, confirming the effectiveness of the proposed method.

Keywords: Motion Vibration and Noise Control
Abstract: Movement of bridge crane payloads through a crowded workspace requires small levels of payload swing for safe operation. The importance of limiting payload swing in such environments was evaluated by studying operator navigation through an obstacle field. Performance under manual operation with and without input shaping was compared with automated traversal using pre-programmed trajectories. These control strategies and test subjects' design and navigation approaches were compared using a small-scale bridge crane. During the navigation tests investigated, the system begins as a single-pendulum and becomes a double-pendulum upon pick up of a payload introducing additional complexity to system control. While implementation of input shaping reduced collisions during manual operation, variance in path selection and shaper design yielded a range of completion times. Implementation of pre-programmed trajectories reduced completion times; however, the lack of human oversight introduced the risk of failing to deposit the payload at the correct location.

Keywords: Fault Detection and diagnosis in Manufacturing, Sensors and Sensing Systems, Artificial Intelligence in Mechatronics
Abstract: Predictive maintenance is an industrial practice to detect component failure ahead in time and before major damage is done to the system. Bearings are susceptible to such damage phenomena and should be replaced before critical failure, therefore early detection proves important. For wide applicability, these detection methods should work on easily available sensor data and have limited computational complexity. This study presents a method for classifying the health of a bearing based on the machines vibrational sensor data and with additional focus on computational complexity. The process involves converting the signals to the frequency spectrum using continuous wavelet transforms, and identifying specific frequency ranges associated with damage phenomena in bearings. Two approaches were used: analyzing the wavelet responses and creating a scalogram image to locate relevant areas. The results obtained on a bearing monitoring data set, created within the Flanders AI Research program, were consistent for both approaches and identified a specific set of scales that resulted in reduced computational load whilst attaining high failure detection rates. Consolidation is achieved by repeating the procedure on two public data sets.

Keywords: Artificial Intelligence in Mechatronics, Machine Learning
Abstract: Geometric parameter errors have a direct impact on the positioning accuracy of industrial robots, making their reduction crucial for enhancing accuracy. Identifying the key geometric parameter errors with the greatest impact on robot accuracy significantly improves its performance. In this paper, we estimate the impact of geometric parameter errors in industrial robots on position accuracy using sensitivity analysis with the Random Forest (RF) method. Firstly, the kinematic error model of the industrial robot is constructed based on the MD-H convention. The principle of RF method is presented, and the geometric parameter errors are randomly sampled by the Latin hypercube sampling (LHS) method, the predictor delta importance (PDI) of each geometric parameter error is calculated. Then, the influence of each geometric parameter error on the position accuracy of the same pose is analyzed. Finally, a simulation experiment is performed with a 6-DOF industrial robot to validate the proposed method's correctness and effectiveness. The results indicate that the precision design of the vital geometric parameter errors could significantly enhance the position accuracy of the industrial robot

Keywords: Mobile Robots, Machine Learning, Control Application in Mechatronics
Abstract: Autonomous mobile robots are used in a wide range of industrial application. Dynamic window approach (DWA) is one of effective local path planning methods considering collision avoidance and kinematic constraints. DWA selects the optical path from path candidates from velocity space by using an evaluation function with fixed weight coefficients. These fixed weight coefficients are designed for the specific environmental situation. Therefore, if the environmental situation such as congestion, road width, and obstacles changes, the evaluation function with fixed weight coefficients may select the inefficient path or path with the collision. To address this issue, this paper proposes the dynamic weight coefficients based on Q-learning for DWA considering environmental situations (DQDWA). Q-learning is one of reinforcement learning methods. The Q-table in DQDWA consists of states of robot and environmental situations, and actions of weight coefficients in DWA evaluation function. By using the learned Q-table, DQDWA dynamically selects weight coefficients and, the optimal path considering environmental situations is generated. The effectiveness of the proposed method was confirmed through simulations.

Keywords: Machine Learning, Control Application in Mechatronics
Abstract: Neural-Network (NN) based compensation is a thoroughly investigated topic in automatic control. However, these approaches include the neural network inside the control loop. This paper proposes an alternative approach where the NN is external to the loop, making decisions on the parameters of a linear compensator in cascade with a plant to be controlled with a feedback system. The proposed model utilizes the adept sequence transduction capabilities of the Transformer architecture. This approach is used to design a discrete controller of any order to maximize available performance. This paper applies this method to simple plants without extreme dynamics and a plant with non-minimum phase and very high quality factor modes.

Keywords: Machine Learning, Sensors and Sensing Systems, Sensor Integration, Data Fusion
Abstract: Developments in sensing and analysis methods have significantly increased the scope of physiological monitoring for healthcare purposes. While the continuous monitoring of physiological measurements enables improved detection and management of many illnesses, accompanying cybersecurity concerns continue to evolve. The large amounts of individualized data necessary to enable learned models for analysis must be sufficiently protected. In addition, the analysis and classification methods themselves should not be vulnerable to attack. This work addresses adversarial individual identification with multiple forms of physiological data, as well as potential performance interruption attacks. The paper proposes a homomorphic encryption scheme to mitigate both of these threats.

Keywords: Genetic Algorithms, Artificial Intelligence in Mechatronics, Fixture and Grasping
Abstract: The manipulation of deformable objects represents an open research topic because of the difficulties in accurately modeling the object behavior in real-world scenarios. This paper presents a trajectory planning framework for the assembly of wiring harnesses for the automotive and aerospace sector, reducing the learning time and simultaneously presenting suitable performance and reliability. A genetic algorithm is used to generate new trajectories according to application constraints. Those trajectories are then executed by the robot and evaluated by means of proper sensor feedback. The proposed framework enable to learn and autonomously improve the task execution, while mantaining a significantly low programming time. Experimental results are reported showing how the robot is capable of optimizing the manipulation of the DLOs gaining experience along the task repetition, while showing high success rate from the very beginning of the learning phase.

Keywords: Intelligent Sensors, Human -Machine Interfaces
Abstract: How can physical computing and interactive fabrics change the way we engage with everyday objects and environments? In this talk, I delve into the transformative potential of applying physical computing principles to wearable technologies and smart textiles, highlighting different breakthroughs such as pocket-based textile sensors capable of detecting user input and recognizing objects carried in our pockets, as well as new touch-sensitive interfaces that leverage different materials like graphene-based fabrics, among others. These types of advancements tease a new era in human-computer interaction, where the seamless integration of wearable technologies, textiles, and physical computing lead to novel, intuitive, and context-aware interactions with the objects and surroundings in our daily lives. As the field is rapidly progressing from fundamental research to commercialization, this talk will showcase the state-of-the-art in the cross-section of domains, as well as describe future research directions and applications that will reshape the way we experience and interact with the world around us.

Keywords: Modeling and Design of Mechatonic Systems, Actuators
Abstract: This study focuses on nonlinear series elastic actuators (SEAs) which have the potential to overcome the performance limit of fixed stiffness SEAs, design a spring unit, and experimentally validate the static characteristic. The result will provide insights into the design of nonlinear SEAs for human assist devices.

Keywords: Control Application in Mechatronics, Identification and Estimation in Mechatronics, Robot Dynamics and Control
Abstract: Soft wearable robots that leverage pneumatic-based actuators are envisioned as a comfortable, portable, and safe alternative to traditionally rigid designs. However, control of such devices is still an outstanding challenge. Most tasks in daily life require adjustments of joint impedance in order to physically interact with dynamic environments. In addition, under conditions of dynamic instability, increased impedance is often required to reduce the effects of neural motion noise. Therefore, it is critical to endow soft wearable robots with impedance modulation capabilities for successful deployment in assistive wearable robot scenarios aimed at augmenting user independence. Our model-based approach leverages the force-deflection characteristics of a single actuator to develop an inverse model for joint impedance of the antagonistic setup. The inverse model takes as input the desired equilibrium and stiffness, and outputs the required pressures of each actuator to achieve the desired impedance.

Keywords: Intelligent Process Automation, Mechatronics in Manufacturing Processes
Abstract: Tank manufacturing for LNG ships is currently mostly manual work due to the size and difficulty of the work. Research on the automation of LNG cargo production using a mobile welding robot is in progress. In this study, software development that determines the operation of a mobile welding robot is introduced.

Keywords: Mechatronics in Manufacturing Processes, Control Application in Mechatronics, Robot Dynamics and Control
Abstract: Tendon-sheath robots have several advantages for entering hazardous confined-space environments given their re- duced size and weight, in addition to being intrinsically-safe when there is a risk of explosion. When teleoperated due to limited sensing in confined-space environments, backlash and stick slip friction in tendon-sheath robots cause hunting-like oscillations around the goal position. The main contributions of this work are to (1) develop an automation method that converges each joint to target values while avoiding backlash and stick slip friction and (2) apply the automation method using traded control to de-hunt the teleoperation. Experimental results show the approach leads to an end effector precision of five-thousands of an inch, an order of magnitude better joint level precision than current tendon- sheath control schemes [1]. Trials of the automated method show a 57.1% reduction in completion time for a precision location task. Initial user studies (N=3) of the method shows a 38.5% reduction in user completion time compared to teleoperation with existing compensation methods.

Keywords: Novel Industry Applications of Mechatroinics, Service Robots, Modeling and Design of Mechatonic Systems
Abstract: This paper proposes a robot-based unmanned automatic charging system that can charge multiple vehicles with a single robot. The robot base is mounted to the ceiling and is designed with a beam and an upper rail to allow horizontal movement. The parking spaces are arranged symmetrically around the beam to minimize the robot's movement while enabling the attachment and detachment of the charging coupler to multiple vehicles.

Keywords: Actuators in Mechatronic Systems, Opto-Mechatronic Sensors, Sensor Integration, Data Fusion
Abstract: Detection of slip during object grasping and manipulation plays a vital role in object handling. Existing solutions largely depend on visual information to devise a strategy for grasping. Nonetheless, in order to achieve proficiency akin to humans and achieve consistent grasping and manipulation of unfamiliar objects, the incorporation of artificial tactile sensing has become a necessity in robotic systems. In this poster, we present a novel physics-informed, data-driven method to detect slip continuously in real time. The GelSight Mini, an optical tactile sensor, is mounted on custom grippers to acquire tactile readings. Our work leverages the inhomogeneity of tactile sensor readings during slip events to develop distinctive features and formulates slip detection as a classification problem. To evaluate our approach, we test multiple data-driven models on 10 common objects under different loading conditions, textures, and materials. Our results show that the best classification algorithm achieves an average accuracy of 99%. We demonstrate the application of this work in a dynamic robotic manipulation task in which real-time slip detection and prevention algorithm is implemented.

Keywords: Automotive Systems, Vehicle Technology
Abstract: This paper presents the design and control of a solar panel cleaning mobile robot carried by a drone. The robot has tracked wheels to stick to the slanted solar panels and move. Control between suction pad and wheel velocities has to be done with care in order not to slip down while moving. Experimental studies of moving on the solar panel were demonstrated to confirm the feasibility.

Keywords: Vehicle Control, Motion Vibration and Noise Control, Control Application in Mechatronics
Abstract: This study proposes an H-infinity control synthesis for solving the actuator saturation problem and conducts an experimental study of H-infinity control for the active suspension system of the quarter car. In the H-infinity control design procedure for the active suspension system, actuator saturation is directly handled by introducing a dynamic model with a saturation function. To improve ride comfort for passengers, the acceleration of the vehicle body is selected as a controlled output, and the H-infinity norm of the transfer function from disturbance to controlled output is optimized. Based on Lyapunov stability theory, the control synthesis problem is formulated as a non-convex bilinear matrix inequality. This design difficulty is overcome by the proposed single-objective distributed quantum-behaved particle swarm optimizer, which efficiently explores the optimal controller that provides the minimum upper limit of the H-infinity norm. The simulation and experimental tests are performed using the Quanser’s active suspension system platform and the road profile generated by trigonometric functions. The results demonstrate the effectiveness of the proposed H-infinity controller.

Keywords: Automotive Systems, Modeling and Design of Mechatonic Systems, Mobile Robots
Abstract: This paper investigates optimization of fuel-efficiency and productivity for automated wheel loaders. A control-oriented model for both the transport phase and bucket loading phase is proposed. The vehicle model includes an automatic gear shift schedule that can be incorporated into the optimization problem. Based on the model, the multi-stage optimization problem is formulated to simultaneously consider all phases of a short cycle with physical constraints. Cycle time and fuel efficiency are used as the weighted performance indexes in a multi-objective cost function. Bucket fill factor is included as a constraint during the bucket loading phase. A nonlinear programming problem is created with collocation using MATLAB and CasADi. The optimization solver IPOPT solves the problem to obtain the optimal state and control trajectories, which can be used as a reference for automated wheel loaders or even as a driver advisory for human-driven wheel loaders.

Keywords: Mobile Robots, Novel Industry Applications of Mechatroinics, Software Design for System Integration
Abstract: There are many benefits for exploring and exploiting underground mines, but there are also significant risks and challenges. One such risk is the potential for accidents caused by the collapse of the pillars, and roofs which can be mitigated through inspections. However, these inspections can be costly and may put the safety of the inspectors at risk. To address this issue, this work presents Rhino, an autonomous robot that can navigate underground mine environments and generate 3D maps. These generated maps will allow mine workers to proactively respond to potential hazards and prevent accidents. The system being developed is a skid-steer, four-wheeled unmanned ground vehicle (UGV) that uses a LiDAR and IMU to perform long-duration autonomous navigation and generation of maps through a LIO-SAM framework. The system has been tested in different environments and terrains to ensure its robustness and ability to operate for extended periods of time while also generating 3D maps.

Keywords: Modeling and Design of Mechatonic Systems, Mobile Robots, Vehicles and Space Exploration
Abstract: Mobile robots can pull payloads far greater than their mass. However, off-road terrain features substantial variation in height, grade, and friction. In addition, temperature changes and precipitation add a time-varying element to the terrain. These effects can cause traction to degrade or fail catastrophically. To maximize tethered payload transport capacity through optimal vehicle traction, unique solutions are required for each surface/condition. This paper presents a system that utilizes a vehicle-mounted, multipurpose manipulator to physically adapt the robot with unique anchors suitable for a particular terrain for autonomous payload transport. Specifically, this work presents "swappable anchors", which can be easily attached/detached to adapt the vehicle using permanent magnets. We present four unique anchor designs, each optimal for a specific surface, and experimentally validate them. The experimental results illustrate how this approach can increase the overall payload capacity of a system on various surfaces by increasing the effective coefficient of friction. We demonstrate how we can use the manipulator to autonomously localize the payload using a visual sensor, attach the payload to the vehicle using a permanent-magnet-based payload key/lock, and enable versatile payload transport capacity.

Keywords: Sensors and Sensing Systems, Mobile Robots, Intelligent Sensors
Abstract: This paper presents a comprehensive modeling technique for optimization of solid-state LiDAR sensor deployment. In particular, a performance measure is developed with physical parameters of flashing LiDAR sensors to describe the pose difference of the LiDAR sensor and target object. An area coverage optimization is then addressed with deployment of LiDAR sensor network (LSN) to demonstrate the effectiveness of the proposed model and the performance measure. An experiment is conducted to verify the proposed resolution criteria of the flash LiDAR sensor and simulations are carried out for validating the developed coverage model for LSN deployment.

Keywords: Mobile Robots, Legged Robots
Abstract: A tracked mobile robot with legs is able to avoid rollover when moving on uneven terrain using its legs. A quadruped tracked mobile robot used in this study can quickly control the legs mounted on the four corners of the body so that they posture properly according to its tipping situation. In this paper, an inertial measurement device is mounted on the robot and posture information is obtained. Based on the obtained posture information, the robot can determine whether it is stable or not. We utilize the normalized energy stability margin to estimate the stability of the robot. The joint angles for the optimal leg posture to recover stability are computed by a simulation. Based on the result of the simulation, the appropriate rollover avoidance motion of the robot was obtained, and its effectiveness was confirmed in the fundamental experiments by the robot.

Keywords: Control Application in Mechatronics, Robot Dynamics and Control, Identification and Estimation in Mechatronics
Abstract: Geared motors are used widely for robots, while the reduction of the backdrivability owing to the amplified friction is a critical issue. The estimation and control of the external torque are necessary to improve the backdrivability. Here, the noise caused by backlash deteriorates the performance of external torque estimation. This study therefore performs the backlash identification of reduction gears using the shaft torsional angular velocity. First, a backlash model is set up based on the torsion angle. Subsequently, the performance of the torsion torque estimation and control is improved by excluding the backlash from the torsion angle based on the identified model. Although the errors in the backlash model are critical for motors with high torsional stiffness, updating the model parameters based on the relative angular velocity values significantly reduces the errors in the backlash model. This study reveals that the accuracy of the torque estimation is improved by compensating for dead zone parameters of the backlash, and the effect is prominent for small external forces. Although the parameters vary depending on how the external force is applied, it is shown that the affect can be suppressed by estimating the parameters online.

Keywords: Modeling and Design of Mechatonic Systems, Mechatronics in Manufacturing Processes, Identification and Estimation in Mechatronics
Abstract: TMECH-01-2023-14786 Ultrasonic tool holders have been widely used in machining hard and brittle materials, but how to control the tool tip oscillation amplitude at the resonance frequency to guarantee the machining accuracy is very important. In this study, a system identification and control of a wireless ultrasonic tool holder was developed. The parameters and size of the piezoelectric ceramics was optimized to ensure that the vibration mode of the tool holder could match the desired resonance frequency. Feedback current of the actuator is used to track the resonance frequency under minimum impedance. A theoretical methodology was applied to obtain the transfer function, and an optimal time domain system identification with bilinear transformation was used to more precisely describe the real tool holder system at the resonance frequency. In order to control the oscillation amplitude of tool tip, an optimal controller with Harris hawks optimizer was used to implement control scheme. From the experimental results, it seen that the amplitude of tool tip can be controlled to desired value within 1 second to fit the requirement in industry.

Keywords: Modeling and Design of Mechatonic Systems, Actuators in Mechatronic Systems, Compuational Models and Methods
Abstract: LISA-Pathfinder is an ESA space mission flown between 2015 and 2017 to demonstrate a technological maturity sufficient for building a gravitational waves telescope in space, such as the Laser Interferometer Space Antenna (LISA). A pair of cubic test masses is hosted inside the LISA-Pathfinder spacecraft and shielded from any force other than the interplanetary gravitational field. The purity of the shielding gives the performance of the mission.

There are a number of aspects that had to be confirmed in-flight. One of them is the transition phase from the launch configuration, when the test masses are locked, to the science free-falling configuration. Each test mass is initially released from the mechanical constraints via a dedicated mechanism and then captured by an electrostatic control system. In fact, each test mass is surrounded by a set of electrodes for actuation and sensing purposes. The performance criterion of the release is the final velocity of the test mass relative to the spacecraft, with an upper threshold set to 5 μm/s. The LISA-Pathfinder first in-flight release velocities highlighted an unexpected dynamics with large linear and angular velocities. The electrostatic control was successful, but only relying on a manual procedure that cannot be considered as baseline for LISA.

This paper helps investigating the in-flight non-compliance by dealing with the modeling of the electrostatic environment around each test mass and its contribution to the release and capture dynamics. The electrostatic model is based on the method of moments, a boundary element numerical technique suitable for estimating forces and capacitances between conductors. We also provide a short overview of the method, which can be used for the analysis of other phenomena within LISA and for the design of future gravitational waves telescopes and space projects.

Keywords: Identification and Estimation in Mechatronics, Control Application in Mechatronics, Vehicle Technology
Abstract: Currently, numerous single-track railway lines are disused due to economic reasons. However, one way they could be reactivated for a bidirectional on-demand service traffic by small vehicles that use only one rail. MONOCABs are such small cabin-like vehicles, stabilized by a system of control moment gyroscopes and a trim mass. They could make an important contribution to improve the mobility offer especially in rural areas. Regarding the MONOCAB, there is currently no reference in comparison with other vehicles. It is mandatory to gain experience before transferring such a new vehicle concept into commercial operation. Especially the safe and robust commissioning of the stabilization control system is crucial and therefore requires an elaborated procedure. At this step, parameters related to the vertical dynamics have to be determined beforehand. This paper presents a comparative investigation of methods to estimate the moment of inertia and gravitational torque constant. Multiple methods in time-domain and frequency-domain are experimentally evaluated and compared with each other. Experimental tests are carried out with a full-scale monorail vehicle.

Keywords: Identification and Estimation in Mechatronics, Actuators in Mechatronic Systems, Fault Detection and diagnosis in Manufacturing
Abstract: One characteristic of rack-and-pinion drives is that they are usually subject to backlash. In non-high precision applications, such as laser cutting, a certain amount of backlash is tolerated. However, changes in the amount of backlash are often related to a fault or damage in the machine. For this reason, it can be useful to monitor the size of the backlash. In this paper, a new method for the determination of backlash in rack-and-pinion drives is introduced and applied to single axes as well as to machines with gantry axes. Since in non-high precision applications there is generally no direct measuring system to detect the output-side position, an additional acceleration sensor is attached to the moving load. Its sensor signal is compared with the motor acceleration obtained from the differentiation of the measured motor speed during a positioning step. With these two acceleration signals, the beginning and the end of the change of the tooth flanks can be identified automatically considering the machine dynamics. From this, the size of the backlash can be determined. It is shown that an automated determination of the backlash is possible even for applications with highly complex machine dynamics.

Keywords: Mechatronics-Enabled Teaching and/or Training, Educational Testbeds and/or Platforms, Mobile Robots
Abstract: Modern-day autonomous vehicles are increasingly becoming complex multidisciplinary systems composed of mechanical, electrical, electronic, computing and information sub-systems. Furthermore, the individual constituent technologies employed for developing autonomous vehicles have started maturing up to a point, where it seems beneficial to start looking at the synergistic integration of these components into sub-systems, systems, and potentially, system-of-systems. Hence, this work applies the principles of mechatronics approach of system design, verification and validation for the development of autonomous vehicles. Particularly, we discuss leveraging multidisciplinary co-design practices along with virtual, hybrid and physical prototyping and testing within a concurrent engineering framework to develop and validate a scaled autonomous vehicle using the AutoDRIVE Ecosystem. We also describe a case-study of autonomous parking application using a modular probabilistic framework to illustrate the benefits of the proposed approach.

Keywords: Virtual Reality and Human Interface, Mechatronics-Enabled Teaching and/or Training, Machine Learning
Abstract: In this study, we propose a virtual reality system for identifying expert-specific skills in a visual inspection task in a refinery by using an eXplainable Artificial Intelligence (XAI) technique. Most previous studies have applied statistical analysis such as t-tests to the mean value of the experimental data, and there is a consequent lack of specificity in the results (i.e., when and where expert skill appears within a long inspection duration). It is thus difficult to provide feedback based on the most important part of the collected experts’ data to the novices. To address this issue, we introduce a Convolutional Neural Network (CNN) with Class Activation Map (CAM) technique, an XAI method, to analyze the experimental data of experienced and novice field operators, and identify the most significant contributors for classifying expert and novice behavior for 120 seconds inspections. The resulting model can classify field operators as expert or novice with an accuracy of 99.1% on average, and visualize the classification criteria as a heat map for each experimental trial. Based on those results, we propose a virtual reality training system for learning expert inspection skills by referencing the CNN results. The contribution of our study is the proposition of a new analytical framework, as well as a training system beyond the limitations of conventional statistical analysis.

Keywords: Modeling and Design of Mechatonic Systems, Mobile Robots
Abstract: Free and structured play should be introduced to children during their childhood in a well-balanced manner. Balls may be used in both plays; therefore, we proposed ball-like robot which will provide children both types of plays. We focused on the jumping of the ball and developed a ball-like jumping robot. In this study, we reported the mechanical design, robot system, performance, and potential of the ball-like robot to jump continuously and turn right/left for controlling the jumping direction. The robot mainly consists of one vibration unit and two compressed springs. Two motions, jumping and turning, are generated by controlling rotation speed of eccentric motors constituting the vibration unit. The experimental results confirmed that the fabricated robot can control the jumping height and average turning angular velocity depending on the rotation speed of the eccentric motors. Furthermore, the robot can be controlled by an operator via commands from a computer with a short delay. The robot can move dynamically. We proposed applications for free and structured play using the ball-like robot from these results.

Keywords: Educational Testbeds and/or Platforms
Abstract: The use of robot-based skill training systems is an emerging topic in nursing education, as many innovative robotic systems have been developed to simulate real patients, offering a safe and self-directed platform for nursing students to learn and practice their skills. Among these training systems, several human patient simulators (HPS) have been proposed to simulate the patient’s performance during patient transfer; however, without an entire motion model and control strategy, most HPS show limited effectiveness in simulating actual patient behavior. Herein, this work presents a novel patient transfer training system that has the potential of improving the practical skills of nursing students. First, we propose a simplified force model for patient transfer motion to estimate the contact force in the absence of wearable sensors. We then reveal the correlation between the nurse’s force and patient’s motion during the transfer through the utilization of the variable admittance model. Finally, we demonstrate the feasibility of the proposed patient transfer training system by performing several experiments on a UR10e robot. To the best of our knowledge, this system is the first patient transfer skills training system that simulates force interaction between nurse and patient using a collaborative robot.

Keywords: Modeling and Design of Mechatonic Systems, Educational Testbeds and/or Platforms, Rapid Prototyping
Abstract: This article presents a novel emotionally expressive robot platform targeting social engagement with children. This platform was implemented in accordance with UNICEF's policy guidance on artificial intelligence (AI) for children, focusing on factors such as safety, transparency, reliability and explainability. The robot prototype is presented from a design and development perspective, outlining all utilized electromechanical components that enable its 11 degrees-of-freedom and sensing functions. Preliminary evaluation results are provided in terms of dependability and expressiveness of basic emotions, thus demonstrating the robot's potential to facilitate trustworthy and secure interactions with children.

Keywords: Biomechatronics, Modeling and Design of Mechatonic Systems, Design Optimization in Mechatronics
Abstract: Exploring granular terrains like loose sand is usually not easy for robots, which must solve the sinking and low-moving efficiency problems. Relying on a unique sidewinding gait, snakes can move efficiently in granular terrains, which brings design insight for continuum robots. However, most previous studies of snake-like sidewinding robots focus on the lateral undulation motion like a sine wave, but few generate the body’s non-zero slope like real snakes. In this article, we analyze the effect of non-zero slope on locomotion efficiency using a Dynamic Resistive Force Model (DRFM) and a material point method (MPM). This work develops a non-wheeled 3D printed snake robot, and the body’s slope can be changed. In addition, the unique structure with soft helix rod makes it have more than 35 degrees of freedom in 35 centimeters in length. In order to reduce weight and complexity, the robot needs only a single motor to achieve the sidewinding gait. Experiments not only confirm that the robot can move more efficiently by changing the slope, which shows the importance of the non-zero slope of the snake robot but demonstrates the designed robot's ability to explore in granular terrains.

Keywords: Modeling and Design of Mechatonic Systems, Actuators in Mechatronic Systems, Compuational Models and Methods
Abstract: Rotary to translational transmission systems play an important role in many applications from engineering to human assistance devices. Although leadscrews and ball-screws are widely available, they suffer mechanical wear/tear problems due to contact friction. Motivated by increasing demands for energy-efficient mechanisms for mobile and wearable robotic systems, this paper presents an analytical method to design a magnetic-leadscrew (MLS) with embedded sensing. MLS is driven by permanent magnets converting magnetic energy to thrust forces while transmitting the rotary-to-translation motion. However, existing designs generally assume an infinitely long MLS, so its magnetic field distribution is axisymmetric and periodic. To relax these assumptions for applications that require maintaining a constant lead over a short travel, the paper formulates the magnetic field and radial/thrust forces of an MLS in closed form using a distributed current source (DCS) method for developing MLS with an embedded field-based sensing system. The sensing method determines the unique solution to the inverse magnetic field model and measures the translation and rotation independently. With the DCS models, a parametric study has been conducted numerically leading to the development of a prototype MLS with embedded sensing, upon which the magnetic field model, sensing system, and algorithm are numerically illustrated and experimentally validated.

Keywords: Modeling and Design of Mechatonic Systems
Abstract: This paper presents the design and analysis of a variable stiffness flexure mechanism with a novel normal-stress electromagnetic stiffness tunable actuator (NESTA). Aiming at real-time stiffness adjustment of flexible mechanisms, the proposed method employs a combination of flexible guiding assembly and the NESTA structure. The static model of the proposed mechanism is developed, including the analysis of the static electromagnetic force and the electromagnetic stiffness, as well as the stiffness of the flexible guide assembly. The electromagnetic stiffness adjustment capability is further evaluated by finite element software COMSOL. The effect of mechanism stiffness adjustment on mechanism characteristics is also analyzed by multiphysics field coupled finite element simulation, which demonstrates the nonlinear characteristics associated with the armature position. The results provide the theoretical and numerical basis for the applications of the NESTA based variable stiffness mechanism.

Keywords: Modeling and Design of Mechatonic Systems, Mechatronics in Manufacturing Processes, Identification and Estimation in Mechatronics
Abstract: An experimental methodology is proposed to characterize the friction behaviour of the unwinding group in a web processing machine. The viscous friction effects from the gearbox are of primary focus in this work as they are dominant during the long and steady operating conditions of the web processing machine. The non-linear change in the viscous friction due to the increase in temperature of the gearbox lubricant is experimentally investigated and characterized. Apart from the viscous friction effects from the gearbox, the friction due to the bearings in the unwinding group is also characterised. The experimentally observed friction torque is represented as a 2D map w.r.t to the rotational speed of the motor and the gearbox lubricant temperature. From the obtained results, the non-linear behaviour of the friction is modelled through curve fit method. Also, a viscous friction coefficient derived from the obtained friction measurements is also presented in the work. The proposed methodology can be applied towards developing dynamic models for model-based design applications, control implementation, parameter identification and deployment of digital twins.

Keywords: Modeling and Design of Mechatonic Systems, Actuators in Mechatronic Systems, Design Optimization in Mechatronics
Abstract: High-performance prostheses are crucial to enable versatile activities like walking, squatting, and running for lower extremity amputees. State-of-the-art prostheses are either not powerful enough to support demanding activities or have low compliance (low backdrivability) due to the use of high speed ratio transmission. Besides speed ratio, gearbox design is also crucial to the compliance of wearable robots, but its role is typically ignored in the design process. This paper proposed an analytical backdrive torque model that accurately estimate the backdrive torque from both motor and transmission to inform the robot design. Following this model, this paper also proposed methods for gear transmission design to improve compliance by reducing inertia of the knee prosthesis. We developed a knee prosthesis using a high torque actuator (built-in 9:1 planetary gear) with a customized 4:1 low-inertia planetary gearbox. Benchtop experiments show the backdrive torque model is accurate and proposed prosthesis can produce 200 Nm high peak torque (shield temperature <60°C), high compliance (2.6 Nm backdrive torque), and high control accuracy (2.7/8.1/1.7 Nm RMS tracking errors for 1.25 m/s walking, 2 m/s running, and 0.25 Hz squatting, that are 5.4%/4.1%/1.4% of desired peak torques). Three able-bodied subject experiments showed our prosthesis could support agile and high-demanding activities.

Keywords: Planning and Navigation, Automotive Systems
Abstract: This paper proposes a new path planning algorithm for robotics called Informed SRRT∗. Compared to conventional RRT∗ algorithms with Euclidean metrics, our algorithm extends the approach by incorporating a local planner from SRRT to satisfy both external and internal constraints. To compute the path to the goal region, we use parameterized cubic curves instead of computationally expensive numerical methods. We add two extra lines at the endpoints of the Bezier spline to leave rooms for the rewiring process. Kinematic constraints require at least three state connections to be tweaked during rewiring. The algorithm always ensures that the path has G2 continuity of curvature within upper-bound constraints. Simulation results demonstrate that the proposed method finds shorter paths than SRRT while maintaining the same iteration of node sampling.

Keywords: Planning and Navigation, Robot Dynamics and Control
Abstract: Unmanned surface vessels (USVs) are widely used in ocean exploration and environmental protection. To ensure that USV can successfully perform its mission, trajectory planning and motion tracking are the two most critical technologies. This paper proposes a novel trajectory generation and tracking method for USV based on optimization theory. Specifically, the USV dynamic model is described with differential flatness, so that the trajectory can be generated by dynamic RRT* in a linear invariant system expression form under the objective of optimal boundary value. We adjust the trajectory through local optimization to reduce the sample number and improve efficiency. The dynamic constraints are considered in the optimization process so that the generated trajectory conforms to the kinematic characteristics of the under-actuated hull, making tracking easier. Finally, motion tracking is added with model predictive control under a sequential quadratic programming problem. Simulated results show that the planned trajectory is more consistent with the kinematic characteristics of USV, and the tracking accuracy remains at a higher level.

Keywords: Automotive Systems, Planning and Navigation, Mechatronics in Manufacturing Processes
Abstract: A novel mechanism to generate non-revisiting uniform coverage (NUC) paths on arbitrarily shaped object surfaces is presented in this work. Given a non-planar surface, non-zero curvature makes traditional homeomorphic fitting of regular template coverage paths from planar regions onto the object surface non-distance-preserving. Any coverage path with a realistic tooling size derived in this way will suffer from overlaps and missing gaps when transformed onto the object surfaces, unable to uniformly cover the target. To overcome this, a discretisation process is adopted to represent the object surface as a uniform unstructured mesh, with resolution set in accordance to the tool size. It is proven that a coverage skeleton path must exist by mesh subdivision refinement which, after a local optimisation step to improve overlap, missing gaps and smoothness, gives rise to template-free superior NUC paths. Extensive simulation examples are presented to prove the validity of the proposed strategy in realistic settings. The proposed scheme is able to achieve 97.9% coverage on benchmark surface tests, outperforming comparable coverage algorithms such as a homeomorphic boustrophedon mapping which can at best achieve 64.6% coverage, or more recent state-of-the-art methods able to reach 94.1% coverage. An accompanying video is supplied with examples, including a real-world implementation of a NUC path tracked by a manipulator. An open-source implementation has been made available.

Keywords: Planning and Navigation, Control Application in Mechatronics, Mobile Robots
Abstract: In last few decades, the coordinated motion of swarm systems which consist of multiple autonomous robots are being intensively examined. These kinds of systems can have various functions such as the creation of the desired formation with physical or non-physical bondings, traveling to a desired position while maintaining the provided formation, and preventing collisions. Especially when looking at recent years, researchers have focused to potential function method to ensure the coordinated motion behavior of swarm systems. In this paper, two different potential function methods and controllers are selected and developed to provide collective behavior, integrated into a decentralized algorithm, implemented at the simulation level, and compared to present a useful guide for future developments on relevant topics. Potential function methods are evaluated and compared within the scope of swarm performance, which is investigated in three stages as gathering individuals, preventing the collisions, and deploying around the target. Thereafter, two different speed controllers are designed for each individual by using PID and sliding mode control methods. Moreover, evaluations of different sliding mode controllers are carried out by using combinations of 2 different sliding surfaces and 3 different switching functions, and the results are compared.

Keywords: Control Application in Mechatronics, Mobile Robots, Novel Industry Applications of Mechatroinics
Abstract: In order to increase productivity and resource efficiency in the construction sector, new approaches are being pursued to automate tasks in the interior fitting of existing buildings. For this purpose, a heterogeneous group of redundant mobile manipulators is used for the time-consuming exact positioning of workpieces. The focus lies hereby on cooperative time optimal trajectory generation for mobile manipulators. To minimize the effort for modeling and formulating the optimal control problem for the different robots, a unified approach based on the Unified Robotics Description Format is developed. The optimal control problem considers the equations of motion of the robots and existing kinematic and dynamic constraints as well as collisions. To enable cooperative trajectories, additional kinematic constraints are imposed on the poses of the end effectors of each mobile manipulator. Different numerical methods, degrees of model abstraction and cost-functionals are investigated with respect to the generated trajectories and the required computation time towards real-time capability for future adaptive trajectory generation using sensor feedback.

Keywords: Planning and Navigation, Mobile Robots, Transportation Systems
Abstract: In recent years, Deep Reinforcement Learning emerged as a promising approach for autonomous navigation of ground vehicles and has been utilized in various areas of navigation such as cruise control, lane changing, or obstacle avoidance. However, most research works either focus on providing an end-to-end solution training the whole system using Deep Reinforcement Learning or focus on one specific aspect such as local motion planning. This however, comes along with a number of problems such as catastrophic forgetfulness, inefficient navigation behavior, and non-optimal synchronization between different entities of the navigation stack. In this paper, we propose a holistic Deep Reinforcement Learning training approach in which the training procedure is involving all entities of the navigation stack. This should enhance the synchronization between- and understanding of all entities of the navigation stack and as a result, improve navigational performance. We trained several agents with a number of different observation spaces to study the impact of different input on the navigation behavior of the agent. In profound evaluations against multiple learning-based and classic model-based navigation approaches, our proposed agent could outperform the baselines in terms of efficiency and safety attaining shorter path lengths, less roundabout paths, and less collisions.

Keywords: Artificial Intelligence in Mechatronics, Intelligent Process Automation, Learning and Neural Control in Mechatronics
Abstract: Path tracking problems are challenging with the absence of dynamic models and information about robot controllers. This paper presents a method of optimizing a motion profile constructed using a set of pre-defined motion primitives and a speed command to track a spatial trajectory with high accuracy, speed, and uniform motion using industrial robots. We use a bi-level optimization approach that optimizes execution accuracy using reinforcement learning and execution speed using bi-section search. We train and evaluate the reinforcement learning policy in simulation for an ABB robot. Experiment results demonstrate that the learned policy reduces the optimization cost to achieve the desired specifications. Additionally, the trained policy can generalize to trajectories not included in the training set.

Keywords: Machine Learning, Neural Networks, Biomechatronics
Abstract: In image-guided surgery, deformation of soft tissues can cause substantial errors in targeting internal targets, since deformation can affect the translation of preoperative image-based surgical plans during surgery. Having a realistic tissue deformation simulator could enhance the accuracy of internal targets localization by giving an accurate estimation of the deformation applied to a preoperative model of the organ. A key challenge is to address the sim-to-real gap between the simulator and the actual intraoperative behaviour of the tissue. The sim-to-real transfer challenge is addressed by formulating the problem as a probabilistic inference over a low-dimensional representation of deformed objects. The proposed method utilizes a generative variational autoencoder structure based on graph neural networks (GNN-VAE) to generate a probabilistic lowdimensional representation of the outputs of a physics-based simulator. To match simulation data to real data, the resultant low-dimensional distribution (i.e., prior distribution) is updated iteratively using an Ensemble Smoother with Multiple Data Assimilation (ES-MDA). The advantages of the proposed method are 1) it only uses simulation data for training the GNN-VAE, and no retraining of GNN-VAE is required intraoperatively, 2) it does not require estimating the mechanical properties of the tissue it is simulating, and 3) is able to work with any physicbased simulator. The proposed framework was verified both in experimental and simulation studies and showed it can reduce the registration error in tissue deformation.

Keywords: Machine Vision, Neural Networks, Machine Learning
Abstract: This paper introduces Deformable Fractional Filters (DFFs) for Convolutional Neural Networks (CNNs). DFFs enhance the efficiency of conventional deformable convolutional filters by introducing a compression mechanism rooted in techniques from fractional calculus. Concretely, our method reduces the parameter overhead requirement of convolutional filters by replacing the kernel with a fractional approximation, which can be trained using only three parameters regardless of the kernel size. DFFs present a compelling use case for the compression of networks that require large kernel sizes. To demonstrate the benefits of DFFs, we report experimental results across a diverse set of computer vision problem domains, including classification, semantic segmentation, medical imaging, and pose estimation. Our experiments illustrate the favorable performance and regularization properties presented by DFFs in comparison with other baseline CNNs

Keywords: Compuational Models and Methods, Neural Networks, Robot Dynamics and Control
Abstract: The field of soft robotics has been experiencing rapid growth, with researchers and engineers showing increasing interest due to the unique capabilities of these robots. Soft robots, characterized by their soft bodies and flexible structures, have demonstrated great potential in addressing real-world challenges across various domains, including medical applications. Effective modeling and control are vital for fully harnessing the potential of soft robots, particularly in applications involving human interaction. However, creating models for soft robots made of soft materials, diverse shapes, and actuators poses significant challenges. Moreover, accurate fault detection in soft robots necessitates precise modeling. This paper introduces a novel machine learning approach, termed deterministic learning, for training a soft robot model using a radial basis function neural network. The research explores the fault detection process by simulating four distinct faults that could impair system control performance, such as diminishing tracking accuracy or inducing instability. Furthermore, the paper examines the identification of fault occurrences during the operation of soft robots.

Keywords: Machine Learning, Machine Vision, Learning and Neural Control in Mechatronics
Abstract: The UniShot 3D precision positioning model was developed using a deep comparison neural network (DCN). This dual-pipeline network extracts features from both the base and inquiry images in real time and predicts the observer’s kinematic movements through internal comparison. We trained the model for transversal and depth movement detections and reported the precision and recall rates through static and dynamic experiments. We also analyzed the feature maps in the convolutional layers at various depths of the model to understand the comparison mechanism of the network. Results showed that the saliency feature patterns of DCNs are distinct from those of image recognition models and that the patterns for the transversal model were distinct from those for the depth model.

Keywords: Neural Networks, Machine Learning, Artificial Intelligence in Mechatronics
Abstract: Robot manipulation of the environment often uses force feedback control approaches such as impedance control. Impedance controllers can be designed to be passive and work well while coupled to a variety of dynamic environments. However, in the presence of a high gear ratio and compliance in manipulator links, non-passive system properties may result in force feedback instabilities when coupled to certain environments. This necessitates an approach that ensures stability when using impedance control methods to interact with a wide range of environments. We propose a method for improving stability and steady-state convergence of an impedance controller by using a deep neural network to map a damping impedance control parameter. In this paper, a dynamic model and impedance controlled simulated system are presented and used for analyzing the coupled dynamic behavior in worst case environments. This simulation environment is used for Nyquist analysis and closed-loop stability analysis to algorithmically determine updated impedance damping parameters that secures stability and desired performance. The deep neural network inputs utilized present impedance control parameters and environmental dynamic properties to determine an updated value of damping that improves performance. In a data set of 10,000 combinations of control parameters and environmental dynamics, 20.3% of all the cases result in instability or do not meet convergence criterion. Our deep neural network improves this and reduces instabilities and failed control performance to 2.29%. The design of the network architecture to achieve this improvement is presented and compared to other architectures with their respective performances.