Keywords: Off-road mobile robots, search and rescue robots, Autonomous Vehicles, Robot Vision and Sensing
Abstract: Multi-robot systems enhance exploration efficiency but encounter challenges in coordination and collaboration. Effective area division is key to balancing workloads, reducing redundancy, and maintaining connectivity for reliable exploration. To address these challenges, this study introduces B-FAME, an adaptive area division algorithm designed for efficient multi-robot exploration and mapping. The algorithm integrates frontier clustering using Dijkstra-based distance matrices, Voronoi-based partitioning, area balancing while maintaining connectivity, and dynamic reallocation of exploration areas as new frontiers emerge. Simulation results illustrate the algorithm's ability to achieve balanced workload distribution among robots, ensuring no overlap and optimized coverage. Furthermore, the dynamic reallocation stage successfully updates partitions during exploration, maintaining coordination without restarting the entire process.

Keywords: Human Vehicle Interaction, Human Movement Modeling, HRI and social robotics
Abstract: This paper presents a novel approach to pedestrian avoidance control based on human imitation using optimal and inverse optimal control techniques. In human-centric environments, where robots and people coexist, ensuring safe operation while maintaining human comfort is a critical challenge. Conventional obstacle avoidance methods typically assume static obstacles, rendering them less effective in dynamic scenarios. To overcome this limitation, we propose a framework that integrates inverse optimal control and Gaussian process regression to identify pedestrian trajectory decision models, which are then incorporated into a collision avoidance strategy based on differential game theory. Experimental evaluations in a pedestrian-robot interaction scenario revealed that, although no statistically significant differences in collision avoidance performance were observed between the proposed and conventional methods, subjective assessments indicated higher comfort (Pleasure) ratings for the proposed method. However, participants also perceived the robot as more dominant in its trajectory execution, which could impact the overall quality of human-robot interaction. These findings suggest that while our approach enhances perceived comfort, further refinement is needed to balance dominance aspects in the robot's movement patterns.

Keywords: Autonomous Vehicles, Service and assistive robotics, Robot Vision and Sensing
Abstract: This paper proposes a sensor fusion method based on Dist-YOLO and LiDAR to enhance the performance of the Dynamic Window Approach (DWA) in dynamic obstacle avoidance, re-ferred to as YOLO-Dynamic Window Approach (Y-DWA). The proposed method employs the Dist-YOLO model to detect pedes-trians and obtain their distance and position information, subse-quently integrated with LiDAR data to compensate for LiDAR's limitations in detecting distant or occluded objects. Specifically, for obstacles detected by LiDAR, a Kalman filter is employed to predict the position and velocity of dynamic obstacles, further enhancing the predictive capability in close-range environments. The DWA generates and selects the optimal path by evaluating the cost of various velocity and steering candidates. Based on this principle, this paper improves the cost function of conventional DWA by incorporating pedestrian detection and dynamic obsta-cle information. This enhancement enables the path planning algorithm to predict and avoid pedestrians or dynamic obstacles in advance, effectively improving both safety and stability in ro-bot navigation.

Keywords: Autonomous Vehicles, Human Vehicle Interaction, Human Movement Modeling
Abstract: This article presents the implementation and experimental evaluation of a driver model that reflects personal driving characteristics for application in automated driving assistance systems. The proposed model incorporates a personal modeling error corrector (PMEC), which estimates and compensates for deviations between predicted and actual lateral displacement during driving. The prediction is generated using model predictive control (MPC) applied to an equivalent two-wheel vehicle model, serving as a baseline representation of standard driving behavior. By integrating the PMEC, the model adapts to personal driving tendencies, improving the accuracy of driver behavior reproduction. A hardware-in-the-loop (HIL) driving simulator with a real steering wheel was used to evaluate the model’s effectiveness in terms of behavioral accuracy and driver workload. Experimental results demonstrated that incorporating the PMEC into the MPC framework enhanced the reproducibility of the driver’s actual operation compared to using MPC alone. Additionally, personalized assistance based on the proposed model contributed to a reduction in driver workload, indicating its potential to improve driving comfort. These findings highlight the benefits of incorporating personal driving characteristics into model-based driver assistance systems and suggest that personalized approaches can enhance the effectiveness of automated driving technologies.

Keywords: Machine Learning and Robot Learning, Robot Vision and Sensing, Autonomous Vehicles
Abstract: The ability to accurately recognize and appropriately respond to traffic lights is essential for improving traffic safety and efficiency, and deep learning has demonstrated outstanding performance in this task. However, collecting annotations for training images remains costly and time-consuming. Therefore, minimizing the number of human annotations required when creating a new dataset is ideal. In this paper, we propose an algorithm that leverages public datasets and the YOLOv10 model to address this challenge. Our approach utilizes labeled public datasets to generate pseudo-labels for unlabeled new data, thereby expanding the training dataset. Furthermore, we introduce a method to reduce data bias and redundancy that may arise during data acquisition. Our method achieves a performance of over 0.629 AP (average precision) by additionally training with 25k frames of unlabeled data. This demonstrates the ability to recognize traffic lights under various environmental conditions, thus meeting the requirements for real-world road scenarios.

Keywords: Intelligent robots for transportation, Autonomous Vehicles, Robot Vision and Sensing
Abstract: The autonomous robot has evolved with the advancement of AI (Artificial Intelligence) technologies, leading to significant improvements in performance. However, these advancements have increased computational demands, requiring higher processing ability and real-time capabilities. In particular, when autonomous robot is performed in a large-scale environment where GPS is unreliable, the computational burden of localization tasks continues to grow. To overcome this challenge, the submap strategy has been introduced, which divides a global map generated through Simultaneous Localization and Mapping into smaller submaps. The submap approach incorporates overlapping regions between submaps to ensure smooth localization during transitions. On the other hand, the submap method also increases storage requirements due to redundant map regions. Moreover, using fixed-size submaps for localization may include unnecessary environmental information, reducing efficiency. This study presents a method that inherently does not allow overlapping regions between submaps, reducing storage requirements while ensuring continuous environmental information during transitions. Furthermore, we propose an approach that dynamically adjusts the provided map area based on the movement direction of agents, improving the efficiency of localization.

Keywords: Entertainment robots, Socio-economic Impacts, HRI and social robotics
Abstract: This paper proposes a novel control framework for robotic swarms capable of turning a musical input into a painting. The approach connects the two artistic domains, music and painting, leveraging their respective connections to fundamental emotions. The robotic units of the swarm are controlled in a coordinated fashion using a heterogeneous coverage policy to control the motion of the robots which continuously release traces of color in the environment. The results of extensive simulations performed starting from different musical inputs and with different color equipments are reported. Finally, the proposed framework has been implemented on real robots equipped with LED lights and capable of light-painting.

Keywords:
Abstract: While the talk’s title may sound oxymoronic, I assure you it’s not.  For example, ongoing meetings of the United Nations in Geneva regarding the Convention on Certain Conventional Weapons consider the many issues surrounding the use of lethal autonomous weapons systems from a variety of legal, ethical, operational, and technical perspectives. Over 80 nations are represented and engaged in the discussion. This talk reprises  and updates the issues the author broached regarding the role of lethal autonomous robotic systems in warfare, and how if they are developed appropriately they may have the ability to significantly reduce civilian casualties in the battlespace. This can lead to a moral imperative for their use, not unlike what Human Rights Watch has attributed regarding the use of precision-guided munitions in urban settings due to the enhanced likelihood of reduced noncombatant deaths. Nonetheless, if the usage of this technology is not properly addressed or is hastily deployed, it can lead to possible dystopian futures. This talk will encourage others to think of ways to approach the issues of restraining lethal autonomous systems from illegal or immoral actions in the context of both International Humanitarian and Human Rights Law, whether through technology or legislation.

Keywords: Robots in the Smart Factory, The role of AI in the implementation of cognitive robots, Robotics for manufacturing
Abstract: With growing interest in smart manufacturing systems around the Industry 5.0 concept, the manufacturing industry is continuously seeking multi-parameter balancing between safety and productivity along with fewer trade-offs. These practices focus on keeping the human-in-loop physically and mentally secure, allowing an uninterrupted flow of human creativity and analytical reasoning into the system. In line with this approach, a novel controller for a Cyber-Physical System (CPS) has been developed. This controller is able to recognize the physiological requirement of the human operator and activates a mobile cobot to take over tasks until the operator goes through a scheduled break time. The mobile cobot is capable of collision avoidance path planning in real-time, which improves safe navigation control in dynamic environments. This entire system prioritizes human factors while maintaining productivity.

Keywords: Robot Vision and Sensing, The role of AI in the implementation of cognitive robots, Industrial robotics
Abstract: Replacement of deteriorated pipework in building is costly and risky. Utilizing AI and robotics for predictive maintenance can greatly reduce expenses and mitigate risks. In this work, we introduce AI and robotics technology for predictive maintenance of iron pipe network. Specifically, AI technology was used to assist in detecting pipe corrosion, and a magnetic crawling robot capable of adhering to and safely navigating iron pipe was used to inspect areas that are difficult for human inspectors to reach. Our proposed human-robot collaborative maintenance framework can significantly enhance inspection efficiency, reduce the risk of work at height and in confined space, and ultimately extend the pipework service life.

Keywords: HRI and social robotics, Industrial robotics, Humanoid Robots
Abstract: Virtual reality (VR) is an emerging industry 4.0 technology to enhance flexibility, efficiency, and quality to ever-increasing economic competition. Autonomous robots can enhance the efficiency significantly in human-robot collaborative (HRC) assembly via VR components. A hybrid automation is taking benefit from the synergistic effect of HRC and the use of continuous simulations offer safe virtual space for validation. However, conventional simulations don’t allow to experience the production system as an end-user in an immersive environment. This paper presents the robot experiment on user interface design in unified framework for HRC assembly in a shared workspace. The experiment aims to verify the effectiveness of visual and haptic cues in various forms that convey the robot intent to human. A virtual laboratory was set up containing sense gloves and virtual input and output devices. The sense gloves were integrated into Unity and the SteamVR to perform the 3D tasks. To ensure real time interaction when capturing virtual objects, the virtual glove from the Unity libraries was the preferred option to generate the required events. Finally, the developed gesture control system allows operators to control an industrial robot via the VR environment.

Keywords: Industrial robotics, New Initiatives and Strategies for New Industries, Socio-economic Impacts
Abstract: Ships will eventually become waste when reaching the end of their operational lives. As part of the sound and safe ship recycling, it is mandatory to inspect ships before dismantling them. The inspection is usually done by experts who perform visual inspection and do tests in any areas deemed to have potential risks to humans. This paper presents a study conducted as part of the SHEREC project, which aims to reduce safety risks to workers in the ship recycling industry by using robotic systems. This paper investigates stakeholder perspectives on using drones for inspecting ships to be dismantled. We interviewed 13 participants who worked in the ship recycling industry to explore the benefits and challenges of using drones for inspecting ships to be dismantled. The results from the interviews indicate three potential benefits of using drones in this context, which are (1) inspecting high structures of a ship, (2) inspecting confined spaces in a ship, (3) making preparations before sending people to take samples. The participants also saw five challenges that may exist when using drones for inspecting ships to be dismantled, such as (1) the need to send people to open all doors prior to the drone inspection, (2) the lack of capability to inspect tight and narrow spaces of a ship, (3) the inability to perform destructive sampling, (4) the potential to disturb asbestos, and (5) the influence of the ship's walls on the drone's signal. Furthermore, the participants also provided four suggestions that could be considered to make drones more suitable for inspecting ships, which are (1) having drones in smaller sizes, (2) equipping drones with gas sensors, (3) installing non-destructive sampling tools on drones, (4) and geo-tagging pictures and videos taken by drones.

Keywords: Intelligent robots for transportation, The role of AI in the implementation of cognitive robots, HRI and social robotics
Abstract: This paper proposes a centralized Large Language Model (LLM)-based Multi-Robot System (MRS) for application of cooperative object transportation. We integrate the LLM into a ROS-based MRS to improve structured task allocation and execution. By leveraging prompt engineering, natural language instructions are converted into hierarchical JSON commands, allowing the central host to efficiently parse and distribute tasks among robots for coordinated multi-robot control. In addition, to achieve efficient and adaptive navigation in different environments, the system integrates A* and Dynamic Window Approach (DWA) algorithms for global path planning and local obstacle avoidance. Simulation and real-world experiments validate the effectiveness of the proposed LLM-based MRS control method. This study contributes to the advancement of MRS in industrial and service applications.

Keywords: Industrial robotics, Off-road mobile robots, search and rescue robots
Abstract: Mining exploration requires robots to traverse unstructured environments with uneven terrain, loose soil, and steep inclines, posing substantial challenges to mobility. Although planetary rovers equipped with passive suspension systems such as the Rocker-Bogie have successfully navigated similarly rugged landscapes, conventional mining robots, often relying on tracked or wheeled mechanisms, can be prone to tipping and immobilization when confronted with obstacles and harsh conditions, also suffering mechanical stress and reduced stability due to frequent inclination shifts. Addressing the need to improve robot stability and adaptability in these complex mining terrains, this study investigates an alternative solution. Here we show that incorporating a Rocker-Bogie suspension into a mining robot significantly improves stability, traction, and adaptability, as demonstrated in simulations conducted on Gazebo. The Rocker-Bogie design reduces angular deviations and lateral oscillations compared with a Tracked System, enabling smoother motion over obstacle-laden terrain; it also maintains continuous ground contact and passively distributes weight to mitigate abrupt changes in orientation. These results underscore the potential for adopting Rocker-Bogie mechanisms to enhance safety and efficiency in both underground and open-pit mining. By applying proven concepts from space robotics, we highlight a promising path toward more robust and versatile mining exploration platforms; ultimately, leveraging advanced suspension systems may enable future robots to operate autonomously in even more extreme and hazardous environments, expanding the reach of mining activities and safeguarding human workers.

Keywords: Robotics for manufacturing
Abstract: The increasing complexity of modern workplaces, especially in High Mix, Low Volume (HMLV) production, demands robotic solutions that are not only flexible but also capable of seamlessly integrating with human workflows. Despite their potential, robots are often met with skepticism, driven by fears of job displacement and reduced human agency. To address these barriers, the ALTO Robotics team employed a design thinking methodology to develop a collaborative robot tailored to real-world workplace needs. By combining agentive technology—where robots act as intelligent assistants—and principles of social robotics that prioritize trust and meaningful interaction, the robot was designed to enhance human-robot collaboration rather than replace human roles. This paper presents the development process, key outcomes, and a user-centered evaluation of the robot's acceptance using the Robotic Social Attribute Scale (RoSAS). Findings highlight the importance of aligning robotic design with human expectations to foster trust, adoption, and effective workplace integration.