Industrial robots play a key role in industrial automation. Robotic arms populate shop-floors: they are used for pick-and-place, assembly, inspection, and many other tasks, to increase the throughput of productive processes and alleviate fatigue and risks of human workers. A huge research effort has been put into the reasoning, planning, and control of robotic manipulators. Nonetheless, industrial implementations often do not exploit at full the great advancements made in these fields. This workshop aims to discuss how recent developments in the planning and control of robot manipulators, on the one hand, and the synergetic integration with results from Artificial Intelligence, on the other, can advance the state of the art and be applied to real-world manufacturing processes. Among the many challenges in the field, the workshop will focus on the following trends that emerged in recent years:
- Human-robot collaboration: collaborative robots are expected to play a key role in the factory of the future. The collaboration of humans and robots is supposed to combine the dexterity and reasoning ability of humans with the precision and continuity of robots. Current industrial solutions often lack smoothness and collaboration results to be discontinuous. This occurs at different decision-making levels. For example, implementations of safety rules according to safety standards (e.g., ISO-TS 15066) stop the robot as soon as human workers enter the robot workspace. Moreover, robot trajectories are often pre-computed and do not adapt to the system changes. Finally, ordering, scheduling, and assignment of tasks do not model human behaviors and preferences, resulting in poor dependability and jeopardizing the overall collaboration experience. Recent advances in task and motion planning addressed this issue in many several ways. Innovative methods have been developed to improve safety, ergonomics, and the efficiency of the process. Nonetheless, a well-established common paradigm is still to come.
- Cognitive manufacturing: a central aspect concerning the integration of AI and Robotics in modern manufacturing scenarios is the enhancement of perception and reasoning capabilities of robotic solutions. AI technologies can indeed endow robot controllers with the cognitive capabilities necessary to understand the state of human operators and the state of the environment and contextualize robot behaviors accordingly. A collaborative robot would, for example, dynamically adapt its behaviors to known skills and monitored physiological state of human workers (e.g., ergonomics, cognitive load, fatigue, etc.) in order to achieve a smooth and natural interaction. Such a higher level of cognition is crucial to systematically put the human-factor in the loop and enable symbiotic, personalized and adaptive interactions between humans and robots.
- Flexible manipulation in challenging scenarios: pick-and-place, sorting, and packaging can be efficiently automatized when they are required to manipulate objects with low variability (similar sizes and shapes) and they are performed in structured environments. However, when it comes to partially structured environments or high-variability, current industrial solutions usually fail because of a lack of flexibility and efficiency. Similarly, manipulation of large and/or deformable objects is still a hard task to perform with robotic manipulators. Examples are those draping processes required in automotive and aerospace (carbon-fiber manipulation) and in the textile industry. Despite these topics have been addressed for a long time by researchers, real-world implementations and successful case studies are rare and only recent research projects are trying to effectively automatize these processes. These new solutions should integrate vision, learning, and planning.
We invite researchers from both industry and academia to contribute to this workshop with papers on their recent advances in these fields, focusing on both theoretical methodology and industrial case studies.
to download the workshop cfp.