Periodic Reporting for period 2 - SOPHIA (Socio-physical Interaction Skills for Cooperative Human-Robot Systems in Agile Production)
Période du rapport: 2021-03-01 au 2022-05-31
To create HRC systems that can radically increase the flexibility and productivity of manufacturing while improving the ergonomics of the workplace, SOPHIA aims to deliver delivering definite advancements in the state of the art for what concerns:
i) Quantitative evaluation and anticipation of human psycho-physical states during work. This entails the development of modular sensory systems and estimation algorithms which can assess human factors (cognitive and physical) and the delivered effort in the short, middle and long term.
ii) Artificial cognition for socially cooperative human-robot interaction. This is to enable timely, natural, and human-in-command interactions between humans and robots on both social and physical levels, representing socio-physical interaction.
iii) Robotic assistants based on adaptive hardware and modular software solutions to continuously interact with human workers and simultaneously manage flexibility demands and ergonomics of the human partner in manufacturing scenarios where state-of-art collaborative robots are not providing effective solutions.
The project SOPHIA responds to these needs by creating a new generation of core robotic technologies for socially cooperative human-robot systems, enabling dynamic state monitoring of the human-robot pair and anticipatory socio-physical behaviours to i) improve human comfort, ergonomics, and trust in automation in hybrid manufacturing environments, and ii) achieve a reconfigurable and resource-efficient production. To this end, the project envisions a large, ambitious work-program with a strong research and innovation dimension to focus on core technology, usability, acceptability, and standardization objectives (see Figure).
i) Quantitative evaluation and anticipation of human psycho-physical states during work. This entails the development of modular sensory systems and estimation algorithms which can assess human factors (cognitive and physical) and the delivered effort in the short, middle and long term.
ii) Artificial cognition for socially cooperative human-robot interaction. This is to enable timely, natural, and human-in-command interactions between humans and robots on both social and physical levels, representing socio-physical interaction.
iii) Robotic assistants based on adaptive hardware and modular software solutions to continuously interact with human workers and simultaneously manage flexibility demands and ergonomics of the human partner in manufacturing scenarios where state-of-art collaborative robots are not providing effective solutions.
The project SOPHIA responds to these needs by creating a new generation of core robotic technologies for socially cooperative human-robot systems, enabling dynamic state monitoring of the human-robot pair and anticipatory socio-physical behaviours to i) improve human comfort, ergonomics, and trust in automation in hybrid manufacturing environments, and ii) achieve a reconfigurable and resource-efficient production. To this end, the project envisions a large, ambitious work-program with a strong research and innovation dimension to focus on core technology, usability, acceptability, and standardization objectives (see Figure).
In this section, we briefly describe the project's progress w.r.t. the Core Technology (CTO), Usability (UO), Acceptability (AO), and Standardization (SO) objectives.
The CTO has seen solid advancements by the work of WP2, led by UT, and WP3 led by INAIL. Our developments in the first reporting period focused on the human movement measurement techniques and sensor fusion, human dynamic modelling and a number of preliminary works on the development of human ergonomics factors and online model calibration tools, and finally model integration. An additional advancement concerns the development of Fellow-Assistant robots that range from simple and modular load compensation mechanisms (Fellow-Assistant WearBots) to collaborative robots (Fellow-Assistant CoBots) with advanced socio-physical interaction skills. These developments have been carried out in WP8 (Fellow-Assistant Robot Bodies), led by IIT.
In parallel to the Fellow-Assistant robot developments, Fellow-Feeling wearables (FFW) that include low-cost worker state monitoring and feedback devices have been designed and developed in part within WP4, led by UNIPI.
The project’s advancements in the development of multi-modal perception and interpretation during the first period has been on the analysis of the underlying requirements in industrial setting (with emphasis on SOPHIA use-cases) and the development of the tools to interpret the capture multi-sensory data to enable classifications of human-robot interaction. This has been the work of WP5 led by UM. The project’s progress in the domain of shared intelligence for collaboration has focused on plan Generation and task decomposition between humans and robots, contingency handling for safety and ergonomics, social interaction for fluent collaboration, and shared control principles for WearBot and CoBot. Finally, during the first year, WP7 contributions have been: I) development of a self-tuning impedance controller that adaptively regulates quasi-static parameters of the robot, by distinguishing between expected interactions and external disturbances, II) creation of adaptive planning strategies that provide the system the ability to cope with unexpected environmental and operational changes, III) design of an optimization-based algorithm which regulates the Cartesian stiffness and the damping of an impedance-controlled robot without relying on force/torque measurements.
Usability Objectives (UO) are to challenge and demonstrate the contribution of SOPHIA’s core technologies to the improved flexibility, ergonomics, and interaction ability of human-plus-robot systems in realistic production use-cases. In the first year of the project, we explicitly steered the project developments through a careful definition and simulation of the project use-cases: VW, HIDRIA, and HKP. Acceptability objectives (AO) have seen advancements through the demonstration of a coherent reduction of cognitive/physical load. Whereas Standardization objectives (SO) were pursued through the analysis of the existing standards in human-robot collaboration and work-related musculoskeletal risks.
The CTO has seen solid advancements by the work of WP2, led by UT, and WP3 led by INAIL. Our developments in the first reporting period focused on the human movement measurement techniques and sensor fusion, human dynamic modelling and a number of preliminary works on the development of human ergonomics factors and online model calibration tools, and finally model integration. An additional advancement concerns the development of Fellow-Assistant robots that range from simple and modular load compensation mechanisms (Fellow-Assistant WearBots) to collaborative robots (Fellow-Assistant CoBots) with advanced socio-physical interaction skills. These developments have been carried out in WP8 (Fellow-Assistant Robot Bodies), led by IIT.
In parallel to the Fellow-Assistant robot developments, Fellow-Feeling wearables (FFW) that include low-cost worker state monitoring and feedback devices have been designed and developed in part within WP4, led by UNIPI.
The project’s advancements in the development of multi-modal perception and interpretation during the first period has been on the analysis of the underlying requirements in industrial setting (with emphasis on SOPHIA use-cases) and the development of the tools to interpret the capture multi-sensory data to enable classifications of human-robot interaction. This has been the work of WP5 led by UM. The project’s progress in the domain of shared intelligence for collaboration has focused on plan Generation and task decomposition between humans and robots, contingency handling for safety and ergonomics, social interaction for fluent collaboration, and shared control principles for WearBot and CoBot. Finally, during the first year, WP7 contributions have been: I) development of a self-tuning impedance controller that adaptively regulates quasi-static parameters of the robot, by distinguishing between expected interactions and external disturbances, II) creation of adaptive planning strategies that provide the system the ability to cope with unexpected environmental and operational changes, III) design of an optimization-based algorithm which regulates the Cartesian stiffness and the damping of an impedance-controlled robot without relying on force/torque measurements.
Usability Objectives (UO) are to challenge and demonstrate the contribution of SOPHIA’s core technologies to the improved flexibility, ergonomics, and interaction ability of human-plus-robot systems in realistic production use-cases. In the first year of the project, we explicitly steered the project developments through a careful definition and simulation of the project use-cases: VW, HIDRIA, and HKP. Acceptability objectives (AO) have seen advancements through the demonstration of a coherent reduction of cognitive/physical load. Whereas Standardization objectives (SO) were pursued through the analysis of the existing standards in human-robot collaboration and work-related musculoskeletal risks.
SOPHIA targets socio-physically adaptive, ergonomic, and reconfigurable production systems that are accessible and understandable by everyone. Such systems are built on the basis of human work processes backed by decision support tools, assistive and interactive robots and smart self-adapting workplace automation. All are combined into European agile and human-centred manufacturing.
Through the innovative tools of WP1 (and WP9) in analysing the usability and acceptability of SOPHIA core technologies, the project will make an impact on a systematic assessment of productivity and ergonomics in agile production systems. With the developments from WP2, the project will make a strong impact on human musculoskeletal dynamic modelling and state estimation theory by exploiting advanced robotics tools and theory. This will enable an effective integration of worker tracking systems in industrial environments.
WP3 achievements will impact online modelling and monitoring of biomechanical risks in manufacturing processes. This will directly reduce workers’ injury risks and improve overall working quality. The Fellow-Feeling Wearables developed in WP4 will impact the design of low-cost, wireless, light-weight monitoring and haptic devices with improved communication and power-autonomy levels. This will enable integrating affordable solutions in real industry situations.
The multi-sensor software library for motion and action recognition that we will develop and validate in WP5, will impact the Fellow-Assistant robots’ intelligence. In particular, the robots will be capable of understanding at best the external environment (human+robot workspace).
The theory and outcomes of WP6 will have a profound impact on the effectiveness and attractiveness of high-tech Fellow-Assistant robots to a large class of workforce, due to the implementation of task allocation, contingency handling and to smart communication interfaces in between, to improve productivity, safety and mutual collaboration. Advanced physical interaction-assistive control and learning strategies of WP7 will profoundly impact the adaptability of robotic systems to varying task conditions and worker states, contributing to their ergonomics and flexibility.
Novel mechatronics and control developments in WP8, along with the use (and integration) of standard well-performing platforms, will have an impact on flexibility, adaptability, and reconfiguration capacity of robots in cross-domain, challenging manufacturing sectors.
Through the innovative tools of WP1 (and WP9) in analysing the usability and acceptability of SOPHIA core technologies, the project will make an impact on a systematic assessment of productivity and ergonomics in agile production systems. With the developments from WP2, the project will make a strong impact on human musculoskeletal dynamic modelling and state estimation theory by exploiting advanced robotics tools and theory. This will enable an effective integration of worker tracking systems in industrial environments.
WP3 achievements will impact online modelling and monitoring of biomechanical risks in manufacturing processes. This will directly reduce workers’ injury risks and improve overall working quality. The Fellow-Feeling Wearables developed in WP4 will impact the design of low-cost, wireless, light-weight monitoring and haptic devices with improved communication and power-autonomy levels. This will enable integrating affordable solutions in real industry situations.
The multi-sensor software library for motion and action recognition that we will develop and validate in WP5, will impact the Fellow-Assistant robots’ intelligence. In particular, the robots will be capable of understanding at best the external environment (human+robot workspace).
The theory and outcomes of WP6 will have a profound impact on the effectiveness and attractiveness of high-tech Fellow-Assistant robots to a large class of workforce, due to the implementation of task allocation, contingency handling and to smart communication interfaces in between, to improve productivity, safety and mutual collaboration. Advanced physical interaction-assistive control and learning strategies of WP7 will profoundly impact the adaptability of robotic systems to varying task conditions and worker states, contributing to their ergonomics and flexibility.
Novel mechatronics and control developments in WP8, along with the use (and integration) of standard well-performing platforms, will have an impact on flexibility, adaptability, and reconfiguration capacity of robots in cross-domain, challenging manufacturing sectors.