Virtual ergonomics for the design of collaborative robots. (Ergonomie en environnement virtuel pour la conception de robots collaboratifs)

The growing number of musculoskeletal disorders in industry could be addressed by the use of collaborative robots, which allow the joint manipulation of objects by both a robot and a person. However the efficiency of a collaborative robot regarding the reduction of musculoskeletal disorders risks is highly task-dependent. Yet, even when designing dedicated systems, the ergonomic benefit provided by the robot is hardly ever quantitatively evaluated, because of the lack of relevant assessment tools. This work aims at developing a generic tool for performing accurate ergonomic assessments of dynamic situations with very little input data. More specifically, it focuses on the development of a methodology to quantitatively compare the ergonomic benefit provided by different collaborative robots, when performing a given task. The word ergonomy refers here to biomechanical factors only. The proposed methodology relies on an evaluation carried out within a digital world, using a virtual manikin to simulate the worker. Indeed, the simulation with a virtual manikin enables easy access to many detailed biomechanical quantities, for different kinds of human morphologies. Besides in the case of collaborative robotics, a virtual instead of a physical mock-up of the robot is used, which can be more easily modified. Therefore ergonomic assessments of the robot-worker system can be performed throughout the design process. Ergonomic indicators which match the requirements of collaborative robotics are defined. Such indicators account for the different biomechanical solicitations which occur during manual activities, performed with or without the assistance of a collaborative robot. The measurement of the proposed ergonomic indicators requires the simulation of the activity with a virtual manikin. To this purpose, a framework for the dynamic simulation of co-manipulation activities is implemented. A strength amplification control law is used for the co-manipulation robot, and the virtual manikin is animated through a LQP optimization technique. The reliability of the proposed measurement framework is then validated. Motion capture based experiments are carried out in order to estimate the realism of the manikin model, and the consistency of the proposed ergonomic indicators. A fully automatic simulation is implemented in order to ensure the usefulness of the manikin-robot simulation, regarding the ergonomic comparison of collaborative robots. The proposed simulation framework allows to estimate a variety of ergonomic indicators while performing a given task. However the high number of indicators makes any kind of conclusion difficult for the user. Hence, a methodology for