4.0 modern factories strive to reorganize operation in the workplace to put Humans back at the center of the activity while preserving their health and providing better work conditions. Collaborative robotics could address these challenges by merging intelligence and skills of Humans and robots to perform complex or demanding tasks. The robot would make the work safer, more ergonomic and more productive. However, human-robot collaborative tasks are so far limited to very simple and separate actions and demand an important learning step to the operator. Scientific barriers have to be overcome to create a full collaboration, where human-robot agents would simultaneously and jointly act toward the task objective.
The project aims at developing a new human-centered generic action-perception scheme to reshape human-robot collaborative works toward an effectively shared activity in 4.0 facilities. This shared-autonomy concept will rethink the levels of collaboration and interaction within the human-robot team in terms of agents’ roles, autonomy, and authority. The project will particularly focus on improving human-robot cooperation in an industrial haptic teleoperation scenario. Haptic teleoperation is a promising method to enable Humans and robots to jointly perform the activity since it naturally combines human high-level intelligence and robot physical capabilities, while maintaining safety and comfort of the human operator. We will rethink human-robot remote interaction to make the robot gain responsibility on simple actions, adapt its behavior with respect to the human intent, and ultimately act as a collaborator toward the task goal. The originality of our approach will be to build upon key psycho-cognitive mechanisms in HR collaboration to develop a human-centered framework which increases ergonomic comfort and intuitiveness of the collaborative task.
Three objectives have to be addressed to develop the novel shared-autonomy method and cover key robotic and cognitive challenges:
1) We seek to study features of human-robot perception-action links and to identify multisensory integration processes involved in human-robot interaction. This human-based approach will constitute the baseline of the later developments and shape our shared-autonomy scheme.
2) We target a shared perception between the different actors, according to their sensorial data and involvement in the task. This shared perception will be based on a multimodal feedback mixture conveying information about the task, the environment, and the collaborators.
3) We aim at merging human-robot commands into a joint action toward the task goal. The human inputs will first be used to infer the operator intent and adapt the robot behavior. Then, the shared action will combine robot skills and human commands into a unified and consistent control objective
All scientific findings will be integrated into a standard, open, and evolutive framework. An experimental collaborative cell will be set up to validate the shared-autonomy framework in augmented haptic teleoperation of the industrial use case. To cover the multidisciplinary challenges of the research program, the consortium relies on a close collaboration between RoBioSS team (Pprime Institute, University of Poitiers), the CeRCA lab (University of Poitiers), and AUCTUS team (INRIA Bordeaux). The synergy of the consortium members will bring transverse and complementary skills needed to reach the project objectives.
Madame Margot Vulliez (Centre Inria de l'université de Bordeaux)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
INRIA Centre Inria de l'université de Bordeaux
Pprime Institut P' : Recherche et Ingénierie en Matériaux, Mécanique et Energétique
Help of the ANR 287,839 euros
Beginning and duration of the scientific project: September 2021 - 48 Months