Multi-Agent Planning and Control for Autonomous On-Orbit Servicing, Assembly, and Manufacturing

NASA envisions On-orbit Servicing Assembly and Manufacturing (OSAM) solutions that involve heterogeneous teams of coordinated service robots. nbsp;It will be infeasible to effectively control such teams using existing teleoperation techniques. nbsp;To address this issue, TRACLabs is developing a multi-agent control framework that builds upon concepts from sliding autonomy and task-level commanding to significantly improve OSAM capabilities. We call this system CARTEL (Coordinated Autonomy to Replace TELeoperation). CARTEL will improve OSAM capabilities by facilitating faster, safer, and less complicated robot motions to achieve NASA#39;s mission goals.This research extendsnbsp;the existing CRAFTSMAN robot planning and control framework developed by TRACLabs to multi-agent systems. This is achieved through an innovative planning and controls approach of treating the multi-agent system as a single agent and by extending the Affordance Template task description language to multi-agent systems. Phase I demonstrated the feasibility of the approach in kinematic simulations of multi-agent OSAM task analogs. Phase II will further develop these capabilities, increase ease of operation, and develop necessary components to bring the developed technology closer to NASA missions.nbsp;nbsp;