2021/06/08 by Maria Vittoria Minniti, Minniti, Maria Vittoria, Ruben Grandia +7 · 1 citation
Engineering · #FOS: Computer and information sciences #Prosthetics and Rehabilitation Robotics #Robotic Locomotion and Control #Robotics (cs.RO) #Vehicle Dynamics and Control Systems
paper · pdf · doi:10.48550/arxiv.2106.04202
openalex publication_date 2021/06/08 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
Modern, torque-controlled service robots can regulate contact forces when\ninteracting with their environment. Model Predictive Control (MPC) is a\npowerful method to solve the underlying control problem, allowing to plan for\nwhole-body motions while including different constraints imposed by the robot\ndynamics or its environment. However, an accurate model of the\nrobot-environment is needed to achieve a satisfying closed-loop performance.\nCurrently, this necessity undermines the performance and generality of MPC in\nmanipulation tasks. In this work, we combine an MPC-based whole-body controller\nwith two adaptive schemes, derived from online system identification and\nadaptive control. As a result, we enable a general mobile manipulator to\ninteract with unknown environments, without any need for re-tuning parameters\nor pre-modeling the interacting objects. In combination with the MPC\ncontroller, the two adaptive approaches are validated and benchmarked with a\nball-balancing manipulator in door opening and object lifting tasks.\n