Reactive control and sensor fusion for mobile manipulators in human robot interaction

摘要: In order to share a workspace with humans, service robot should be able safely interact within an unstructured environment. this context, the shall adapt its behavior and react environment changes human activities. The robots based on motion planning are not fast enough, so we propose reactive trajectory controller track targets, activities prevent event like collisions. reliability of proposed is recent fusion techniques identify mouvements detect forces associated events. We employ 6D force/torque sensor estimate inertial parameters manipulated objects, then used complement visual tracking process compute contact between end-effector analysed classified by using learning different events, such as grasping object or collision This work, conducted part European projects DEXMART SAPHARI, ANR ASSIST ICARO, has been integrated validated Jido PR2 platforms LAAS-CNRS.