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

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