This paper takes into explication the requirements to plan trajectories for taking a roving target and manipulating it with two arms' mobile robots with 7 degree-of-freedom provided by a Barrett hand and a gripper in an environment that takes the barrier problems on consideration. For the robots in charge, the algorithm has as function to determine the necessary trajectory to catch the object, based on an observation: when the object starts moving, the algorithm uses also some variables like the start configuration, the object's speed and the robot's model and the scene. This algorithm is an expansion of RRT-JT which exploits the Jacobian-based gradient descent to coach a 7-DoF WAM robotic arm with a dynamic base and a robot gripper, in sort to generate the shortest time path for the capture procedure, while avoiding obstacles. Finally, we present empirical studies on tracking and capturing the strolling object with our mobile robot and we compare the results of two robots.
展开▼
