This paper focuses on planning strategies for object-interception, within the context of an active prediction planning and execution (APPE) strategy. Herein, the goal is to bring the robot end-effector to a pre-grasping location, so that a fine-motion grasping-strategy can be used. This objective is optimized within the APPE strategy by solving two problems: (i) selection of an optimal rendezvous-point; and, (ii) planning of a time-optimal robot-trajectory to this rendezvous-point. Problem (i), addressed herein, is formulated as a minimization of the robot-target interception-time. An adaptive-planning algorithm is proposed, where only the exact amount of computation time required is used to find the optimal rendezvous-point. Within this strategy, uncertainties must be expected and tolerated. The level of uncertainty provides a convergence constraint for the planning algorithm. The operation of this planning strategy is demonstrated via computer simulation. Experimental work using this approach is also cited.
展开▼
