Characterizing Performance of Sensorless Fine Positioning Moves in the Presence of Grasping Position Uncertainty

In this paper, we characterize the performance of sensorless fine positioning moves in the presence of initial position uncertainty of a part being grasped by a robot. More importantly, we are interested in approaches that work for low-cost robots which provide coarse positional accuracies. Accordingly, we also characterize how the parameters used by the robot’s motion planner influence the overall performance. We present a planning algorithm that performs sensorless manipulation in a two-step process comprising transport and fine-positioning. We report results from experiments with a Baxter robot to illustrate our approach.