Planning velocity profiles from task-level constraints and environment uncertainties

Abstract

A method for parameterizing robot trajectories in the presence of uncertainties is presented. The planning process is defined as a problem of constrained optimization and the concept of a task's difficulty is used as an optimization criterion. The task difficulty, as defined by the authors, comprises the combined effects of velocity and uncertainty, mimicking human perception of difficulty in positioning tasks. The success probability is used as a constraint necessary for planning tasks with contradicting requirements. This planning paradigm is demonstrated with an experiment that contains opposing requirements: reaching the obstacle in a given time, but without exceeding certain maximal impact force. The planner is implemented on a real system.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>