An Enhanced Optimization Approach for Generating Smooth Robot Trajectories in the Presence of Obstacles

Abstract

The basic concept of this work is the integration of path planning in environments with obstacles and motion execution under robot dynamic constraints. Instead of tracking an exactly planned geometric path, which limits not only the search space of the optimal trajectory but also the flexibility for on-line modification, a robot (mobile or manipulators) realizes its motion under the guidance of a sequence of grossly selected crucial points called subgoals. B-splines are used to generate smooth trajectories through these subgoals. To optimize the motion time, first, a multicriterion cost function is designed which integrates motion time, trajectory length and curvature. Then a hybrid optimization algorithm, with which the local minima can be explicitly detected, is developed. Experiments show that optimization results are considerably improved using the cost function and the optimization algorithm in comparison with other published approaches. I. Introduction A. Problem Description A...