Finding Optimal Manipulator Arm Shapes to Avoid Collisions in a Static Environment

Abstract

In situations of a confined workplace with a lot of obstacles and a complicated required trajectory of the endpoint of an industrial or collaborative robot, it may be impossible to find a suitable robot and its position within the workplace to fulfill the given task. In some cases, it could be favorable to design a custom manipulator arm with an unusual kinematic structure or shapes of some of its links. This article presents a novel way of finding the optimal lengths and shapes of two crucial links of a manipulator arm, where the target lengths are as short as possible to reduce mass, and the shape in the form of a Bézier curve is chosen to avoid collisions. The chosen type of kinematic structure of the manipulator arm is fixed and is based on the most typical structure of existing industrial robots, with six degrees of freedom. Two algorithm variants were proposed; one method uses iterations to find the solution based on in-depth collision analysis, and the second method uses the particle swarm optimization algorithm. Both methods were implemented in a simulation system and verified in several testing workplaces.