Collision-free object manipulation planning in complex environments

Abstract

This paper presents a novel algorithm for robots to reach its target in a complex and clutter environment through minimal obstacle removal. Since robotic manipulation has been focused only in the industrial field, improving the efficiencies in the service domains are still an ongoing challenge. For complex and clutter service environments (i.e. supermarkets), relocation of obstacles are needed to reach the target. Previous studies have focused on different methods to reach the target, but only focuses on target reach rather than the number of obstacles that are removed. In this study, modified Vector Field Histogram plus and Rapidly-exploring Random Tree are used to develop the proposed algorithm for task and motion planning. This novel algorithm reduces the number of relocated obstacles in the cluttered and complex environments. From this, the time to reach the target is reduce due to minimal obstacle removal. In the simulation, the proposed method reduced up to 37% for the number of obstacle relocation and 32% reduced performance time. Additionally, through real robot experiment, the proposed algorithm has reduced up to 34% for the number of obstacles relocation and 36% reduced performance time compared to other existing methods. Therefore, this study provides an optimal algorithm for object manipulation in complex and clutter environment. The work presented here has profound implications for the use of robotic manipulation in service domains.