Vision-Aided Path Planning Using Low-Cost Gene Encoding for a Mobile Robot

摘要

Path planning is intrinsically regarded as a multi-objective optimization problem (MOOP) that simultaneously optimizes the shortest path and the least collision-free distance to obstacles. This work develops a novel optimized approach using the genetic algorithm (GA) to drive the multi-objective evolutionary algorithm (MOEA) for the path planning of a mobile robot in a given finite environment. To represent the positions of a mobile robot as integer-type genes in a chromosome of the GA, a grid-based method is also introduced to relax the complex environment to a simple grid-based map. The system architecture is composed of a mobile robot, embedded with the robot operating system (ROS), the ArUco system and a laptop, executing the algorithms of path planning and image processing. Both simulations and experimental results are presented to verify the feasibility of the proposed method. In applications, this work can be employed in a commercial ball-collecting or an object-carrying robot.