Route planning based on improved probabilistic roadmap and harmony search

Abstract

Aiming at obstacle avoidance during marine navigation, improved probabilistic roadmap and harmony search are proposed to design routes for ships. The proposed algorithm makes use of an improved probabilistic roadmap at first. It sets and expands nodes around obstacles and the line which connects the starting point and ending point. All nodes are set and connected in phases according to the motion characteristics of obstacles and ships. A complete path network map is generated by making use of a relatively small number of nodes, based on which initial global routes can be generated by choosing nodes. Then the initial global routes are optimized by employing an improved harmony search. The motion characteristics of obstacles result in the multimodal solution space. Slight changes of the nodes can make the new routes unworkable. To avoid situations like that, only the routes which satisfy the optimization conditions can be updated by overlapping, nodes elimination and fine-tuning. Experimental results show that compared with other algorithms, the proposed algorithm can generate global routes with higher quality. The number of unworkable routes generated in the optimization process is less than that of other algorithms. The proposed algorithm is more applicable and stable.