Multi-level MQA-A* Path Planning Algorithm Based on Adjacent Node Aggregation

Xiaoning YANG,Chengqi CHENG,Bo CHEN,Xiaochong TONG,Jing HE

DOI: https://doi.org/10.3969/j.issn.1672-1586.2018.01.015

2018-01-01

Abstract:A multiscale quarter aggregation A*(MQA-A*) algorithm was developed as an improved A* algorithm to plan path efficiently on large and complex grid maps. The algorithm builds high-level abstract grid maps layer by layer via adjacent grid aggregation, generates a rough path on the high-level map, and refines the rough path down to the original grid map based on the intra-connection information within high-level aggregation grids. MQA-A* was demonstrated in several scenarios, which varied in the complexity of obstacle distribution. The experiments showed that MQA-A* was able to improve the computational efficiency significantly in reducing memory usage and decreasing algorithm CPU time. Particularly, the algorithm can obtain much better performance as the proportion of examined grids increases.