A lightweight GA-HP algorithm for multi-UAVs coverage path planning in unknown environment

Under unknown environment, the efficiency of most algorithms is largely limited by frequent segmentation of special shaped areas and the assumption of sparse obstacles. These features increase unnecessary computational consumption, resource requirements, and realization difficulty, which seriously affect the practicality of algorithms. Therefore, a genetic allocation-hybrid planning (GA-HP) algorithm is proposed for coverage path planning of multiple unmanned aerial vehicles (multi-UAVs) in unknown environment, which is based on a centralized allocation-distributed planning framework. Firstly, a center-edge model is constructed, which calculates task time through the edges and centers of areas. Secondly, a proposed adaptive elite genetic method achieves area allocation without segmentation, which designs elitism and adaptive parameters to improve the efficiency and adequacy of optimization. Furthermore, a proposed hybrid method generates artificial potential fields, which achieves guidance and reduce overlapping searches. In addition, the edge point detection is designed to improve the coverage integrity. Finally, the performance of the algorithm is validated through experiments and analysis. Compared with existing algorithms, the GA-HP algorithm has lighter computational load and higher task execution efficiency in obstacle dense scenes. Meanwhile, the proposed algorithm responds quickly to emergency, which improves practicality.