An Effective Path Planning Method Based on VDWF-MOIA for Multi-Robot Patrolling in Expo Parks

Expo parks are characterized by dense crowds and a high risk of accidents. A multi-robot patrolling system equipped with multiple sensors can provide personalized services to visitors and quickly locate emergencies, effectively accelerating response times. This study focuses on developing efficient patrolling strategies for multi-robot systems. In expo parks, this requires solving the multiple traveling salesman problem (MTSP) and addressing multi-robot obstacle avoidance in static environments. The main challenge is to plan paths and allocate tasks effectively while avoiding collisions and balancing workloads. Traditional methods often struggle to optimize task allocation and path planning at the same time. This can lead to an unbalanced distribution of patrol tasks. Some robots may have too much workload, while others are not fully utilized. In addition, poor path planning may increase the total patrol length and reduce overall efficiency. It can also affect the coordination of the multi-robot system, limiting its scalability and applicability. To solve these problems, this paper proposes a multi-objective immune optimization algorithm based on the Van der Waals force mechanism (VDWF-MOIA). It introduces an innovative double-antibody coding scheme that adapts well to environments with obstacles, making it easier to represent solutions more diversely. The algorithm has two levels. At the lower level, the path cost matrix based on vector rotation-angle-based obstacle avoidance (PCM-VRAOA) calculates path costs and detour nodes. It effectively reduces the total patrol path length and identifies optimal obstacle avoidance paths, facilitating collaborative optimization with subsequent task allocation. At the higher level, a crossover operator inspired by the Van der Waals force mechanism enhances solution diversity and convergence by enabling effective crossover between antibody segments, resulting in more effective offspring. The proposed algorithm improves performance by enhancing solution diversity, speeding up convergence, and reducing computational costs. Compared to other algorithms, experiments on test datasets in a static environment show that the VDWF-MOIA performs better in terms of total patrol path length, load balancing metrics, and the hypervolume (HV) indicator.