Task planning method for multirobot collaboration in substation inspection

Efficient equipment inspection is crucial for ensuring the safety of substations. Although inspection robots are widely used, existing single-robot methods often suffer from low efficiency and limited resource utilization. To address these issues, this paper investigates multi-robot task coordination and planning. First, a unified intelligent inspection system is proposed to enable seamless data interaction and collaboration. Next, a task point assignment and path optimization model is developed to minimize energy consumption and time costs through an improved genetic algorithm. Additionally, a deep Q-learning network is employed for collaborative path planning, leveraging the optimized inspection sequences. A 500kV substation is used as a case study to validate the proposed method. The results show that the proposed collaborative approach significantly reduces time costs and improves overall efficiency compared to traditional single-robot schemes, while also demonstrating high adaptability in handling faults and unexpected situations.