Mine SAR online planning based on grey system theory and proximal policy optimization

The confined space and complex layout of underground mines create significant challenges for personnel and large equipment during rescue operations. Additionally, the pathways in underground mines can lead to communication disruptions. To address these issues, this article proposes an online path-planning algorithm specifically designed for the unmanned ground vehicle used in mining environments for search and rescue operations. Firstly, the proposed algorithm described a data analysis algorithm based on the even Grey model alongside a target search decision-making process and a multi-objective decision analysis-based target tracking process. It employs the Even Grey model to analyze real-time environmental data. Then, the proposed algorithm enhances online path planning by integrating a near-end strategy optimization algorithm with multi-objective decision analysis. This combination allows the unmanned ground vehicle to effectively navigate and engage dynamic targets, even in unknown environments where targets may not be visible. Finally, simulations in various mine scenarios were conducted to evaluate the feasibility of the proposed algorithm. The performance of the algorithm was compared with multiple online planning algorithms to validate its superior capabilities. The experimental results demonstrate that the proposed search planning algorithm can successfully navigate and capture targets within the designated search area, enabling continuous searching of multiple targets.