Using robots to replace manual inspection in complex underground environments is an important direction for intelligent coal mine production. To address the coupled problem of task allocation, path planning, and charging replenishment in continuous multi-robot inspection of underground mine tunnels, this paper proposes a charging aware two-stage task planning method. First, an underground tunnel environment model with task points, irregular obstacles, and charging stations is constructed. Based on this model, a multi-robot task allocation model considering road-environment factors and charging feasibility is established, and an improved bidirectional ant colony optimization (IBACO) algorithm is adopted to obtain the inspection task sequence for each robot. Second, a path planning model considering charging constraints is developed by integrating road environment factors, obstacle avoidance, travel time, and energy related cost. An improved honey badger–fly optimization algorithm (IHBAFOA) is then used to generate safe inspection paths. Simulation results show that IBACO⌃IHBAFOA can support continuous multi-robot inspection and achieves lower path cost and relatively good stability compared with the comparison algorithms. Compared with ACO⌃HBAFOA, ACO⌃IGWO, ACO⌃DPSO, ACO⌃AGASA, ACO⌃PSOGSA, and ACO⌃PIOFOA, the proposed method achieves an average cost function value of 6800.3, corresponding to a reduction of approximately 7.20% to 11.16%, with a standard deviation of 6.50. Meanwhile, its average running time is 120.0 s, which is not the shortest among all compared algorithms, indicating that computational efficiency still needs further improvement. These results demonstrate that the proposed method can provide a feasible task planning scheme for continuous multi-robot inspection in underground mine tunnels while considering road-environment effects, obstacle avoidance, and charging constraints.
Science
