考虑平台特性的多层建筑物内履带式无人平台运动规划

To solve the navigation problem of the unmanned vehicles in multi-storey buildings, a motion planning framework considering the characteristics of vehicles is proposed. Based on the characteristics of tracked vehicles, the primitive scheme of zero-radius steering is adopted and the Voronoi Path is introduced, which improves the flexibility and safety of the global planner in a narrow environment. Then, the smooth global path is obtained through piecewise cubic Hermite interpolation. Based on model prediction with respecting to the kinematic model of tracked vehicles, the Wavefront Value is introduced to improve the robustness of the local planning algorithm in the case of inaccurate obstacle positioning, and combined with the Finite State Machine to implement the autonomous navigation task between multiple floors. Finally, the simulation and real vehicle experiment are performed. The results show that the proposed algorithm can better adapt to the characteristics of narrow indoor space and also prove its feasibility in the actual environment.