Multi-layered integrated path planning for solving path mismatch problem

This paper proposes an integrated path planner which provides global waypoints from a multi-layered merged map and a local path planner (LPP) using a built-in sensor. The LPPs can avoid obstacles within the range of built-in sensors, but it is impossible to generate an efficient path out of the sensor range. The GPPs can generate an efficient global path with a built-in global map in a target region, but fidelity is very low because of the absence of local sensors. A simple combination of a GPP and a LPP causes a path mismatch problem due to the difference of the information acquired from the local sensor and the information of the preliminary global map. When some wrong global waypoints which are caused by the low fidelity information or the change of the terrain are given to UGVs, the proposed method can escape a standstill of UGVs using the multi-layered map and the adaptive layer transition. When there is no problem to follow global waypoints, the proposed method tries to search global waypoints in a detailed layer map for acquiring more optimal paths as much as possible. The results of simulations show the better performance of the multi-layered integrated path planner by comparing with other algorithms.