Reliability-Based Mission Planning of Off-road Ground Vehicle under Uncertain Environment

View Video Presentation: https://doi.org/10.2514/6.2021-0677.vid Mission planning plays an essential role in enabling off-road autonomous ground vehicles (AGVs) to fulfill the assigned tasks with less time or economic cost. However, the presence of heterogeneous uncertainty sources in terrain conditions leads to the variability of AGV mobility prediction, which poses new challenges in estimation of mission reliability and achievement of good trade-off between cost and reliability during mission planning. This paper proposes a reliability-based path planning model based on two reliability metrics, namely state mobility reliability and mission mobility reliability. The spatial-dependent uncertainty in terrain conditions is first propagated to the vehicle mobility prediction, which is then used to generate the stochastic mobility map. Based on the generated stochastic mobility map, the vehicle mobility reliability can be analyzed according to the reliability metrics. Building upon the reliability-based mission planning model, a reliable and robust rapidly-exploring random tree (R2-RRT*) algorithm is proposed to incrementally search for the final path which is robust to uncertain environment and meets reliability target. Even though the developed approach is applied to off-ground AGVs, it can also be applied to mission planning of aircraft in uncertain environment. A case study is used to demonstrate the efficacy of the proposed R2-RRT* algorithm.