Model Predictive Path Integral Control for Articulated Construction Vehicle on Irregular Sloped Terrain

An autonomous construction vehicle (ACV) requires a high level of safety because of irregular terrain and complex site conditions. In particular, uneven sloped terrain increases the risk of vehicle rollover, making autonomous control that considers terrain geometry essential. This study proposes a model predictive path integral (MPPI)-based control and path planning method to reduce rollover risk and support safe navigation toward the target in environments containing irregular slopes and obstacles. The proposed method is developed for an articulated construction vehicle and uses a geometric approach to estimate the vehicle roll angle from terrain normal vectors. The estimated roll angle is incorporated into the MPPI cost function to evaluate candidate trajectories for vehicle stability. Simulation results demonstrate that the proposed method reduces rollover risk by avoiding high-risk sloped terrain and generates stable and safe trajectories compared with conventional MPPI. These results support the effectiveness of the proposed method for improving the safety of ACVs in unstructured construction environments.