Energy-Optimal Ground Vehicle Trajectory Planning on Deformable Terrains
Nathan Goulet, Beshah Ayalew
Abstract
Automated ground vehicles can reduce downtime and increase safety in a wide variety of applications that occur in energy resource constrained unstructured environments (e.g. military, agricultural, mining, and disaster response). In these applications, it is imperative to plan trajectories that minimize energy consumption. To this end, we propose a framework where an energy cost-to-go map of the environment is coupled with a model predictive controller that uses higher fidelity models to capture important aspects of off-road energy consumption, including terramechanics, elevation change, and associated vehicle dynamics. The resulting control framework is compared against common tracking control formulations and a reduction in energy consumption by more than 8% is observed for the evaluated scenarios. Furthermore, the computational challenges of the approach are discussed and some approximations are offered.