Nonplanar Model Predictive Control for Autonomous Vehicles with Recursive Sparse Gaussian Process Dynamics
🤖 AI Summary
This work addresses the degradation in trajectory tracking accuracy of autonomous vehicles operating on non-planar terrain due to complex vehicle dynamics. To this end, the authors propose a geometry-aware recursive sparse Gaussian process (RSGP) model that enables online, adaptive learning of vehicle dynamics in three-dimensional terrain. The RSGP model is seamlessly integrated into a non-planar Model Predictive Path Integral (MPPI) controller, forming an end-to-end real-time control framework. Extensive evaluations in a customized Isaac Sim simulation environment demonstrate that the proposed approach significantly improves trajectory tracking accuracy on challenging unstructured terrains while maintaining strong real-time performance and adaptability.
Application Category
📝 Abstract
This paper proposes a nonplanar model predictive control (MPC) framework for autonomous vehicles operating on nonplanar terrain. To approximate complex vehicle dynamics in such environments, we develop a geometry-aware modeling approach that learns a residual Gaussian Process (GP). By utilizing a recursive sparse GP, the framework enables real-time adaptation to varying terrain geometry. The effectiveness of the learned model is demonstrated in a reference-tracking task using a Model Predictive Path Integral (MPPI) controller. Validation within a custom Isaac Sim environment confirms the framework's capability to maintain high tracking accuracy on challenging 3D surfaces.
Problem
Research questions and friction points this paper is trying to address.
nonplanar terrain
autonomous vehicles
vehicle dynamics
trajectory tracking
model predictive control
Innovation
Methods, ideas, or system contributions that make the work stand out.
nonplanar MPC
recursive sparse Gaussian Process
geometry-aware modeling
autonomous vehicles
MPPI control
🔎 Similar Papers
No similar papers found.
