Cautious NMPC with Gaussian Process Dynamics for Autonomous Miniature Race Cars

Abstract

This paper presents an adaptive high performance control method for autonomous miniature race cars. Racing dynamics are notoriously hard to model from first principles, which is addressed by means of a cautious nonlinear model predictive control (NMPC) approach that learns to improve its dynamics model from data and safely increases racing performance. The approach makes use of a Gaussian Process (GP) and takes residual model uncertainty into account through a chance constrained formulation. We present a sparse GP approximation with dynamically adjusting inducing inputs, enabling a real-time implementable controller. The formulation is demonstrated in simulations, which show significant improvement with respect to both lap time and constraint satisfaction compared to an NMPC without model learning.

References in corpus (3)

• Optimization-Based Autonomous Racing of 1:43 Scale RC Cars
• Path Tracking of Highly Dynamic Autonomous Vehicle Trajectories via Iterative Learning Control
• Online Constrained Model-based Reinforcement Learning

Cited by in corpus (23)

• Autonomous Vehicles on the Edge: A Survey on Autonomous Vehicle Racing
• Stochastic data-driven model predictive control using Gaussian processes
• Online learning-based Model Predictive Control with Gaussian Process Models and Stability Guarantees
• TUM autonomous motorsport: An autonomous racing software for the Indy Autonomous Challenge
• A data-efficient geometrically inspired polynomial kernel for robot inverse dynamics
• VisuoSpatial Foresight for Multi-Step, Multi-Task Fabric Manipulation
• Gaussian Process-based Stochastic Model Predictive Control for Overtaking in Autonomous Racing
• DDK: A Deep Koopman Approach for Dynamics Modeling and Trajectory Tracking of Autonomous Vehicles
• Learning-based Model Predictive Control for Safe Exploration
• Safe Learning-based Gradient-free Model Predictive Control Based on Cross-entropy Method
• Learning Model Predictive Control with Error Dynamics Regression for Autonomous Racing
• Learning-Based On-Track System Identification for Scaled Autonomous Racing in Under a Minute
• BayesRace: Learning to race autonomously using prior experience
• Hybrid Gaussian Process Modeling Applied to Economic Stochastic Model Predictive Control of Batch Processes
• VisuoSpatial Foresight for Physical Sequential Fabric Manipulation
• Convergence Certificate for Stochastic Derivative-Free Trust-Region Methods based on Gaussian Processes
• ABC-LMPC: Safe Sample-Based Learning MPC for Stochastic Nonlinear Dynamical Systems with Adjustable Boundary Conditions
• Reference design for closed loop system optimization
• Learning-based Event-triggered MPC with Gaussian processes under terminal constraints
• Learning self-triggered controllers with Gaussian processes
• Distributionally Robust Trajectory Optimization Under Uncertain Dynamics via Relative Entropy Trust-Regions
• Prediction with Approximated Gaussian Process Dynamical Models
• Gaussian Process-Based Model Predictive Control for Overtaking