Contextual Tuning of Model Predictive Control for Autonomous Racing
This work addresses the challenge of suboptimal behavior in autonomous racing when environmental conditions change, offering a more efficient tuning method for robotics and control systems, though it is incremental as it builds on existing learning-based MPC and Bayesian optimization techniques.
The paper tackles the problem of data-efficient controller tuning for autonomous racing by adapting both the model and objective simultaneously using contextual Bayesian optimization, resulting in optimized lap times across different environmental conditions with fewer data compared to standard methods, as demonstrated in over 3,000 laps on a 1:28 scale RC race car platform.
Learning-based model predictive control has been widely applied in autonomous racing to improve the closed-loop behaviour of vehicles in a data-driven manner. When environmental conditions change, e.g., due to rain, often only the predictive model is adapted, but the controller parameters are kept constant. However, this can lead to suboptimal behaviour. In this paper, we address the problem of data-efficient controller tuning, adapting both the model and objective simultaneously. The key novelty of the proposed approach is that we leverage a learned dynamics model to encode the environmental condition as a so-called context. This insight allows us to employ contextual Bayesian optimization to efficiently transfer knowledge across different environmental conditions. Consequently, we require fewer data to find the optimal controller configuration for each context. The proposed framework is extensively evaluated with more than 3'000 laps driven on an experimental platform with 1:28 scale RC race cars. The results show that our approach successfully optimizes the lap time across different contexts requiring fewer data compared to other approaches based on standard Bayesian optimization.