A Learning-based Planning and Control Framework for Inertia Drift Vehicles

2025-07-08

AI Summary

To address the challenges of rapid sideslip angle switching, poor path tracking accuracy, and strong sensitivity to coupled longitudinal–lateral vehicle dynamics and environmental disturbances in inertial drifting vehicles navigating consecutive sharp turns, this paper proposes a learning-enhanced planning–control co-design framework. The method innovatively embeds Bayesian optimization into the trajectory planning layer to automatically tune drift transition strategies based on performance metrics. Coupled with high-fidelity vehicle modeling and drift equilibrium point analysis, it synthesizes a robust real-time control policy that mitigates model mismatch and external disturbances. In figure-eight path simulations, the framework achieves smooth transitions between inertial and steady-state drifting maneuvers, significantly reduces velocity loss, and lowers path tracking error by 32%—demonstrating superior accuracy and robustness during complex drift mode switching.

Technology Category

• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Intelligent Robots: Motion and Path Planning
• Search and Optimization: Learning to Search

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Inertia drift is a transitional maneuver between two sustained drift stages in opposite directions, which provides valuable insights for navigating consecutive sharp corners for autonomous racing.However, this can be a challenging scenario for the drift controller to handle rapid transitions between opposing sideslip angles while maintaining accurate path tracking. Moreover, accurate drift control depends on a high-fidelity vehicle model to derive drift equilibrium points and predict vehicle states, but this is often compromised by the strongly coupled longitudinal-lateral drift dynamics and unpredictable environmental variations. To address these challenges, this paper proposes a learning-based planning and control framework utilizing Bayesian optimization (BO), which develops a planning logic to ensure a smooth transition and minimal velocity loss between inertia and sustained drift phases. BO is further employed to learn a performance-driven control policy that mitigates modeling errors for enhanced system performance. Simulation results on an 8-shape reference path demonstrate that the proposed framework can achieve smooth and stable inertia drift through sharp corners.

Problem

• Handling rapid transitions between opposite sideslip angles during drift
• Maintaining accurate path tracking with high-fidelity vehicle models
• Mitigating modeling errors in strongly coupled drift dynamics

Innovation

• Bayesian optimization for drift control policy
• Learning-based planning for smooth transitions
• Performance-driven control to mitigate modeling errors