FSDP: Fast and Safe Data-Driven Overtaking Trajectory Planning for Head-to-Head Autonomous Racing Competitions

2025-03-08

AI Summary

This work addresses real-time, safe, and dynamically feasible trajectory planning for overtaking in resource-constrained autonomous racing. The proposed method features: (1) a sparse Gaussian process-based opponent motion prediction model, significantly improving estimation accuracy and computational efficiency; and (2) a two-layer quadratic programming (QP) optimization framework—where the upper layer generates high-quality initial trajectories via polynomial fitting, and the lower layer formulates a model predictive control (MPC) problem in the Frenet coordinate system to jointly ensure safety, dynamical feasibility, and real-time performance. Experiments on the F1TENTH platform demonstrate an 8.93% improvement in overtaking success rate, robustness to higher opponent velocities, smoother generated trajectories, and a 74.04% reduction in computation time compared to Predictive Spliner.

Technology Category

• Planning, Routing, and Scheduling: Model-Based Reasoning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Generating overtaking trajectories in autonomous racing is a challenging task, as the trajectory must satisfy the vehicle's dynamics and ensure safety and real-time performance running on resource-constrained hardware. This work proposes the Fast and Safe Data-Driven Planner to address this challenge. Sparse Gaussian predictions are introduced to improve both the computational efficiency and accuracy of opponent predictions. Furthermore, the proposed approach employs a bi-level quadratic programming framework to generate an overtaking trajectory leveraging the opponent predictions. The first level uses polynomial fitting to generate a rough trajectory, from which reference states and control inputs are derived for the second level. The second level formulates a model predictive control optimization problem in the Frenet frame, generating a trajectory that satisfies both kinematic feasibility and safety. Experimental results on the F1TENTH platform show that our method outperforms the State-of-the-Art, achieving an 8.93% higher overtaking success rate, allowing the maximum opponent speed, ensuring a smoother ego trajectory, and reducing 74.04% computational time compared to the Predictive Spliner method. The code is available at: https://github.com/ZJU-DDRX/FSDP.

Problem

• Generates safe overtaking trajectories for autonomous racing.
• Improves computational efficiency and accuracy using sparse Gaussian predictions.
• Ensures kinematic feasibility and safety with a bi-level quadratic programming framework.

Innovation

• Sparse Gaussian predictions enhance computational efficiency
• Bi-level quadratic programming for trajectory generation
• Model predictive control in Frenet frame ensures safety