Closed-Loop Refinement and Execution for Learned Driving Planners

2026-09-30

AI Summary

This study addresses the issue of learned driving planners causing vehicle stagnation, conflicts, or abrupt braking during closed-loop execution due to accumulated errors. To mitigate this, we propose CLRE, a hierarchical control framework that enhances robustness without retraining the upstream planner by optimizing reference trajectories via receding-horizon control and filtering feasible candidates. Furthermore, it introduces a prediction-conditioned oriented bounding box (OBB) feasibility test alongside a saturated braking law to effectively guarantee execution safety. Evaluated on the Bench2Drive benchmark, this plug-and-play approach elevates the driving score to 56.42 and achieves a route completion rate of 72.23%, while significantly reducing collision events. Overall, the proposed method substantially improves the closed-loop performance of frozen planners.

Abstract

Learning-based driving planners are usually trained and evaluated in open loop against logged trajectories. In closed loop, a trajectory with small displacement error can still stall the vehicle, steer it into a conflict with surrounding agents, or be executed with abrupt braking. We introduce Closed-Loop Refinement and Execution (CLRE), a hierarchical receding-horizon control framework designed to mitigate these failure modes while leaving the upstream planner frozen and adding no new learned model. The upper layer treats the nominal trajectory as a reference and solves a finite-horizon optimal control problem that trades route progress against interaction with predicted agents. Solving it from several initializations gives a candidate set, and a prediction-conditioned oriented-bounding-box (OBB) feasibility test retains only candidates whose minimum predicted OBB clearance over the horizon meets a threshold. The lower layer executes the lowest-cost survivor, or a route-centerline backup when none remains, through the tracking controller supplied with the planner, augmented by a range-based speed bound and a saturated proportional braking law. In closed-loop simulation on 126 Bench2Drive routes with VAD as the upstream planner, CLRE raises the driving score from 43.41 to 56.42 and route completion from 57.27 to 72.23, and reduces collision events from 70 to 53.

Problem

• closed-loop driving
• learned driving planners
• trajectory execution safety
• autonomous driving

Innovation

• Closed-Loop Refinement
• Receding-Horizon Control
• Oriented-Bounding-Box Feasibility
• Hierarchical Framework
• Driving Planner