DriveVer: Lightweight Trajectory Evaluator as Test-Time Verifier for Autonomous Driving
2026-07-01 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors found that current self-driving car models often struggle to improve because training them more gets too expensive and slow. To fix this, they made DriveVer, a small tool that checks and improves planned driving paths while the car is driving, without needing extra training. DriveVer looks at different possible routes and predicts which are safest, making small adjustments if needed. Tests showed that DriveVer helps make driving plans safer and better, adding only a little extra computing work so it still works in real time.
autonomous drivingend-to-end modelstrajectory planningtest-time verificationmulti-view visual representationkinematic featuresNAVSIM benchmarkreal-time inferencedual-head architecturegeometric refinement
Authors
Chong He, Yuechen Luo, Fang Li, Shaoqing Xu, Fuxi Wen
Abstract
End-to-end autonomous driving models often encounter performance bottlenecks, as training-time scaling leads to high computational costs and diminishing marginal returns. Existing planners typically adopt a one-shot generation paradigm, lacking secondary validation and active correction mechanisms to detect and revise suboptimal or unsafe trajectories during inference. To address this issue, we propose DriveVer, a lightweight, plug-and-play Test-Time Verifier that leverages the test-time scaling paradigm to enable autonomous driving systems to validate and refine trajectories without costly and heavy training. We construct a dedicated trajectory dataset based on the NAVSIM benchmark through condition-driven clustering and balanced sampling according to ego-vehicle states and navigation commands. Employing a dual-head architecture, DriveVer efficiently fuses candidate trajectories with multi-view visual representations and ego-vehicle kinematic features to simultaneously predict a safety confidence score and an absolute geometric refinement vector. Extensive experiments on the NAVSIM benchmark show that DriveVer significantly improves the performance of base planning models. Notably, as an extremely compact model with only 34M parameters, DriveVer introduces minimal computational overhead, achieving competitive results while maintaining real-time inference efficiency.