Test-Time Trajectory Optimization for Autonomous Driving

• URL: http://arxiv.org/abs/2606.07170v1
• Date: Fri, 05 Jun 2026 11:39:27 GMT
• Title: Test-Time Trajectory Optimization for Autonomous Driving
• Abstract summary: End-to-end planners for autonomous driving typically generate a set of candidate trajectories, score each one, and return the highest-scoring candidate.<n>We treat the scorer as a learned trajectory-level reward function and search for trajectories that maximize it.<n>Our method, TOAD, runs the Cross-Entropy Method at test time, warm-started from the planner's proposals.
• Score: 77.87162309994567
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: End-to-end planners for autonomous driving typically generate a set of candidate trajectories, score each one, and return the highest-scoring candidate. However, the scorer is applied only after the proposals are generated and cannot influence the set of trajectories: a weak set of candidates limits planning performance regardless of the scorer's quality. We instead treat the scorer as a learned trajectory-level reward function and search for trajectories that maximize it. Our method, TOAD, runs the Cross-Entropy Method at test time, warm-started from the planner's proposals. It requires no retraining and is plug-and-play for existing planners. Across six base planners, TOAD improves results on NAVSIM-v1 (94.7 PDMS), NAVSIM-v2 (56.3 EPDMS), and the closed-loop HUGSIM benchmark. The code will be made publicly available via the project page: https://valeoai.github.io/TOAD/.

Related papers

• Slow Brain, Fast Planner: Latency-Resilient VLM-Augmented Urban Navigation [37.49678227133038]
  Learning-based planners for sidewalk navigation can generate diverse candidate trajectories in real time.<n>But their scoring functions often fail to select the best trajectory in challenging situations.<n>We propose a VLM-Planner interface that uses a VLM to select a candidate index from the planner's proposal set and then fuse it with the planner's initial output.
  arXiv  Detail & Related papers  (2026-06-18T16:40:07Z)
• DIRECT: When and Where Should You Allocate Test-Time Compute in Embodied Planners? [57.585275546688116]
  Vision-Language Models (VLMs) are increasingly deployed as high-level planners for embodied agents.<n>We argue that choosing when and where to spend test-time compute is central to bringing frontier performance to the real world.<n>We introduce DIRECT, a routing framework that uses multimodal scene context to allocate compute per prompt.
  arXiv  Detail & Related papers  (2026-06-10T17:58:49Z)
• CLOVER: Closed-Loop Value Estimation and Ranking for End-to-End Autonomous Driving Planning [11.2767413780552]
  CLOVER is a framework for end-to-end autonomous driving planning.<n>A generator produces diverse candidate trajectories, and a scorer predicts planning-metric sub-scores.<n>The generator is refined toward teacher-selected top-$k$ and vector-Pareto targets.
  arXiv  Detail & Related papers  (2026-05-14T17:32:18Z)
• HAD: Combining Hierarchical Diffusion with Metric-Decoupled RL for End-to-End Driving [51.268878540511054]
  We propose HAD, an end-to-end planning framework with a Hierarchical Diffusion Policy.<n>We show that HAD achieves new state-of-the-art performance on both NAVSIM and HUGSIM.
  arXiv  Detail & Related papers  (2026-04-04T04:12:47Z)
• LLaViDA: A Large Language Vision Driving Assistant for Explicit Reasoning and Enhanced Trajectory Planning [28.59507336524504]
  Trajectory planning is a fundamental yet challenging component of autonomous driving.<n>We propose LLaViDA, a Large Language Vision Driving Assistant that leverages a Vision-Language Model (VLM) for object motion prediction.<n>On the NuScenes benchmark, LLaViDA surpasses state-of-the-art end-to-end and other recent VLM/LLM-based baselines in open-loop trajectory planning task.
  arXiv  Detail & Related papers  (2025-12-20T04:38:35Z)
• Autonomous Vehicle Path Planning by Searching With Differentiable Simulation [55.46735086899153]
  Planning allows an agent to safely refine its actions before executing them in the real world.<n>In autonomous driving, this is crucial to avoid collisions and navigate in complex, dense traffic scenarios.<n>Here we propose Differentiable Simulation for Search (DSS), a framework that leverages the differentiable simulator Waymax as both a next state predictor and a critic.
  arXiv  Detail & Related papers  (2025-11-14T07:56:34Z)
• Tree-Guided Diffusion Planner [31.664192839205608]
  Planning with pretrained diffusion emerged as a promising approach for test-time guided control problems.<n>We propose a zero-time test-time planning framework that balances exploration and exploitation through structured trajectory generation.
  arXiv  Detail & Related papers  (2025-08-29T17:27:44Z)
• AlphaZeroES: Direct score maximization outperforms planning loss minimization [61.17702187957206]
  Planning at execution time has been shown to dramatically improve performance for agents in both single-agent and multi-agent settings. A family of approaches to planning at execution time are AlphaZero and its variants, which use Monte Carlo Tree Search together with a neural network that guides the search by predicting state values and action probabilities. We show that, across multiple environments, directly maximizing the episode score outperforms minimizing the planning loss.
  arXiv  Detail & Related papers  (2024-06-12T23:00:59Z)
• Path Planning in a dynamic environment using Spherical Particle Swarm Optimization [0.0]
  A Dynamic Path Planner (DPP) for UAV using the Spherical Vector-based Particle Swarm optimisation technique is proposed in this study. The path is constructed as a set of way-points that stands as re-planning checkpoints. Path length, Safety, Attitude and Path Smoothness are all taken into account upon deciding how an optimal path should be. Four test scenarios are carried out using real digital elevation models. Each test gives different priorities to path length and safety, in order to show how well the SPSO-DPP is capable of generating a safe yet efficient path segments.
  arXiv  Detail & Related papers  (2024-03-19T13:56:34Z)
• Time-Optimal Planning for Quadrotor Waypoint Flight [50.016821506107455]
  Planning time-optimal trajectories at the actuation limit of a quadrotor is an open problem. We propose a solution while exploiting the full quadrotor's actuator potential. We validate our method in real-world flights in one of the world's largest motion-capture systems.
  arXiv  Detail & Related papers  (2021-08-10T09:26:43Z)
• It Is Not the Journey but the Destination: Endpoint Conditioned Trajectory Prediction [59.027152973975575]
  We present Predicted Conditioned Network (PECNet) for flexible human trajectory prediction. PECNet infers distant endpoints to assist in long-range multi-modal trajectory prediction. We show that PECNet improves state-of-the-art performance on the Stanford Drone trajectory prediction benchmark by 20.9% and on the ETH/UCY benchmark by 40.8%.
  arXiv  Detail & Related papers  (2020-04-04T21:27:13Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.