ReflexDiffusion: Reflection-Enhanced Trajectory Planning for High-lateral-acceleration Scenarios in Autonomous Driving

2026-01-14

AI Summary

Existing diffusion-based trajectory planners struggle to simultaneously satisfy vehicle dynamics and road geometry constraints under high lateral acceleration scenarios due to imbalanced training data, often yielding unsafe trajectories. This work proposes ReflexDiffusion, a gradient-driven reflection enhancement mechanism integrated during the inference phase of diffusion models. By computing gradients between conditional and unconditional noise predictions at each denoising step, the method dynamically amplifies critical constraint signals—such as road curvature and lateral dynamics—to reinforce physical feasibility without requiring retraining. The approach is architecture-agnostic and can be directly incorporated into existing diffusion planners. Evaluated on the nuPlan Test14-hard benchmark, ReflexDiffusion improves driving scores by 14.1% over the current state-of-the-art in high lateral acceleration scenarios, significantly enhancing trajectory safety and stability under extreme conditions.

Technology Category

• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Computer Vision: Diffusion Models for Vision
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Generating safe and reliable trajectories for autonomous vehicles in long-tail scenarios remains a significant challenge, particularly for high-lateral-acceleration maneuvers such as sharp turns, which represent critical safety situations. Existing trajectory planners exhibit systematic failures in these scenarios due to data imbalance. This results in insufficient modelling of vehicle dynamics, road geometry, and environmental constraints in high-risk situations, leading to suboptimal or unsafe trajectory prediction when vehicles operate near their physical limits. In this paper, we introduce ReflexDiffusion, a novel inference-stage framework that enhances diffusion-based trajectory planners through reflective adjustment. Our method introduces a gradient-based adjustment mechanism during the iterative denoising process: after each standard trajectory update, we compute the gradient between the conditional and unconditional noise predictions to explicitly amplify critical conditioning signals, including road curvature and lateral vehicle dynamics. This amplification enforces strict adherence to physical constraints, particularly improving stability during high-lateral-acceleration maneuvers where precise vehicle-road interaction is paramount. Evaluated on the nuPlan Test14-hard benchmark, ReflexDiffusion achieves a 14.1% improvement in driving score for high-lateral-acceleration scenarios over the state-of-the-art (SOTA) methods. This demonstrates that inference-time trajectory optimization can effectively compensate for training data sparsity by dynamically reinforcing safety-critical constraints near handling limits. The framework's architecture-agnostic design enables direct deployment to existing diffusion-based planners, offering a practical solution for improving autonomous vehicle safety in challenging driving conditions.

Problem

• autonomous driving
• trajectory planning
• high-lateral-acceleration
• long-tail scenarios
• vehicle dynamics

Innovation

• ReflexDiffusion
• trajectory planning
• diffusion models
• high-lateral-acceleration
• inference-time optimization