# A Unified Candidate Set with Scene-Adaptive Refinement via Diffusion for End-to-End Autonomous Driving

Source: [A Unified Candidate Set with Scene-Adaptive Refinement via Diffusion for End-to-End Autonomous Driving](<https://www.sota2.com/research/paper/6982d294a8b01e76cc8e9378>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-06T05:00:04.454Z

Authors: Zhengfei Wu, Shuaixi Pan, Shuohan Chen, Shuo Yang, Yanjun Huang

Published: 2026-02-03T05:14:08.000Z

## Abstract

End-to-end autonomous driving is increasingly adopting a multimodal planning paradigm that generates multiple trajectory candidates and selects the final plan, making candidate-set design critical. A fixed trajectory vocabulary provides stable coverage in routine driving but often misses optimal solutions in complex interactions, while scene-adaptive refinement can cause over-correction in simple scenarios by unnecessarily perturbing already strong vocabulary trajectories.We propose CdDrive, which preserves the original vocabulary candidates and augments them with scene-adaptive candidates generated by vocabulary-conditioned diffusion denoising. Both candidate types are jointly scored by a shared selection module, enabling reliable performance across routine and highly interactive scenarios. We further introduce HATNA (Horizon-Aware Trajectory Noise Adapter) to improve the smoothness and geometric continuity of diffusion candidates via temporal smoothing and horizon-aware noise modulation. Experiments on NAVSIM v1 and NAVSIM v2 demonstrate leading performance, and ablations verify the contribution of each component. Code: https://github.com/WWW-TJ/CdDrive.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [End-to-End Autonomous Driving on NAVSIM v1](<https://www.sota2.com/research/sota/end-to-end-autonomous-driving-on-navsim-v1.md>): End-to-end Autonomous Driving; NAVSIM v1; NC: 0.987.

- [Autonomous Driving on NAVSIM v2](<https://www.sota2.com/research/sota/autonomous-driving-on-navsim-v2.md>): Autonomous Driving; NAVSIM v2; NC: 98.3.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2602.03112>)

[PDF](<https://arxiv.org/pdf/2602.03112.pdf>)
