Research map / Label-efficient and robust learning / Recognition and long-tail
Hydra-NeXt: Robust Closed-Loop Driving with Open-Loop Training
Zhenxin Li, Shihao Wang, Shiyi Lan, Zhiding Yu, Zuxuan Wu, Jose M. Alvarez
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add diffusion policy pi_dp instead of discrete control decoder for trajectory refinement (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Driving Score: 60.15 -> 65.89 | Improved |
| run pi_dp at 10 Hz instead of 2 Hz (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Driving Score: 61.95 -> 65.89 | Improved |
| run pi_ctrl at 2 Hz instead of 10 Hz (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Driving Score: 64.66 -> 65.89 | Improved |
| increase proposal number N from 5 to 20 (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Driving Score: 65.52 -> 66.09 | No clear effect |
| increase proposal number N from 5 to 20 (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Success Rate (%): 46.36 -> 46.08 | No clear effect |
| add control decoder ensembled with trajectory (ablation) | Hydra-MDP · Bench2Drive (CARLA v2 protocol) | Success Rate (%): +12.7% | Improved |
| add multi-step control prediction as auxiliary supervision (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Driving Score: +1.8 DS | Improved |
| apply binary focal loss to brake prediction (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Success Rate (%): +1.8% | Improved |
| add post-processing of final control output such as slowing in turns (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Success Rate (%): +2.5% | Improved |
| add nearest neighbor matching to diffusion policy proposals instead of plain ensembling (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Driving Score: around 3 DS -> 65.9 DS | Improved |
| add trajectory guidance to pi_dp candidate selection instead of random selection (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Driving Score: +5.36 DS and +4.28 SR | Improved |
| use DDIM scheduler (20 steps) with flash attention instead of DDPM (100 steps) (ablation) | Hydra-NeXt · Bench2Drive (CARLA v2 protocol) | Driving Score: 65.89 -> 64.87; latency 528.3 -> 243.9 ms | Mixed |
Later papers that compared with this method
4 later accepted papers report Hydra-NeXt as a baseline in their results tables; 4 of them report a better number on most of the results they share with it.
- DriveSuprim: Towards Precise Trajectory Selection for End-to-End Planning (AAAI 2026 · better on 3 of 3 shared results)
- AlignDrive: Aligned Lateral-Longitudinal Planning for End-to-End Autonomous Driving (NeurIPS 2026 · better on 2 of 2 shared results)
- SparseDriveV2: Scoring is All You Need for End-to-End Autonomous Driving (ECCV 2026 · better on 2 of 2 shared results)
- VADv2: End-to-End Autonomous Driving via Probabilistic Planning (ICLR 2026 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- DriveAdapter — DriveAdapter: Breaking the Coupling Barrier of Perception and Planning in End-to-End Autonomous Driving (ICCV 2023) · better on 5 of 5 shared results
- Transfuser — TransFuser: Imitation With Transformer-Based Sensor Fusion for Autonomous Driving (TPAMI 2022) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Bench2Drive: Towards Multi-Ability Benchmarking of Closed-Loop End-To-End Autonomous Driving (NeurIPS 2024 · similarity 0.87)
- RAD: Training an End-to-End Driving Policy via Large-Scale 3DGS-based Reinforcement Learning (NeurIPS 2025 · similarity 0.87)
- NAVSIM: Data-Driven Non-Reactive Autonomous Vehicle Simulation and Benchmarking (NeurIPS 2024 · similarity 0.86)
- Model-Based Policy Adaptation for Closed-Loop End-to-end Autonomous Driving (NeurIPS 2025 · similarity 0.86)
- Diffusion-Based Planning for Autonomous Driving with Flexible Guidance (ICLR 2025 · similarity 0.86)
- DriveHierarchy: A Benchmark for Diagnosing VLM Driving Capabilities from Open-Loop Understanding to Closed-Loop Execution (NeurIPS 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
