![[2509.21961] FlowDrive: moderated flow matching with data balancing for trajectory planning](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2509.21961] FlowDrive: moderated flow matching with data balancing for trajectory planning
Summary
The paper presents FlowDrive, a trajectory planning method that addresses data imbalance in driving datasets by using moderated flow matching and data balancing techniques.
Why It Matters
As autonomous driving technology advances, ensuring reliable performance in diverse scenarios is crucial. This research tackles the challenge of data imbalance, which can lead to unsafe driving behaviors by improving the training of trajectory planners. The findings could enhance the safety and efficiency of autonomous systems.
Key Takeaways
- FlowDrive improves trajectory planning by addressing data imbalance in training datasets.
- The method uses moderated flow matching to enhance trajectory diversity while maintaining scene consistency.
- FlowDrive achieves state-of-the-art results on benchmark tests, outperforming existing learning-based planners.
Computer Science > Robotics arXiv:2509.21961 (cs) [Submitted on 26 Sep 2025 (v1), last revised 17 Feb 2026 (this version, v2)] Title:FlowDrive: moderated flow matching with data balancing for trajectory planning Authors:Lingguang Wang, Ömer Şahin Taş, Marlon Steiner, Christoph Stiller View a PDF of the paper titled FlowDrive: moderated flow matching with data balancing for trajectory planning, by Lingguang Wang and 3 other authors View PDF HTML (experimental) Abstract:Learning-based planners are sensitive to the long-tailed distribution of driving data. Common maneuvers dominate datasets, while dangerous or rare scenarios are sparse. This imbalance can bias models toward the frequent cases and degrade performance on critical scenarios. To tackle this problem, we compare balancing strategies for sampling training data and find reweighting by trajectory pattern an effective approach. We then present FlowDrive, a flow-matching trajectory planner that learns a conditional rectified flow to map noise directly to trajectory distributions with few flow-matching steps. We further introduce moderated, in-the-loop guidance that injects small perturbation between flow steps to systematically increase trajectory diversity while remaining scene-consistent. On nuPlan and the interaction-focused interPlan benchmarks, FlowDrive achieves state-of-the-art results among learning-based planners and approaches methods with rule-based refinements. After adding moderated guidance and light post-...
