An Uncertainty-Weighted Decision Transformer for Navigation in Dense, Complex Driving Scenarios
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916952228757504 |
|---|---|
| author | Zhang, Zhihao Peng, Chengyang Zhu, Minghao Yurtsever, Ekim Redmill, Keith A. |
| author_facet | Zhang, Zhihao Peng, Chengyang Zhu, Minghao Yurtsever, Ekim Redmill, Keith A. |
| contents | Autonomous driving in dense, dynamic environments requires decision-making systems that can exploit both spatial structure and long-horizon temporal dependencies while remaining robust to uncertainty. This work presents a novel framework that integrates multi-channel bird's-eye-view occupancy grids with transformer-based sequence modeling for tactical driving in complex roundabout scenarios. To address the imbalance between frequent low-risk states and rare safety-critical decisions, we propose the Uncertainty-Weighted Decision Transformer (UWDT). UWDT employs a frozen teacher transformer to estimate per-token predictive entropy, which is then used as a weight in the student model's loss function. This mechanism amplifies learning from uncertain, high-impact states while maintaining stability across common low-risk transitions. Experiments in a roundabout simulator, across varying traffic densities, show that UWDT consistently outperforms other baselines in terms of reward, collision rate, and behavioral stability. The results demonstrate that uncertainty-aware, spatial-temporal transformers can deliver safer and more efficient decision-making for autonomous driving in complex traffic environments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_13132 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | An Uncertainty-Weighted Decision Transformer for Navigation in Dense, Complex Driving Scenarios Zhang, Zhihao Peng, Chengyang Zhu, Minghao Yurtsever, Ekim Redmill, Keith A. Robotics Artificial Intelligence Autonomous driving in dense, dynamic environments requires decision-making systems that can exploit both spatial structure and long-horizon temporal dependencies while remaining robust to uncertainty. This work presents a novel framework that integrates multi-channel bird's-eye-view occupancy grids with transformer-based sequence modeling for tactical driving in complex roundabout scenarios. To address the imbalance between frequent low-risk states and rare safety-critical decisions, we propose the Uncertainty-Weighted Decision Transformer (UWDT). UWDT employs a frozen teacher transformer to estimate per-token predictive entropy, which is then used as a weight in the student model's loss function. This mechanism amplifies learning from uncertain, high-impact states while maintaining stability across common low-risk transitions. Experiments in a roundabout simulator, across varying traffic densities, show that UWDT consistently outperforms other baselines in terms of reward, collision rate, and behavioral stability. The results demonstrate that uncertainty-aware, spatial-temporal transformers can deliver safer and more efficient decision-making for autonomous driving in complex traffic environments. |
| title | An Uncertainty-Weighted Decision Transformer for Navigation in Dense, Complex Driving Scenarios |
| topic | Robotics Artificial Intelligence |
| url | https://arxiv.org/abs/2509.13132 |