NavThinker: Action-Conditioned World Models for Coupled Prediction and Planning in Social Navigation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914404330635264 |
|---|---|
| author | Hu, Tianshuai Gong, Zeying Kong, Lingdong Mei, XiaoDong Ding, Yiyi Zeng, Qi Liang, Ao Li, Rong Zhong, Yangyi Liang, Junwei |
| author_facet | Hu, Tianshuai Gong, Zeying Kong, Lingdong Mei, XiaoDong Ding, Yiyi Zeng, Qi Liang, Ao Li, Rong Zhong, Yangyi Liang, Junwei |
| contents | Social navigation requires robots to act safely in dynamic human environments. Effective behavior demands thinking ahead: reasoning about how the scene and pedestrians evolve under different robot actions rather than reacting to current observations alone. This creates a coupled prediction-planning challenge, where robot actions and human motion mutually influence each other. To address this challenge, we propose NavThinker, a future-aware framework that couples an action-conditioned world model with on-policy reinforcement learning. The world model operates in the Depth Anything V2 patch feature space and performs autoregressive prediction of future scene geometry and human motion; multi-head decoders then produce future depth maps and human trajectories, yielding a future-aware state aligned with traversability and interaction risk. Crucially, we train the policy with DD-PPO while injecting world-model think-ahead signals via: (i) action-conditioned future features fused into the current observation embedding and (ii) social reward shaping from predicted human trajectories. Experiments on single- and multi-robot Social-HM3D show state-of-the-art navigation success, with zero-shot transfer to Social-MP3D and real-world deployment on a Unitree Go2, validating generalization and practical applicability. Webpage: https://hutslib.github.io/NavThinker. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_15359 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | NavThinker: Action-Conditioned World Models for Coupled Prediction and Planning in Social Navigation Hu, Tianshuai Gong, Zeying Kong, Lingdong Mei, XiaoDong Ding, Yiyi Zeng, Qi Liang, Ao Li, Rong Zhong, Yangyi Liang, Junwei Robotics Social navigation requires robots to act safely in dynamic human environments. Effective behavior demands thinking ahead: reasoning about how the scene and pedestrians evolve under different robot actions rather than reacting to current observations alone. This creates a coupled prediction-planning challenge, where robot actions and human motion mutually influence each other. To address this challenge, we propose NavThinker, a future-aware framework that couples an action-conditioned world model with on-policy reinforcement learning. The world model operates in the Depth Anything V2 patch feature space and performs autoregressive prediction of future scene geometry and human motion; multi-head decoders then produce future depth maps and human trajectories, yielding a future-aware state aligned with traversability and interaction risk. Crucially, we train the policy with DD-PPO while injecting world-model think-ahead signals via: (i) action-conditioned future features fused into the current observation embedding and (ii) social reward shaping from predicted human trajectories. Experiments on single- and multi-robot Social-HM3D show state-of-the-art navigation success, with zero-shot transfer to Social-MP3D and real-world deployment on a Unitree Go2, validating generalization and practical applicability. Webpage: https://hutslib.github.io/NavThinker. |
| title | NavThinker: Action-Conditioned World Models for Coupled Prediction and Planning in Social Navigation |
| topic | Robotics |
| url | https://arxiv.org/abs/2603.15359 |