Search papers, labs, and topics across Lattice.
DreamFlow, a DRL-based local navigation framework, addresses the limitations of onboard sensing in cluttered environments by extending the robot's perceptual horizon using conditional flow matching (CFM). The CFM module learns a probabilistic mapping between local height map latent representations and broader spatial representations, conditioned on navigation context, allowing the robot to predict unobserved environmental features. Experimental results in simulation and real-world quadrupedal robot deployments demonstrate that DreamFlow outperforms existing methods in latent prediction accuracy and navigation performance.
Robots can now "hallucinate" the world beyond their sensors using conditional flow matching, enabling more robust navigation in cluttered environments.
Local navigation in cluttered environments often suffers from dense obstacles and frequent local minima. Conventional local planners rely on heuristics and are prone to failure, while deep reinforcement learning(DRL)based approaches provide adaptability but are constrained by limited onboard sensing. These limitations lead to navigation failures because the robot cannot perceive structures outside its field of view. In this paper, we propose DreamFlow, a DRL-based local navigation framework that extends the robot's perceptual horizon through conditional flow matching(CFM). The proposed CFM based prediction module learns probabilistic mapping between local height map latent representation and broader spatial representation conditioned on navigation context. This enables the navigation policy to predict unobserved environmental features and proactively avoid potential local minima. Experimental results demonstrate that DreamFlow outperforms existing methods in terms of latent prediction accuracy and navigation performance in simulation. The proposed method was further validated in cluttered real world environments with a quadrupedal robot. The project page is available at https://dreamflow-icra.github.io.
