Search papers, labs, and topics across Lattice.
Zhejiang University
5
0
6
A latency-resilient fusion layer transforms slow VLM outputs into real-time trajectory scoring, cutting average distance error by 30% in complex navigation scenarios.
WalkOCC achieves superior sidewalk occupancy prediction by leveraging unpaired monocular images, eliminating the need for costly 3D annotations.
Pedestrian interactions and long-horizon navigation are critical pain points for visual navigation models, with SidewalkBench revealing that scaling training with synthetic data could be the key to overcoming these challenges.
FlowPilot achieves a remarkable 42% success rate in complex sidewalk navigation tasks, showcasing the power of human-preference flow policies in real-world applications.
Autonomous driving policies that ace open-loop tests can still crash and burn in the real world, because they learn to ignore the reactive nature of closed-loop environments.
