Structured Perception Representations for Off-road Dynamics Learning
Autonomously navigating off-road terrain at high speed demands accurate control, which often relies on dynamics models that capture complex interactions between the robot and terrain.
ProofPaper ↗
Key points
- In off-road regimes, terrain interaction is a major component of the vehicle's dynamics.
- Prior work attempts to capture this effect by training learned models of the vehicle's dynamics, conditioned on its perception of the local environment.
- Our approach, SpatialBEV, exploits this spatial structure.
- At each step, we query the bird's-eye-view (BEV) features around the predicted state, conditioning the next prediction on current geometric and semantic cues.
Sources (1)
- [1]Structured Perception Representations for Off-road Dynamics LearningarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 16, 04:02 AM
Autonomously navigating off-road terrain at high speed demands accurate control, which often relies on dynamics models that capture complex interactions between the robot and terrain.
In off-road regimes, terrain interaction is a major component of the vehicle's dynamics.
Extractive summary: sentences quoted from the sources.