A self-supervised transformer and PointNet matcher for sparse, noisy consumer radar point clouds improves radar-inertial odometry position accuracy by 14 to 19 percent on average in real flights.
A novel radar point cloud gen- eration method for robot environment perception,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Learning Point Correspondences In Radar 3D Point Clouds For Radar-Inertial Odometry
A self-supervised transformer and PointNet matcher for sparse, noisy consumer radar point clouds improves radar-inertial odometry position accuracy by 14 to 19 percent on average in real flights.