A transformer network estimates body-frame velocity and uncertainty from raw 4D radar spectral cubes, fused with IMU in a pose graph to achieve lower relative pose error than classical baselines on indoor sequences.
Learning a depth covariance function,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
UNRIO: Uncertainty-Aware Velocity Learning for Radar-Inertial Odometry
A transformer network estimates body-frame velocity and uncertainty from raw 4D radar spectral cubes, fused with IMU in a pose graph to achieve lower relative pose error than classical baselines on indoor sequences.