An IterIEKF algorithm for quadruped odometry, relying on proprioceptive kinematic constraints, outperforms vanilla IEKF and SO(3) Kalman filters in accuracy and consistency on simulations and real datasets.
Legged robot state estimation using invariant neural-augmented kalman filter with a neural compensator
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
-
Iterated Invariant EKF for Quadruped Robot Odometry
An IterIEKF algorithm for quadruped odometry, relying on proprioceptive kinematic constraints, outperforms vanilla IEKF and SO(3) Kalman filters in accuracy and consistency on simulations and real datasets.