Measuring robots’ real-time velocity correctly is important for locomotion control. Inertial Measurement Unit (IMU) is widely used for velocity measurement. Limited by the bias and random error, IMU alone often can’t meet the requirement. This paper makes use of Extended Kalman Filter (EKF) to fuse kinematics and IMU, and inhibits the drift successfully. We calibrate the bias and recognize the random errors of IMU. Then the forward kinematics of legs is established and the EKF algorithm for velocity estimation is designed based on IMU and kinematics. Finally, the presented algorithm is validated in simulation and on a quadruped robot based on hydraulic driver in trotting gait.
Velocity Estimation for Quadrupeds Based on Extended Kalman Filter
Key Engineering Materials ; 621 ; 525-532
2014-08-11
8 pages
Article (Journal)
Electronic Resource
English
EKF , Velocity Estimation , IMU , Kinematics , Quadruped
VELOCITY ESTIMATION FROM CYCLE PEDALING BY USING EXTENDED KALMAN FILTER ALGORITHM
European Patent Office | 2021
|Extended Kalman Filter vs. Error State Kalman Filter for Aircraft Attitude Estimation
British Library Conference Proceedings | 2011
|LOS Rate Estimation Using Extended Kalman Filter
Springer Verlag | 2019
|Sideslip angle estimation using extended Kalman filter
Automotive engineering | 2008
|