Online Kinematic and Dynamic-State Estimation for Constrained Multibody Systems Based on IMUs

Area of Science:

• Multibody Dynamics and Control Systems
• Robotics and Mechatronics
• State Estimation and Sensor Fusion

Background:

• Estimating kinematic and dynamic states of mechanisms from noisy measurements presents significant challenges.
• Existing methods struggle with closed-loop, constrained mechanisms due to interdependent state variables.
• Inertial Measurement Units (IMUs) offer a promising, albeit noisy, data source for state estimation.

Purpose of the Study:

• To develop and validate novel online estimation techniques for kinematic and dynamic states of constrained mechanisms.
• To adapt Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF) for real-time state estimation in multibody systems.
• To address limitations of traditional estimators when applied to systems with dependent coordinates.

Main Methods:

• Derivation of mechanism states (position, velocity, acceleration) from IMU signals using multibody kinematics.
• Integration of generic multibody dynamic equations into EKF and UKF variants.
• Application of estimators on manifolds of allowed positions and velocities by estimating independent coordinates.

Main Results:

• Successful online estimation of kinematic and dynamic states for a planar four-bar linkage using IMUs.
• Experimental validation confirms the accuracy and real-time capability of the proposed EKF and UKF approaches.
• The method effectively handles closed-loop, constrained mechanisms, overcoming limitations of previous techniques.

Conclusions:

• The proposed Kalman filtering approach provides an effective solution for real-time online state estimation in complex multibody systems.
• Estimating independent coordinates on constraint manifolds enables robust state tracking for mechanisms previously difficult to model.
• This work advances the field of multibody dynamics by offering a practical and experimentally validated real-time estimation methodology.