A Discrete-Time Extended Kalman Filter Approach Tailored for Multibody Models: State-Input Estimation

Area of Science:

• Mechatronics
• System Dynamics
• Control Engineering

Background:

• Model-based force estimation is gaining traction in mechatronics.
• It leverages high-fidelity models with affordable sensors for efficient analysis.
• Accurate state and input estimation is crucial for mechatronic systems.

Purpose of the Study:

• To present an inverse input load identification methodology.
• To combine high-fidelity multibody models with a Kalman filter for state-input estimation.
• To address the challenge of redundant state descriptions in multibody models.

Main Methods:

• A novel linearization framework for time-discretized equations.
• Utilizing a Kalman filter-based estimator with multibody models.
• Experimental validation on a slider-crank mechanism.

Main Results:

• The framework accurately estimates input torque and system states.
• It demonstrates stability and computational efficiency.
• The method relies solely on readily available angular motor velocity data.

Conclusions:

• The proposed methodology offers an accurate and efficient state-input estimation strategy.
• It successfully handles redundant state descriptions in multibody models.
• Experimental validation confirms the framework's effectiveness for mechatronic systems.