An iterative neural network approach applied to human-induced force reconstruction using a non-linear electrodynamic...

Area of Science:

• Biomechanics
• Sport Engineering
• Structural Engineering

Background:

• Human-induced force analysis is crucial in various fields like biomechanics and structural engineering.
• Replicating ground reaction forces (GRF) provides insights into human movement, athletic performance, injury rehabilitation, and structural vibrations.

Purpose of the Study:

• To present an experimental method for accurately replicating human-generated GRF using an electrodynamic shaker.
• To address the challenge of signal reproduction in nonlinear and non-invertible systems.

Main Methods:

• An iterative neural network and an inversion-free approach were developed to optimize the drive signal for the shaker.
• The system minimizes the error between the shaker's output force and the reference GRF signal.

Main Results:

• The neural network successfully updated shaker parameters to achieve desired force replication.
• Excellent results were obtained in both time and frequency domains, demonstrating high-fidelity signal reproduction.

Conclusions:

• The proposed iterative neural network approach effectively replicates human-induced GRF with high accuracy.
• This method offers a valuable tool for biomechanical research, performance optimization, injury assessment, and engineering applications.