Ground Force Precision Calibration Method for Customized Piezoresistance Sensing Flexible Force Measurement Mat

Area of Science:

• Biomechanics
• Sensor Technology
• Machine Learning

Background:

• Force plates are crucial in biomechanics for measuring ground reaction forces during locomotion and static postures.
• Existing force measurement systems face normalization challenges due to hardware and signal processing limitations.
• Accurate weight distribution measurement is essential for various applications, including gait analysis and clinical assessments.

Purpose of the Study:

• To develop a large-area force mat utilizing piezoresistive sensing technology.
• To implement a deep learning-based calibration method for accurate weight measurement.
• To address and overcome the normalization issues inherent in current force measurement hardware and signal processing.

Main Methods:

• Development of a large-area force mat based on piezoresistive sensing principles.
• Application of a deep neural network (DNN) regression model for weight measurement calibration.
• Utilizing test data comprising point-wise weight measurements and center of pressure variables for model training and prediction.

Main Results:

• The developed force mat system achieved high accuracy in weight measurement calibration.
• Average weight errors were found to range from a minimum of 0.06% to a maximum of 3.334%.
• The deep learning approach offers a simpler and more effective calibration compared to traditional resistance-to-weight ratio methods.

Conclusions:

• The novel piezoresistive force mat combined with deep learning provides a robust solution for accurate weight measurement.
• This method effectively normalizes measurements, overcoming limitations of conventional force plate systems.
• The study demonstrates a significant advancement in sensor technology for precise biomechanical data acquisition.