Classification of Standing and Walking States Using Ground Reaction Forces

Area of Science:

• Robotics
• Biomechanics
• Wearable Technology

Background:

• Wearable robots, including gait rehabilitation robots, necessitate precise real-time wearer state classification (standing vs. walking).
• Existing methods using ground reaction force (GRF) can be prone to errors, particularly with shuffling movements.

Purpose of the Study:

• To develop and validate a novel technique for real-time classification of standing and walking states using an insole device.
• To improve the accuracy of state classification for wearable robot control and activity monitoring.

Main Methods:

• An insole device equipped with force sensing resistors was used to measure GRF.
• A new technique utilizing waveform length of the center of pressure changes within a time window was developed.
• The technique was applied to a conventional threshold method and an artificial neural network (ANN) model for classification.

Main Results:

• The developed waveform length technique significantly reduced classification errors compared to conventional GRF methods.
• Real-time classification of standing and walking states was successfully achieved using the ANN model.
• The insole device demonstrated effectiveness in distinguishing between standing and walking states, even with shuffling movements.

Conclusions:

• The novel insole-based GRF measurement and waveform length analysis technique enables accurate real-time classification of wearer states.
• This technology is suitable for enhancing wearable robot operations, such as gait rehabilitation.
• The insole device offers potential for remote monitoring of daily living activities.