Inertial Sensor-Based Instrumented Cane for Real-Time Walking Cane Kinematics Estimation

Area of Science:

• Biomedical Engineering
• Gerontology
• Rehabilitation Technology

Background:

• Falls are a leading cause of injury in the elderly.
• Impaired balance and mobility are key indicators of fall risk.
• Existing fall risk assessment methods often lack real-time feedback capabilities.

Purpose of the Study:

• To develop a real-time fall risk feedback system.
• To utilize an inertial sensor-based instrumented cane for data acquisition.
• To estimate cane kinematics, specifically contact phase and orientation.

Main Methods:

• Developed a convolutional neural network for estimating the cane's contact phase.
• Compared and validated various algorithms for cane orientation estimation using optical motion capture.
• Integrated sensor data processing onto a single-board computer for real-time operation.

Main Results:

• The proposed convolutional neural network model for cane contact phase prediction demonstrated superior accuracy compared to previous models.
• The Madgwick filter was identified as the most effective algorithm for cane orientation estimation.
• The system successfully achieved real-time estimation of both cane contact phase and orientation.

Conclusions:

• The developed instrumented cane system provides accurate, real-time estimation of key cane kinematics.
• This technology has the potential to enhance fall risk assessment and prevention strategies for elderly individuals.
• Real-time feedback from the system can aid in timely interventions to mitigate fall risks.