A Multi-Sensor Matched Filter Approach to Robust Segmentation of Assisted Gait

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Gait Analysis

Background:

• Assistive devices (ADs) like canes are crucial for individuals with mobility impairments.
• Instrumenting ADs offers non-invasive monitoring of gait and reliance on the device.
• Accurate sensor data processing and gait segmentation are essential for analyzing AD usage.

Purpose of the Study:

• To develop and validate a highly accurate, multi-sensor-based gait segmentation algorithm.
• To create a robust algorithm capable of handling diverse walking conditions and terrains.
• To improve the extraction of relevant gait information from sensor data collected by canes.

Main Methods:

• A novel multi-sensor matched filter (MSMF) algorithm was developed.
• The algorithm combines matched filtering based on loading information with angular rate reversal and peak detection.
• The MSMF algorithm was tested using a hybrid sensorized cane on various terrains with 30 healthy participants.

Main Results:

• The proposed MSMF algorithm demonstrated high accuracy and reliability in segmenting gait events.
• Performance was compared against variations of the gyroscope peak detection (GPD) algorithm.
• The algorithm proved robust across different walking conditions and terrains.

Conclusions:

• The developed multi-sensor gait segmentation algorithm offers a significant advancement for analyzing mobility data.
• This technology can enhance the monitoring of individuals using assistive devices.
• Accurate gait event segmentation is a critical step towards better understanding and supporting mobility impairments.