Description:
This technology employs a waist-mounted sensor and algorithm that can accurately monitor gait and predict falls. The sensor uses a tri-axial accelerometer to classify gait events into stationary, sit-to-stand or stand-to-sit transitional, walking, running, jumping, and stand-to-kneel or kneel-to-stand transition. The eight tracked features include Energy, Entropy, Mean, Variance, Windowed Mean Difference, Variance trend, etc. These are then implemented in a tree type structure to sort different activities of daily living (ADL).
Reference Number: D-0837
Market Applications:
• Physical Therapy/ Medical Monitoring
• Geriatrics
• Exercise Technology
Features, Benefits & Advantages:
Even though different individuals at different speeds and styles performed the activities, it was observed that the system classified different activities with high accuracy, with the overall accuracy of the system being 98%. The system can also be used in a broad market, and may prove to be effective in monitoring elderly people. This algorithm may shed light on the incidents prior to the fall, and help to reduce the chance of falling.
• Early detection of falls
• Tracking of other life activities
Intellectual Property:
Provisional Application # 61946881 was filed on 03/02/2014
Development Stage:
Investigational patient trials have been conducted and in-vitro proof of concept has been completed.
Researchers:
Tim Dallas, Electrical Engineering, Texas Tech University
Publications:
Feature selection and activity recognition system using a single triaxial accelerometer:
IEEE Trans Biomed Eng. 2014 Jun;61(6):1780-6. doi: 10.1109/TBME.2014.2307069. Epub 2014 Mar 28.
Key Words:
Motion sensors, fall detection, accelerometer, medical monitoring