Physical Therapy Training & Progress Remote Monitoring

INTRODUCTION

What Problem are you going to solve?

Physical therapy, beginning with gentle, stretching exercises, often begins soon after surgery on any part of arms or legs. Practice with the artificial limb may begin as soon as 10 to 14 days after surgery.

The determination of how well the patient meeting the recovery goal is subjective via the observation done by the care provider, a person. The length of the recovery process can be long or short without a scientific set of data. This can lead to a poor decision to discharge the patient too soon or too late that can impact the patient long-term activities and medical cost. In addition, during the COVID-19 pandemic, with the limitation of in-person visit, the traditional physical therapy can be impacted greatly.

What are you going to build to solve this problem? How is it different from the existing solutions? Why is it helpful?

With an AI/ML implemented system, the patient's Arms/Legs training and progress remote monitoring solution during rehabilitation will be a great medical tool to assist both patients and health care providers. It will let patient reach the recovery condition better and with confident based on the collective measured data. The data can also be kept as the record and stored in the hospital/clinic database for future use when needed.

So far, I haven't seen a remote solution to help the physical therapy at home or at the clinic yet. It'll be much more useful in this COVID-19 pandemic difficult time.

How does the solution work? What are the main features? Please specify how you use the Microchip development-IoT-Kit in your solution?

To support the demo of this solution within both of the time budget and self-funding budget to meet the submission timeline, a system including one or couple nodes of Microchip remote sensor units and a demo software will be built. This project will demo the walking therapy for the case of post leg/knee surgery.

How it works:

The Microchip sensor nodes will be attached to near upper ankles of left and/or right legs via straps. Patient can walk on a treadmill or hallway. The data from sensors will be streamed to the monitoring application wirelessly via Bluetooth connection. The AI analysis application will analyze and classify the signal data using a demo physical therapy ML algorithm to determine the recovery status.

The features:

This application will show the progress of the therapy, the wellness of the leg being practiced and the estimation of recovery percentage referred to the normal walking person. In practical, the data will also be sent to the therapist office for remote reviewing and report generation. Therefore, the physical therapy sessions can be done in any room of a clinic or at patient home without a need of in-person care at some point in the progress.

The use of the Microchip IoT Dev Kit and additional sensor board:

The Microchip IoT Dev Kit will help to collect sample data from one or more of the available sensors (i.e. temperature or/and light) during the evaluation of the developing application using Microchip API and tools for the IoT kit. The test data will be used to test and to build a test model using Google Cloud services. Any issues arose via Microchip API and SDK during this evaluation, I'll work with Microchip technical support to resolve and to move forward.

Upon a successful initial build process, additional Microchip activity sensor node with motion measurement capabilities needs to be purchased to complete this project. The full name of the hardware is Microchip’s ULP Connected Wearable Activity Monitor Demonstration board. A specific demo physical therapy data analysis and ML algorithm will be built base on the data from these nodes. A leg wellness classification will also be built for the project. The classification can be done locally as an IoT edge device or via internet through Google Cloud services. In this demo, a smart phone is used to communicated to the activity sensor module to collect the therapy training data. The data files can be saved at a local memory or remote network / cloud storages. For the recorded data processing, a laptop is used to run the project data analysis application to determine the therapy session results and to create chart. In practical, this would be the best process flow for the physical therapy remote monitoring used for the rehabilitation.

Below is my demo video to describe the make of this project with working results.