Smart Brace for Monitoring Patients with Scoliosis using a Multimodal Sensor Board Solution

Abstract

The aim of this study is to develop a platform to monitor compliance with brace treatment in patients with scoliosis. Scoliosis is a curvature of the spine that frequently occurs in adolescents. Nonoperative treatment with a thoracolumbosacral orthosis (TLSO) is widely used. However, a brace that is not worn correctly is not effective at controlling scoliosis, regardless of the duration of brace wear. As a solution for monitoring these patients, we developed a low power multi-modal sensor board capable of: 1) logging pressure distribution inside the brace using analog pressure sensors and 2) detecting different activities that the patient is involved in using accelerometer sensor. We employ the two modalities of signals recorded from the brace to achieve high precision compliance monitoring system. Our data processing algorithm suite includes a two-stage data classification design. In the first stage, we detect six predefined activities including: standing, sitting, walking, running, lying down, and climbing the stairs using an embedded motion sensor. In the second stage, we detect four levels of brace tightness based on features extracted from internal force sensors and activity specific models. Our results demonstrated high levels of accuracy for activity and tightness level classification.