Rede de área corporal sem fio para rastreamento inercial 3D em tempo real para apoio à reabilitação pós acidente vascular encefálico

Abstract

The use of sensors has become increasingly common and important in our lives. The various destinated uses of theWireless Body Area Network (WBAN) such as body tracking is a proof of that. Body motion Traking by itself has several purposes like video games, clinical diagnosis, rehabilitation, athlete’s training, biomechanical analysis, and others. The body motion tracking can be relevant in situations like post-stroke patients performing rehabilitation due to partial or total loss of body movements. In this case, a WBAN of inertial sensors can be used to perform body motion tracking. Then, the data provided by the sensors can be used in biomechanical and kinematic analysis of the movements executed by the patient. Even though there are many options to perform tracking of human body movements a continuous and real-time tracking of body movements is still a challenge. To overcome some of these limitations, the aim of this project is to propose an inertial sensors wireless network model, which allows the synchronism of a number of sensors. The main use of this network is tracking the movements of the upper limbs during the rehabilitation of post-stroke patients. The network prototype has two types of nodes called central and sensor. The central node is the data concentrator node that controls the other nodes in the network. The sensor node uses inertial sensors (accelerometer and gyroscope) and a magnetic sensor (magnetometer) to capture the movements. The network prototype has the purpose of transmitting information reliably, synchronized, in real-time, with a high data rate speed. For this, microprocessors and radiofrequency transceivers are also used. To validate the model, tests were performed to evaluate the network latency, reliability, and synchronism. The results show the implementation of a hardware and software model of an inertial sensor network capable of tracking the movements of the upper limbs in a synchronous and real-time way. The sensor network was configured in order to make inertial data acquisition at a rate of 100Hz. The network latency value resulted in only 7.4% of the data sampling period. After the visual analysis of the results through graphics generated in real-time by the application on the computer, the reliability of the data and the synchronism between the nodes were also adequate to the main purpose of the work.