Study And Design Of Arduino Based Lightweight Motion Capture System In WBAN

Recent advances in micro-electromechanical systems(MEMS), wireless communication, low-power miniaturized sensors, system-on-chip design have made it possible for a Wireless Body Area Network(WBAN) which can be deployed transparently on human body, whilst a novel motion capture(Mo-cap) technology which is based on MEMS motion sensors rather than traditional optics or electromagnetism is in focus. WBAN is a collection of low-power, miniaturized wireless sensor nodes that support long-term and real-time monitoring of human body’s physiological status and motion state. WBAN can be applied to healthcare, interactive entertainment, sports training, etc. Mo-cap is substantially monitoring of target object’s motion state, and can also be applied to healthcare, interactive entertainment, sports training, etc which has a great similarity with WBAN. A wireless Mo-cap system which adopt MEMS motion sensors to track the motion state of human body can be regarded as a WBAN.Given the features of a WBAN such as cost-effective, low complexity, low-power,lightweight, constrained deployment, etc, this paper proposed an Arduino based lightweight motion capture system(ALMo Cap Sys for short) to realize the long-term monitoring of human motion state in a WBAN which can be used for future human motion analysis. First a system architecture for ALMo Cap Sys was designed based on analysis of the motion data flow in a WBAN. Then a simplified human skeleton tree hierarchy model was raised based on the study of features of human model for Mo-cap in a WBAN. On the basis of the skeleton model, a solution for motion sensor nodes’ deployment and a solution for the nodes’ sequential wireless communication were formed. After an insight into three types of sensor in MARG(Magnetic, Angular Rate, and Gravity) motion sensor array, each type of sensor’s error model was well formed and simplified, then experiment showed the reliability of the simplified error model. A complementary filter based direction cosine matrix(CDCM for short) sensor fusion algorithm was derived after study of three kinds of posture representation methods and two categories of filters, then experiment showed the reliability of CDCM. At last, a human motion data processing platform which contains functions of data parsing, data calibration, animation demonstration, data storage, etc was put on table. By using the platform, a series of intuitional Mo-cap experiments were conducted and verified thereliability of the whole ALMo Cap Sys, which has laid a good foundation for subsequent human motion analysis.