Sensor Data Transmission And Processing Of Inertia Motion Capture System

Human motion capture system is widely used in the field of human-computer interaction, motion analysis, model analysis, virtual reality, animation, intelligent monitoring system, and game production. Generally, human motion capture system makes up of motion parameter raw data capture and transmission system, motion parameter analysis system and human posture matching system. Human motion capture system in this thesis bases on MEMS inertial sensors and wireless sensor networks. Human motion parameter raw data capture system is the basic part of the whole system. This thesis proposes a method called Pipelined Data Collection, which makes full use of limited data transmission speed in WSN and achieve as high as possible data refresh rate. Now the system whose performance transcends the similar commercial products, works at 120Hz data refresh rate, while the average lost rate of data package is less than 1% and the data delay is no more than 40ms. Motion parameter analysis system is also very important for a human motion capture system. On research of the feature of acceleration, magnetic and gyroscope data, this thesis proposes some data preprocessing methods which are suitable to the actual system. Finally, this thesis also proposes an efficient orientation algorithm called Compensation Fusion algorithm. Comparing with the Kalman filter algorithm, it achieves similar accuracy (static accuracy 1°RMS and dynamic accuracy 2°RMS) but needs less computation. So it is suitable for Sensor nodes who has limited computing resources.