Online Attitude Solution And Application Research Of Motion Sensor Signal Based On DSP

Attitude measurement system is an inertial navigation system to measure attitude anglesof the carrier using inertial devices, and it has a wide range of attitude applications frommilitary navigation, geography exploration, oil drilling platforms to game equipment andembedded devices. Aiming at the defects of the previous numerals inertial devices andwith large volume, lack of the ability to eliminate coordinate errors and poor real-time ofattitude measurement system, this thesis uses DSP technology to collect, store, calculate,compensate, correct and posture analyze the action signals which are collected from theintegrated acceleration sensor and magnetic sensor. This system takes advantage of DSP’sstrong ability of data calculation and high real-time characteristics, to preliminary achievecarrier’s attitude online measurement and analysis, to apply the attitude angles to gesturerecognition and to realize online gesture recognition.Firstly, DSP receives the motion sensor’s signal by using the extended asynchronousserial circuit,configures the DSP’s external memory interface (EMIFA) and.cmd file ofCCS software, store the signal on SDRAM;Secondly, Process the format of the received signal and convert the data, and thencomplete the filtering process for the data to make the data more smooth and stable. Thencalculate the attitude angles on the DSP system, compensate and correct attitude anglesonline;Thirdly, Research on online gesture recognition on the DSP system. Use the dynamictime warping algorithm for gesture recognition, and by using attitude projection,concentrate the matched distance of distorted on the a-axis component to reduce theinfluencing factors on algorithm matched, improve the gesture recognition rate. Inaddition, adopts FLASH storage technology to realize fast access for the template data toimprove the running speed of the algorithm recognition.The last part covers attitude tests and gesture recognition experiments. The experimentsdisplay the attitude angles data calculated and corrected on the DSP system. The resultsshow that the measurement accuracy of attitude angles is rather high and the data are stable; The results of online gesture recognition on the DSP are displayed by VC++interface, and the results indicate that the recognition speed is faster and the accuracy israther higher, by contrasting the data of the recognition effects before and after attitudeprojection, it shows that gesture recognition is significantly improved compared to theprevious methods.