Gesture recognition for wearable motion sensors based on DSP

The gesture recognition algorithms based on acceleration sensor have problems for real-time and algorithms performance,aiming at this problem,a 6-axis AMI602 gesture recognition system of motion sensors combining the acceleration and geomagnetic is proposed: in hardware,a wireless is applied to collect the data and DSP(TMS320F2812) with its strong processing ability to realize the algorithm of gesture recognition,the real time of its system identification is enhanced.In algorithms,the acceleration characteristic values to do gesture projection is put foruard,and the new acceleration characteristic is gotten,dynamic time warping(DTW) algorithm is used to match the templates,which improves the recognition rate of motion.Through the real machine verification in the DSP system,the method in the recognition rate and delay have improved significantly compared to previous methods.