Motion-recognition system with dual-processor using threshold-based algorithm

Due to the decline of physical function, the elderly may suffer from falls which are the major cause ofstroke, paralysis or even death. This paper presents a motion-monitoring device based on MTK+STM32 dual-processor. The device uses GPS/WIFI positioning in order to solve the problem of location in case of different conditions.Low-power mode based on acceleration detection can reduce system power consumption.The algorithm of threshold-based Euler angle and maximum resultant acceleration has been tested to push the accuracy up to about 95%. The experiments show that the current consumption is less than 3mA under the low-power mode. The accuracy of GPS can achieve about 7m. The practice has showed that it can be well implemented to the wearable positive health-protective equipment.