基于加速度传感器的跌倒检测与报警系统设计

Design of Fall Detection Technology and Alert System Based on Accelerometer

设计了一种基于MMA7260加速度传感器的跌倒实时检测与报警系统,用以实现老人跌倒后可以被及时发现并救助;系统采用嵌入式ARM Cortex M3内核的STM32微处理器,搭载了GPRS无线通信和GPS卫星定位模块;对跌倒和日常生活运动特征参数进行了研究分析,提出了基于SVM阈值法的三次判别算法;实验结果表明,系统对大部分运动状态都能正确识别,误报、漏报率较低,可有效的区分跌倒和日常生活行为,针对老人测试准确性达到97.5%以上,具有较高的稳定性、可靠性和识别率。

In order to realize that the aged can be descovered and rescued timely while they fall , a real-time fall detection and alert system based on MMA7260 accelerometer was designed. This system was made up of ARM Cortex M3 based MCU STM32, GPRS wireless communication and GPS satellite positioning module. This paper analyzed fall and daily life movement characteristic parameters, and proposed the three discrimination algorithm based on SVM threshold method. Results show system has correct recognition for most motion state and has lower misinformation rate, can distinguish between fall and daily life behavior effectively, according to the aged＇ s test, accuracy up to 97.5% above, so system has high stability, reliability and recognition rate.