A Real-Time Human Posture Recognition System Using Internet of Things (IoT) Based on LoRa Wireless Network

摘要

Posture recognition technologies based on the Internet of Things (IoT) are widely required to care industry, such as fall detection of elder persons. In order to realize the low-power transformation of motion information in wide-area, Long Range (LoRa) is used in this paper to develop a human posture recognition system. The system is integrated by an mpu-9250 sensor, a LoRa Shield board and an Arduino Mega master control board, which collect human posture data and transmit them to the cloud server remotely. Combined with a random forest algorithm, real-time human posture movement data is carried out to recognize and classify human posture movement. The posture recognizing accuracy calculated by random forest algorithm is the higher than that of other classic machine learning algorithms. This way, our proposed real-time human posture recognition system is able to assist care industry to automatically monitor real-time posture situations of elder persons.