Study On Algorithm And Routing For Fall Detection In Wireless Sensor Networks

Making full use of advanced technical to achieve the goal of real-time monitoring of falls among elders and fast reporting to monitoring center is very important in modern aging society. Because of the merits of wireless sensor network like real-time monitoring, self-organization and wide coverage area, fall detection based on wireless sensor network attracts widespread attention.Fall detection itself has some high demands for fall recognition algorithm, it requires distinguishing fall events through large amount of activities of daily life in a rapid and accuracy way. Furthermore, report the fall events to monitoring center through the resource-constrained wireless sensor network after the fall occurs is another important issue. In this paper, we focused on the topics on fall recognition and event report routing algorithm, hoping to enhance the fall detection true positive rate and prolong the network lifetime. The major works are as follows:(1) Relative theories about fall detection and wireless sensor network routing algorithms are first reviewed. A fall data acquisition model based on tri-axis accelerometer has been proposed, and mobile sink node has been firstly introduced in wireless fall detection sensor network. We innovatively introduced traditional wireless sensor network into detecting of fall events.(2) In order to improve the recognition rate for fall events and reduce the false alarms for activities of daily life, we introduced a multi-sensors based post fall detection algorithm using heart rate sensor attached in the system. Simulation results show that our multi-sensor system has better performance than former single accelerometer system. As a result, it raised true positive rate, while lowered false negative rate.(3) In order to alleviate the formation of hot spot, sink mobility technology is introduced into wireless sensor network. An energy-efficient unequal clustering algorithm adopting mobile sink node called MSSE(Mobile Sink based Stable Election) is proposed. Simulation results show that MSSE algorithm outperforms LEACH and the usage of mobile sink node significantly improves the performance of the sensor network.