Study And Application Of Safety Monitoring System For Smart Home Based On Wireless Sensor Networks

Using wireless sensor networks, smart home system can implement capturing the location and status of objects in real time, operating the devices remotely and monitoring human's activities. However, it is one of the most difficult problems that how to analyze and fuse these information in order to apply in the smart home system. Aiming to the security problems about persons and objects in smart home, this paper deeply studies the technogies and methods about information capturing, mining and fusion in the applications such as human activities recognition, fire dection and remote monitoring, then presents the key technologies of safty monitoring sytem for smart home based on wireless sensor networks, finally proposes the fire dection method based on Bayesian networks and D-S evidence theory, and the human activities dection model based on Bayesian belief networks. The works of this paper may provide the theory basis and method for the remote safety montoring and controlling in smart home environments. The outstanding contributions and innovations of the paper are as follows.1. Investigated the key technologies such as wireless sensor networks, data fusion and mining, and intelligent control in safety monitoring system for smart home according to its characteristics and requirements. And deeply explored the requirements of these techonologies applied in the safety monitoring system for smart home.2. Proposed a kind of fire dection technology based on wireless sensor networks, and constructed the fire dection method based on Bayesian network and D-S evidence theory, then designed and developed the fire monitoring sytem. This system utilizes the smoking, gas and temperature information captured by the nodes of sensors and judges the fire status using the information fusion technology, by which the fire alarm function can be implemented for smart home.3. Proposed the mining model for human activities modes based on wireless sensor networks that recognizes the human daily behaviors from a lot of status changement data of objects using Bayesian belief network approach. Our research builds the basis for the research of remote activity monining in smart home environments. In order to solve the difficulty in the structute learning of the Bayesian belief network, an edge-encoded genetic algorithm is proposed to improve the correctness and efficiency of the learning algorithm.Finally, this paper discussed the scheme of the safety monitoring system for smart home based on wireless sensor networks and proposed the system architecture and network protocol of the safety monitoring system. The summaries and future research works are presented in the final chapter.