A Study Of Optimization Base On Context Knowledge About Data Collection In Smart Environment

With the development of wireless communication technology, embedded technology, distributed computing technology and so on, smart environment is coming into a period of rapid development, and it is widely used in the environmental monitoring, smart home, and medical care and so on. And how to realize fast deployment, dynamic adjustment and real-time efficient collection of the accurate data is directly related to the effects of various applications in smart environment. Therefore, fast deployment and dynamic adjustment and efficient data collection technology research is very important to the application and development of smart environment.At present, in order to solve the deployment problem in smart environment, researchers have proposed many system structure optimization methods, like using service oriented mechanism, structure based on an intelligent gateway, also a relatively good mature technology such as OSGi. In these methods, the main idea is to model the services that smart environment can provide to reduce the degree of coupling among these functions, and then fast restructure these modules to realize the fast deployment and dynamic adjustment of smart environment. However these methods don’t do well in real-time applications or improving the effect of node utilization.While in recent years, much research effort has been devoted to efficient data gathering in WSNs and various types of data gathering mechanisms have been proposed for large-scale sensor networks. Some researchers improve the data collection rate in physical means such as increasing the number of sink node, receiving channel and using mobile sink, and other research works are concentrated on data compression, data backward prediction and distributed data collection to reduce the communication traffic in the WSN. Unfortunately the hardware methods don’t solve the problem essentially, and both kinds of methods have trouble in real-time applications.Our work on the fast deployment and dynamic adjustment of smart environment and improving the data collection effect mainly include the following parts:1) Analyzing the main function of the smart environment system, and then having proposed a smart environment system with smart layer structure based on the realization features of the functions that a smart environment provides.2) Establishing a model of hardware resources of the sensor node, and analyzing and modeling the common services of the functions that the smart environment offers, finally giving a plan about decomposing the user’s requirements to get the node resource allocation scheme and completing task allocation algorithm based on the above scheme.3) Having proposed and designed DPHT(Data Prediction based on HMM according to context aware knowledge), a real-time distributed predicted data collection system to solve the congestion and data loss caused by too many connections to sink node in elder care scenarios.