Design Of Cell Phone Based House-Hold Surveillance System In Ambient Intelligence Environment

Ambient intelligence (AmI) is a concept that humans are surrounded by computing and networking technology unobtrusively embedded in their surroundings. It was first developed by the ISTAG advisory group to the European in 2001. It came up with a brand new interaction method between humans and computers and focused on mobile computing, embedded computing and ubiquitous computing. When the researches of AmI are deployed faster and faster, Beijing University of Technology Embedded System and Software Institute undertook the National 211 Project"Ambient Intelligence and Embedded System". The"Cell Phone based House-hold Video Surveillance System"described in the dissertation is a subsystem of the"Ambient Intelligence and Embedded System".According to AmI concept, principle and technical, we designed a mobile video surveillance system for house-hold and office oriented application. It takes intelligent cell phones as surveillance terminals, takes GPRS link as network carrier, and the surveillance server is the manager of IP cameras and is also the bridge which connect cell phones and the other intelligent systems within AmI environment. The system represents both"view"and"control"concept in surveillance application. It not only provides users the surveillance images but also enables them to get other intelligent services in AmI environment. As well as resolving surveillance application's cost, flexibility, expandability problems in house-hold usage, the significance of the system is also create a new method by which users can access local area AmI services from wide area environment by a cell phone.In the system we built a cell phones and cameras'many-to-many surveillance structure, and designed special mechanism about cameras management and surveillance requisition management. After researching GPRS network in depth, we designed the pure characters based communication protocol which was fit for GPRS network with limited bandwidth and asymmetry upload and download channels. For the surveillance images, we designed the Frame Indexing Mechanism and the Variable Step S-Frame Selecting Mechanism to optimize the performance and data consistency, and to fit in with the unstable GPRS network. In addition, system designed many other mechanisms to optimize the viewing performance considering the features of cell phone devices, such as smooth effect of PTZ controlling and the images'auto adaptive scaling. Finally, we proved the stability, error tolerance and the efficiency under stress, and then we pointed out the remaining problems and the possible solutions by analysis.