Wireless Sensor Network (wsn) Implicit Security Technology Research And Implementation

Due to the small volume and low energy consumption, wireless sensor network(WSN) attracts more researchers’ attention. Anti-Theft is one of the novel applicationsof WSN. There are many inevitable limitations in traditional anti-theft technologies,such as dead angle, easily being found and bad anti-jamming. Aiming to theseshortcomings, Hidden Anti-theft System (HAS for short) is proposed for the first timein this paper. The system detects human motion according to the RSSI changing whichis attenuated by human (Shadowing effect). The HAS framework is established in thispaper and some key technologies in framework are studied. The main contributions areas follows:(1) According to experiments of single communication link, it is found thathuman body absorbs more electromagnetic wave than other common physical media.The change of RSSI is obvious. Anti-theft task is, to detect body motion in monitoredarea. The monitored links are abstracted into edges in undirected graph. That convertsthe task into models. Balanced Edges Partitioned by Vertices Problem (B-EPVP forshort) is formally defined for the first time, to solve balance the agent distribution loadproblem. GD and PD algorithms are proposed to solve B-EPVP, and proved to begreedy.(2) The result of static environment (no body motion) experiment shows thatRSSIs of all links are stable in different stable ranges. In order to obtain all stable RSSIranges and thresholds, K-Mode Set Problem (KMSP for short) is introduced. HASchecks the received RSSI samples: a sample is considered as abnormity if the samplevalue is beyond the stable range of the link. A trade-off abnormity handling algorithmis introduced. The performance can be well accepted by users. The scanning is moreintensive when two channels are working simultaneously. In order to minimize the overlapped area in every time slot in monitored area, and to maximize the scanningarea in every time slot, Unrelated2Vertices Sequence Problem (U2VSP) is formallydefined, and solved by SG algorithm.(3) This paper discusses the construction and communication of multiplemonitored areas, studies the problems of communication collision in adjacent subnet,data query and alarm routing. The system identifies user by communicating with thesuper node which the user holds. HAS is implemented in framework with real motes.The system achieves a good performance. The experiment results show that dual-layerscanning is better than single-layer scanning in response time. No false negative hasbeen found, and the false positive is very low.