Speaker: Peng Xie Day: Wednesday, 12/8/2004 Room: ITE 336 Time: 4:10pm-4:50pm Title: Statistical En-route Filtering and Sensor Networks ---------------------------------------------------------------------------- In the large sensor network, the sensor node is vulnerable to security attack. A compromised node can inject false message into the sensor network (false positives), stop generating the report of real event(false negatives) and drop the message pass through it. The false positive attack not only causes false alarm but also depletes the scarce resource of the sensor network, the power energy. In this paper, we propose a security mechanism, Statistical En-route Filtering (SEF) to solve the false positives. SEF assigns each node a small numbers of keys. When some event occurs, then nodes generates the report and authenticates it by its own keys. When the report is forwarded to the sink, each node along the path can verify the validity of the report probabilistically and drop the forged report. The sink can detect all the forged reports. The performance of SEF is analyzed theoretically and showed that SEF is energy saving and effective. The result from simulation shows that SEF can reduce energy by 50% or more in many cases and drop 80-90% of injected false report. back to CSE-TS page