| Summary: | 碩士 === 國立臺北大學 === 資訊管理研究所 === 94 === Most applications of WSNs sensors are deployed over unfavorable environment, which could incur probably sensor nodes work or communication improperly. Therefore, the fault estimation model using Bayesian Belief Network (BBN) is proposed and based on it to construct fault map. In this paper, we develop the strategies that assist the existing geographical routing protocols to deal with the fault problem and consider the routing problem based on the fault map via two types of data delivery models: query-driven and event-driven. For query-driven model, a novel algorithm, called relay node selection algorithm (RNSA) is proposed, which selects a few relay nodes as temporary destinations. For event-driven model, we propose two fault-aware forwarding rules (FFRs): per-node and per-path, to control the dependability of routing. Simulation shows that the success rate of data transmission originated from base station can be raised substantially and the hop count is also reduced via the deployment of relay nodes. Also, the dependability of routing can also be controlled effectively by FFRs.
|
|---|
