The Researchment Of Collaborative Detection Model And Coverage In Wireless Sensor Networks

The basic task of wireless sensor networks is to detect required informations inappointed regions, so the problems of coverage and detection become the basic problemin the design and application of WSN. Researchers have get lots of achievements incoverage and detection problems. But there are following three inadequacies existed:(1)most of the researches adopts pure computational geometry, unit (the ball) model andtwo value detection model, ignoring the environmental noise and signal attenuation.(2)most of the receiving signals of the sensor nodes use idealized assumption ofindependent and identically distributed, without considering the correlation betweenmeasurements of neighboring sensor nodes.(3) most of the researches just focused onthe two-dimensional design, but practical applications had been extended to3D space.For these problems, this thesis proposes the collaborative sensing model.Considering the noise, this model make a pair of nodes as the basic sensing unit and usedifferent detector complete the fusion and decision to exploit the possible correlation toget better coverage. By the Pearson criterion, for a given false alarm rate, calculate thedetector test statistics and the threshold.By the Monte Carlo simulation get the model’sdistribution of detection-probability and draw it out. Compare the coverage performanceinfluenced by the detectors and distances of the model. Simulation results show that inclose arrangement case, the collaborative sensing model by two nodes has covered moreefficiency than the traditional single point sensing model. Based on this model,Proposed the coverage algorithm.Combined with the existed coverage algorithms based on square and regularhexagon,cover the two-dimensional space by the models. The compare of the volumeratio of polyhedrons can fill the space shows that, the coverage based on TruncatedOctahedron is the most effective one in three-dimensional space. Finally, proposed thecovering algorithm of three-dimensional space based on the model.