Wavelet-based Compression Of The Data Collection Algorithm. Wsns

Structural Health Monitoring (Structural Health Monitoring, referred to as SHM) isone of the most representative application of wireless sensor networks, take bridgestructural health monitoring system for example, such applications often require toarrange a large number of sensor nodes in high-density to cover the entire bridge,furthermore, in order to obtain real-time monitoring information, sensor nodes need tocollect the sensor data frequently, this process consume a lot of node energy. It isdifficulty to maintain a long-time working for the sensor nodes whose energy is finite atthis situation, especially when the data to collect is multi-dimensional, this problemgoes worse. The theme of the research in this paper, is how to collect the sensor datamore energy effectively, and the contribution as follows:1. Compressive data collection algorithm based on wavelet-segment constant(W-SDC) is proposed, the algorithm gets lower quantity of messages transmissionduring data collection, it can greatly reduce communication energy consumption inlarge-scale network and prolong the life cycle of the Wsns. Traditional data collectionmethods achieve the total messages transmission at O(N~2 ), compressed samplingmethod is O(MN ), this algorithm will reduce the level of the total transmission to O(mN ), where . Meanwhile, the TOSSIM simulation experiments analysisshow that, W-SDC algorithm reduce the total messages transmission effectively, and thereconstruction error is acceptable.2. Conventional Haar wavelet compression techniques, by filtering through thethreshold, will ignores the wavelet coefficients which are not that important, butcommonly in this part of the discarded wavelet coefficients, there are still somecoefficients that affect the wavelet reconstruction greatly should be retained. Therefore,in order to reduce the reconstruction error of wavelet compression, this paper presents the probability wavelet algorithm based on affect elements—AEWavelet. According tothe amount of influence on leaf nodes of wavelet coefficient in the wavelet errortree, the affect elements AE of the wavelet coefficients are calculated. The algorithmconstruct the random rounding probability of wavelet coefficients based on affectelements, and then randomly choose which wavelet coefficients to retain with thecalculated probability. Through theoretical analysis, it proofs that the algorithm iseffective and correct in the selection of wavelet coefficients. Meanwhile, the experimentanalysis shows: the idea of random rounding, greatly reducing the Haar waveletreconstruction error and improve the accuracy of the wavelet reconstruction.