Research On Compression Awareness In Wireless Sensor Networks

In the 21th century the wireless sensor network (WSN) has been widely used in various fields, with the increase of scale, the processing capacity of the wireless sensor network node and low battery energy are exposed to the public. And the compression sensing (CS) theory can solve the bug of the sensor node. CS is applied to the wireless sensor network, not only can reduce the burden of data collection, but also can make the required less storage. There are different between the CS theory and traditional sampling theory. The CS can get original signal by reconstructing the sparse signals, which don’t have to satisfy the Nyquist sampling theory. Based on compressed sensing theory in the application of WSN as the research object, the main work is as follows.According to the different kind of signals which had been transformed and reconstructed, one dimensional time-domain pulse signal and bridge load harmonic signal are as the original object, and two kinds of signals would be reconstructed by the convex optimization algorithm and the greedy algorithm.And then the acoustic signal as the actual test in the experiment, the Telosb node and electret microphone wireless acoustic signal are made of the acquisition hardware system. At the same time, in the research of acoustic signal reconstruction in this paper, the BP, OMP, StOMP have been used to reconstruct the acoustic signal respectively. In the same base case, analysis of three algorithm accuracy and reconstruction time.At last through a series of simulation and experiment, the minimum error of reconstruction algorithm is BP and the fastest refactoring algorithm is StOMP. In practical engineering applications, the accuracy of signal reconstruction is affected by selecting suitable transformation and reconstruction algorithm.