Information Acquisition Methods In Underwater Sensor Networks Based On Compressed Sensing

With the continuous development and utilization of marine resources andgradual developing of marine economy, underwater sensor network is becominga hot area of research gradually. In the wireless sensor network, particularly inunderwater environment, the wireless network resources such as sensor energyand the network bandwidth is limited severely and of great importance.Meanwhile, all of sensor nodes are deployed at the bottom of ocean or suspendedin the deep area of that, which makes it really difficult to recharge the nodeenergy. Considering the specificity of underwater communication environmentdepicted above, it’s necessary to find out an innovative and effective underwaterinformation collection scheme, which is the core problems.While in recent years,the revolutionary theory of compressed sensing technology, provides aninnovative research breakthrough solution.Just as what compressed sensing theory points out: as long as the originalsignal can be sparsely expressed in a sparse domain, we can measure the originalsignal using a random measurement matrix which doesn’t correlate with thesparse matrix, and finally within the permissible error, we can reconstruct theoriginal data information accurately and precisely utilizing the observed values.Considering the characteristics that most natural signals can be compressed in asparse matrix,thus we do not need to collect the sensory data of each sensor nodein underwater sensor networks. In substitute, the sink node collect the sensorydata packets whose number is far less than that of sensor nodes according tocertain routing rules, furthermore, the sink node uses the values of data observedto reconstruct the original data value of each sensor node.This paper mainly studies the information acquisition scheme based oncompressed sensing theory in the two-dimensional and three-dimensionalunderwater sensor network, and single-hop transmission and multi-hoptransmission networks are discussed as well. The compressed sensing theory iscombined with the clustering protocols of sensor network, the number oftransmitted data packets greatly reduced based on compressed sensing theory andthe balanced energy consumption achieved through the clustering protocols. Asfor the three-dimensional multihop networks, a distributed randomized multi-hopcompressed sensing scheme is proposed. Theoretical analysis and simulationresults show that, compared with the conventional scheme of informationcollection, the proposed scheme possesses great advantages in saving thenetwork energy consumption, prolonging the network lifetime, saving network bandwidth and so on.Specific research work makes some sense in providing significant technicalreferences for the support of the development of large-scale sensor networks, theexploration of marine resources and marine career developing.