Energy-efficient Transmission Methods In Marine Monitoring Sensor Networks Based On Compressed Sensing

Oceans are important bases to maintain the survival and progress of human society. Underwater Sensor Networks(USNs), as key facilities, play an important role to explore the ocean world. Concerning the complicated architecture of three-dimensional marine monitoring sensor networks and the adverse environment of underwater acoustic channels, designing energy-efficient data transmission algorithms has become core problems to the research of USNs. The theory of Compressed Sensing(CS), which can achieve signal acquisition with effective compression, provides an innovative research breakthrough solution to the energy-limited USNs.According to CS, original signal can be reconstructed precisely from far few random projections within permissible error, if the signal to be collected is sparse. Due to the temporal-spatial correlation of most natural signals, sink node can use the collected random measurements, whose number is far fewer than original information, to reconstruct original network map based on certain routing rules. Its application can significantly reduce the energy consumption of communication and the requirement of channel bandwidth. This paper focuses on designing energy-efficient data transmission algorithms by studying how CS can be applied in three-dimensional marine monitoring sensor networks.Firstly, the complex underwater acoustic channels environment is analyzed. Based on the attenuation, noise and bandwidth models of underwater acoustic channels, a single-hop transmission energy consumption model is established, which defines the evaluation method of routing algorithms in energy consumption. Furthermore, the performance advantage of CS and multi-hop routing strategy are analyzed based on the model.Secondly, three-dimensional and multi-hop routing algorithms based on distributed CS are designed to extend the network lifetime and three routing schemes are discussed. The routing schemes of them can be effectively combined with the design of observation matrix in CS. Simulation results based on the measured data show that compared with the traditional information transmission scheme, the proposed schemes significantly excel in energy saving and can effectively reconstruct the original data in the meantime. Besides, the performance analysis and comparison of the above algorithms are completed to select the optimal algorithm.Finally, taking CS as technology background, the clustering protocol and inter-cluster multi-hop routing are combined to design a clustering, uneven-layered and multi-hop routing scheme based on CS, CS-CULM. With CS as data fusion technology, the inter-cluster transmission is fulfilled by an improved three-dimensional clustering protocol NEW-LEACH, then the selected optimal algorithm is adopted to forward the fused packets to sink node. Experimental results based on the measured data show that CS-CULM can achieve better performance in balancing energy consumption while ensuring the original data reconstruction precision. Moreover, the problem that original signal in three-dimensional marine monitoring sensor networks is sparse in subsection frequency domain is analyzed and solved to optimize CS-CULM algorithm, it achieves the target of improving the data reconstruction precision.