Cross-layer Optimization For Video Coding In Wireless Sensor Networks

With the widely application of wireless sensor networks, more and more researchers pay attention to the key technologies in wireless sensor networks. Video transmission account for an increasing proportion in wireless sensor networks, and it is different with other traffics because of large amount of data with real-time. Wireless sensor networks must adapt to this development as a ubiquitous technology in the future. So, the primary task is to find an energy efficient algorithm to achieve energy-saving in the network video transmission. A new technology called compressed sensing which integrated sampling and compression appear in recent years. Distributed video coding with compressed sensing has characteristics of simple encoding and complex decoding, then this coding scheme is very suitable for energy-constrained wireless sensor networks.This paper introduced the application of distributed compressive video sensing in wireless sensor networks, analyzed the coding theory and simulation method. Then, we proposed a simple cross-layer optimization scheme to improve reconstruction video quality and reduce network energy consumption. The main contents of this article are as follows:1) describe state of the art in wireless sensor network including the problems to be solved, the basic theory of compressed sensing, and the implementation of distributed compressive video sensing in wireless sensor networks.2) knowledge on the use of the compressed sensing theory, transfer video by distributed compressive video sensing, and joint decoding of compressed data in the decoder to restore the original image. Finally, prove that this encoding method can effectively reduce the amount of stream data.3) analyze the quality of the reconstruction video under different compression sampling rate, quantization parameters and channel quality, then compare the difference between the distributed compressive video sensing and the traditional coding, and the performance on energy-saving and error resilience is analyzed. Finally, propose a cross-layer strategy which adjusting the sampling rate according to the channel conditions is proposed and this strategy can effectively improve the reconstructed video quality and reduce network energy consumption.