Optimization on data object compression and replication in wireless multimedia sensor networks

Wireless Multimedia Sensor Networks (WMSNs) have brought unprecedented potentials for applications requiring ubiquitous access to multimedia contents such as still images. However, new challenges have arisen due tothe extra sensor capacity and various requirements of multimedia objects in network processing. In this paper, we consider a large-scale WMSN comprising multiple storage nodes and many multimedia sensor nodes. In particular, we investigate the Optimal Compression and Replication (OCR) of multimedia data objects. In sharp contrast to earlier research, we integrate both computation and communication energy consumption as a joint optimization problem. We prove that the problem is NP-hard if storage nodes have limited storage capacities. We proposed a solution based on Lagrangian relaxation interwoven with the subgradient method. Extensive simulations are conducted to evaluate the performance of the proposed solution.