Game-theoretic solutions through intelligent optimization for efficient resource management in wireless visual sensor networks

• Network resources are allocated among the nodes of a DS-CDMA visual sensor network.
• Two variants of the Nash bargaining solution (n.NBS, c.NBS) are utilized.
• n.NBS (c.NBS) treats each node (class of nodes) as equally advantaged.
• The resulting optimization problems are solved using particle swarm optimization.
• NBS variants outperform schemes that minimize average or maximum video distortion.

We propose a quality-driven cross-layer optimization scheme for wireless direct sequence code division multiple access (DS-CDMA) visual sensor networks (VSNs). The scheme takes into account the fact that different nodes image videos with varying amounts of motion and determines the source coding rate, channel coding rate, and power level for each node under constraints on the available bit rate and power. The objective is to maximize the quality of the video received by the centralized control unit (CCU) from each node. However, since increasing the power level of one node will lead to increased interference with the rest of the nodes, simultaneous maximization of the video qualities of all nodes is not possible. In fact, there are an infinite number of Pareto-optimal solutions. Thus, we propose the use of the Nash bargaining solution (NBS), which pinpoints one of the infinite Pareto-optimal solutions, based on the stipulation that the solution should satisfy four fairness axioms. The NBS results in a mixed-integer optimization problem, which is solved using the particle swarm optimization (PSO) algorithm. The presented experimental results demonstrate the advantages of the NBS compared with alternative optimization criteria.