Convex Optimized Average Consensus Weights for Data Aggregation in Wireless Sensor Networks

The energy effectiveness of the executed algorithms is one of the most critical design requirements placed on the wireless sensor networks as the operation of this technology is conditioned by the lifetime of each sensor node. Therefore, many experts from the field have been intensively dealing with this topic from numerous views over the past decades. In this paper, we address efficient data aggregation in wireless sensor networks. More specifically, distributed consensus-based data aggregation with bounded execution is examined in this technology. We focus our attention on the Convex Optimized average consensus weights, which are considered the best-performing consensus-based algorithm in many scenarios, as seen in the literature. In the experimental part, their performance is evaluated for different initial configurations of the implemented stopping criterion in random graphs of various connectivity. We examine their estimation precision and convergence rate and compare their performance with the Best-constant weights (lately identified as the optimal algorithm for the implemented stopping criterion). The goal of this scientific contribution is to verify whether the Convex Optimized weights are applicable in wireless sensor networks, whether they are the best-performing distributed consensus algorithm also in this technology, and identify their optimal initial configuration for the implemented stopping criterion.