Robust In-Network Data Processing for Target Tracking in WSNs

摘要

Target tracking is becoming a common and important application of large-scale wireless sensor networks. Tree-based in-network processing for tracking data is an efficient technique to prevent packet overflowing and minimize communication energy consumption. But existing tree-based approaches have a serious problem which is non-robustness under unreliable link conditions in wireless sensor networks. In this paper, we propose a Conditional Multicasting (CM) technique and an h-distance Common Ancestor Tree (h-CAT) algorithm for robust in-network processing for target tracking data via tree topology. CM is a hop by hop data loss inference based multicasting scheme, and h-CAT is a special tree for CM to maximize data recovery with the use of redundancies while minimizing the existence of redundancies within networks. CM and h-CAT provide higher successful data arrival rate than existing tree-based approaches. Moreover, CM and h-CAT employ low extra communication cost and minimal increase of traffic overhead.