Reducing congestion in real-time multi-party-tracking sensor network applications

This paper presents a framework for a congestion reduction mechanism in multiple objects tracking applications. The multiple objects tracking application has a real-time characteristic, through which all packets should be delivered in as timely fashion as possible. The proposed framework consists of sink nodes, backbone nodes, mobile nodes and agent nodes. The location information of the mobile node is delivered to the sink node using both localization and routing techniques, which consider the network congestion. The proposed congestion reduction technique avoids network congestion by decreasing the number of transmitting packets based on the idea that network congestion occurs when the total data in transit is greater than the total channel capacity. The decrement is achieved by both merging and attachment operations, which consolidate several small packets into one large packet in intermediate backbone nodes. The mechanism also considers a time-constrained delivery to assure the user's service request. The simulation is conducted using a TOSSIM simulator, and the result shows that the proposed mechanism improves performance in terms of the successful delivery ratio and the delay time.