Traffic-aware Congestion Control and Rate Adaptation Protocol for Wireless Multimedia Sensor Networks

Abstract

Wireless multimedia sensor networks (WMSNs) are a new breed of low power wireless sensor networks (WSNs) that are aimed at delivering both scalar and multimedia data. WMSNs inherit the miniature size, low power, low processing and short range wireless communication traits from WSNs. Addition of multimedia content in WMSNs demands fulfillment of various QoS parameters like low end-to-end delay, acceptable jitter rate and higher throughput. Over the past few years, considerable research efforts are made to satisfy these QoS requirements by proposing new traffic-aware medium access and routing algorithms. However, the issues of rate adaptation and congestion control are not exhaustively researched. Congestion in any form of wireless networks is a major source of packet drops resulting in decreased network performance. In WMSNs congestion can occur due to sudden transfer of multimedia content requiring high throughout. Loss of packets in WMSNs occurs due to high traffic flow but link failure or mobility can also cause packet drops. It is important to adjust rate of information delivery based on network conditions which can be observed on hop-by-hop and end-to-end basis. In WSNs end-to-end strategies are discouraged because of spatial correction of data in a single flow. But in WMSNs spatial correlation is almost negligible in case of multimedia data. Hence end-to-end mechanisms can also be adopted for both congestion and rate control. In this thesis, we have investigated the issue of rate adaptation and congestion control for WMSNs. A traffic?aware congestion control protocol is proposed that operates on end-to-end principle at the transport layer. The proposed protocol uses burst losses to detect congestion at the destination and directs source nodes to adjust reporting rate accordingly. The proposed protocol is evaluated using simulation analysis which shows significant improvement in terms to achieve throughput.