QoS-enabled Video Streaming in Wireless Sensor Networks

Abstract

Recent advances in wireless communications technology and low-power, low-cost CMOS imaging sensors enable a new sensing modality employing ubiquitous distributed video sensing. In many video sensing applications multimedia streams are often required by the base station to provide detailed information of the target spot in the sensor field. However, limited bandwidth, unstable network environment, and data transmission interference prevent the large-scale deployment of such applications without new strategies for energy and data capacity management. In this paper we propose a dynamic path formation algorithm based on our data path throughput estimation model. Coupled with a distributed TDMA packet scheduling scheme, this path formation algorithm can establish throughput-aware video delivery path between source and destination. OPNET simulation results indicate that the throughput estimation is accurate and our proposed TDMA scheme is preferred for streaming applications. We also explore the use of distributed mobile base stations as an option to improve the egress of video data streams from the system.