Enhanced scheme for adaptive multimedia delivery over wireless video sensor networks

Abstract

Lately Video Sensor Networks (VSN) are increasingly being used in the context of smart cities, smart homes, for environment monitoring, surveillance, etc. In such system, the trade-off between Quality of Service (QoS) and energy consumption is always a big issue. As the wireless transmission part plays the dominant role in power consumption, many researches propose energy saving schemes based on the adjustment of duty cycle by adaptively switching between wake-up/sleep state of nodes. However, the main drawback of this method is that it affects streaming quality in terms of throughput and delay. Therefore, one of the most important challenges when designing an energy-aware VSN is to keep the balance between energy consumption and video delivery quality. This paper proposes an Enhanced scheme for Adaptive Multimedia Delivery (eAMD) that dynamically adjusts the wake-up/sleep duration of video sensor nodes based on the node remaining battery levels and network performance. A Markov Decision Process (MDP)-based framework is used to formulate the problem and an innovative algorithm based on Q-Learning is proposed to find the optimal policy for video sensor nodes. Using both a systematic and algorithmic approach, our proposed system architecture and algorithms hold the potential to improve the trade-off between video streaming quality and energy efficiency in comparison with other state-of-the-art adaptive video based algorithms.