An adaptive strategy for mobile ad hoc media streaming

Abstract

Mobile devices are increasingly popular and many of them are capable of handling multimedia content. Users enjoy the ability to access their collection of media objects anywhere. Wireless connectivity is often integrated into these handhelds, therefore providing the opportunity to stream multimedia content among mobile and ad hoc peers. An important consideration is to transfer a multimedia object in its entirety. This is often challenging since the transmission time for such an object can be considerable due to an unfavorable combination of a large object size and limited available bandwidth. We previously introduced a novel strategy to improve the probability of success to stream a video sequence based on studying the minimum buffer size. This strategy takes advantage of layered video encoding schemes such as scalable video coding (SVC) or multiple description coding (MDC). The technique adaptively selects the number of layers to be streamed to deliver more frames before the wireless link disconnects while keeping the video quality high. In our current study, we simplify this strategy by using the streaming probability alone to dynamically adjust the number of layers to be delivered. Our proposed technique improves the prediction accuracy by incorporating the 802.11 Auto-Rate Fallback (ARF) scheme along with two popular mobility models: the random waypoint and the random walk mobility models. While ARF-which steps down the sending rate when consecutive transmission errors occur-is implemented in all hardware that follows the popular IEEE 802.11 standard, it is not commonly modeled in existing work. In addition, our approach can retransmit missing layers if peers reconnect after a link break, hence improving the rendering quality. We have performed extensive simulations to validate our technique and the results show an improvement in streaming probability as well as the number of layers that are transmitted. © 2006 IEEE.