A Geo-Adaptive JavaScript DASH Player
Garson Zhong, Ayub Bokani · 2014
In Dynamic Adaptive Streaming over HTTP (DASH), the video bitrate is adapted to the network's condition based on client's feedback. Typically, a high bitrate is selected for next video chunk when the average of previously observed bandwidth samples is high and vice versa. An unexpected drop in the network throughput causes video freezing as the adaptation is done based on network's status in previous moment. In a vehicular environment, this problem is more likely to occur due to rapid movements of client to new locations where the network throughput can be very volatile. In this paper, we present a modified JavaScript DASH player that enables the streaming client to optimize the streaming performance by taking advantage of pre-collected bandwidth statistics in different locations. In this video player, Markov Decision Process (MDP) has been considered as the underlying optimization framework.