Abstract
This paper presents a fast retry limit adaptation method for video streaming applications over IEEE 802.11e distributed networks. The method enables each source to adapt the number of retransmissions associated to each video packet by relating the perceived distortion to the drop probability and the acceptable delay to the expiration time, without asking the destination for feedback distortion/delay information. The resulting framework, which is based on a simplified but accurate evaluation of the network statistics and of the distortion introduced by the loss of a specific packet, provides a closed-form, and hence computationally cheap, estimation of the retry limit. Furthermore, with respect to most of the existing solutions, the proposed strategy accounts for the impact of the higher priority voice access category (AC), in order to improve the reliability of the retry limit adaptation in the presence of contending ACs. The method is validated by a simulation platform including the physical communication chain and the 802.11e medium access control layer, and its performance is compared to that obtained from an existing solution and from the optimum theoretical settings