Research On Adaptive Forward Error Correction Algorithm For Video Transmission Over Wireless Network

With the continuous development of wireless network technology, wireless multimedia technology has been more and more widely used, and also brought great changes to human life. People increase the demands of video quality transmitted over wireless network. As one of the key technologies of wireless networks, the Wireless Mesh Network(WMN) obtains more and more attention, video transmission over a wireless Mesh network has become a research hotspot. The wireless Mesh network has characteristics of multi-hop, topology changed and self-organization, which result in greater difficulty and more serious packet loss for video transmission. Therefore,appropriate error control technique is needed to ensure the success of video transmission. Forward error correction coding is one of error control methods, it can recover lost data during transmission by adding additional redundant packets. Since the loss of packets in the wireless network is gusty and fluctuant, so it is difficult for traditional forward error correction technology to achieve the best results in wireless video transmission, since less redundant packets could not restore the lost data, while excessive redundant packets may consume too much bandwidth. Thus redundant packets generated by error correction mechanism should be able to adjust automatically based on the channel conditions and network load. Research on adaptive forward error correction technology has great significance.Firstly, the thesis summarizes the research status of home and abroad about the adaptive forward error correction algorithm and introduces several common error control methods.We analyze simply the performance of static forward error correction.The experimental results prove the shortcomings of Enhanced Adaptive Forward Error Correction(EAFEC) algorithm and the necessity of improving the algorithm.Secondly, in EAFEC, the smoothing factor of queue length is a constant, for this question, we analyze the relationship between the changed weight value of queue length and the number of redundant packets and the peak signal to noise ratio in different packet error rates. We propose an solution to tune the weight value of queue length based on the channel state, combined with the research on the effects of weight value on the algorithm performance. We run the proposed algorithm and analyze the results over wireless Mesh network.Finally, for the question of the queue length and retransmission times are divided into primary and secondary factor artificially in original EAFEC algorithm, weintroduce the balance parameter to balance the proportion of the queue length and the retransmission times in determining the number of redundant packets to solve this issue. And the proportion of num_FEC1 and num_FEC2 in determining num_FEC finally is adjusted according to the network load and channel conditions. We build the wireless Mesh network in NS-2 to test the algorithm proposed in this paper. The results show that the Balance Parameter Introduced for Enhanced Adaptive Forward Error Correction(BEAFEC) algorithm achieves better performance in such as packet loss, the number of redundancy and PSNR. The improved algorithm can recover lost video data well in the case of adding fewer redundant packets. The values of the most PSNR obtained are more than 40 d B in the case of packet loss rate is 0.3. It improves the video quality greatly and allows better visual experience for users.