Adaptive Channel Coding and Diversity for Image Transmission in Wireless Enviroment

• Main Authors: Chi yung Lin, 林器勇
• Other Authors: 賴坤財
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/17176876803371491715

碩士 === 國立臺灣科技大學 === 電子工程系 === 93 === Multimedia communications over wireless channels have become a popular topic in recent mobile communication. It is necessary that we apply data compression and channel coding techniques in image transmission, when we transmit images in wireless fading channel, which typically suffer serious signal degradation and have only limited bandwidths. However, different degrees of importance of bits in a codeword is expectable. Therefore, we should design adaptive channel coding that takes this into account so that the decoded image at the receiver has the best quality. Trellis-coded quantization (TCQ) and scalar quantization (SQ) have obvious corresponding relations between codewords and reconstruction values. We can estimate the received image PSNR when knowing the BER of the various bits in a codeword. In the part of channel coding, we select a powerful coding technology – turbo code. Besides the standard turbo code structure, we extent it to multiple turbo codes. We also apply the diversity technology. We construct the table of BERs by turbo code, and adjust the energy of each bit base on the importance of bit and channel condition. We use unequal error protection by assigning different channel code rates and bit energy allocation to different data bits. Moreover, because of system setup and communication equipment, a wireless channel with rate constraint and energy constraint would be considered. We use an exhaustive search method to find the channel code combination and energy allocation in a particular channel condition. We search the corresponding BER and estimate the corresponding PSNR to achieve the best image PSNR in this channel with rate and energy constraint. Simulation results show that adaptive channel coding have better PSNR than fixed channel coding, when the signal-to-noise ratio in channel is low.