The Research On BATS Codes And Its Applications

With the development of communication technology,wireless multi-hop networks supporting high-speed data transmission will gradually replace traditional single-hop networks and become the mainstream of future wireless communication networks.In the process of wireless transmission,noise,interference,attenuation and other factors cause packet loss on each link in the multi-hop network,which affects the reliability of transmission.However,the automatic request for retransmission,traditional erasure coding,and network coding can not effectively solve the above problems.BATS codes is a class of concatenated codes based on fountain code and random linear network coding.It inherits the characteristics of fountain code,such as rateless,low coding and decoding complexity,and also has the high throughput characteristics of network coding.So that it can guarantee the reliability of data transmission in multi-hop networks.Therefore,the theoretical research and application design of BATS codes in multi-hop networks have great significance.Firstly,this paper briefly introduces the basic theoretical knowledge involved in BATS codes,and then introduces the coding and decoding principles and degree distribution design methods of BATS codes in detail,and analyzes the excellent performance of BATS codes in multi-hop networks.After that,through simulation,the effects of various coding parameters,degree distribution,decoding algorithms and other factors on the performance of BATS codes are analyzed in turn.Secondly,this paper analyzes two shortcomings of the BATS codes: large coefficient overhead and poor short codeword performance,and then proposes corresponding optimization design methods.In order to improve the reliability of short code transmission of BATS codes,this paper designs optimal BATS codes based on important information feedback and coding statistics for the application scenarios with and without feedback channel.The optimization principles of the two schemes are almost the same: by transmitting the packet with the highest degree to the decoder,the decoding process of the belief propagation algorithm is promoted to increase the probability of successful recovery of the source information.Aiming at the problem of large overhead of coefficients,this paper proposes a BATS codes optimization scheme based on a pseudo-random sequence.Through fixing the initial value of the pseudo-random sequence generator to synchronize the encoded information,so that the transmission of encoded information is avoid and the effectiveness of transmission is improved.Finally,combined BATS codes with image coding,this paper proposes an image transmission scheme based on wavelet SPIHT image coding and sliding window BATS codes.The compressed data stream generated by SPIHT coding is encoded and decoded in the form of sliding window to achieve image transmission with high QoE.Through simulation analysis,it is found that the randomness of the BATS code may cause the reliability of image transmission to be not guaranteed.In this regard,this paper improves the sliding window structure of the BATS code,and proposes a sliding window BATS code based on encoding statistics.By increasing the encoding probability of unencoded compressed packets,the effective recovery probability of compressed data at the decoder is improved,so as to realize the reliable transmission of images.