Reliable Transmission Using Fountain Codes And Their Combination With Network Coding

Fountain Codes are an efficient and reliable transmission method over the erasure channel. With comparable low complexity in both encoding and decoding, and superb data recovery performance, fountain codes is capable of carrying out broadcasting of data with much higher efficiency and less packets to be transmitted, compare to traditional ways of broadcasting. Hence, fountain codes are especially suitable for multimedia broadcasting services.We will discuss how to promise reliable data transmission using fountain codes like LT codes over the AWGN channel. Based on the existed problems, we proposed a transmission system which is consisted of IBRC (Improved Binary Repetition Codes) codes concatenated with LT codes. We proposed a decoding algorithm dedicated to a special set of IBRCs, which makes this set of IBRCs not only capable of correcting transmission errors aroused by channel interference, but also able to identify incorrect decoding result before they are sent to the LT decoder to avoid consecutive decoding errors in the LT decoding process, just as CRC does in the traditional ways of transmission. Hence our system could maintain the reliability of data transmission using LT codes over the AWGN channel without involving CRC techniques. Moreover, the calculation complexity of the proposed decoding algorithm is very low which is only consisted of additions between vectors, making IBRCs quite competitive channel coding schemes. We test the performance of the proposed system on data recovery, and find this transmission system effective according to the simulation results.We continue to analyze the combination strategies between fountain codes and network coding over a multi-hop network. We proposed a new combination method in which the relay nodes in the network performs opportunistic network coding, so that the LT coding process which is used to be carried out at the transmitter is now performed throughout the data transmission process over relay nodes in different layers, and receivers are able to recover the source data by LT decoding. We refer to the fundamental theories of fountain codes to explain the principles of the proposed combination scheme and also forecast its performance variation based on the knowledge of LT coding. The simulation results validate the effectiveness of the proposed strategy and validate our presumption of its performance. At the same time, this strategy is easy to be carried out and put into practice, with still possibly huge improvement in data recovery performance.