Investigation On Codec Technology And Application Of Rateless Code In Wireless Communication System

With the rapid development of wireless communication technology, more and more atten-tion has been paid to the efficient and reliable data transmission technology. In face of the increasingly complex communications environment, the transmitter is difficult to pre-know the exact channel state information in many cases. Then, how to adaptively choose the appropriate code rate to protect the data? Rateless code provides us a new way to solve the problem. Rate-less code, as a kind of efficient and reliable error correction technology, attribute the character of code rate adaption to the channel state, does not require the feedback signal during the transmis-sion, and provide a new way of research and design the adaptive, efficient and reliable multi-user wireless communication networks. In this dissertation, the codec technology and the applications of rateless code in wireless communication are intensively investigated, and the contents of this work are listed as follows:For the codec technology, we investigate the coding structure and the decoding algorithm of rateless code for noisy channels. From the perspective of coding structure, by analyzing the problem of applying the structure similar to LT code over noisy channels, we point out that the existence of degree 1 parity nodes connecting to noisy channel leads the problem of error floor. Then, we design a rateless encoding structure named AR (Accumulate Rateless) code, which utilize the accumulator to eliminate the degree 1 parity nodes and is different with the serial con-catenated structure such as Raptor code. With linear encoding and decoding complexity, AR code effectively solves the problem of error floor, and can approximate the performance of the Raptor code. From the perspective of decoding algorithm, we take Raptor code, the most popular rate-less code at current, as an example, and summarize the characteristics and the deficiencies of the existing decoding algorithms. For the characteristics of decoding rateless code, we proposed the serial storage belief propagation algorithm. Then, utilizing the Gaussian approximation density evolution, we prove that it requires fewer decoding iterations. By comparing the complexity of the serial and parallel decoding algorithm in each decoding iteration, we prove that our algorithm has lower decoding complexity without the loss of bit error rate performance.For the applications of rateless code, we investigate the combination of rateless code and the multiple access technology. We propose a new multiple access technology:Coded-Rateless Multiple Access (C-RMA). In the C-RMA system, the transmitter of each user works in the same mode without the need of coordination during the transmission, just as simple as the ALOHA system. From the perspective of the investigation over erasure channel, each user utilize its own access controller to adjust the access probability in order to achieving the maximum system throughput. Our research shows that the best access probability for each user is related to the total number of users serving in the system. Simulation results show that, as a kind of distributed random access technology, C-RMA achieves good performance of the system throughput. Next. from the perspective of the investigation over noisy channel, we investigate the problem of rate-less transmission and the recovery algorithm of each user. The randomness encoding structure of each user provide a natural feature to distinguish the messages of each user. By iterative decod-ing and interference cancelation, the common receiver can effectively overcome the interference between different users. In the simulation, we validate the good performance of C-RMA technol-ogy comparing with IDMA technology. With the simplicity and good throughput performance, C-RMA technology has a strong competitiveness in the distributed communication system in which the transmitters are difficult to pre-know the exact channel state information.Finally, the problem of rateless multiple access is further studied. We provide a new dis-tributed random multiple access technology named non-coding rateless multiple access (NC-RMA). We regard all the data of the users in the multiple access system as a whole, and regard the problem of separating the user data as a kind of distributed joint decoding problem. We re-gard the superposition of the users'signals as a kind of physical layer coding in infinite domain, propose the corresponding decoding algorithm, and propose to utilize the user diversity gain to provide convenience to the decoding. However, during the iterative decoding, the dependence will deteriorate the performance of iterative decoding. Taking into account the independent re-quirement of the messages in the iterative decoding process and the gain brought by the signal combining, we propose the iterative recovery algorithm for NC-RMA system. Then, by extrinsic information transfer technology, we analyze the impact of the signal combining in the iterative recovery algorithm. Simulations verify the good performance of NC-RMA system in both two working modes. NC-RMA system can be easily distributed designed, and achieve the maximum throughput by controlling the access probability of each user.