Study On Adaptive Demodulation And M-Ary Spread Spectrum Based On Sparse Graph Codes

Rate adaptive(RA) schemes provide an elegant solution to the problem of transmitting data over a time-varying channel, which adjust data rate according to the channel state. Traditional rate adaptive schemes require either that the transmitter needs knowledge of the channel state from the receiver by means of a feedback path, or that the schemes need a second-layer ARQ protocol, which will definitely lead to much cost. The adaptive demodulation(ADM) schemes using rateless codes in this paper can avoid above pitfalls which preserve advantageous properties of traditional rate adaptive schemes.Spread spectrum technique has many good properties, but in a direct sequence spread spectrum system, the chip rate of spread spectrum sequence should be rather high. Also, the system has no coding gain with a rather low achievable information transmission rate. A soft spread spectrum system encodes information by low rate forward error correction codes which increases the achievable information transmission rate and obtains a high coding gain. In this paper, we study several soft spread spectrum schemes combined with LDPC codes resulting in excellent performance.The main progress in this paper include:1. Two ADM schemes are researched. The transmitter encodes information using rateless codes and the receiver would demodulate the symbols at a nonfixed rate according to the decision regions set(DRS) and maximum a posteriori probability(MAP) criteria respectively. The decoder begins to decode as soon as the the number of received demodulated bits is?1+??K. The receiver with a fixed number of demodulated bits and the transmitter with a fixed number of encoded bits are studied and simulated. Simulation results show that in the former case, the bit error rate(BER) performance of ADM schemes is much better than that of the non-ADM schemes and the ADM scheme based on MAP performs better than the ADM scheme based on DRS, and in the latter case, the BER performance gets worse with the increase in the number of the nondemodulated bits when Es/N0 is in some region.2. Several schemes of soft spread spectrum are represented. In order to raise the reliability of the spread spectrum systems, schemes of LDPC codes concatenated with spread spectrum codes are adopted. Hadamard codes are used as spread spectrum codes so that correlation operation in the receiver can be simplified by using fast Hadamard transformation(FHT). In addition, the performance of nonbinary multiplicatively repeated LDPC(MR-LDPC) codes are simulated. In order to show the superiority of the soft spread spectrum schemes, binary LDPC codes concatenated with direct sequence spread spectrum systems are researched. The research results show that soft spread spectrum systems can achieve a lower bit error rate and a higher coding gain. Also, schemes of soft spread spectrum using nonbinary LDPC codes perform much better than that using binary LDPC codes.3. Another contribution of this paper is that we provide several construction schemes of low rate nonbinary codes. Nonbinary Raptor-like codes are the promotion of binary Raptor codes on the finite field. Nonbinary MR-LDPC codes are a class of nonbinary low rate codes with excellent performance. It shows that both codes perform well.