Research On OFDM Signal Detection Algorithms For LED Communication System

Compared with other communications, satellite communication can provide full-time services because it has the advantages of unrestricted geography, wide coverage and high capacity. Therefore, it is the powerful supplement and support for terrestrial wireless communications. Especially, as low earth orbit (LEO) satellite has shorter delay, it is more suitable for real-time communication traffic. However, with the surge demand in high-speed multimedia serves and the increasing requirements of users for quality of service, spectrum resources become more and more scarce. Therefore, it is important that an efficient transmission scheme is adopted in LEO satellite communication systems.According to OFDM (Orthogonal Frequency Division Multiple) transmission technology applied to LEO satellite communication, this thesis researches on signal detection algorithms on both uplink and downlink. Though OFDM technology has high spectrum efficiency and combats channel fading, the signals on the links of LEO satellite are subject to Doppler shift, multi-user interference and other factors, resulting in frequency offset. It is necessary to compensate the frequency offset of the transmitted signals to complete the signal detection.For unlink, based on PIC (Parallel Interference Cancellation) algorithm, this thesis proposes two improved algorithms. One is a PIC algorithm based on minimum linear error. It estimates error using the minimum linear error method, and then eliminates the influence of the all carrier frequencies in a parallel way. Another is the optimum weight PIC algorithm, which improves the accuracy by identifying the most accurate value of the coefficient in parallel computing. Both of the two algorithms can well eliminate the frequency offset. For downlink, based on Schmidl&Cox algorithm and Kim algorithm, this thesis pays attention to frequency compensation algorithms in both time and frequency domains and proposes a least mean square algorithm, which has satisfactory performances of bit error rate (BER). According to complex channel conditions, a frequency offset compensate algorithm based on Frequency Space Alternating Generalized Expectation Maximization is proposed, which can reduce the error symbol in the transmission channel. When unknown channel knowledge is too much, a novel blind estimation algorithm independent of prior knowledge is introduced, which is based on the decision feedback equalizer structure and recursive least square. Simulation results demonstrate that its BER is still able to keep lower to guarantee high quality of transmitted signals in satellite channel.