| In many application scenarios, the signal to noise ratio(SNR) of received signal is too low, for the factors of transmitted power, external environment and so on. This article mainly focuses on the frame synchronization under extremely low SNR condition, including the design of new sync words and new frame synchronization detection algorithm and its demonstration and test on SDR platform.By the principle of Kronecker product, we combine 48 bits optimization sequence, 13 bits barker code sequence and 13 bits interleaved sequence of “-1 / +1” to obtain 624 bits compound sync word, which not only has good aperiodic correlation, providing more reliable decision criterion for the algorithm, but also widen the gap between optimum decision value of received data and adjacent points, facilitating the selection of optimum decision value.The structure of frame synchronization is the periodical insertion of sync words, and the corresponding algorithm is the parallel correlation frame synchronization detection and judgment, which are based on “the maximum-correlation decision rule”. The basic process of frame synchronization detection is as follows. Firstly, we deserialize the output data streams of matched filter. Secondly, we apply parallel multi-window sliding correlation operations on all data. Thirdly, we select the maximum value from the relative values, which correspond to the continuous samples points in a frame, to locate the frame starting point. Finally, we propose a “modified decision rule” in multi-frame consecutive judgment to improve the performance of synchronization detection at the cost of computation complexity.To analyze the performance of frame synchronization detection completely, we analyze “the maximum-correlation decision rule” and “the modified decision rule” from two aspects, “the detection of frame starting point” and “the selection of decision value”. The results show while taking “the maximum-correlation decision rule”, “the detection of frame starting point” performs well when SNR is higher than-15.0dB, and the performance deteriorates rapidly as SNR decreasing. Meanwhile, “the selection of decision value” also worsen fast. Although appropriate increase of “+1 /-1” in the sync words can improve the result, the effect is limited. Thus the parallel correlation detection algorithm can detect the frame starting point effectively, but it is hard to select the optimum decision value. Fortunately, “the modified decision rule” can solve this problem with extra complexity.PC, baseband board of Virtex-6 ML605 and radio frequency board of FMC30 RF are used to set up SDR platform, simulate the process of communication and verify the frame synchronization detection algorithm. We insert sync words periodically to framing and adopt pre-adding of Gaussian sequence with corresponding SNR to simulate noise. Root ascending cosine roll-offs filters are used for shaping filter. The modulation mode is Binary Phase Shift Keying(BPSK). At the receiving end, following coherent demodulation and matched filtering, we employ the parallel correlation detection algorithm based on “maximum-correlation decision rule” to conduct the frame synchronization detection. After the accomplishment of frame synchronization, we use sync words to estimate the SNR of received signal and analyze the performance of detection. |
PDF Full Text Request