A low complexity packet detection algorithm for CSS in AWGN channel

In this paper, we proposed a new packet detection algorithm with high performance and low hardware complexity for chirp-spread-spectrum (CSS) in AWGN channel. There are two typical conventional packet detection algorithms. First, the conventional algorithm based on mixed-correlation (CABMC) has a good performance by using matched filter to remove unused band noise, but the algorithm has a high hardware complexity. Second, the conventional algorithm based on auto-correlation (CABAC) has a simple hardware structure, but the algorithm does not guarantee a good performance because of additional noise in the CSS. The proposed algorithm detects packet by gathering energies after accomplishing differential multiplication between the outputs of adjacent matched filters which have rotated almost on the same phase. This algorithm reduced the number of registers and keep up high performance. Also, this algorithm can detect packet by using just one symbol while conventional algorithms need at least two symbols. Simulation results show that the proposed algorithm detects packet in the higher SNR only about 0.25 dB as compared with the CABMC and reduces the number of registers about 83.70% as compared with the CABMC. And the proposed algorithm detects packet in the lower SNR about 3.52 dB as compared with the CABAC and reduces the number of registers about 59.25% as compared with the CABAC.