| With the rapid development of digital signal processing theory, algorithms and large scale integrated circuits, digital receiver has become the inevitable trend for the development of receiver system. In the digital receiver, signal detection is the primary task of signal processing, and its performance is the precondition and guarantee for comprehensive performance of the whole receiver system. Therefore, to improve the system communication quality in a bad communication condition, the first thing is to boost the signal detection performance. The thesis is mainly to study the signal detection and adaptive threshold technique in digital receiver. The main contents of the thesis are displayed as follows:1) As the traditional energy detection scheme has a poor noise resisting ability, and doesn't making full use of the characteristics of preamble sequences, a cumulative scheme based on circle sequence is adopted to improve its capability of noise-tolerance. At the same time, a function deduction between the improved amount of SNR (signal to noise ratio) of adjudication variable and signal frequency offset are provided in details in the thesis. Simulation results show that the cumulative scheme can raise SNR of peak-value to certain extent. Meanwhile, we analyze and make contrast of computation and detection performance of the energy detection scheme, delayed correlation method and conjugate multiplication program, and the conjugate multiplication scheme comes out to have the best performance.2) Whereas the problems of CFAR (constant false alarm ratio) adaptive threshold technique, that low detection probability in low SNR and lack enough restriction of false alarm event in high SNR, an improved adaptive threshold detection scheme is proposed in the thesis. To improve the detection performance, correlation signal power and average signal power in two adjacent pseudorandom code periods are multiplied correspondingly. Meanwhile, instead of the fixed threshold, an adaptive threshold varying with the noise power and SNR is adopted to adjudge the peak-value. Moreover, an exiguity fixed threshold is employed to reduce the false alarm event in high SNR. Theory analysis and simulation results show that the improved scheme has a better acquisition performance than the original one.3) Compared with conventional detection technique based on the correlation peak-value of received signal and local pseudorandom code, the thesis proposes a new signal detection technique based on the ML estimation of SNR. Simulation results show that, the new detection scheme can obtain a lower false alarm probability than the traditional one while has the same detection probability. Due to its sensitivity to phase-offset, the new scheme needs combining with the traditional technique to detect signal. In addition, signal detection technique is studied under large frequency offset and an improved frequency-compensation strategy is adopted in the signal detection scheme to reduce its computation complexity.4) During the PN code acquisition in DSSS (Direct Sequence Spread-spectrum System), a serial-parallel acquisition structure is introduced to reduce the conflict between acquisition time and circuit complexity and then the whole performance of acquisition system is improved. Under the strong interference of communication environment, differentially non-coherent detection technique and accumulation method are used in the thesis. They can increase the SNR of peak-value effectively and improve the detection performance in low SNR. To solve problems of hard implementation and poor performance of acquisition for long PN code, a dual-folding code acquisition scheme is adopted to reduce the implementation complexity while grantee the acquisition performance. |
