Research On Key Techniques Of Nonlinear Dynamical System Based Communication Signal Detection

Weak signal detection has been widely used in communications, radar, sonar and other fields. Most of the traditional methods of weak signal detection are based on linear system, and they can not perform well under very low signal-to-noise ratio (SNR) scenarios. In order to detect the weak signal more accurately under very strong noise environments, new theories and methods are being explored. In recent years, the advancement of nonlinear science provides a new idea to analyze and solve such problems. In fact, there exist some specific phenomenons in nonlinear systems, such as resonance, high order harmonic, chaos and so on. In contrast to suppressing noise in linear systems, the noise can be utilized in nonlinear systems.So far, to the best of our knowledge, nonlinear systems have only been applied to detect the weak signal consisting of single tone, such as sine wave. However, communication signals are usually bandpass, and may have a variety of frequency compents and phases. Meanwhile, the randomness of information source will lead to the discontinuity of symbol waveform. In this dissertation, the stochastic resonance and Duffing oscillator of the nonlinear dynamics are proposed to be applied in the weak communication modulation signals detection system. The main contributions of this dissertation are summarized as follows:Firstly, weak communication signals detection is investigated based on an adaptive stochastic resonance system. The stability of numerical solution is discussed for adaptive bistable stochastic resonance system. Then, a novel energy detection algorithm and periodogram-based energy detection algorithm are applied for the detection of BFSK, QPSK, MSK and 16QAM modulation signals under both synchronous and asynchronous environments. Moreover, the influence of noise uncertainty is also concerned. Compared with conventional energy detection algorithm, the sumulations results show that both of these two algorithms can significantly improve the detection performance.Secondly, Duffing chaotic system is used to detection weak communication signals based on the characteristic that the system state of Duffing chaotic system is sensitive to the signal amplitude’s variation. Four factors that affect the state transtion of Duffing chaotic system are investigated, which include step size, initial phase, noise and signal frequency. As the key point of Duffing oscillator based weak signal detection is the effective discrimination of system status, two quantitative status identification methods are further investigated, which are based on power spectrum characteristics and pseudo Hamiltonian respectively. These two methods are investigated as follows:● Based on the significant discrepancy of power spectrum associate with the output signal’s low frequency component when the system is in the chaos state and the great periodical state, power spectrum characteristics based system state transition detection is investigated in this dissertation. Simulations are performed to verify the performance of detecting BFSK, QPSK, MSK and 16QAM modulation signals.● Based on the significant discrepancy of pseudo Hamiltonian associate with the output signal when the system is in the chaos state and the great periodical state, pseudo Hamiltonian based system state transition detection algorithm is studied. Simulations are also performed to detect weak BFSK, QPSK, MSK and 16QAM modulation signals.Compared with the energy detection algorithm, the results show that both of these two algorithms can also significantly improve the detection performance.