Study Of Signal Detection Based On Duffing Oscillator With Band-limited Noise

Signal detection is a critical section in communication.Both signal reception to resist frequency deviation between cooperators and signal detection to find signal between competitors need to detect the accurate carrier wave frequency firstly.The requirement of modern communication is lower SNR,less priori information and more complex signal modulation.Therefore,nonlinear methods are brought to replace traditional linear methods to realize signal detection in low SNR.Nonlinear methods can resist strong noise and detect signal in lower SNR,but they are still in theory.Duffing oscillator,as a typical application of chaos theory,is an useful nonlinear system that can be used to detect carrier wave frequency.However,traditional study on Duffing oscillator is about broadband noise and the core analyzed is only one signal reception.Therefore,we,based on the narrowband Gaussian noise with high power spectral density,modeling the interference among signals with some sine signals,study signal detection methods utilizing Duffing oscillator to work in real communication environment.Firstly,the basic theory of Duffing oscillator is studied and difference oscillators are analyzed to get better parameters and thresholds.Then,the interference of noise to traditional signal detection methods based on Duffing oscillator is analyzed and the ability of Duffing oscillator to resist noise is proved.After this,noise is put inside the function of Duffing oscillator and the new state of Duffing oscillator in narrowband Gaussian noise with high power spectral density is received.General cycle state and general chaos state are defined through inferring the changing rule of cycle probability and amplitude of driving force in theory.With these basement,general intermittent chaos theory is put forward and signal detection method based on it is set.Finally,the platform of signal detection is built and the feasibility of single signal and multi signals detection is proved in the environment of narrowband Gaussian noise with high power spectral density.Also,with the comparison between this method and other traditional methods is set to prove its superiority.