Weak signal detection in chaos using adaptive Neuro-Fuzzy Inference System

Classical statistic detection theory regards chaotic noise as random signal, which weakens the performance of signal detection. Based on chaotic dynamic mechanism, detecting weak signal embedded in chaos is discussed in this paper. This method uses the Adaptive Neuro-Fuzzy Inference System (ANFIS) to establish a prediction model of chaotic background for the ANFIS's powerful ability of learning and local predictability of chaos, and then combines the Fast Fourier Transform (FFT) algorithm to process predictive error to detect the weak harmonic signal. This method can improve the signal-to-noise ratio of system output. The chaotic series of Lorenz system was utilized as chaotic background for experiments. Experiment results were achieved from the computer simulation, which shows that we can not only detect the presence of the weak signal, but also determine the frequency of the detected signal using the proposed method.