An adaptive stochastic resonance detection method with a knowledge-based improved artificial fish swarm algorithm

The selection of the parameters plays a decisive role in the detection performance of parameter- tuning stochastic resonance (SR) system. In order to solve the problems that the existing adaptive stochastic resonance (ASR) system is easy to fall into local optimum and slow convergence rate in parameter optimization, an ASR detection method with knowledge-based improved artificial fish swarm algorithm (AFSA) is proposed. The lens based Opposition-based Learning strategy (lensOBL) is introduced into the AFSA, micro and macro adjustments can be made to enhance the global search ability, especially in the later stage of the algorithm. In addition, in order to speed up the convergence rate of the optimization, the relationship between the system structure parameters and the SR phenomenon produced is considered to guide the optimization process of the algorithm. Compared with the ASR methods based on the AFSA, the proposed method can obtain better parameters and improves the detection performance of the system. Additionally, to solve the problem of detecting multifrequency weak signals submerged in strong noise and many frequency bandwidths, firstly, the SR of the weak signal of large parameters is realized through adjusting the damping coefficient and the system shape parameters, which solves the problem that the traditional SR is only suitable for small parameter weak signal detection. Then, parallel ASR detection systems are designed based on the idea of the proposed method to identify multifrequency weak radio signals. Finally, simulation data and experimental analysis show that, compared with other adaptive parameter-tuning SR methods, the proposed method enhances the convergence speed and accuracy of optimization, has good robustness, and improves the diversity of fish swarm. It enables the optimal detection of multifrequency weak signal better in strong noise background, and broadens the potential applications of weak signal detection methods based on the principle of SR.