To solve the problem that weak signals are difficult to detect accurately in low signal-to-noise ratios, this paper presents a method to achieve effective detection of weak signals, applying the method of stochastic resonance to bistable systems. The principle of the method is that by transferring part of the noise energy to the signal energy, enabling the detection of weak signals at low signal-to-noise ratios. It makes it easier to extract the signal at the receiving end. This model designs a parametrized conditioning system based on the factors influencing the output power spectrum and output SNR of a stochastic resonant system. Based on the experimental results, the influence of parameters a and b on the model can be analysed, and the optimal noise intensity range of the system can be found. At the receiving end of the system, the constellation diagram and BER are used as a measure of system performance. Simulation experiments show that stochastic resonance can effectively enhance the energy of weak signals under low signal-to-noise conditions, and the demodulation performance of the system is significantly better than that of the system without the use of stochastic resonance.