Stable high-precision network cooperative detection method

The invention provides a stable high-precision network cooperative detection method. The method comprises the steps: for different networks, according to the network and different external environments during detection, selecting a parameter which can make a decision most accurately through optimization; clustering network nodes according to different parameters of node distribution distance, environment and equipment difference, independently selecting an initial condition and a detection model in each cluster, completing parameter estimation by using an ATC or CTA multi-task diffusion strategy, and performing detection judgment by using the estimated parameters to obtain a final judgment result. According to the invention, in the same detection task, according to the method, multiple initial conditions and detection models can be selected for detection, the defects of poor detection precision and instability caused by the fact that a single initial condition and detection model is not matched with the envir