Radar recognition of meteorological objects

The peculiarities of radar recognition of meteorological objects are considered, when interperiod fluctuations of reflected signals can be described by an autoregressive model. The recognition algorithms based on use of autoregressive or grid filters are presented, parameters of which are determined by these models and estimated on the base of learning samples. The decision of an object type is made up on the base of dispersion minimum of the filter prediction error, when the samples of intensity fluctuations of scattered signals entered into the filter input that corresponds to an object under recognition. The results of the experimental study of radar recognition efficiency of some meteorological objects using the radar are developed. Finally, acceptable practice results for radar recognition of meteorological objects were obtained by Monte Carlo method, when average probability of correct recognition exceeds 0.8. The solving of the radar recognition problem is of great interest for different practical applications, including aircraft flight safety, prevention of hail damage, squalls etc.