Radar Echoes Refined Modeling And Parameter Estimation For Bird Targets

With the continuous growth of the flights,the pressure of bird strikes prevention is increasing,and bird strikes have become a major hidden danger threatening flight safety and people’s lives.Radar is not affected by weather factors and is currently the main mean of bird targets detection.The Doppler generated by the relative motion between the bird targets body flapping wings and the radar can provide effective features for bird targets recognition.Due to the interference of environmental factors and the maneuvering characteristics for non-cooperative bird targets,it is difficult to observe the characteristics of bird targets echoes in all directions.Research on the fine modeling and parameter estimation of bird targets echoes can provide data sources and theoretical basis for the detection and identification of bird targets.Firstly,the flapping wing movement model for birds is combined with the array antenna radar model to establish the mathematical model of the bird echoes for the array antenna radar.The characteristics of the bird echoes are analyzed in the time-frequency domain to verify the validity of the echoes model.Secondly,to improve the modeling accuracy,a method for fined modeling of bird echoes considering the actual electromagnetic scattering characteristics is proposed.Multi-Layer Fast Multipole Method is used to calculate the radar cross section of birds in three typical flapping wing attitudes.The flapping wing attitudes are calculated and combined with the signal propagation model for the array antenna.Then the fined modeling of the bird echoes for every channel at any observation point is realized.Finally,aiming at the problem of bird targets parameter estimation,the autocorrelation function method is used to estimate the flapping frequency,and then the forearm and upper arm of the bird targets are estimated by the matching pursuit algorithm.Considering that the matching pursuit algorithm needs to build a dictionary,which leads to the problem of lower computational efficiency,so the method considering the time-frequency concentration index is used to estimate the arm length,which can greatly reduce the calculation time.On the premise that the parameter estimation accuracy is contained,the computational efficiency can be improved.