Target Position Estimation Method Research For Compact Array Dual-Frequency HF Radar

The compact High-Frequency(HF)oceanographic radar with monopole/crossed-loop antenna as receiving sensor has the advantages of small footprint and convenient maintenance.Its inversion capabilities for ocean surface parameters such as wind,wave and current have been extensively studied.However,the target detection property of the compact HF oceanographic radar has not been fully explored.Among them,target position estimation problem includes direction of arrival(DOA)estimation and range estimation.Its estimation accuracy is very important to target monitoring performance,which is the weakness of compact HF radar with the wide-beam antenna.This thesis aims to improve the target azimuth and position estimation performance of compact HF radar from the perspectives of receiving antenna array and dual-frequency radar.The research work is as follows1.The azimuth estimation performance of dual monopole/crossed-loop antenna array is analyzed.The small antenna aperture and pattern distortion restrict the multiple signal classification(MUSIC)algorithm performance of monopole/crossed-loop antenna.Its inability to estimate the azimuth of two coherent signals also limits the application of HF radar in omnidirectional detection environments such as floating platforms.At the cost of increasing the footprint of only one antenna,the dual antenna array can increase the antenna aperture and also has the potential to realize the coherent DOA estimation by using decorrelate algorithm.But its DOA estimation performance needs to be analyzed and verified.Therefore,the theoretical analysis and simulation on the DOA estimation accuracy and the coherent DOA estimation performance of the dual monopole/crossed-loop antenna array are conducted in this thesis.Firstly,the theoretical advantages of the dual monopole/crossed-loop antenna array compared to the monopole/crossed-loop antenna are derived from the theoretical MUSIC error formula,and the experimental data results prove it.In addition,the two-sources DOA estimation performance of the dual monopole/crossed-loop antenna array is verified and analyzed through simulation,and a coherent signals case of the shore-based radar experimental data is used to test its application ability.2.A target position estimation method for dual-frequency HF radar based on support vector regression(SVR)is proposed.Compared with radar operating at single frequency,dual-frequency radar can overcome the occlusion effect of sea clutter on targets and monitor ship targets of different sizes.The traditional estimation method of dual-frequency radar adopting weighted fusion is simple and practical,but its performance is limited by the estimation accuracy of single-frequency radar data.In this thesis,a target position estimation method for dual-frequency HF radar based on SVR is proposed.Through the SVR algorithm,the mapping model between the azimuth and range of target with the characteristics of dualfrequency radar data is directly constructed.Then,the model is used to estimate the range and azimuth of the target from new radar data and achieve position estimation.The experimental data results show that the proposed method fully combines the advantages of dual-frequency radar and the characteristics of machine learning.The results of the proposed method have the most superior estimation performance compared with the results of single-frequency radar using traditional estimation algorithms and SVR algorithms.Finally,the impact of the training sample number on the SVR algorithm performance is discussed and analyzed.