Research On Terrain Following/Terrain Avoidance Radar Target Detection Methods

Obstacles in the flight path or close to it often pose a significant threat to low-flying air vehicle.Being equipped with a terrain following(TF)/ terrain avoidance(TA)radar allows a low-flying air vehicle to detect terrain profile data and sense the unmarked obstacle of digital map in real time,which significantly contributes to path planning and obstacle avoidance of air vehicle.Therefore,TF/TA radar is widely used in the field of navigation and collision avoidance of low-flying air vehicle.In order to better explore the topographic information in front of the low-flying air vehicle,signal detection technique is adopted in TF/TA radar.Clutter characteristic analysis,background power estimation,constant false alarm rate(CFAR)design,high-probability detection and high-precision location are the essences of signal detection of TF/TA radar.They are also the hot issues and frontier problems in the field.For these problems,this dissertation is engaged in theoretical analysis,method research and real experiment.The main contents of the dissertation are as follows:1.Ground clutter distribution model and probability of false alarm of TF/TA radar are deduced from space-time relation of TF/TA radar signal acquisition.They reveal the quantitive relations of clutter distribution,detection threshold and probability of false alarm.It establishes the foundation for the design of high-preformance signal detection algorithm in ground clutter environment.2.Clutter power estimation method is proposed based on positive-definite matrix space.The method can avoid large discrete and spiky clutters under limited training sample conditions,and achieve high-precision of estimation for background clutter power.Meanwhile,the method also can solve the problem of excessive and fluctuant detection threshold encountered in existing detection methods,and obtain high-probability of detection in intense non-homogeneous clutter environments.3.Non-homogeneous clutter automatic censoring method is proposed based on ordered data difference.The method can reject the unwanted samples in the reference window,and solve the problem of excessive detection threshold caused by large discrete and spiky clutters.Meanwhile,the method also can achieve robust CFAR detection performance without any prior information about the background environment.4.The echo signal of TF/TA radar is modeled as a two-dimensional space-variant convolution-like model,and maximum likelihood based detection method is proposed.The method can solve the difficult detection problem of multiple obstacles located in one single beam and same slant distance in complicated background scene,and achieve high-probability of detetion and high-precision location of multiple obstacles.The aforementioned signal detection methods of TF/TA radar are verified by simulation and real experiment,which can obtain high-probability of detection and high-precision location of obstacle for low-flying air vehicle.