基于模型转换频率估计的低空目标分类

为解决传统机械扫描预警雷达的低空目标分类问题, 提出一种利用模型转换频率估计的飞鸟与无人机目标分类方法。首先, 建立可能的无人机和飞鸟目标运动模型, 进而对目标轨迹进行多模型滤波和平滑处理。然后, 通过估计目标的模型转换频率实现目标分类。仿真结果验证了所提方法的鲁棒性和有效性, 证明经过平滑处理的目标分类结果优于仅经过滤波处理。针对低空预警雷达在机场和海岸环境中采集的真实数据, 在无人机频繁改变飞行方向模拟飞鸟机动性特征的情况下, 所提方法仍然可以较好地跟踪和区分无人机与飞鸟目标, 大幅降低了雷达系统的虚警率。

Low-altitude target classification based on model conversion frequency estimation

In order to solve the problem of low altitude target classification of traditional mechanical scanning early warning radar,a method for target classification of flying bird and unmanned aerial vehicle(UAV)based on model conversion frequency estimation is proposed.Firstly,the possible UAV and flying bird target motion models are established,the multi-model filtering and smoothing are applied to the target trajectory.And then the target classification is realized by the model conversion frequency of the target estimation.Simulation results verify the robustness and effectiveness of the proposed method,and prove that the target classification result after smoothing is better than that only after filtering.According to the real data collected by low altitude early warning radar in the airport and coastal environment,the proposed method can still track and distinguish UAV and bird targets,and greatly reduce the false alarm rate of the radar system when UAV frequently changes its flight direction to simulate the mobility characteristics of flying bird.