基于LSTM的无人机轨迹识别技术研究

Research on UAV Trajectory Recognition Based on LSTM

无人机目标逐渐成为威胁低空空域安全的重要因素,针对低空空域对目标识别的实时性和准确性需求,提出基于目标轨迹进行无人机识别的方法。构造目标飞行的轨迹特征,使用雷达探测到的目标飞行数据训练模型,对无人机目标进行识别。实验结果表明,使用递归神经网络(RNN)的长短期记忆(LSTM)结构建模分析目标的轨迹信息,具有精度高、易实现的特点。

Drone targets are becoming an import threat to the safety of low-altitude airspace.In view of the real-time and accuracy requirements of low-altitude airspace for target identification,amethodfor UAV recognition based on target trajectory is proposed.Construct the trajectory characteristics of the target flight,use the radar to detect the target flight data training model,and identify the target of the UAV.The experimental results show that the trajectory information of the target is highly accurate and easy to implement by using the Long Short Term Memory structure modeling of the recursive neural network.