基于PSO-ELM的卫星导航欺骗式干扰检测CSCD

近年来,卫星导航系统在军事监测、精细农业、交通监控、资源勘探、灾害评估等领域得到了广泛应用,但由于卫星导航信号结构公开且到达地面时强度微弱,卫星导航系统极易受到各种各样的干扰,其中欺骗式干扰因具有较强的隐蔽性,对卫星导航系统构成巨大的安全威胁。传统的欺骗式干扰检测方法大多采用单一参数进行检测,具有一定局限性。考虑到欺骗干扰源在欺骗过程中会引起一系列参数变化,构造了一个多参数输入的卫星导航欺骗式干扰检测模型,即将多个特征参数作为极限学习机(ELM)的输入,并通过训练和学习将真实信号与欺骗信号区分开,从而实现欺骗式干扰检测。同时,利用粒子群优化(PSO)算法优化ELM中的输入权值矩阵和隐层偏置,改善由于网络参数随机生成导致分类精度低的问题。仿真实验证明了该方法在卫星导航欺骗干扰检测方面的可行性和有效性。

Deceptive Jamming Detection of Satellite Navigation Based on PSO-ELM

In recent years, satellite navigation systems have been widely used in military monitoring, precision agriculture, traffic monitoring, resource exploration, disaster assessment and other fields. However, due to the open structure of satellite navigation signals and the weak signal strength when they reach the ground, satellite navigation systems are extremely vulnerable to various problems. Among all kinds of interferences, deceptive interference poses a huge security threat to the satellite navigation system due to its strong concealment. Traditional deceptive interference detection methods mostly use a single parameter for detection, which has certain limitations. Considering that the deceptive interference source can cause a series of parameter changes during the deception process, this paper constructs a multi-parameter input satellite navigation deceptive jamming detection model, which uses multiple characteristic parameters as the input of the extreme learning machine (ELM), and through training and learning, distinguishes real signals from deceptive ones, so as to achieve deceptive interference detection. At the same time, the particle swarm optimization (PSO) algorithm is used to optimize the input weight matrix and hidden layer bias in ELM to solve the problem of low classification accuracy due to the random generation of network parameters. The simulation experiment proves the feasibility and effectiveness of this method in the detection of satellite navigation deceptive jamming.