Research On Parameter Identification Method Of UAV Spread Spectrum Communication

The low-cost and versatile of unmanned aerial vehicle(UAV)bring convenience to people’s lives,but also bring severe challenges to public safety and social order.Micro UAV prevention and protection has become a popular research direction in the current public safety field.Among the prevention methods for UAVs,electromagnetic interference is a low-cost and highly reliable method.The key technique of this method is whether it can accurately predict the hopping frequency of non-cooperative UAVs.In general,small civilian UAVs use frequency hopping spread spectrum communication as the uplink communication method,this paper proposes to combine time-frequency analysis and echo state network to predict the frequency hopping spread spectrum communication signals of non-cooperative UAVs.First,use the time-frequency analysis method to perform time-frequency analysis on the UAV measurement and control signal to obtain the time-frequency analysis matrix.The time-frequency analysis methods used in this part include short-time Fourier transform,spectrogram method,Wegener Wiley distribution,Pseudo-Wigner-Wigner distribution and smooth pseudo-Wigner-Willi distribution,and perform algorithm optimization according to the identification effect of frequency hopping signal.Draw time-frequency three-dimensional distribution diagrams,time-frequency variation ridge diagrams,etc.according to the data distribution in the time-frequency analysis matrix,and calculate UAV frequency hopping measurement and control signal parameters including frequency hopping frequency,frequency hopping period,and frequency hopping time information.Then we use the identified hopping frequency sequence as training data to train the echo state network(ESN),the short-term memory function of the ESN is used to realize the single-step prediction of the UAV frequency hopping frequency.This work provides the possibility to use the electromagnetic interference to the non-cooperative UAVs.Finally,in order to further explore the relationship between the single-step prediction success rate of the echo state network and the size of the training data and the size of the reserve pool,this article carried out two exploratory experiments:first,change the training data size and the size of the reserve pool,and count the echo state network traverses the full-period frequency hopping frequency sequence under different parameter conditions for the rate of single-step successful prediction;second,change the length of the frequency hopping frequency sequence,and consider the training data size required for the echo state network to achieve 100%single-step successful prediction rate and study the trends of two values.The simulation analysis results show that among a large number of time-frequency analysis methods,the spectrogram method is both accurate and efficient,and is more suitable for the identification of UAV frequency-hopping measurement and control signals in this paper.In addition,the identification accuracy of the UAV frequency hopping spread spectrum measurement and control signal parameter identification algorithm given in this article will be significantly improved with the increase of the signal-to-noise ratio,and the zero-error identification of UAV frequency hopping frequency can be achieved when the signal-to-noise ratio is greater than or equal to-8dB.When the signal-to-noise ratio is-8dB,for the hopping frequency sequence with a period length of 1023,using the 140 continuous frequency sequence data,i.e.,only 13.7%of the whole period,to train the ESN can achieve accurate enough prediction.Our test results also show that,with the frequency hopping frequency period increasing,the ratio of the training data size with respect to the hopping frequency period required to train the ESN gradually decreases.