Research On Digital Domain Self-Interference Cancellation Technology For Jammer

With the rapid development of electronic information technology,the traditional singlefunction electronic warfare equipment can no longer meet the operational requirements.In the face of complex battlefield electromagnetic environment,the integrated and universal highefficiency electronic warfare equipment that can continuously perceive battlefield situation is the key development trend.As the key equipment of electronic warfare,if the jammer can transmit and receive at the same time,it can greatly improve the work efficiency,strengthen the real-time performance of jamming,and promote the integration and coordination of reconnaissance and jamming.However,in order to achieve real simultaneous transmission and reception,it is necessary to solve the problem of coupling self-interference between the transmitter and receiver of the jammer,so as to avoid the influence of high-power selfinterference on normal signal reception.Aiming at these problems,this paper mainly studies the cancellation technology of jammer self-interference in digital domain,including linear and nonlinear self-interference cancellation.Firstly,considering the linear self-interference problem of jammer in different interference scenarios,a linear cancellation method based on adaptive filtering algorithm is designed.Aiming at the problem that the uncorrelated target signal aggravates the random fluctuation of the estimation coefficient when canceling the noise self-interference,the influence of the interference-to-noise ratio and the algorithm step size on the steady-state error is discussed and analyzed.Aiming at the problem that the related target signal is mistakenly cancelled and the coefficient converges multiple times when canceling the forward self-interference,a method of estimating the self-interference channel by using large-bandwidth Gaussian white noise is derived.The simulation results show that the high interference-to-noise ratio and small step length are beneficial to reduce the influence of the target signal during cancellation.The selfinterference channel can be estimated to a large extent by using the large bandwidth white noise training sequence.The training coefficient can effectively suppress the self-interference within the working bandwidth,and the cancellation ratio can reach 50 d B.Secondly,considering the main nonlinear components brought by the power amplifier in the actual transceiver link,a nonlinear self-interference cancellation method based on deep learning is adopted.Aiming at the self-interference broadband characteristics of jammers,the network input form containing high-order expansion terms is used to characterize more highorder dimension features.The network structure combines the advantages of convolutional network and bidirectional long short-term memory network,and plays the role of feature fusion and bidirectional time-dependent learning respectively.The simulation verifies that this network has excellent fitting ability for nonlinear self-interference of different modeled power amplifiers,and is insensitive to the nonlinear order and memory depth of the power amplifier.When simulating the cancellation of nonlinear broadband noise self-interference of the power amplifier modeled by Wiener,compared to the traditional memory polynomial modeling cancellation method,this network method improves the cancellation ratio by more than 10 d B.Finally,this paper designed an FPGA experimental platform to simulate the simultaneous operation of the transmitter and receiver of the jammer,in order to verify the linear selfinterference cancellation method based on the adaptive filtering algorithm,and additionally connected the power amplifier at the transmitter to verify the deep learning self-interference cancellation method in the nonlinear scenarios.The results show that the linear and nonlinear self-interference cancellation methods used in this paper can achieve in-band cancellation ratios of more than 50 d B and 46 d B,respectively.