Construction Of Target Signal Detection And Extraction System Based On Heterogeneous Computing Platform

Signal detection and extraction in complex electromagnetic environments has been an enduring challenge in the field of communication engineering.With the development of mobile wireless communication,the objective demand is increasing,and various signals are gradually showing more complex and diversified features,such as increasing signal bandwidth,more complex transient characteristics,and increasing carrier frequency,which pose more severe challenges to signal detection systems.The traditional signal detection and extraction algorithm and hardware computing platform can no longer meet the highperformance and real-time requirements of non-cooperative dynamic reception and analysis in complex electromagnetic signal environment,therefore,for the key technologies in the process of signal detection and extraction,this paper conducts in-depth research in the following three aspects.1.For the broadband multi-signal detection problem,a Wideband multi-signal detection algorithm based on noise base flattening signal edge identification is proposed.The method separates the noise floor from the signal by means of double gradient thresholding,so as to improve the detection probability while reducing the probability of false detection.Theoretical analysis and experimental simulation show that the algorithm can effectively overcome the problem of false detection caused by colored noise and the problem of missed detection caused by excessive differences between signals,and has a certain degree of robustness in complex electromagnetic environments.In addition,the algorithm has a fine frequency resolution,which can accurately detect and identify two or more consecutive signals with narrow protection intervals and output accurate bandwidth and carrier frequency estimation results.2.For the non-uniform channelization filtering and extraction of broadband signals,a multi-scale non-uniform channelization filtering and extraction method based on sub-signal bandwidth distribution is proposed.When constructing the channelization structure,multiple channel division number scales are preset,and the most suitable one is selected for signal channelization segmentation and reconstruction processing according to the carrier frequency and bandwidth distribution of sub-signals in the input signal.Under the condition that each sub-signal can be extracted separately,the edge invalid bandwidth of the reconstructed signal is reduced as much as possible,thus reducing the complexity of the system and improving the overall computational efficiency,while facilitating the subsequent processing of the signal.3.To address the problems of low computational efficiency,large size and high-power consumption of the traditional hardware platform,a heterogeneous computing platform integrating RF front-end,Zynq embedded hardware platform and upper PC is built,and the proposed algorithm is sunk into the platform for implementation.The advantages of the small size and low power consumption of the Zynq platform,the efficient parallel computing acceleration capability of the FPGA,and the powerful general-purpose serial computing capability of the ARM and x86 architectures are integrated,the logical layout of the modules is rationalized and the resources are flexibly allocated,which significantly improves the overall execution efficiency of the system.Based on the heterogeneous computing platform,this paper completes the construction of signal acquisition,broadband multi-signal detection,non-uniform channelization extraction and data communication modules,and verifies the correctness of the system and the effectiveness of the detection and extraction algorithms through practical tests,finally realizing the successful construction of the target signal detection and extraction system.