Research On Interference Of Electromagnetic Signals To Signal Classification System

With the rapid development of science and technology and electronic products,more and more intelligent products have complex structures and high integration.These products have extremely harsh electromagnetic environments due to factors such as mixed frequency bands,high power,and dense distribution.Therefore,it is very important to conduct electromagnetic interference research and analysis on intelligent systems.In this paper,the research on the electromagnetic environment effect of the unmanned vehicle system with a high degree of intelligence is carried out first,and the survival and adaptability of the unmanned vehicle system in the complex electromagnetic environment are verified.Then,the millimeter-wave radar in the unmanned vehicle system is further selected as the research object,and the human motion echo signal classification system is developed by using the deep learning algorithm and the millimeter-wave radar hardware to identify human motion,and the classification in the complex electromagnetic environment is studied.Finally,through algorithm improvement,the electromagnetic anti-interference ability of the classification system is improved.The main work of this paper is as follows:(1)Conduct electromagnetic interference tests on unmanned vehicle systems to verify their normal operation in complex electromagnetic environments.The electromagnetic interference test of the unmanned vehicle system is carried out by simulating the complex electromagnetic environment common in life through electrostatic discharge and GTEM chamber.By analyzing the work logs and sensor data of the tested unmanned vehicle system,it was found that due to electromagnetic interference,the perception module of the unmanned vehicle system would experience abnormal lidar data,millimeter-wave radar crashes,and camera transmission cable connection interruptions.Cause unmanned vehicles to build abnormal map,abnormal stop,collision and other faults.(2)The human motion sample library is constructed through the millimeter-wave radar experiment platform.First of all,considering the representativeness of human actions and their practicability in real scenes,this paper designs nine human actions of walking left,walking right,walking forward,walking backward,running left,running right,running forward,running backward,and jumping..Then,nine kinds of human motions were collected through the millimeter-wave radar platform.After data processing,the radar signals were converted into two kinds of feature maps: range-time map and Doppler map,and a human motion sample library was constructed through standardized processing.Finally,through the design of the same-frequency interference experiment,the simulation method is used to simulate the electromagnetic mutual interference scene built by the radar,and the human body action electromagnetic interference sample library is constructed.(3)Through algorithm improvement,a dual-stream feature fusion anti-jamming recognition algorithm is proposed.Using the constructed two-dimensional human motion Doppler map and distance-time map as training data sets,the spatial pyramid pooling algorithm and the channel attention mechanism are used to improve the VGG16 network and the Res Net50 network respectively,which improves the feature extraction ability of the network.Then,the feature fusion module based on the self-attention mechanism is adopted to fuse the distance feature and Doppler frequency feature output by the feature extraction network to make the feature more significant.Finally,the robustness of the algorithm in the scene of electromagnetic interference is verified through experiments.