Research On Application Of Metalens And Millimeter-wave Radar Interference Analysis In Autonomous Driving Environment Perception

Recently,with the rapid increase of vehicle quantity and prosperous evolution of autopilot technique,sensors such as lidar and millimeter-wave radar have been deployed in a large number of autonomous vehicles for real-time perception of the surrounding road environment.Aiming at the problem of large size of traditional focusing lens in lidar receiving optical system,metasurface provides a new scheme for miniaturization design of large field focusing lens.In addition,in the face of complex electromagnetic environment,serious safety issues may arise when millimeter-wave radar,the core sensor of autonomous vehicles,is subject to electromagnetic interference in practice.Therefore,it is of great significance to study metalens design and millimeter-wave radar interference for improving autonomous driving environment perception technology.On the basis of the generalized Snell’s law,a single-layer dynamic tunable metalens model combining GST phase change material with metasurface is designed in this paper.By controlling the crystallization level of GST phase change material in each cell,the proposed model with a compact size of 11.04×11.04 μm2 and a thickness of 240 nm can provide the required varying phase distribution for wide-angle incident wave focusing.The focusing effect of the proposed model at different incident angles is analyzed from four perspectives by the electromagnetic simulation software:incident plane wave conversion effect,focal length,FWHM(Full Width at Half Maximum),and focusing efficiency.Simulation results show that the model can focus a wide-angle incident wave from-30° to 30° to a single point.The proposed model has advantages of compact size and wide-angle focusing.Simultaneously,on account of the time-consuming issue that a large amount of electromagnetic simulation is needed in the metalens design,this paper proposes a fast prediction method of the crystallization level of GST phase-change material based on random forest regression model.This method can be used to quickly determine the crystallization level of GST phase-change material in the metalens unit cell according to the incident angle and the required phase shift.The results show that MAPE and RMSE of the random forest regression model are 4.59%and 0.0131496 respectively,which verifies the accuracy of the prediction results.In view of the problem that millimeter-wave radar has insufficient risk resistance to deliberate jamming and has potential safety hazards,this paper builds a simulation platform of deliberate sinusoidal interference based on the working principle of millimeter-wave radar.This platform can be used to analyze the mechanism and influence of deliberate sinusoidal interference signal.Result comparison indicates the experimental noise floor level in different interference scenarios are consistent with simulation results and the errors are not more than 5 dB,which verifies the effectiveness of the proposed simulation platform.