Study On Noise And Target Identification Of Road Detection Radar

With the steady development of the global economy,automobiles have become the main transportation in life,and road traffic safety has increasingly become the focus of social attention.As an effective road traffic monitoring tool,millimeter-wave radar has the characteristics of strong anti-interference ability and working around the clock,becoming the core sensor in the field of road traffic gradually.Accurate speed,range and angle measurement,target tracking errors reduction,and recognition accuracy improvement are the main development directions of millimeter wave traffic monitoring radars.The noise performance of the radar directly determines the target detection sensitivity.In this paper,the main noise sources of the 77 GHz silicon-based millimeter wave radar system are analyzed,focusing on the influence of radar system noise on the target speed and distance resolution,and further the Gradient Boosting Decision Tree(GBDT)recognition method based on the one-dimensional range profile of the target is used to accurately identify and analyze the detection target.The main research contents and results are as follows:1.Different noise sources in the 77 GHz radar system and their influence on the target performance parameters are studied.The noise in the noise power spectrum mainly comes from the phase noise and the receiver noise of the radar system.Utilizing MATLAB to model the two types of noise separately,the simulation results show that the in-band phase noise raises the noise floor near the target and reduces the range resolution,weakening the radar’s ability to detect weak signals;the receiver noise reduces the target range resolution,but the measurement The distance accuracy has almost no effect.On the basis of theoretical research,the noise power spectrum of the two radar transceiver systems was compared and measured.The experimental results show that Texas Instruments’ AWR1243 has a noise floor of-81 d Bm.The signal leakage at the transmitter and receiver makes the noise floor at close distances up to12 d Bm,the target range resolution for RCS of 1 under the bandwidth of 75 MHz is 2m;the noise floor of Calterah’s CAL77A2T4 R is-68 d Bm,and the range resolution of 3m.The range accuracy of both radars is within 0.5m.2.Aiming at the vehicle tracking and recognition in the road monitoring scene,the recognition accuracy of different target recognition methods is compared and found that the GBDT target recognition method based on one-dimensional range profile has the best accuracy.For the 77 GHz millimeter-wave radar,electromagnetic simulation and actual measurement data are used to obtain the one-dimensional range profile of different vehicle models,and the expressive characteristics of the one-dimensional range profile are further experimented as training and test data for different recognition models,and compared with the identification methods of logistic regression(LR)and factorization machine(FM).The results show that the accuracy of GBDT target recognition is higher than logistic regression(LR)and factorization machine(FM).When the data division ratio of the training set and the test set is 0.2 and the learning rate is 0.1,the accuracy of GBDT target recognition based on one-dimensional range profile is as high as 92.3%.Meet the actual road inspection needs.