Compressed Sensing DOA Estimation In Vehicle-mounted Millimeter-wave Radar Communication Integration System

The vehicle-mounted millimeter-wave radar communication integration system integrates radar detection and wireless communication functions,and is widely used in intelligent transportation systems and electronic countermeasures.In the vehicle networking environment,the vehicle-mounted millimeter-wave radar communication integration system combines MIMO technology and OFDM technology to use millimeter waves while achieving vehicle-to-vehicle communication.The radar senses the surrounding vehicle position information（distance,speed,and azimuth）.Therefore,the signal source DOA estimation algorithms for high-resolution,real-time signal in a vehicle-mounted millimeter-wave radar integration communication system have a certain application prospect,which provides a guarantee for unmanned driving technology and vehicle reasonable obstacle avoidance.In order to reduce the sampling rate of the DOA estimation of the array signal in the integrated system of millimeter-wave radar,the sparse multipath in the millimeter-wave communication system and the sparseness of the spatial signal source relative to the entire airspace are considered.The direction of arrival angle estimation is combined.The dissertation focuses on the DOA estimation algorithm based on compressive sensing greedy algorithm.（1）When the array manifold matrix is divided by equal angle,in view of the large computational complexity and poor real-time performance of ₁l norm convex optimization process for DOA estimation based on compressed sensing sparse reconstruction,in this paper,the greedy algorithm is used to replace the ₁l norm to approximate the ₀l norm optimization problem of DOA estimation.The proposed SAMP algorithm has higher DOA estimation accuracy in a single snapshot,which greatly reduces the computational complexity.At the same time,DOA estimation for coherent signal sources is also applicable and DOA estimation can be performed under the condition of unknown signal source number.（2）When the array manifold matrix is divided by equal sinusoidal,an iterative regularization sparsity adaptive matching pursuit（IR-SAMP）algorithm is proposed.The IR-SAMP algorithm uses regularization method and backtracking screening,and eliminates the inappropriate atoms in the backtracking stage,so as to better approximate the signal source sparsity.When the SNR is low,in order to improve the DOA estimation performance of the IR-SAMP algorithm,the paper then proposes the MMV-SAMP algorithm and the MMV-IR-SAMP algorithm under the multi-measurement vector model,and the proposed MMV-IR-SAMP algorithm still has high DOA estimation accuracy at low SNR.Simulation experiments show that when the array manifold matrix is divided by equal angle,the proposed SAMP algorithm for DOA estimation has a lower complexity with a single snapshot,and the DOA estimation accuracy is higher.This method is also suitable for DOA estimation of coherent signal sources.When the array manifold matrix is divided by equal sinusoidal,the IR-SAMP algorithm proposed in this paper increases the regularization process under a single snapshot and improves the DOA estimation accuracy.At the same time,compared with the IR-SAMP algorithm,the proposed MMV-IR-SAMP algorithm and MMV-SAMP algorithm improve the DOA estimation performance under low SNR,and MMV-IR-SAMP algorithm is better than MMV-SAMP algorithm.