基于压缩感知的5G雷达目标距离和速度联合估计方法

Joint Estimation of Target Range and Velocity for 5G Radar Based on Compressive Sensing

随着无线通信5G/6G技术的发展,利用5G雷达实现城市环境下低空目标探测已成为新兴业务,而目标参数估计更是5G雷达目标探测的重要任务之一。目前基于5G雷达的目标参数估计采用基于子空间投影的高分辨处理方法,该方法需要大量的采样数据,计算复杂度比较高,导致目标参数估计性能在工程应用中还不太理想,而且对于相关性比较高的目标,还需解相关处理。本文利用目标距离的时域稀疏性和目标多普勒的频域稀疏性,将目标距离和多普勒估计归结于稀疏表示问题,从而提出一种基于压缩感知的5G雷达目标距离和速度联合估计方法。所提方法首先通过在过完备基的基础上实现阵列协方差矩阵的稀疏表示,然后构建协方差向量的L1-范数约束优化问题,最后利用内点法高效求解二阶锥规划问题,通过联合寻找阵列协方差向量的稀疏系数来实现目标参数的估计。所提算法具有较高的目标参数估计分辨率和较强的相关信号源处理能力。而且所提算法在构建优化问题时还给出了明确的误差抑制准则,该准则使其即使在低信噪比(signal-to-noise ratio,SNR)的情况下也具有统计鲁棒性。仿真实验结果表明,所提算法能够有效地实现5G雷达目标距离和速度的高分辨联合估计,同时不需要解相干处理;而且所提算法在较低信噪比环境下能够稳健地实现目标参数估计。

With the development of 5G/6G technology in wireless communication,the use of 5G radar to detect low-altitude targets in urban environments has given rise to new business activities,and target parameter estimation has been one of the important tasks of 5G radar target detection.Currently,the target parameter estimation for 5G radar uses the high-resolution processing method based on subspace projection.Subspace-based methods require a large segment of consecutive samples and suffer from high computational complexity.The target parameter estimation performance of subspace-based methods is not ideal for practical engineering.Moreover,high-correlation targets necessitate correlation processing.In this study,the target range and Doppler estimation are reduced to sparse representation using the sparsity of the target range in a time domain and the sparsity of the target Doppler in a frequency domain.Further,a joint estima‐tion method of target range and velocity for 5G radar based on compressed sensing was proposed to achieve radar sens‐ing along with wireless communication.First,a signal model was formulated to represent the target echo,while the transmit waveform made use of orthogonal frequency division multiplexing communication signals.Subsequently,the proposed method achieved a sparse representation of the array covariance matrix based on an overcomplete basis.Then,an L1-norm-constrained optimization problem was constructed for the covariance vector.Finally,the second-order cone programming problem was efficiently solved to determine the sparse vector coefficient using the interior point method.The objective parameter estimation was achieved by jointly searching for the sparse coefficients of the array covariance vector.Compared with the subspace-based methods,the proposed method has high resolution and can estimate coherent signals based on an arbitrary array.More importantly,the proposed algorithm also provides a clear error suppression cri‐terion when constructing optimization problems,rendering it statistically robust even in low signal-to-noise ratio(SNR)situations.Simulation results show that the proposed algorithm can achieve a jointly high-resolution estimation of target range and velocity for 5G radar without the need for decoherence processing.Moreover,the proposed algorithm can real‐ize the target parameter estimation robustly in low-SNR environments.