Scholarlink

<p>With the continuous improvement of autonomous driving capabilities, millimeter wave radar has gradually become one of the key sensors in the automotive field. However, the presence of excessive radar clutter points and sparse effective point clouds severely hampers accurate target detection and reduces the radar’s ability to perceive the surrounding environment, thereby limiting its performance in autonomous driving systems. To address these issues, a high-resolution imaging algorithm based on point cloud is proposed. Firstly, an improved adaptive target detection algorithm based on SO-CFAR is presented to mitigate the problem of target information being overshadowed by clutter. Secondly, a novel DOA joint estimation algorithm is proposed to improve the density and quality of the effective point cloud. The algorithm integrates Conventional Beamforming (CBF), Deterministic Maximum Likelihood Estimation (DML), and Alternating Projection (AP) techniques to accurately determine the azimuth and elevation information of target. To further address noise interference and facilitate target clustering, the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm is employed. Finally, the effectiveness of the proposed method is verified using the TI AWR2243 four-chip cascade radar to collect actual scene data and generate high-resolution point cloud images.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

4D millimeter wave radar high-resolution imaging method based on point cloud

  • 期刊论文

  • 发表日期：

    2026年3月24日

查看全文
  • Jinlong Zhou,
  • Decheng Bao,
  • Renjing Gao

摘要

With the continuous improvement of autonomous driving capabilities, millimeter wave radar has gradually become one of the key sensors in the automotive field. However, the presence of excessive radar clutter points and sparse effective point clouds severely hampers accurate target detection and reduces the radar’s ability to perceive the surrounding environment, thereby limiting its performance in autonomous driving systems. To address these issues, a high-resolution imaging algorithm based on point cloud is proposed. Firstly, an improved adaptive target detection algorithm based on SO-CFAR is presented to mitigate the problem of target information being overshadowed by clutter. Secondly, a novel DOA joint estimation algorithm is proposed to improve the density and quality of the effective point cloud. The algorithm integrates Conventional Beamforming (CBF), Deterministic Maximum Likelihood Estimation (DML), and Alternating Projection (AP) techniques to accurately determine the azimuth and elevation information of target. To further address noise interference and facilitate target clustering, the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm is employed. Finally, the effectiveness of the proposed method is verified using the TI AWR2243 four-chip cascade radar to collect actual scene data and generate high-resolution point cloud images.