一种三维建模的车载雷达数据压缩评估方法  

A Three-Dimensional Modeling Approach for Evaluating Data Compression in Vehicle-Mounted Radar Systems

作　　者：彭家辉 马文豪 

机构地区：[1]上海大学特种光纤与光接入网重点实验室,上海200444 [2]上海毫感科技有限公司,上海201203

出　　处：《工业控制计算机》2025年第2期15-18,共4页Industrial Control Computer

基　　金：上海市科委科技创新行动计划(22511103304)资助。

摘　　要：在复杂交通环境下,车载FMCW毫米波雷达会生成大量数据,因此研究数据有损压缩技术对雷达性能的影响至关重要。然而,目前国内外尚缺乏适用于评估这类有损压缩技术的指标。提出了两种三维参考模型,并将其作为评估指标,用于进行有损压缩误差分析。研究结果显示,由信噪比、压缩比、速度误差或多普勒噪声误差构建的三维参考模型,相较于传统的有损压缩误差评估指标更为直观有效,能够有效展示不同压缩比数据误差对物体识别与检测的影响。这些模型可用于评估为了追求压缩比而带来的精度损失对雷达系统的影响程度,为检测整个FMCW毫米波雷达系统的精度提供了有效的工具。In complex traffic environments,automotive FMCW millimeter-wave radar generates substantial volumes of data,making the study of lossy compression techniques and their impact on radar performance crucial.However,there is currently a lack of suitable metrics for evaluating such lossy compression techniques both domestically and abroad.This paper proposes two three-dimensional reference models as evaluation metrics for conducting error analysis of lossy compression.The findings reveal that the 3D reference model,constructed based on signal-to-noise ratio,compression ratio,velocity error,or Doppler noise error,offers greater intuitiveness and efficacy compared to traditional indices for evaluating lossy compression errors.They effectively demonstrate the impact of data errors at different compression ratios on object recognition and detection.

关 键 词：毫米波雷达 调频连续波 有损压缩 

分 类 号：TN959[电子电信—信号与信息处理] U463.75[电子电信—信息与通信工程]