面向室内复杂布局的EKF融合定位技术应用研究

打开文本图片集

关键词：宽带MIM0毫米波雷达；视觉系统；融合定位技术；目标识别；扩展卡尔曼滤波；加权平均算法中图分类号： 7；TP391.4 文献标识码：A 文章编号：2096-4706（2025）19-0016-06

Abstract： Focusing on the problem poor accuracy in target positioning trajectory tracking in complex indoor layouts，researchers conduct in-depth exploration EKF fusion positioning technology.This paper proposes apositioning method based on thefusion broadb MIMO milimeter wave radar visual information，which leverages the advantages millimeter wave radar to accuratelyobtain target information has strong cluter suppresion capability.Aimingat its shortcomingsinargetRecogitionaninovativemultimodaldatafusioarchitectureisostructed.Tepixel-leveldepfusion isachievedthroughtheWeightedAverageAlgorithm，thedataisoptimizedviatheEKFalgorithmtosolvethedrawbacks traditionalmethods.Experimentshowthatthekeyindicatorsthismethodaresignificantlyimproved，providingaeliable solutionfor indoor positioning tracking， promoting the development multi-sensor fusion technology.

Keywords：broadb MIMO milimeter wave radar; visual system; fusion positioning technology; Target Recognition; ExtendedKalmanFilter;WeightedAverageAlgorithm

0 引言

宽带MIMO （Multiple-Input Multiple-Output）毫米波雷达融合了宽带、MIMO及毫米波雷达特性，它发射 77GHz 调频连续波信号，经混频信号处理、快速傅里叶变换与二维恒虚警检测，精准获取目标的距离、速度、角度信息，实现室内目标定位与跟踪[-3]，但雷达在行人与类似障碍物识别上存在不足，而视觉系统可直观获取环境信息，擅长颜色感知与目标分类，可有效检测行人，定位精度较高，但易受光照干扰，复杂场景识别困难[46]

基于宽带MIMO毫米波雷达与视觉融合的定位技术，结合了雷达的主动探测性、高灵敏度和视觉系统的可视性，充分发挥了两者的优势[。（剩余7530字）