基于UKF的毫米波雷达手势识别算法研究

蒋留兵; $2; 车 俐

引用本文:	蒋留兵，朱柏青，车俐. 基于UKF的毫米波雷达手势识别算法研究[J]. 雷达科学与技术, 2023, 21(4): 364-374.[点击复制]
	JIANG Liubing, ZHU Boqing, CHE Li. Research on Millimeter Wave Radar Gesture Recognition Algorithm Based on UKF[J]. Radar Science and Technology, 2023, 21(4): 364-374.[点击复制]

【打印本页】【下载PDF全文】【查看/发表评论】【下载PDF阅读器】【关闭】

←前一篇|后一篇→

过刊浏览高级检索

本文已被：浏览 3846次下载 519次	码上扫一扫！
分享到：微信更多字体:加大+\|默认\|缩小-
基于UKF的毫米波雷达手势识别算法研究
蒋留兵，朱柏青，车俐
1. 桂林电子科技大学信息与通信学院，广西桂林 541004;2. 桂林电子科技大学计算机与信息安全学院，广西桂林 541004;3. 广西无线宽带通信与信号处理重点实验室，广西桂林 541004

摘要:

本文提出了一种基于无迹卡尔曼滤波（UKF）的毫米波雷达干扰抑制的手势识别方法。首先根据原始雷达信号设置的采样点与线性调频信号数量，估计了目标的距离与多普勒参数。之后针对实际场景中类目标干扰较多的情况，设计了一套完整的基于UKF的场景类目标抑制方法，接着利用卷积神经网络（CNN）对不同手势距离?多普勒特征谱图进行提取和识别。实验结果表明，该抑制方法有效地解决了类目标干扰给手势识别带来的困扰，手势识别的平均准确率为98.74%，经过抑制干扰算法后准确率相较于干扰抑制之前提升了7.29%。

关键词: 毫米波雷达手势识别无迹卡尔曼滤波类目标干扰抑制

DOI：DOI:10.3969/j.issn.1672-2337.2023.04.002

分类号:TN958.94

基金项目:国家自然科学基金（No.61561010）；广西创新驱动发展专项资助（No.桂科 AA21077008)；“广西无线宽带通信与信号处理重点实验室”2022 年主任基金项目资助（No.GXKL06220102，GXKL06220108）；桂林电子科技大学研究生教育创新计划资助项目（No.2022YXW07）；桂林电子科技大学研究生教育创新计划资助项目（No.2022YCXS080）

Research on Millimeter Wave Radar Gesture Recognition Algorithm Based on UKF

JIANG Liubing, ZHU Boqing, CHE Li

1. School of Information and Communication，Guilin University of Electronic Technology, Guilin 541004, China;2. School of Computer and Information Security, Guilin University of Electronic Technology, Guilin 541004, China;3. Guangxi Key Laboratory of Wireless Broadband Communication and Signal Processing, Guilin 541004, China

Abstract:

In this paper, we proposed a millimeter?wave radar gesture recognition method based on unscented Kalman filter （UKF）. Firstly, the distance and Doppler parameters of the target are estimated according to the sampling points set by the raw radar signals and the number of LFM signals. After that, a complete set of UKF?based scene target?like interference suppression method is designed for the situation that there are many target?like interference in the actual scene. Then, convolution neural network （CNN） is used to extract and identify the distance?Doppler feature spectrum of different gestures. The experimental results show that the suppression method effectively solves the problem of gesture recognition caused by target?like interference. The average accuracy of gesture recognition is 98.74%, and the accuracy is 7.29% higher than that before interference suppression.

Key words: millimeter wave radar gesture recognition unscented Kalman filter target⁃like interference suppression