[1]高 鹏,张 岩,唐新余,等.结合注意力机制的雷达多信号动作识别方法[J].计算机技术与发展,2023,33(01):157-164.[doi:10. 3969 / j. issn. 1673-629X. 2023. 01. 024]
　GAO Peng,ZHANG Yan,TANG Xin-yu,et al.Radar Multi-signal Action Recognition Method Based on Attention Mechanism[J].,2023,33(01):157-164.[doi:10. 3969 / j. issn. 1673-629X. 2023. 01. 024]
点击复制

结合注意力机制的雷达多信号动作识别方法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
33
期数:
2023年01期
页码:
157-164
栏目:
人工智能
出版日期:
2023-01-10

文章信息/Info

Title:
Radar Multi-signal Action Recognition Method Based on Attention Mechanism
文章编号:
1673-629X(2023)01-0157-08
作者:
高 鹏1; 2; 3; 4 ; 张 岩1; 3 ; 唐新余1; 2; 3; 4 ; 王 蒙1; 2; 3; 4 ; 季文飞4
1. 中国科学院 新疆理化技术研究所,新疆 乌鲁木齐 830011;
2. 中国科学院大学,北京 100049;
3. 中国科学院 新疆民族语音语言信息处理重点实验室,新疆 乌鲁木齐 830011;
4. 江苏中科西北星信息科技有限公司,江苏 无锡 214135
Author(s):
GAO Peng1; 2; 3; 4 ; ZHANG Yan1; 3 ; TANG Xin-yu1; 2; 3; 4 ; WANG Meng1; 2; 3; 4 ; JI Wen-fei4
1. The Xinjiang Technical Institute of Physics and Chemistry,Chinese Academy of Sciences,Urumqi 830011,China;
2. University of Chinese Academy of Sciences,Beijing 100049,China;
3. Xinjiang Laboratory of Minority Speech and Language Information Processing,Chinese Academy of Sciences,Urumqi 830011,China;
4. Jiangsu CAS Nor-West Star Information Technology Co. ,Ltd. ,Wuxi 214135,China
关键词:
人体动作识别; 毫米波雷达; 多信号特征; 卷积神经网络; 注意力机制
Keywords:
human motion recognition; millimeter wave radar; multi-signal characteristics; convolutional neural network; attention mechanism
分类号:
TP391. 4
DOI:
10. 3969 / j. issn. 1673-629X. 2023. 01. 024
摘要:
雷达动作特征谱图对走、摔等较为宽幅的人体动作表征效果差,单一特征和不匹配特征数据结构的分类方法会降低动作识别的性能。 针对以上问题,提出一种结合注意力机制的雷达多信号特征动作识别方法。 首先,使用配置时分复用模式的多输入多输出毫米波雷达采集动作回波,将回波处理成短时能量、频率质心、相位变化( 水平、俯仰) 四维时序信号特征;然后,根据信号特征数据结构设计了多信号序列融合分类网络,该网络由 1DCNN 对信号抽取高维特征,再将特征送入 GRU 以充分提取时序规律,并引入 Attention 机制对重要特征映射加权赋予 GRU 隐含状态不同的权重,最终通过SoftMax 层完成动作分类;最后,在实际采集的雷达多信号数据集上进行实验,结果表明,多信号序列特征可以充分表征人体动作,所设计的网络收敛速度快,对 8 种不同的动作分类,平均正确率达到了 98. 5% 。
Abstract:
The radar activity feature spectrum has poor characterization effect on wide human actions such as walking and falling. Thesingle action feature and mismatch feature data structure can reduce the performance of activity recognition. Aiming at above problems,aradar multi-signal feature extraction method combining attention mechanism is proposed. Firstly,such method 　uses the multiple inputmultiple output millimeter-wave radar with time-division multiplexing mode to collect the action echo which is processed into four-dimensional time series signal 　features of short-term energy,frequency centroid and phase change ( horizontal,pitch) . Then,a multi-signalsequence fusion classification network is designed according to the signal feature data structure. In this network,1DCNN extracts high-dimensional features from signals,and then sends the features to GRU to fully extract timing rules. Besides,the Attention mechanism isintroduced to map important features and assign different weights to hidden states of GRU,and action classification is completed throughSoftMax layer. Finally,experiments on the radar multi-signal data set collected in practice show that the multi-signal features sequencecan fully characterize human actions,and the designed network has a fast convergence speed. The average accuracy of 8 different actionsis 98. 5% .

相似文献/References:

[1]宫法明,马玉辉.基于时空双分支网络的人体动作识别研究[J].计算机技术与发展,2020,30(09):23.[doi:10. 3969 / j. issn. 1673-629X. 2020. 09. 005]
　GONG Fa-ming,MA Yu-hui.Research on Human Action Recognition Based on Space-time Double-branch Network[J].,2020,30(01):23.[doi:10. 3969 / j. issn. 1673-629X. 2020. 09. 005]
[2]赵越坤,罗素云,魏 丹,等.基于毫米波雷达和视觉的目标检测方法[J].计算机技术与发展,2023,33(06):35.[doi:10. 3969 / j. issn. 1673-629X. 2023. 06. 006]
　ZHAO Yue-kun,LUO Su-yun,WEI Dan,et al.Target Detection Method Based on Millimeter-wave Radar and Vision[J].,2023,33(01):35.[doi:10. 3969 / j. issn. 1673-629X. 2023. 06. 006]
[3]朱联祥,牛文煜,仝文东,等.基于混合注意力机制的视频人体动作识别[J].计算机技术与发展,2023,33(09):105.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 016]
　ZHU Lian-xiang,NIU Wen-yu,TONG Wen-dong,et al.Video Human Action Recognition Based on Hybrid Attention Mechanism[J].,2023,33(01):105.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 016]

更新日期/Last Update: 2023-01-10