基于扩散模型的室内定位射频指纹数据增强方法

doi:10.11959/j.issn.1000-436x.2023200

通信学报 ›› 2023, Vol. 44 ›› Issue (11): 201-212.doi: 10.11959/j.issn.1000-436x.2023200

• 学术论文 • 上一篇

基于扩散模型的室内定位射频指纹数据增强方法

艾浩军¹^,², 曾维珂¹^,², 陶荆杰¹^,², 徐锦盈¹^,², 常含笑¹^,²

¹ 武汉大学国家网络安全学院，湖北武汉 430072
² 武汉大学空天信息安全与可信计算教育部重点实验室，湖北武汉 430072

修回日期:2023-10-10 出版日期:2023-11-01 发布日期:2023-11-01
作者简介:艾浩军（1972− ），男，湖北汉川人，博士，武汉大学副教授、博士生导师，主要研究方向为普适计算、计算机视觉、无线感知等
曾维珂（2000− ），女，江西南昌人，武汉大学硕士生，主要研究方向为室内定位、计算机视觉
陶荆杰（1999− ），男，江西南昌人，武汉大学硕士生，主要研究方向为普适计算、机器学习
徐锦盈（1999− ），女，湖北武汉人，武汉大学硕士生，主要研究方向为室内定位、计算机视觉
常含笑（2000− ），男，河南济源人，武汉大学硕士生，主要研究方向为机器学习、计算机视觉
基金资助:
国家自然科学基金资助项目(61971316);国家重点研发计划基金资助项目(2016YFB0502204)

Radio frequency fingerprint data augmentation for indoor localization based on diffusion model

Haojun AI¹^,², Weike ZENG¹^,², Jingjie TAO¹^,², Jinying XU¹^,², Hanxiao CHANG¹^,²

¹ School of Cyber Science and Engineering, Wuhan University, Wuhan 430072, China
² Key Laboratory of Aerospace Information Security and Trusted Computing Ministry of Education, Wuhan University, Wuhan 430072, China

Revised:2023-10-10 Online:2023-11-01 Published:2023-11-01
Supported by:
The National Natural Science Foundation of China(61971316);The National Key Research and Development Program of China(2016YFB0502204)

摘要/Abstract

摘要：

射频指纹室内定位方法通过在离线阶段采集足量信号指纹建立密集指纹库保证定位精度。为降低指纹采集成本，提出一种基于扩散模型的射频指纹数据增强方法（FPDiffusion）。首先建立指纹序列的时序图表示，通过高斯加噪方法实现扩散模型的前向过程，反向过程采用U型自编码器网络，根据射频指纹特性设计了网络的损失函数，最后给出了基于稀疏指纹生成稠密指纹的计算过程。实验结果表明，在仅有少量有标签指纹的情况下， FPDiffusion方法在K-近邻（KNN）和卷积神经网络（CNN）算法上的定位误差降低率分别达到76%和28%，在KNN上的定位精度相比高斯过程回归（GPR）和GPR-GAN方法有显著提升。

关键词: 扩散模型, 数据增强, 射频指纹, 室内定位

Abstract:

The radio frequency fingerprint indoor localization method ensures the accuracy by collecting a sufficient amount of fingerprints in the offline state to build a dense fingerprint database.A data augmentation method called FPDiffusion was proposed based on diffusion model to reduce the cost of fingerprint acquisition.Firstly, a temporal graph representation of the fingerprint sequence was constructed, the forward process of the diffusion model was accomplished by adding Gaussian noise, and a U-Net was utilized for the reverse process.The loss function of the network was designed according to the characteristics of radio frequency fingerprints.Finally, the computational process for generating dense fingerprints based on sparse fingerprints was presented.Experimental results demonstrate that FPDiffusion achieves 76% and 28% localization error reduction on K-nearest neighbor (KNN) and convolutional neural network (CNN) respectively, and significantly improves localization accuracy on KNN compared to Gaussian process regression (GPR) and GPR-GAN when only a small amount of labeled fingerprints is available.

Key words: diffusion model, data augmentation, radio frequency fingerprint, indoor localization

中图分类号:

TP391

艾浩军, 曾维珂, 陶荆杰, 徐锦盈, 常含笑. 基于扩散模型的室内定位射频指纹数据增强方法[J]. 通信学报, 2023, 44(11): 201-212.

Haojun AI, Weike ZENG, Jingjie TAO, Jinying XU, Hanxiao CHANG. Radio frequency fingerprint data augmentation for indoor localization based on diffusion model[J]. Journal on Communications, 2023, 44(11): 201-212.

图/表 18

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

表1

图11

图12

图13

图14

表2

图15

表3

应用3种指纹增强方法的ALE和Rd对比"

定位模型	GPR		GPR-GAN		FPDiffusion
定位模型	ALE/m	Rd	ALE/m	Rd	ALE/m	Rd
KNN	3.14	14%	3.12	24%	$3 . 01$	$76 %$
CNN	2.74	25%	2.73	28%	$2 . 73$	$28 %$

表3

参考文献 27

[1]	赵增华, 童跃凡, 崔佳洋 . 基于域自适应的 Wi-Fi 指纹设备无关室内定位模型[J]. 通信学报, 2022,43(4): 143-153.
	ZHAO Z H , TONG Y F , CUI J Y . Device-independent Wi-Fi fingerprinting indoor localization model based on domain adaptation[J]. Journal on Communications, 2022,43(4): 143-153.
[2]	周牧, 王斌, 田增山 ,等. 室内BLE/MEMS跨楼层融合定位算法[J]. 通信学报, 2017,38(5): 1-10.
	ZHOU M , WANG B , TIAN Z S ,et al. Indoor BLE and MEMS based multi-floor fusion positioning algorithm[J]. Journal on Communications, 2017,38(5): 1-10.
[3]	WANG R R , LI Z H , LUO H Y ,et al. A robust Wi-Fi fingerprint positioning algorithm using stacked denoising autoencoder and multi-layer perceptron[J]. Remote Sensing, 2019,11(11): 1293.
[4]	AI H J , SUN X , TAO J J ,et al. DRVAT:exploring RSSI series representation and attention model for indoor positioning[J]. International Journal of Intelligent Systems, 2022,37(7): 4065-4091.
[5]	VASWANI A , SHAZEER N , PARMAR N ,et al. Attention is all you need[C]// Proceedings of the 31st International Conference on Neural Information Processing Systems. New York:ACM Press, 2017: 6000-6010.
[6]	TALVITIE J , RENFORS M , LOHAN E S . Distance-based interpolation and extrapolation methods for RSS-based localization with indoor wireless signals[J]. IEEE Transactions on Vehicular Technology, 2015,64(4): 1340-1353.
[7]	SINHA R S , LEE S M , RIM M ,et al. Data augmentation schemes for deep learning in an indoor positioning application[J]. Electronics, 2019,8(5): 554.
[8]	GOODFELLOW I , POUGET-ABADIE J , MIRZA M ,et al. Generative adversarial networks[J]. Communications of the ACM, 2020,63(11): 139-144.
[9]	NJIMA W , CHAFII M , CHORTI A ,et al. Indoor localization using data augmentation via selective generative adversarial networks[J]. IEEE Access, 2021,9: 98337-98347.
[10]	LAN T , WANG X M , CHEN Z K ,et al. Fingerprint augment based on super-resolution for Wi-Fi fingerprint based indoor localization[J]. IEEE Sensors Journal, 2022,22(12): 12152-12162.
[11]	SUGASAKI M , SHIMOSAKA M . Robustifying Wi-Fi localization by between-location data augmentation[J]. IEEE Sensors Journal, 2022,22(6): 5407-5416.
[12]	KINGMA D P , WELLING M . Auto-encoding variational Bayes[J]. arXiv Preprint,arXiv:1312.6114, 2013.
[13]	SOHL-DICKSTEIN J , WEISS E A , MAHESWARANATHAN N ,et al. Deep unsupervised learning using nonequilibrium thermo dynamics[C]// Proceedings of the 32nd International Conference on International Conference on Machine Learning. New York:ACM Press, 2015: 2256-2265.
[14]	HO J , JAIN A , ABBEEL P . Denoising diffusion probabilistic models[C]// Proceedings of the 34th International Conference on Neural Information Processing Systems. New York:ACM Press, 2020: 6840-6851.
[15]	DHARIWAL P , NICHOL A . Diffusion models beat GANs on image synthesis[J]. arXiv Preprint,arXiv:2105.05233, 2021.
[16]	HO J , SALIMANS T . Classifier-free diffusion guidance[J]. arXiv Preprint,arXiv:2207.12598, 2022.
[17]	ROMBACH R , BLATTMANN A , LORENZ D ,et al. High-resolution image synthesis with latent diffusion models[C]// Proceedings of 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE Press, 2022: 10674-10685.
[18]	NICHOL A , DHARIWAL P , RAMESH A ,et al. GLIDE:towards photorealistic image generation and editing with text-guided diffusion models[J]. arXiv Preprint,arXiv:2112.10741, 2021.
[19]	CHEN K M , CHANG R Y . A comparative study of deep-learning-based semi-supervised device-free indoor localization[C]// Proceedings of IEEE Global Communications Conference (GLOBECOM). Piscataway:IEEE Press, 2022: 1-6.
[20]	KUMAR S , HEGDE R M , TRIGONI N . Gaussian process regression for fingerprinting based localization[J]. Ad Hoc Networks, 2016,51: 1-10.
[21]	ZOU H , CHEN C L , LI M X ,et al. Adversarial learning-enabled automatic Wi-Fi indoor radio map construction and adaptation with mobile robot[J]. IEEE Internet of Things Journal, 2020,7(8): 6946-6954.
[22]	CHAWLA N V , BOWYER K W , HALL L O ,et al. SMOTE:synthetic minority over-sampling technique[J]. Journal of Artificial Intelligence Research, 2002,16: 321-357.
[23]	INOUE H . Data augmentation by pairing samples for images classification[J]. arXiv Preprint,arXiv:1801.02929, 2018.
[24]	ZHANG H , CISSE M , DAUPHIN Y N ,et al. Mixup:beyond empirical risk minimization[J]. arXiv Preprint,arXiv:1710.09412, 2017.
[25]	NICHOL A , DHARIWAL P . Improved denoising diffusion probabilistic models[J]. arXiv Preprint,arXiv:2102.09672, 2021.
[26]	ISOLA P , ZHU J Y , ZHOU T H ,et al. Image-to-image translation with conditional adversarial networks[C]// Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE Press, 2017: 5967-5976.
[27]	HENIGHAN T , KAPLAN J , KATZ M ,et al. Scaling laws for autoregressive generative modeling[J]. arXiv Preprint,arXiv:2010.14701, 2020.

参数符号	参数取值	参数说明
β_t	0.999	前向过程中，为防止结尾的一些轮次方差过大设置的一个上限值
s	0.008	前向过程余弦式噪声策略中的小偏移量
λ	0.001	反向过程总损失函数中L_VLB 的权重，防止方差对模型训练的影响过大，同时对L_VLB应用梯度停止策略
ep	47 000	反向过程网络训练轮次
lr	1×10^-4	学习率
bs	512	批量大小
T	1 000	扩散步骤数最大值

基于扩散模型的室内定位射频指纹数据增强方法

Radio frequency fingerprint data augmentation for indoor localization based on diffusion model

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献 27

相关文章 15

Metrics

推荐阅读 0

[1]	王志恒, 徐彦彦. 室内定位隐私保护综述[J]. 通信学报, 2023, 44(9): 188-204.
[2]	汤凌韬, 王迪, 刘盛云. 面向非独立同分布数据的联邦学习数据增强方案[J]. 通信学报, 2023, 44(1): 164-176.
[3]	谢良波, 李宇洋, 王勇, 周牧, 聂伟. 基于自适应蝙蝠算法的室内RFID定位算法[J]. 通信学报, 2022, 43(8): 90-99.
[4]	赵增华, 童跃凡, 崔佳洋. 基于域自适应的Wi-Fi指纹设备无关室内定位模型[J]. 通信学报, 2022, 43(4): 143-153.
[5]	谢良波, 刘西西, 王勇, 周牧, 田增山. 基于RFID载波相位的室内EKF定位算法[J]. 通信学报, 2022, 43(3): 124-134.
[6]	言洪萍, 周强, 王世豪, 姚旺, 何刘坤, 王良民. 基于SDN的实际网络流中Tor网页复合特征提取方法[J]. 通信学报, 2022, 43(3): 76-87.
[7]	孙大洋, 石文孝, 张定国. 基于室内地图推演与信号穿墙修正的定位优化方法[J]. 通信学报, 2022, 43(10): 146-156.
[8]	王慧强, 高凯旋, 吕宏武. 高精度室内定位研究评述及未来演进展望[J]. 通信学报, 2021, 42(7): 198-210.
[9]	谢良波, 李升, 周牧, 李泽, 田增山, 王亚, 付长友. 基于散射体信息的室内NLOS多站协作定位算法[J]. 通信学报, 2021, 42(5): 63-74.
[10]	周牧, 龙玥辛, 蒲巧林, 王勇, 何维. 基于CSI张量分解的室内Wi-Fi指纹定位方法[J]. 通信学报, 2021, 42(11): 159-171.
[11]	刘桂岐, 钱志鸿, 李华亮, 孙佳妮, 冯一诺, 王雪. 基于有效AP选择和多分类LDA的室内定位算法[J]. 通信学报, 2021, 42(11): 109-120.
[12]	杨国伟, 黄兆标, 樊冰, 周雪芳, 毕美华. 基于可见光通信的室内定位与定向系统[J]. 通信学报, 2020, 41(12): 162-170.
[13]	乐燕芬,汤卓,盛存宝,施伟斌. 基于多分布密度位置指纹的高效室内定位算法研究[J]. 通信学报, 2019, 40(1): 172-179.
[14]	何杰,吴雅南,段世红,徐丽媛,吕家慧,徐诚,齐悦. 人体对UWB测距误差影响模型[J]. 通信学报, 2017, 38(Z1): 58-66.
[15]	周牧,王斌,田增山,张千坤. 室内BLE/MEMS跨楼层融合定位算法[J]. 通信学报, 2017, 38(5): 1-10.

定位模型		ALE/m		Rd
定位模型	E_l	E_m	E_s	Rd
KNN	3.17	2.96	3.01	76%
CNN	2.83	2.47	2.73	28%