基于大语言模型的超宽带雷达多任务学习方法

饶翀; 姜夕康; 郭嘉航; 李蕾; 张琳

doi:null

您当前的位置：

首页 >

文章列表页 >

基于大语言模型的超宽带雷达多任务学习方法

更新时间：2026-03-03

- 基于大语言模型的超宽带雷达多任务学习方法
- UWB-LLM: ultra-wideband radar multi-tasking based on large language model
- 物联网学报 2026年
- 作者机构：
  
  1.北京邮电大学人工智能学院，北京 100876
  2.北京市大数据中心，北京 101160
- 作者简介：
  
  张琳，zhanglin@bupt.edu.cn
- 基金信息：
  
  国家自然科学基金资助项目(61971056)
- DOI：
  中图分类号：
- 收稿：2025-07-14，
  
  修回：2025-11-17，
  
  录用：2026-02-09，
- 稿件说明：
移动端阅览
饶翀, 姜夕康, 郭嘉航, 等. 基于大语言模型的超宽带雷达多任务学习方法[J/OL]. 物联网学报, 2026.

RAO Chong, JIANG Xikang, GUO Jiahang, et al. UWB-LLM: ultra-wideband radar multi-tasking based on large language model[J/OL]. Chinese Journal on Internet of Things, 2026.
饶翀, 姜夕康, 郭嘉航, 等. 基于大语言模型的超宽带雷达多任务学习方法[J/OL]. 物联网学报, 2026. DOI：

RAO Chong, JIANG Xikang, GUO Jiahang, et al. UWB-LLM: ultra-wideband radar multi-tasking based on large language model[J/OL]. Chinese Journal on Internet of Things, 2026. DOI：

摘要

超宽带雷达通过无线感知实现人数检测和生命体征监测，现有方法依赖统计特征或轻量级神经网络，在多种任务间的迁移与泛化能力有限。鉴于大语言模型的优异跨模态学习能力，提出了基于大语言模型的超宽带雷达多任务微调框架UWB-LLM，结合混合专家与低秩适应方法实现将雷达信号的时序特征映射至大语言模型嵌入空间并高效微调。在自采集数据集和公开数据集上，分别对人数检测任务及呼吸、心电图与连续血压三种生命体征信号估计任务进行训练和评估。实验结果表明UWB-LLM在人数检测准确率和三种生命体征信号估计任务的相关系数较现有算法平均提升了37.62%、7.47%、18.16%和14.70%（相关代码与数据集已开源

）。

Abstract

Ultra-wideband (UWB) radar enables people counting and vital signs monitoring with wireless sensing. However

existing methods rely on statistical features or lightweight neural networks

which h

ave limited ability to migrate and generalize across multiple tasks. In view of the powerful cross-modal learning capability demonstrated by large language models (LLMs)

a UWB radar multi-task learning framework based on LLM

UWB-LLM

is proposed. This approach successfully maps the temporal features of radar signals into the embedding space of LLMs

and by integrating the mixture of experts (MoE) mechanism and low-rank adaptation (LoRA) strategy

UWB-LLM performs parameter-efficient fine-tuning (PEFT) for multi-task learning. Experiments were conducted on both self-collected and publicly available datasets for people counting task as well as respiration

electrocardiogram (ECG) and continuous blood pressure estimation tasks

respectively. Compared to the state-of-the-art algorithms

UWB-LLM achieves average accuracy improvements of 37.62% for people counting and 7.47%

18.16%

and 14.70% for the three vital sign estimation tasks (The relevant code and datasets are available

关键词

Keywords

references

刘洋 , 董安明 , 禹继国 , 等 . 基于CNN-BiGPU的复杂连续人体活动Wi-Fi感知方法 [J ] . 物联网学报 , 2023 , 7 ( 4 ): 153 - 167 .

LIU Y , DONG A M , YU J G , et al . A Wi-Fi sensing method for complex continuous human activities based on CNN-BiGPU [J ] . Chinese Journal of Internet of Things , 2023 , 7 ( 4 ): 153 - 167 .

DI DOMENICO S , PECORARO G , CIANCA E , et al . Trained-once device-free crowd counting and occupancy estimation using WiFi: A Doppler spectrum based approach [C ] // 2016 IEEE 12th International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob) . Piscataway : IEEE Press , 2016 : 1 - 8 .

吴彤 , 李业深 , 黄镇煌 , 等 . 一种行人遮挡下的UWB非视距传播识别方法 [J ] . 物联网学报 , 2023 , 7 ( 4 ): 63 - 71 .

WU T , LI Y S , HUANG Z H , et al . A UWB NLOS identification method under pedestrian occlusion [J ] . Chinese Journal on Internet of Things , 2023 , 7 ( 4 ): 63 - 71 .

CHOI J W , YIM D H , CHO S H . People counting based on an IR-UWB radar sensor [J ] . IEEE Sensors Journal , 2017 , 17 ( 17 ): 5717 - 5727 .

YANG X , YIN W , LI L , et al . Dense people counting using IR-UWB radar with a hybrid feature extraction method [J ] . IEEE Geoscience and Remote Sensing Letters , 2019 , 16 ( 1 ): 30 - 34 .

YANG X Z , YIN W F , ZHANG L . People counting based on CNN using IR-UWB radar [C ] // 2017 IEEE/CIC International Conference on Communications in China (ICCC) . Piscataway : IEEE Press , 2017 : 1 - 5 .

BAO R , YANG Z . CNN-based regional people counting algorithm exploiting multi-scale range-time maps with an IR-UWB radar [J ] . IEEE Sensors Journal , 2021 , 21 ( 12 ): 13704 - 13713 .

CHOI J H , KIM J E , KIM K T . Deep learning approach for radar-based people counting [J ] . IEEE Internet of Things Journal , 2022 , 9 ( 10 ): 7715 - 7730 .

WANG D Y , YOO S , CHO S H . Experimental comparison of IR-UWB radar and FMCW radar for vital signs [J ] . Sensors , 2020 , 20 ( 22 ): 6695 .

SCHELLENBERGER S , SHI K , STEIGLEDER T , et al . A dataset of clinically recorded radar vital signs with synchronised reference sensor signals [J ] . Scientific Data , 2020 , 7 : 291 .

CHEN J , ZHANG D , WU Z , et al . Contactless electrocardiogram monitoring with millimeter wave radar [J ] . IEEE Transactions on Mobile Computing , 2024 , 23 ( 1 ): 270 - 285 .

OYAMADA Y , KOSHISAKA T , SAKAMOTO T . Experimental demonstration of accurate noncontact measurement of arterial pulse wave displacements using 79-GHz array radar [J ] . IEEE Sensors Journal , 2021 , 21 ( 7 ): 9128 - 9137 .

QIU Y , MA X , LI X , et al . Non-Contact blood pressure estimation from radar signals by a stacked deformable convolution network [J ] . IEEE Journal of Biomedical and Health Informatics , 2024 , 28 ( 8 ): 4553 - 4564 .

李鑫尧 , 李晶晶 , 朱磊 , 等 . 资源受限的大模型高效迁移学习算法研究综述 [J ] . 计算机学报 , 2024 , 47 ( 11 ): 2491 - 2521 .

LI X Y , LI J J , ZHU L , et al . Efficient transfer learning of large language models with limited resources: a survey [J ] . Chinese Journal of Computers , 2024 , 47 ( 11 ): 2491 - 2521 .

ALAYRAC J B , DONAHUE J , LUC P , et al . Flamingo: a visual language model for few-shot learning [C ] // Proceedings of the 36th International Conference on Neural Information Processing Systems (NIPS '22) . Red Hook : Curran Associates Inc. , 2022 : 23716 – 23736 .

Large Generative AI in Models in Telecom (GenAINet) , IEEE ComSoc Emerging Technology Initiative . Large-Scale AI in Telecom: Charting the Roadmap for Innovation, Scalability, and Enhanced Digital Experiences [R ] . 2025 .

LIU L , CUI G , WAN C , et al . ECG-LLM: leveraging large language models for low-quality ECG signal restoration [C ] // 2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) , Piscataway : IEEE Press , 2024 : 3537 - 3542 .

LI Z , ZHENG W L , LU B L . Gram: a large-scale general EEG model for raw data classification and restoration tasks [C ] // 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , Piscataway : IEEE Press , 2025 : 1 - 5 .

LIU B , LIU X , GAO S , et al . LLM4CP: adapting large language models for channel prediction [J ] . Journal of Communications and Information Networks , 2024 , 9 ( 2 ): 113 - 125 .

LIU X , GAO S , LIU B , et al . LLM4WM: Adapting LLM for Wireless Multi-Tasking [J ] . IEEE Transactions on Machine Learning in Communications and Networking , 2025 , 3 : 835 - 847 .

SHENG Y , HUANG K , LIANG L , et al . Beam prediction based on large language models [J ] . IEEE Wireless Communications Letters , 2025 , 14 ( 5 ): 1406 - 1410 .

LIU Q , WU X , ZHAO X , et al . When MOE meets LLMs: parameter efficient fine-tuning for multi-task medical applications [C ] // Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR '24 ), New York : Association for Computing Machinery , 2024 : 1104 - 1114 .

SINGH A , GAO X , YAVARI E , et al . Data-based quadrature imbalance compensation for a CW Doppler radar system [J ] . IEEE Transactions on Microwave Theory and Techniques , 2013 , 61 ( 4 ): 1718 - 1724 .

CHOWDHURY F A , HOSAIN M K , ISLAM M S B , et al . ECG waveform generation from radar signals: A deep learning perspective [J ] . Computers in Biology and Medicine , 2024 , 176 : 108555 .

LIU S , JOHNS E , DAVISON A J . End-to-end multi-task learning with attention [C ] // 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , Piscataway : IEEE Press , 2019 : 1871 - 1880 .

CAKONI D , STORRER L , CORNELIS B , et al . Outdoor group counting based on micro-Doppler signatures obtained with a 77GHz FMCW radar [C ] // 2024 21st European Radar Conference (EuRAD) , Piscataway : IEEE Press , 2024 : 376 - 379 .

MACH T K T , PHAM C T , LE M . UWB impulse radar for people counting with convolutional neural network on microcontrollers [J ] . IEEE Sensors Journal , 2024 , 24 ( 9 ): 1 5643– 15650 .

TODA D , ANZAI R , ICHIGE K , et al . ECG signal reconstruction using FMCW radar and convolutional neural network [C ] // 2021 20th International Symposium on Communications and Information Technologies (ISCIT) , Piscataway : IEEE Press , 2021 : 176 - 181 .

YAMAMOTO K , HIROMATSU R , OHTSUKI T . ECG signal reconstruction via Doppler sensor by hybrid deep learning model with CNN and LSTM [J ] . IEEE access , 2020 , 8 : 130551 - 130560 .

CHEN J , ZHANG D , WU Z , et al . Contactless electrocardiogram monitoring with millimeter wave radar [J ] . IEEE Transactions on Mobile Computing , 2024 , 23 ( 1 ): 270 - 285 .

WU Y , NI H , MAO C , et al . Contactless reconstruction of ECG and respiration signals with mmWave radar based on RSSRnet [J ] . IEEE Sensors Journal , 2024 , 24 ( 5 ): 6358 - 6368 .

GADDI M , PONZINA F , ASGARINEJAD F , et al . HyperECG: ECG signal inference from radar with hyperdimensional computing [C ] // 2024 IEEE 24th International Conference on Bioinformatics and Bioengineering (BIBE) , Piscataway : IEEE Press , 2024 : 1 - 5 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

暂无数据