基于PYNQ-Z2的机械通气后遗症预测模型嵌入式实现研究

金紫怡; 朱之晨; 杜江; 陈仪香

doi:10.20193/j.ices2097-4191.2025.0061

PDF(6997 KB)

集成电路与嵌入式系统 ›› 2025, Vol. 25 ›› Issue (12) : 33-39. DOI: 10.20193/j.ices2097-4191.2025.0061

智能嵌入式系统软硬件协同设计与应用专栏

基于PYNQ-Z2的机械通气后遗症预测模型嵌入式实现研究

金紫怡 ¹ ,
朱之晨 ² ,
杜江 ³ ,
陈仪香 ¹^,^*

作者信息 +

Research on embedded implementation of mechanical ventilation sequelae prediction model based on PYNQ-Z2

JIN Ziyi ¹ ,
ZHU Zhichen ² ,
DU Jiang ³ ,
CHEN Yixiang ¹^,^*

Author information +

文章历史 +

摘要

介绍了PVAC模型的嵌入式部署,旨在预测急性呼吸衰竭患者发生呼吸机相关后遗症(VAC)的风险。PVAC模型通过USMOTE(0.9)算法处理不平衡数据,并结合AdaBoost分类器实现了71.11%的准确率和68.89%的精确度。为了克服现有AI医疗系统依赖云端服务器的局限性,采用PYNQ-Z2开发板实现了PVAC模型的完全嵌入式部署。该方案具有离线独立运行、硬件加速提升计算效率和成本优势三大特点。实验结果表明,软硬件协同方案相比纯软件方案,总执行时间从46.3 ms显著降低至10.2 ms,提速幅度达到78%,ARM处理器的负载从98%大幅降至28%,而模型预测准确率仅下降0.2%,基本保持原有性能水平,不仅验证了PVAC模型嵌入式化的可行性,还为其他医疗AI应用的本地化部署提供了参考。未来可进一步优化决策树结构,利用FPGA动态可重构特性支持更复杂的模型,扩展对时序信号的处理能力,开发低功耗模式来延长设备使用时间、提升系统的实用性、扩大适用范围。

Abstract

This paper presents the embedded deployment of the PVAC model to predict the risk of ventilator-associated complications (VAC) in patients with acute respiratory failure. The PVAC model employs the USMOTE (0.9) algorithm to address imbalanced data and integrates an AdaBoost classifier, achieving an accuracy of 71.11% and a precision of 68.89%. To overcome the limitations of existing AI medical systems that rely on cloud servers, we implemented a fully embedded deployment of the PVAC model using the PYNQ-Z2 development board. This solution offers three key advantages: offline standalone operation, hardware acceleration for improved computational efficiency, and cost-effectiveness. Experimental results demonstrate that the hardware-software co-design approach significantly reduces the total execution time from 46.3 ms to 10.2 ms, achieving a speedup of 78%. Meanwhile, the ARM processor's workload decreases dramatically from 98% to 28%, with only a 0.2% drop in prediction accuracy, effectively preserving the model's original performance. This study not only validates the feasibility of embedding the PVAC model but also provides a reference for the localized deployment of other medical AI applications. Future work may focus on further optimizing the decision tree structure, leveraging the dynamic reconfigurability of FPGAs to support more complex models, extending the capability to process temporal signals, and developing low-power modes to extend device usage time, thereby enhancing the system's practicality and applicability.

导出引用

金紫怡, 朱之晨, 杜江, 等. 基于PYNQ-Z2的机械通气后遗症预测模型嵌入式实现研究[J]. 集成电路与嵌入式系统. 2025, 25(12): 33-39 https://doi.org/10.20193/j.ices2097-4191.2025.0061

JIN Ziyi, ZHU Zhichen, DU Jiang, et al. Research on embedded implementation of mechanical ventilation sequelae prediction model based on PYNQ-Z2[J]. Integrated Circuits and Embedded Systems. 2025, 25(12): 33-39 https://doi.org/10.20193/j.ices2097-4191.2025.0061

中图分类号： TP23 (自动化装置与设备)

参考文献

列表( 原文顺序 | 文献年度倒序 | 文中引用次数倒序 ) 可视化分析

[1]	N K ADHIKARI, R A FOWLER, S BHAGWANJEE, et al. Critical care and the global burden of critical illness in adults[J]. The Lancet, 2010, 376(9749):1339-1346. https://doi.org/10.1016/S0140-6736(10)60446-1 https://linkinghub.elsevier.com/retrieve/pii/S0140673610604461 本文引用 [1]

[2]

URNER

, P

JUNI

, B

HANSEN

, et al. Time-varying intensity of mechanical ventilation and mortality in patients with acute respiratory failure:a registry-based, prospective cohort study[J]. The Lancet Respiratory Medicine, 2020, 8(9):905-913.

https://doi.org/10.1016/S2213-2600(20)30325-8

https://linkinghub.elsevier.com/retrieve/pii/S2213260020303258

本文引用 [1]

[3]	S RAOOF, M H BAUMANN, C C S COLLABORATIVE. An official multisociety statement:Ventilator-associated events:the new definition[J]. Critical Care Medicine, 2014. 本文引用 [1]

[4]	张佳, 徐静. 基于SMOTE算法的急性缺血性脑卒中患者发生吞咽功能障碍风险预警模型的构建[J]. 医学信息, 2025, 38(12):6-11. ZHANG J, XU J. Development of a Risk Warning Model for Dysphagia in Patients with Acute Ischemic Stroke Based on the SMOTE Algorithm[J]. Journal of Medical Information, 2025, 38(12):6-11 (in Chinese). 本文引用 [1]

[5]	R A BAUDER, T HASANIN.T M KHOSHGOFTAAR, Data sampling approaches with severely imbalanced big data for medicare fraud detection[C]// 2018 IEEE 30th international conference on tools with artificial intelligence (ICTAI).IEEE, 2018:137-142. 本文引用 [1]

[6]	ZHU Z C, DU J, KANG N, et al. A Model for the Prediction of Ventilator-Associated Conditions in Patients with Acute Respiratory Failure[J]. Medical Artificial Intelligence, 2024, 1(1):1-10. 本文引用 [1]

[7]	N S KUMAR, K N RAO, A GOVARDHAN, et al. Undersampled k-means approach for handling imbalanced distributed data[J]. Progress in Artificial Intelligence, 2014, 3(1):29-38. https://doi.org/10.1007/s13748-014-0045-6 http://link.springer.com/10.1007/s13748-014-0045-6 本文引用 [1]

[8]	孙愉婷. 基于ZYNQ的光谱图像处理与加速设计[D]. 济南: 山东大学, 2024.DOI:10.27272/d.cnki.gshdu.202.0044044. SUN Y T. Design of Spectral Image Processing and Acceleration Based on ZYNQ[D]. Jinan: Shandong University, 2024.DOI:10.27272/d.cnki.gshdu.2024.004404 (in Chinese). 本文引用 [1]

[9]	A LIU, J GHOSH, C MARTIN. Generative oversampling for mining imbalanced datasets[J]. DMIN, 2007, 7:66-72. 本文引用 [1]

[10]

V CHAWLA

, K

W BOWYER

, L

O HALL

, et al. Smote: synthetic minority over-sampling technique[J]. Journal of artificial intelligence research, 2002, 16:321-357.

https://doi.org/10.1613/jair.953

https://www.jair.org/index.php/jair/article/view/10302

本文引用 [1] 摘要

An approach to the construction of classifiers from imbalanced datasets is described. A dataset is imbalanced if the classification categories are not approximately equally represented. Often real-world data sets are predominately composed of ``normal'' examples with only a small percentage of ``abnormal'' or ``interesting'' examples. It is also the case that the cost of misclassifying an abnormal (interesting) example as a normal example is often much higher than the cost of the reverse error. Under-sampling of the majority (normal) class has been proposed as a good means of increasing the sensitivity of a classifier to the minority class. This paper shows that a combination of our method of over-sampling the minority (abnormal) class and under-sampling the majority (normal) class can achieve better classifier performance (in ROC space) than only under-sampling the majority class. This paper also shows that a combination of our method of over-sampling the minority class and under-sampling the majority class can achieve better classifier performance (in ROC space) than varying the loss ratios in Ripper or class priors in Naive Bayes. Our method of over-sampling the minority class involves creating synthetic minority class examples. Experiments are performed using C4.5, Ripper and a Naive Bayes classifier. The method is evaluated using the area under the Receiver Operating Characteristic curve (AUC) and the ROC convex hull strategy.

[11]

许巾一, 陈仪香, 李凯旋. 异构分布式嵌入式系统的优化设计方法[J]. 微纳电子与智能造, 2020, 2(1):45-55.DOI:10.19816/j.cnki.10-1594/tn.2020.01.045.

J Y

, CHEN

Y X

, LI

K X

. An Optimized Design Method for Heterogeneous Distributed Embedded Systems[J]. Micro/Nano Electronics and Intelligent Manufacturing, 2020, 2(1):45-55.DOI:10.19816/j.cnki.10-1594/tn.2020.01.045 (in Chinese).

本文引用 [1]

[12]	H HE, Y BAI, E A GARCIA, et al. Adasyn:Adaptive synthetic sampling approach for imbalanced learning[C]// 2008 IEEE international joint conference on neural networks (IEEE world congress on computational intelligence).IEEE, 2008:1322-1328. 本文引用 [1]

[13]	WANG M, YAO X, CHEN Y. An imbalanced-data processing algorithm for the prediction of heart attack in stroke patients[J]. IEEE Access, 2021, 9:25394-25404. https://doi.org/10.1109/Access.6287639 https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=6287639 本文引用 [1]