The Library
Optimising deep learning at the edge for accurate hourly air quality prediction
Tools
Tools
Tools
Wardana, I. Nyoman Kusuma , Gardner, J. W. and Fahmy, Suhaib A. (2021) Optimising deep learning at the edge for accurate hourly air quality prediction. Sensors, 21 (4). 1064. doi:10.3390/s21041064 ISSN 1424-8220.
|
PDF
WRAP-Optimising-deep-learning-edge-accurate-hourly-air-quality-prediction-2021.pdf - Published Version - Requires a PDF viewer. Available under License Creative Commons Attribution 4.0. Download (17Mb) | Preview |
|
![[img]](https://wrap.warwick.ac.uk/style/images/fileicons/application_pdf.png) |
PDF
WRAP-Optimising-deep-learning-edge-accurate-hourly-air-quality-prediction-2021.pdf - Accepted Version Embargoed item. Restricted access to Repository staff only - Requires a PDF viewer. Download (13Mb) |
Official URL: https://doi.org/10.3390/s21041064
Abstract
Accurate air quality monitoring requires processing of multi-dimensional, multi-location sensor data, which has previously been considered in centralised machine learning models. These are often unsuitable for resource-constrained edge devices. In this article, we address this challenge by: (1) designing a novel hybrid deep learning model for hourly PM2.5 pollutant prediction; (2) optimising the obtained model for edge devices; and (3) examining model performance running on the edge devices in terms of both accuracy and latency. The hybrid deep learning model in this work comprises a 1D Convolutional Neural Network (CNN) and a Long Short-Term Memory (LSTM) to predict hourly PM2.5 concentration. The results show that our proposed model outperforms other deep learning models, evaluated by calculating RMSE and MAE errors. The proposed model was optimised for edge devices, the Raspberry Pi 3 Model B+ (RPi3B+) and Raspberry Pi 4 Model B (RPi4B). This optimised model reduced file size to a quarter of the original, with further size reduction achieved by implementing different post-training quantisation. In total, 8272 hourly samples were continuously fed to the edge device, with the RPi4B executing the model twice as fast as the RPi3B+ in all quantisation modes. Full-integer quantisation produced the lowest execution time, with latencies of 2.19 s and 4.73 s for RPi4B and RPi3B+, respectively. View Full-Text
10 selecting the best model, we optimise the model for edge devices, using Raspberry Pi 3 Model B+ 11 (RPi3B+) and Raspberry Pi 4 Model B boards (RPi4B). The lite version produced 4 times smaller
12 file size compared to the original version. From the lite version, further size reduction can be 13 achieved by implementing different post-training quantisations. About a 47% reduction can be
14 achieved by dynamic range quantisation, about 45% by full integer quantisation, and about 35% 15 by float16 quantisation. A total of 8272 hourly samples were continuously executed directly at the
16 edge. The RPi4B executed these data two times faster compared to the RPi3B+ in all quantisation 17 modes. Full-integer quantisation produced the most effective time execution, with latencies of
18 2.19 seconds and 4.73 seconds for RPi4B and RPi3B+, respectively.
| Item Type: | Journal Article | ||||||
|---|---|---|---|---|---|---|---|
| Subjects: | Q Science > QA Mathematics > QA76 Electronic computers. Computer science. Computer software Q Science > QH Natural history T Technology > TD Environmental technology. Sanitary engineering |
||||||
| Divisions: | Faculty of Science, Engineering and Medicine > Engineering > Engineering | ||||||
| Library of Congress Subject Headings (LCSH): | Air quality , Air quality management, Air -- Pollution -- Measurement , Machine learning , Edge computing, Air quality -- Computer simulation | ||||||
| Journal or Publication Title: | Sensors | ||||||
| Publisher: | MDPI AG | ||||||
| ISSN: | 1424-8220 | ||||||
| Official Date: | 4 February 2021 | ||||||
| Dates: |
|
||||||
| Volume: | 21 | ||||||
| Number: | 4 | ||||||
| Article Number: | 1064 | ||||||
| DOI: | 10.3390/s21041064 | ||||||
| Status: | Peer Reviewed | ||||||
| Publication Status: | Published | ||||||
| Access rights to Published version: | Open Access (Creative Commons) | ||||||
| Date of first compliant deposit: | 1 February 2021 | ||||||
| Date of first compliant Open Access: | 1 February 2021 | ||||||
| RIOXX Funder/Project Grant: |
|
||||||
| Related URLs: |
Request changes or add full text files to a record
Repository staff actions (login required)
 |
View Item |
Downloads
Downloads per month over past year
