Optimising Deep Learning at the Edge for Accurate Hourly Air Quality Prediction
Open accessThe DOI matches its Crossref record; 41 of the 41 references it cites were found in a registry or an index.
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.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures


![Figure 3 Post-training optimisation methods provided by TensorFlow [46].](https://figures.cochat.ai/paper-figures/072c025c2c854ac7aa37eead2073d8f1/optimising-deep-learning-at-the-edge-for-accurate-hourly-air-quality-prediction-figure-3.jpg)













Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Pruning techniques allow removal of redundant structures of network and parameters [15,16].”
From TinyML: Enabling of Inference Deep Learning Models on Ultra-Low-Power IoT Edge Devices for AI Applications (2022) · cites this paper as [16]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“For instance, using quantization techniques, which are conversion techniques that convert float-point numbers to minimize precision numbers, intending to shrink the size of the DL model with minimal degradation of accuracy.”
From TinyML: Enabling of Inference Deep Learning Models on Ultra-Low-Power IoT Edge Devices for AI Applications (2022) · cites this paper as [16]
The cited source says what the claim attributes to it.
In the citing paper
“Furthermore, deployment of pre-trained DL models into tiny edge devices is now possible, after performing some techniques to compress DL models and optimize the inference.”
From TinyML: Enabling of Inference Deep Learning Models on Ultra-Low-Power IoT Edge Devices for AI Applications (2022) · cites this paper as [16]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 1×
“Long Short-Term Memory (LSTM) [45] is a structural modification of the Recurrent Neural Network (RNN) that adds memory cells in the hidden layer so that it can be implemented to control the flow of information in time-series data.”
BackgroundLeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
(2014) · Cited by 33,419
Edge computing: Vision and challenges
Shi, W., Cao, J., Zhang, Q., Li, Y., & Xu, L. (2016) · IEEE Internet of Things Journal · doi:10.1109/jiot.2016.2579198 · Cited by 8,733 · cited 1×
“Moreover, edge computing can address energy-aware and bandwidth saving applications [5].”
BackgroundFine Particulate Air Pollution and Hospital Admission for Cardiovascular and Respiratory Diseases
Dominici, F., Peng, R. D., Bell, M. L., Pham, L., McDermott, A., Zeger, S. L., & Samet, J. M. (2006) · JAMA · doi:10.1001/jama.295.10.1127 · Cited by 2,634 · cited 1×
“It may lead to lung cancer [20], affect cardiovascular diseases [21] and even result in death [22].”
BackgroundDeep Learning With Edge Computing: A Review
Chen, J., & Ran, X. (2019) · Proceedings of the IEEE · doi:10.1109/jproc.2019.2921977 · Cited by 1,540 · cited 5×
“Both ASICs and FPGAs are generally more energy-efficient than conventional CPUs and GPUs [4].”
BackgroundMethodInfluentialA Survey on the Edge Computing for the Internet of Things
Yu, W., Liang, F., He, X., Hatcher, W. G., Lu, C., Lin, J., & Yang, X. (2018) · IEEE Access · doi:10.1109/access.2017.2778504 · Cited by 1,520 · cited 1×
“Edge computing refers to the deployment of computation closer to data sources (edge) [1], rather than more centrally as is the case with cloud computing.”
BackgroundConvergence of Edge Computing and Deep Learning: A Comprehensive Survey
Wang, X., Han, Y., Leung, V. C. M., Niyato, D., Yan, X., & Chen, X. (2020) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2020.2970550 · Cited by 1,498 · cited 1×
“For data processing and information inference, it is also possible to embed intelligence at edge devices, which can be enabled by machine learning (ML) algorithms [6,7].”
BackgroundOutdoor Particulate Matter Exposure and Lung Cancer: A Systematic Review and Meta-Analysis
Hamra, G. B., Guha, N., Cohen, A., Laden, F., Raaschou-Nielsen, O., Samet, J. M., Vineis, P., Forastiere, F., Saldiva, P., Yorifuji, T., et al. (2014) · Environmental Health Perspectives · doi:10.1289/ehp/1408092 · Cited by 1,023
Evaluating Critical Security Issues of the IoT World: Present and Future Challenges
Frustaci, M., Pace, P., Aloi, G., & Fortino, G. (2018) · IEEE Internet of Things Journal · doi:10.1109/jiot.2017.2767291 · Cited by 764 · cited 1×
“It can address latency, privacy and scalability issues faced by cloud-based systems [2,3].”
Background
Research works printed but not linked: 7
- Hinton, G., Vinyals, O., and Dean, J. (2015). Distilling the Knowledge in a Neural Network. arXiv.
- Xu, J., Chen, L., Lv, M., Zhan, C., Chen, S., and Chang, J. (2021). HighAir: A Hierarchical Graph Neural Network-Based Air Quality Forecasting Method. arXiv.
- Zhang, Q., Lam, J.C., Li, V.O., and Han, Y. (2020). Deep-AIR: A Hybrid CNN-LSTM Framework forFine-Grained Air Pollution Forecast. arXiv.
- Chen, H.C., Putra, K.T., and Chun-WeiLin, J. (2021). A Novel Prediction Approach for Exploring PM2.5 Spatiotemporal Propagation Based on Convolutional Recursive Neural Networks. arXiv.
- Wu, H., Judd, P., Zhang, X., Isaev, M., and Micikevicius, P. (2020). Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation. arXiv.
- Li, J., and Alvarez, R. (2021). On the quantization of recurrent neural networks. arXiv.
- Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., Corrado, G.S., Davis, A., Dean, J., and Devin, M. (2016). TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems. arXiv.
Cited by
Air-pollution prediction in smart city, deep learning approach
Bekkar, A., Hssina, B., Douzi, S., & Douzi, K. (2021) · Journal of Big Data · doi:10.1186/s40537-021-00548-1 · Cited by 296
“Four recently published models are suggested, such as AC-LSTM by [53], LSTM-FC by [33], XGBoost by [54], and CNN-LSTM by [55], which are evaluated for comparing the performance of the proposed model.”
MethodBourechak, A., Zedadra, O., Kouahla, M. N., Guerrieri, A., Seridi, H., & Fortino, G. (2023) · Sensors · doi:10.3390/s23031639 · Cited by 258
“[24] On-device airquality prediction Prediction CNN, LSTM Dataset from University of California–Irvine (UCI) Machine Learning Repository page Edge devices(RPi3B+, RPi4B) Posttraining quantization Hardware accelerator TensorFlow Lite RMSE, MAE execution time Low-complexity model latency Accuracy degradation”
MethodAlajlan, N. N., & Ibrahim, D. M. (2022) · Micromachines · doi:10.3390/mi13060851 · Cited by 224
“Pruning techniques allow removal of redundant structures of network and parameters [15,16].”
MethodJouini, O., Sethom, K., Namoun, A., Aljohani, N. F., Alanazi, M. H., & Alanazi, M. N. (2024) · Technologies · doi:10.3390/technologies12060081 · Cited by 202
“…consisting of various devices such as NVIDIA Jetson Nx and Jetson Nano to collect water sensor data and to perform real-time video-based fish detection using DL. Wardana et al. [77] introduced a new DL model that integrates one-dimensional CNN and LSTM network to anticipate a short-term hourly PM2.”
Deep-learning architecture for PM2.5 concentration prediction: A review
Zhou, S., Wang, W., Zhu, L., Qiao, Q., & Kang, Y. (2024) · Environmental Science and Ecotechnology · doi:10.1016/j.ese.2024.100400 · Cited by 135
A Hybrid Spatiotemporal Deep Model Based on CNN and LSTM for Air Pollution Prediction
Tsokov, S., Lazarova, M. K., & Aleksieva-Petrova, A. (2022) · Sustainability · doi:10.3390/su14095104 · Cited by 117
“Since the values of the wind direction are strings, these values are converted to numbers before normalization using the strategy suggested in [42] by replacing the string with degrees from 0 to 360.”
InfluentialAlwaisi, Z., Kumar, T., Harjula, E., & Soderi, S. (2024) · Internet of Things · doi:10.1016/j.iot.2024.101398 · Cited by 112
Edge AI for Real-Time Anomaly Detection in Smart Homes
Reis, M. J. C. S., & Serôdio, C. (2025) · Future Internet · doi:10.3390/fi17040179 · Cited by 93
“Previous studies have demonstrated that full-integer quantization can achieve latencies as low as 2.19 s for deep learning models, while optimized inference frameworks can lead to 79% energy savings and significant execution time reductions [23,24].”
InfluentialMampitiya, L. I., Rathnayake, N., Leon, L. P., Mandala, V., Azamathulla, H. M., Shelton, S., Hoshino, Y., & Rathnayake, U. (2023) · Environments · doi:10.3390/environments10080141 · Cited by 78
TinyML Models for a Low-Cost Air Quality Monitoring Device
Wardana, I. N. K., Fahmy, S. A., & Gardner, J. W. (2023) · IEEE Sensors Letters · doi:10.1109/lsens.2023.3315249 · Cited by 43
“Theimmensevolumeof collectedspatiotemporaldatahasprovidedabetteropportunityto applymachinelearning(ML)techniquesinairqualityareas,such asaircontaminantprediction[4],[5],missingdataimputation[6],[7], andclassificationtasks[8].”
Self-citationEstimation of missing air pollutant data using a spatiotemporal convolutional autoencoder
Wardana, I. N. K., Gardner, J. W., & Fahmy, S. A. (2022) · Neural Computing and Applications · doi:10.1007/s00521-022-07224-2 · Cited by 39
“In this study, root mean square error ( RMSE ) and mean absolute error ( MAE ) are used, following work done by [63].”
MethodSelf-citationSassi, M. S. H., & Fourati, L. C. (2022) · Computer Networks · doi:10.1016/j.comnet.2022.108904 · Cited by 31
Split BiRNN for real-time activity recognition using radar and deep learning
Werthen‐Brabants, L., Bhavanasi, G., Couckuyt, I., Dhaene, T., & Deschrijver, D. (2022) · Scientific Reports · doi:10.1038/s41598-022-08240-x · Cited by 25
“Wardana et al. 18 use a hybrid CNN and LSTM model for use on an edge device, such as a Raspberry Pi.”
MethodAkyol, S., Yıldırım, M., & Alataş, B. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.106768 · Cited by 20
Multistation collaborative prediction of air pollutants based on the CNN-BiLSTM model
Lu, Y., & Li, K. (2023) · Environmental Science and Pollution Research · doi:10.1007/s11356-023-28877-z · Cited by 19
“Wardana et al. (2021) predicted day-ahead air quality with a CNN-LSTM.”
BackgroundAlajlan, N. N., & Ibrahim, D. M. (2024) · Journal of Advanced Research in Applied Sciences and Engineering Technology · doi:10.37934/araset.47.2.94120 · Cited by 15
“For instance, in the previous studies [29,49-51, 53”
BackgroundResultSalcedo, E. (2024) · Journal of Imaging · doi:10.3390/jimaging10120326 · Cited by 15
Predicting Beijing Air Quality Using Bayesian Optimized CNN-RNN Hybrid Model
Tu, Z., & Wu, Z. (2022) · Asia Conference on Algorithms, Computing and Machine Learning (CACML) · doi:10.1109/cacml55074.2022.00104 · Cited by 12
A Novel Optimized Deep Learning Model for Canola Crop Yield Prediction on Edge Devices
Valarezo-Plaza, S., Torres-Tello, J. W., Singh, K. D., Shirtliffe, S. J., Deivalakshmi, S., & Ko, S. (2024) · IEEE Transactions on AgriFood Electronics · doi:10.1109/tafe.2024.3414953 · Cited by 11
“This approach effectively reduces the size of a model with minimal impact in accuracy [18].”
BackgroundHarnessing deep learning for air pollution forecasting: trends, techniques, and future prospects
Lawrence, S., & Bhathmanabhan, S. (2026) · Artificial Intelligence Review · doi:10.1007/s10462-026-11496-8 · Cited by 11
Collaborative Learning at the Edge for Air Pollution Prediction
Wardana, I. N. K., Gardner, J. W., & Fahmy, S. A. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3341116 · Cited by 13
“Pearson’s correlation coefficient between two series is described as follows [2], [44]: Authorized licensed use limited to the terms of the applicable license agreement with IEEE.”
MethodInfluentialSelf-citationAir Quality Prediction and Multi-Task Offloading based on Deep Learning Methods in Edge Computing
Sun, C., Li, J., Sulaiman, R., Alotaibi, B. S., Elattar, S., & Abuhussain, M. A. (2023) · Journal of Grid Computing · doi:10.1007/s10723-023-09671-0 · Cited by 10
“The training process can be delayed by high dimensions of irrelevant features, which force the employment of costly and time-consuming computing equipment [1].”
BackgroundMethodInfluentialTang, X., Wu, N., & Pan, Y. (2023) · Applied Sciences · doi:10.3390/app132312794 · Cited by 10
Torres-Tello, J. W., & Ko, S. (2022) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas48785.2022.9937940 · Cited by 10
“The results here obtained are aligned to the ones found in literature that show the important reduction in size and inference times of DL models when using pruning and quantization, while maintaining good performance [5], [9], [15].”
BackgroundMethodResultInfluentialAryanto, I. K. A. A., Maneetham, D., & Crisnapati, P. N. (2023) · Applied Sciences · doi:10.3390/app132312953 · Cited by 9
A novel bifold-attention-LSTM for analyzing PM2.5 concentration-based multi-station data time series
Pranolo, A., Zhou, X., & Mao, Y. (2024) · International Journal of Data Science and Analytics · doi:10.1007/s41060-024-00666-y · Cited by 9
Saini, J., Dutta, M., & Marques, G. (2022) · Internet of Things · doi:10.1016/j.iot.2022.100628 · Cited by 12
Exploring Hybrid Models For Short-Term Local Weather Forecasting in IoT Environment
Toại, T. K., Šenkeřík, R., Vo, H. T. X., Vo, H. M., Ulrich, A., Musil, M., & Zelinka, I. (2023) · MENDEL · doi:10.13164/mendel.2023.2.295 · Cited by 8
“A few researchers localized dataset points by combining sensors and satellite imagery [43, 44], which may be quite complicated in real applications.”
BackgroundMethodAir quality index prediction with optimisation enabled deep learning model in IoT application
Sivakumar, S. (2024) · Environmental Technology · doi:10.1080/09593330.2024.2409993 · Cited by 8
“In general, DL-based approaches predict air contaminants more accurately [1,12].”
InfluentialA Soil pH Sensor and a Based on Time-Series Prediction IoT System for Agriculture
Aguiar, S., & Barros, E. N. S. (2023) · Brazilian Symposium on Computing System Engineering (SBESC) · doi:10.1109/sbesc60926.2023.10324263 · Cited by 7
“Other studies, such as [14] and [15], have also addressed the use of LSTMs in edge computing for pollutant prediction and resource use prediction, respectively, showing the effectiveness of these architectures in devices of reduced computational capacity.”
BackgroundPengembangan Model Prediksi Curah Hujan di Kota Denpasar Menggunakan Metode LSTM dan GRU
Suranata, I. W. A. (2023) · Jurnal Sistem dan Informatika (JSI) · doi:10.30864/jsi.v18i1.603 · Cited by 6
“Selain itu, untuk analisis lebih lanjut guna mengetahui relasi antara atribut target dengan atribut lainnya, dilakukan perhitungan untuk mendapatkan nilai matriks korelasi dengan Rumus 2 [10].”
BackgroundSmart indoor air quality monitoring for enhanced living environments and ambient assisted living
Saini, J., Dutta, M., & Marques, G. (2023) · Advances in computers · doi:10.1016/bs.adcom.2023.10.008 · Cited by 6
Gonçalves, R., & Ribeiro, V. M. (2023) · Energy · doi:10.1016/j.energy.2023.130045 · Cited by 6
IoT-enhanced infant incubator monitoring system with 1D-CNN temperature prediction model
Aryanto, I. K. A. A., Maneetham, D., & Crisnapati, P. N. (2024) · Indonesian Journal of Electrical Engineering and Computer Science · doi:10.11591/ijeecs.v34.i2.pp900-912 · Cited by 6
Li, J., Feng, Y., & Ge, J. (2025) · Chem & Bio Engineering · doi:10.1021/cbe.5c00042 · Cited by 5
Graph Learning-based Regional Heavy Rainfall Prediction Using Low-Cost Rain Gauges
Salcedo, E. (2024) · IEEE Latin American Conference on Computational Intelligence (LA-CCI) · doi:10.1109/la-cci62337.2024.10814868 · Cited by 5
“Edge computing-based solutions are being explored to minimise sensor data transmission to a central server and maximise local processing [18], [19].”
Psaropa, M. X., Kontogiannis, S. I., Lolis, C. J., HATZIANASTASSIOU, N., & Pikridas, C. (2025) · Applied Sciences · doi:10.3390/app15137432 · Cited by 7
A Hybrid Machine Learning Framework for Multi-Pollutant Air Quality Assessment in Urban Environments
Mustafa, M., Akhtar, M., Ahmad, A., Javaid, F., Haldar, B., & Nisar, B. (2026) · Sustainability · doi:10.3390/su18042148 · Cited by 3
Sarkar, N., Keserwani, P. K., Ranjan, R., & Govil, M. C. (2026) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-026-10555-1 · Cited by 4
Compact Edge-AI Architecture for Real-Time Decision Making in Smart Devices
Soni, M., Singh, T., Wawage, P., Tonde, D., Borkar, P. S., & Nadaf, J. S. (2026) · International Conference on Emerging Smart Computing and Informatics (ESCI) · doi:10.1109/esci68015.2026.11493408 · Cited by 3
“The quantization techniques such as the 8-bit and mixed-precision inference techniques cut the memory consumption and energy expenditure by a large margin [8].”
Aryanto, I. K. A. A., Maneetham, D., & Triandini, E. (2024) · International Journal of Power Electronics and Drive Systems/International Journal of Electrical and Computer Engineering · doi:10.11591/ijece.v14i2.pp2293-2312 · Cited by 3
“Through the analysis performed by the LSTM model, certain patterns in the data can be identified and predicted [9], [10].”
MethodHenna, S., Amjath, M., & Yar, A. (2025) · AIMS environmental science · doi:10.3934/environsci.2025024 · Cited by 3
Air Pollution Monitoring Approach using Atomic Orbital Search Algorithm with Deep Learning Driven
(2023) · Global NEST Journal · doi:10.30955/gnj.005373 · Cited by 3
Advancements in Air Pollution Prediction and Classification Models: Exploring Emerging Frontiers
Mittal, A., Arora, S., Sharma, S., & Garg, A. (2024) · International Conference on Advanced Computing and Communication Systems (ICACCS) · doi:10.1109/icaccs60874.2024.10716970 · Cited by 3
Development of Smart Sensor for IoT Based Environmental Data Analysis Through Edge Computing
Chatterjee, A. (2022) · Communications in computer and information science · doi:10.1007/978-3-031-10551-7_1 · Cited by 2
Analyzing inference workloads for spatiotemporal modeling
Jain, M., Agostini, N. B., Ghosh, S., & Tumeo, A. (2024) · Future Generation Computer Systems · doi:10.1016/j.future.2024.107513 · Cited by 3
Physical-Chemical Properties Prediction of Chao Phraya River using Deep Learning Methods
Suranata, I. W. A., Aryanto, I. K. A. A., Maneetham, D., Wardana, I. N. K., & Crisnapati, P. N. (2024) · International Conference on Smart Computing and Communications (ICSCC) · doi:10.1109/icscc62041.2024.10690816 · Cited by 2
“Lastly, 1D-CNN is a convolutional neural network adapted to process one-dimensional sequential data, such as time series [12][13].”
MethodSelf-citationMeng, X., Xie, C., Tang, X., & Pan, Y. (2025) · Discover Applied Sciences · doi:10.1007/s42452-025-07891-5 · Cited by 2
Hybrid artificial neural network algorithm for air pollution estimation
Kumbhar, V., Sohi, S. S., Jayaram, V., Sreelekshmy, P. G., Shukla, S. K., & Abhilash, K. S. (2022) · International Journal of Health Sciences · doi:10.53730/ijhs.v6ns5.9080 · Cited by 2
A Moderation effect of atmospheric conditions on PM 2.5 traffic related concentration
Roşca, E., RUSCA, F. V., Dinu, O. M., Burciu, Ş., Oprea, C. C., Ilie, A., & Coman, C. (2023) · Romanian Journal of Transport Infrastructure · doi:10.2478/rjti-2023-0007 · Cited by 2
Lestari, I. G. A. N., & Mahendra, I. N. D. A. (2023) · Jurnal Sistem dan Informatika (JSI) · doi:10.30864/jsi.v17i2.565 · Cited by 2
“Persamaan untuk masing-masing teknik evaluasi dijelaskan pada rumus 10 – 13 [12][13].”
Mustamin, S. B., Atnang, M., Sahriani, S., Fajar, N., Sari, S. K., Pahlawan, M. R., & Amrullah, M. (2024) · Journal of Scientific Insights · doi:10.69930/jsi.v1i3.183 · Cited by 2
“This hybrid model is optimized for (Wardana et al., 2021) edge computing devices , demonstrating the potential of the technology for more efficient applications in the field.”
Design and Implementation of an Air Quality Testing System Based on STC12C5A
Wu, F., Liu, F., Feng, Y., & Cai, Z. (2021) · International Journal of Circuits Systems and Signal Processing · doi:10.46300/9106.2021.15.110 · Cited by 1
Enhanced short-term prediction of urban PM2.5 concentrations by improved hybrid deep learning
Zhou, Y., Lyu, Y., Dang, X., Bol, R., Zhang, P., Yu, N., & Zhang, Y. (2025) · Journal of Environmental Health Science and Engineering · doi:10.1007/s40201-025-00954-0 · Cited by 1
Machine Learning for Energy Optimization in Computing Systems
Bansal, M., Jain, N., Pathak, N., & Sharma, N. (2025) · Wiley eBooks · doi:10.1002/9781394383658.ch19 · Cited by 1
Fahim, M. M., Yesmin, S. H., Islam, S., Faruque, M. P. B., Salam, M. A., Uddin, M. M., Islam, S., Ahmed, T., Binyamin, M., & Karim, M. R. (2026) · IEEE International Conference on Quantum Photonics, Artificial Intelligence & Networking (QPAIN) · doi:10.1109/qpain69676.2026.11545925 · Cited by 1
“The first barrier is computational cost—models with hundreds of millions of parameters require high-end GPUs to perform inference, which makes them unsuitable for distributed edge networks or IoT-based climate monitoring systems [11], [12], [13].”
Utilizing deep learning models for accurate prediction of air pollution levels
Jain, H. (2023) · International Journal of Engineering in Computer Science · doi:10.33545/26633582.2023.v5.i2a.91 · Cited by 1
Artificial Neural Network in the Field of Environment
(2022) · Machine Learning Theory and Practice · doi:10.38007/ml.2022.030408
M, O. N., Adehi, M. U., & Aimufua, G. (2025) · American Journal of Innovation in Science and Engineering · doi:10.54536/ajise.v4i1.2828
“After training the model, we adopted the post-training quantisation method to reduce the size of the model by rounding off its parameter to the nearest 8-bit integers (Wardana et al. , 2021).”
Liu, H., Li, Y., & Duan, Z. (2025) · Engineering Applications of Computational Methods · doi:10.1007/978-981-96-5777-3_1
Air Suspended Particle Measurement [ASPM] in downhill areas using Deep Learning Technique
Gagan, B. J., & Muruganandam, N. S. (2024) · International Conference on Control, Computing, Communication and Materials (ICCCCM) · doi:10.1109/iccccm61016.2024.11039979
Lightweight Federated Learning with Genetic Optimization for PM 2.5 Forecasting in IoT NetworksPreprint
Nazari, H., Farjami, Y., & TaeiZadeh, A. (2025) · Research Square · doi:10.21203/rs.3.rs-7633900/v1
Quantifying uncertainty and improving reliability of time-series based deep learning models
Werthen‐Brabants, L. (2023) · Ghent University Academic Bibliography (Ghent University)
Применение квантованных моделей машинного обучения для встраиваемых систем в реальных задачах
Ачкасов, А. В., Yagodkin, A., Владимирович, М. Ф., & Makarenko, P. V. (2026) · Международный журнал Программные продукты и системы · doi:10.15827/0236-235x.153.070-079
TinyML: Applications, Algorithms, Co-design and Implementations
Ghazal, O. G., Rahman, T., & Shafik, R. (2026) · Springer eBooks · doi:10.1007/978-3-032-06286-4_6
Predictive Air Quality Monitoring System with Adaptive Sampling and 4G LTE
Whyshnavi, P., Kogulshiyam, S., Yesindan, R., Valluvan, R., Thiruvaran, T., & Mukunthan, T. (2026) · International Conference on Advanced Research in Computing (ICARC) · doi:10.1109/icarc68737.2026.11453560
“This has led to the development of various adaptive sampling strategies, ranging from simple threshold-based methods [4] to sophisticated machine learning controllers that maintain data accuracy while extending battery life [5], [6], [7], [8].”
Munandar, G. M., Fatkhurrokhman, T., & Hamdi, L. F. (2026) · Jurnal Teknik Industri Jurnal Hasil Penelitian dan Karya Ilmiah dalam Bidang Teknik Industri · doi:10.24014/jti.v12i1.39516
“TensorFlow Lite and edge-oriented optimization further enable trained models to be converted into formats suitable for mobile or edge deployment [23], [24], [25].”
Bao, Q., Sun, B., Chen, X., Zhang, Q., Ye, J., & Zhao, X. (2026) · International Journal of Computational Intelligence Systems · doi:10.1007/s44196-026-01334-8
“Wardana et al. [13 ] crafted a composite deep learning framework that integrates a one-dimensional convolutional neural network (CNN) with long short-Term Memory (LSTM) to forecast hourly PM 2.5 levels, optimizing the model’s performance running on edge devices.”
InfluentialAir quality analysis based on threshold values and anomaly detection using machine learning models
БОРОДІЙ, І., & Osukhivska, H. (2026) · COMPUTER-INTEGRATED TECHNOLOGIES EDUCATION SCIENCE PRODUCTION · doi:10.36910/6775-2524-0560-2026-63-03
Das, R., Acharyya, A., Majumdar, S., & Srivastava, G. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3707055
“Edge-oriented ML research [24], [25], [26] investigates lightweight inference strategies under hardware constraints but rarely reports device-level metrics.”
RainAI: A Hybrid 1DCNN-LSTM Model for Rainfall Prediction with KNN-Based Imputation
Lestari, I. G. A. N., Saryanti, I. G. A. D., Maneetham, D., Wardana, I. N. K., & Aryanto, I. K. A. A. (2026) · Evergreen · doi:10.5109/7434182
Self-citationAn overview of machine learning methods on microcontrollers and IoT devices in the TinyML paradigm
Achkasov, A. V., Yagodkin, A., & Lemza, S. (2026) · Modeling of systems and processes · doi:10.12737/2219-0767-2026-19-2-39-58
Ramadan, M. N. A., AlKhedher, M. A., & Alhindawi, R. (2026) · Indoor Air · doi:10.1155/ina/1015833
Metadata
- Authors
- Wardana, I. N. K., Gardner, J. W., & Fahmy, S. A.
- Affiliations
- University of Warwick (GB)
- Corresponding author
- Wardana, I Nyoman Kusuma
- Venue
- Sensors
- Publisher
- MDPI AG
- Published
- 2021-02-04
- Volume
- 21
- Issue
- 4
- Pages
- 1064
- ISSN
- 1424-8220 (ISSN-L)
- PMC
- PMC7913936
- OpenAlex
- W3126841703
- Cited by
- 61
- References
- 41
- Citations by year
- 2026: 10 · 2025: 8 · 2024: 14 · 2023: 16 · 2022: 11 · 2021: 2
- FWCI
- 3.48
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Sensors (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Warwick Research Archive Portal (University of Warwick) (acceptedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- King Abdullah University of Science and Technology Repository (King Abdullah University of Science and Technology) (submittedVersion, other-oa)
- Funders
- Lembaga Pengelola Dana Pendidikan (Ref: S-1027/LPDP.4/2019)
- Topics
- Physical Sciences › Environmental Science › Environmental Engineering › Air Quality Monitoring and Forecasting
- Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Keywords
- air quality forecasting, PM2.5, edge computing, CNN-LSTM hybrid model, model compression, Raspberry Pi, inference latency, sensor networks
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.