Deep Convolutional Neural Network with RNNs for Complex Activity Recognition Using Wrist-Worn Wearable Sensor Data
Open accessThe DOI matches its Crossref record; 77 of the 77 references it cites were found in a registry or an index.
Abstract
Sensor-based human activity recognition (S-HAR) has become an important and high-impact topic of research within human-centered computing. In the last decade, successful applications of S-HAR have been presented through fruitful academic research and industrial applications, including for healthcare monitoring, smart home controlling, and daily sport tracking. However, the growing requirements of many current applications for recognizing complex human activities (CHA) have begun to attract the attention of the HAR research field when compared with simple human activities (SHA). S-HAR has shown that deep learning (DL), a type of machine learning based on complicated artificial neural networks, has a significant degree of recognition efficiency. Convolutional neural networks (CNNs) and recurrent neural networks (RNNs) are two different types of DL methods that have been successfully applied to the S-HAR challenge in recent years. In this paper, we focused on four RNN-based DL models (LSTMs, BiLSTMs, GRUs, and BiGRUs) that performed complex activity recognition tasks. The efficiency of four hybrid DL models that combine convolutional layers with the efficient RNN-based models was also studied. Experimental studies on the UTwente dataset demonstrated that the suggested hybrid RNN-based models achieved a high level of recognition performance along with a variety of performance indicators, including accuracy, F1-score, and confusion matrix. The experimental results show that the hybrid DL model called CNN-BiGRU outperformed the other DL models with a high accuracy of 98.89% when using only complex activity data. Moreover, the CNN-BiGRU model also achieved the highest recognition performance in other scenarios (99.44% by using only simple activity data and 98.78% with a combination of simple and complex activities).
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
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
“Previous studies, including [10, 18, 23, 24], have amalgamated convolutional neural networks (CNN), recurrent neural networks (RNN), and attention mechanisms;”
From Adaptive temporal attention mechanism and hybrid deep CNN model for wearable sensor-based human activity recognition (2025) · cites this paper as [10]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“This leads to more reduced error rates in the confusion matrix (Fig. 8) when compared to CNN-BiGRU [10] and CNN-LSTM [24].”
From Adaptive temporal attention mechanism and hybrid deep CNN model for wearable sensor-based human activity recognition (2025) · cites this paper as [10]
In the citing paper
“Furthermore, employing data standardization or normalization methods can enhance input quality and expedite model training [10, 31].”
From Adaptive temporal attention mechanism and hybrid deep CNN model for wearable sensor-based human activity recognition (2025) · cites this paper as [10]
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 2×
“However, the RNNs suffered from the vanishing gradient problem [45] that made them hard to train with acceptable performance.”
BackgroundThe Elements of Statistical Learning
Hastie, T., Tibshirani, R., & Friedman, J. (2009) · Springer Series in Statistics · doi:10.1007/978-0-387-84858-7 · Cited by 25,970
Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
Chung, J., Gulcehre, C., Cho, K., & Bengio, Y. (2014) · arXiv · doi:10.48550/arxiv.1412.3555 · Cited by 11,008 · cited 2×
“The GRU is a basic version of the LSTM that does not have a separate memory cell in its structure [52].”
BackgroundBidirectional recurrent neural networks
Schuster, M., & Paliwal, K. K. (1997) · IEEE Transactions on Signal Processing · doi:10.1109/78.650093 · Cited by 10,172 · cited 1×
“The BiLSTM was introduced in 1997 by Schuster and Paliwal to increase the amount of knowledge available in the LSTM network [49].”
MethodOn the Properties of Neural Machine Translation: Encoder–Decoder Approaches
Cho, K., van Merrienboer, B., Bahdanau, D., & Bengio, Y. (2014) · Proceedings of SSST-8, Eighth Workshop on Syntax, Semantics and Structure in Statistical Translation · doi:10.3115/v1/w14-4012 · Cited by 7,099 · cited 1×
“[51] introduced the gate recurrent unit (GRU) network, a novel RNN-based model, in 2014.”
BackgroundA survey of cross-validation procedures for model selection
Arlot, S., & Célisse, A. (2010) · LillOA (Université de Lille (University Of Lille)) · Cited by 3,669
A survey of cross-validation procedures for model selection
Arlot, S., & Celisse, A. (2010) · Statistics Surveys · doi:10.1214/09-ss054 · Cited by 3,522 · cited 1×
“To split the data for training and testing, several techniques can be utilized, such as k-fold cross validation [68].”
MethodThe Vanishing Gradient Problem During Learning Recurrent Neural Nets and Problem Solutions
Hochreiter, S. (1998) · International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems · doi:10.1142/s0218488598000094 · Cited by 2,939 · cited 1×
“LSTMs are very popular in the natural language processing domain where they are used for word prediction, language translation, etc., including in the HAR domain [46,47].”
BackgroundDeep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition
Ordóñez, F., & Roggen, D. (2016) · Sensors · doi:10.3390/s16010115 · Cited by 2,761
Factorizing personalized Markov chains for next-basket recommendation
Rendle, S., Freudenthaler, C., & Schmidt-Thieme, L. (2010) · Proceedings of the 19th international conference on World wide web · doi:10.1145/1772690.1772773 · Cited by 2,692
Cited by
Items 101–144 of 144
TinyML based Deep Learning Model for Activity Detection
Kayarvizhy, N., Gera, B. M., Sharma, B., & Dakshinamurthy, S. (2023) · International Journal on Recent and Innovation Trends in Computing and Communication · doi:10.17762/ijritcc.v11i11s.8081 · Cited by 4
Improving Performance and Explainability of Sensor-Based Human Activity Recognition
Mekruksavanich, S., Jantawong, P., & Jitpattanakul, A. (2024) · Joint International Conference on Digital Arts, Media and Technology with ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering (ECTI DAMT & NCON) · doi:10.1109/ectidamtncon60518.2024.10480073 · Cited by 2
“However, accurately and reliably recognizing intricate real-world human activities remains challenging [8], [9].”
BackgroundSelf-citationMekruksavanich, S., & Jitpattanakul, A. (2025) · Joint International Conference on Digital Arts, Media and Technology with ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering (ECTI DAMT & NCON) · doi:10.1109/ectidamtncon64748.2025.10962017 · Cited by 2
Self-citationMekruksavanich, S., Jantawong, P., Hnoohom, N., & Jitpattanakul, A. (2021) · International Computer Science and Engineering Conference · doi:10.1109/icsec53205.2021.9684642 · Cited by 4
“When a time series-based solution is needed, recurrent neural networks, particularly LSTM models and convolutional neural network (CNN) models have been used effectively in various applications [12]–[15].”
MethodSelf-citationA Hybrid Deep Neural Network for Classifying Transportation Modes based on Human Activity Vibration
Mekruksavanich, S., Jantawong, P., You, I., & Jitpattanakul, A. (2022) · International Conference on Knowledge and Smart Technology · doi:10.1109/kst53302.2022.9729079 · Cited by 2
“Afterward, researchers and urban planners recognized that tracking and assessing mobility regarding living in the city could further support the development of innovative center operations and smart transportation systems [3].”
BackgroundSelf-citationKnowledge Mapping of Human Activity Recognition Techniques for Assistive Living
Agarwal, P., & Alam, M. (2023) · International Journal of Sensors Wireless Communications and Control · doi:10.2174/2210327913666230911113149 · Cited by 6
A Novel Human Activity Recognition Model
Zeng, X., Huang, M., Zhang, H., Ji, Z., & Ганчев, И. (2023) · International Conference on Mathematics and Computers in Sciences and Industry (MCSI) · doi:10.1109/mcsi60294.2023.00024 · Cited by 2
Nandakishora, Y., Christydass, S. P. J., Bhargav, K., & Kumar, S. S. (2025) · Wiley eBooks · doi:10.1002/9781394302895.ch8 · Cited by 1
Advancing Human Activity Recognition Using Ensemble Deep CNN-GRU Network
Khan, M. H., Ashfaq, N., Asif, A., & Farid, M. S. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-031-91235-1_8 · Cited by 1
Human Activity Recognition in the Classroom Using Low-Cost Sensors
Suescun-Ferrandiz, S., Cazorla, M., & Gomez‐Donoso, F. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-02725-2_38 · Cited by 1
Zhang, M., Liu, H., Ye, Z., Liu, L., & Zhang, X. (2026) · Applied Bionics and Biomechanics · doi:10.1155/abb/7560773 · Cited by 1
Analisis Kinerja Algoritma Deep Learning pada Pengolahan Data Kompleks
Khodijah, S., & Rizki, C. A. (2025) · Journal of Electrical Engineering Research · doi:10.64803/joeer.v1i3.20 · Cited by 1
“Demikian juga, fusi CNN dengan LSTM (Long Short-Term Memory) telah terbukti sangat efektif dalam pengenalan tindakan olahraga dan analisis aktivitas manusia yang kompleks, mengingat kedua arsitektur ini mampu menangkap hubungan temporal dalam data [6].”
Research on Navigation and Positioning of Electric Inspection Robot Based on Improved CNN Algorithm
Long, Y., Du, M., Luo, X., Li, S., & Liu, Y. (2022) · Journal of Sensors · doi:10.1155/2022/9369543 · Cited by 1
“In addition to the traditional and relatively perfect technologies such as machinery, electricians and electronics, communication, automation, and sensors, it also adds deep learning neural network to improve the ability of system fault identi fi cation [9, 10].”
BackgroundMeng, B., & Yang, F. (2022) · Wireless Communications and Mobile Computing · doi:10.1155/2022/7542104 · Cited by 1
“It is unique in environmental monitoring and can monitor temperature, humidity, noise, light intensity, pressure, and various components of the land [9], as well as the size of objects.”
BackgroundRetractedThe Deep Learning-based Human Activity Recognition Using Smart Wearable Sensors : A Tutorial
Mekruksavanich, S., & Jitpattanakul, A. (2022) · Research Briefs on Information and Communication Technology Evolution · doi:10.56801/rebicte.v8i.133 · Cited by 1
“The resultant collection of forecasted labels is then evaluated using state-of-the-art HAR criteria (i.e., Accuracy, Recall, Precision, and F1-score).”
MethodSelf-citationA Lightweight Neuromorphic CNN for Human Activity Recognition on Edge Device
Agarwal, P., & Alam, M. (2023) · International Conference on Intelligent Computing and Control Systems (ICICCS) · doi:10.1109/iciccs56967.2023.10142696 · Cited by 1
“Several Deep Learning algorithms, such as CNN (discussed in [14], [15][16]), RNN (described in [17][18]), and Autoencoders (mentioned in [19]), have been utilized in HAR.”
MethodLang, L., Thuente, D. J., & Ma, X. (2023) · Journal of Intelligent & Fuzzy Systems · doi:10.3233/jifs-231691 · Cited by 1
RetractedGesture Recognition and Interpretation through Wearable Sensors and Recurrent Neural Networks
Bhansali, P., Shukla, P., Purohit, M., & Mehta, S. (2024) · MIT Art, Design and Technology School of Computing International Conference (MITADTSoCiCon) · doi:10.1109/mitadtsocicon60330.2024.10575132 · Cited by 1
Mekruksavanich, S., Phaphan, W., & Jitpattanakul, A. (2024) · International Conference on Telecommunications and Signal Processing (TSP) · doi:10.1109/tsp63128.2024.10605952 · Cited by 1
“Convolutional neural networks (CNNs) and Recurrent neural networks (RNNs) have proven highly effective in capturing both spatial and temporal patterns within sensor data, leading to accurate activity identification [20–22].”
BackgroundSelf-citationNatively Neuromorphic LMU Architecture for Encoding-Free SNN-Based HAR on Commercial Edge Devices
Fra, V., Leto, B., Pignata, A., Macii, E., & Urgese, G. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72359-9_28 · Cited by 3
Unveiling Human Activity Patterns in Smart Cities Through a CNN-LSTM Approach
Ameur, I., Ameur, M. E. A., Ameur, M., & Degha, H. E. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-031-82112-7_4 · Cited by 2
Non-uniform spiking touch sensor for classification of movement patterns using neural networks
Markiewicz, M., Janusz, S., & Brzozowski, I. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.128866 · Cited by 1
CADDI: An in-Class Activity Detection Dataset using IMU data from low-cost sensorsPreprint
Márquez-Carpintero, L., Suescun-Ferrandiz, S., Pina-Navarro, M., Cazorla, M., & Gomez‐Donoso, F. (2025) · arXiv (Cornell University) · Cited by 1
Enhancing Health Monitoring using Efficient Hyperparameter Optimization
Singhal, R. (2022) · Journal of Artificial Intelligence and Capsule Networks · doi:10.36548/jaicn.2022.4.004
Manjula, S., & Begum, M. S. (2024) · Signals and communication technology · doi:10.1007/978-3-031-47942-7_40
A Survey on Deep Learning Based Human Activity Recognition System
Thomas, A. L., & Judith, J. E. (2024) · Communications in computer and information science · doi:10.1007/978-3-031-73065-8_10
Optimizing Human Gliding Performance Using Wearable Nano-biosensors
Hou, X. (2024) · Lecture notes on data engineering and communications technologies · doi:10.1007/978-3-031-71619-5_25
Remote Monitoring of Human Pulse of Long-Distance Runners Considering Mobile Wearable Sensors
Li, C., & Hu, Y. (2025) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-031-90084-6_22
Intelligent Capturing Method of Sports Training Track Data Based on Wearable Sensor
Hao, J., & Chen, Z. (2025) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-031-90084-6_20
Encoding Human Activity Recognition: Local Coordinate Coding on Multimodal Sensory Data
Noor, K., Haider, T., Aziz, S., Zulfiqar, Z., Khan, M. H., & Farid, M. S. (2024) · International Conference on Innovative Computing (ICIC) · doi:10.1109/icic63915.2024.11116285
“Several deep network architectures have been proposed for human activity recognition (HAR) like Convolutional Neural Networks (CNNs) [22], Recurrent Neural Networks (RNNs) [23], Gated Recurrent Units (GRUs) [24], and hybrid models [25].”
Precision Exercise Recognition Through Machine Learning: Insights from Comparative Analysis
Raja Karthikeya, M., Sooraj, R., Keerthana, R., Bhavana, S., & Kenchetty, P. (2025) · IEEE World Conference on Applied Intelligence and Computing (AIC) · doi:10.1109/aic66080.2025.11212184
“• Long Short-Term Memory (LSTM) LSTM [12] is a type of RNN [13] designed to handle sequential data by maintaining long-term dependencies.”
Wearable Multi-Sensor System for Human Activity Recognition using Fusion-based Deep Learning
Mitra, P., Quadri, H., Bella, K. P., Kiron, V., Maity, M., & Priyakumar, D. (2025) · IEEE India Council International Subsections Conference (INDISCON) · doi:10.1109/indiscon66021.2025.11254669
“While prior research ideas were focused on classifying simple human activities this [4] aims to develop a hybrid This research work is funded by IHub-Data, IIIT Hyderabad deep learning method CNN-BiGRU for complex human activity (CHA) recognition using wrist-worn wearable sensors.”
Ma, H. (2026) · International Conference on Communication, Information, and Digital Technologies (CIDT) · doi:10.1117/12.3091595
“The most obvious limitation is the inability to generalize across diverse biomechanical characteristics and the reliance on large-scale annotated datasets [6].”
Chawaphan, P., Maneetham, D., & Crisnapati, P. N. (2026) · Engineering Technology & Applied Science Research · doi:10.48084/etasr.15555
“This demonstrates that positional and directional cues alone can consistently differentiate various wheelchair stair-related states [20].”
Mekruksavanich, S., & Jitpattanakul, A. (2026) · Joint International Conference on Digital Arts, Media and Technology with ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering (ECTI DAMT & NCON) · doi:10.1109/ectidamtncon67592.2026.11460035
“Conversely, recurrent models such as long short-term memory (LSTM) networks can represent temporal dependencies, yet they may overlook critical local patterns in sensor signals [11], [12].”
Self-citationMekruksavanich, S., Phaphan, W., & Jitpattanakul, A. (2026) · Joint International Conference on Digital Arts, Media and Technology with ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering (ECTI DAMT & NCON) · doi:10.1109/ectidamtncon67592.2026.11460058 · Cited by 1
“As the need to recognize complex and composite activities continues to grow, deep learning methods have become increasingly favored because of their demonstrated effectiveness in modeling complex data interactions [20].”
Self-citationConception d'un dispositif médical intelligent multi-capteur pour la détection de la mobilité faible
Muhoza, A. C. (2024) · theses.fr (ABES) · doi:10.70675/cab923d6z3a5dz4b8az9d1bz89e63c403a17
Enhancing Wearable-Based Elderly Activity Recognition Through a Hybrid Deep Residual Network
Mekruksavanich, S., & Jitpattanakul, A. (2026) · Machine Learning and Knowledge Extraction · doi:10.3390/make8040107
Self-citationTransfer Learning Based Recognition of Human Activity Using Wearable Data
Bathala, N. R., & Sahukara, K. V. (2026) · Journal of Electrical Engineering and Technology · doi:10.1007/s42835-026-02812-x
“Deep learning has revolutionized HAR in recent years by automatically identifying features and utilizing neural network architectures such as conventional neural networks and recurrent neural networks [15, 16].”
Shorfuzzaman, M., Hanzla, M., Alabdullah, B. I., Alonazi, M., Almotiri, J., & Jalal, A. (2026) · Bioengineering · doi:10.3390/bioengineering13080866
“Complementary wearable-sensor architectures combining convolutional and recurrent models have further improved recognition of complex daily activities [8,9].”
Enhancing human activity recognition through GAN-augmented data and multi-branch CNN-LSTM networks
Shahid, A., Khan, M. H., Nisar, M. A., & Farid, M. S. (2026) · Knowledge-Based Systems · doi:10.1016/j.knosys.2026.116783 · Cited by 1
Shen, Z., Gao, B., & Shi, M. (2026) · World Congress on Medical Physics and Biomedical Engineering, September 7 - 12, 2009, Munich, Germany · doi:10.1007/978-3-032-20291-8_18
Optimizing Human Activity Recognition with Ensemble Deep Learning on Wearable Sensor Data
Ashfaq, N., Sadiq, I., Junaid, M., Khan, M. H., & Farid, M. S. (2024) · International Journal of Innovations in Science and Technology · doi:10.33411/ijist/1069
Francies, M. L. (2026) · Journal of Smart Sensors and Computing · doi:10.64189/ssc.26213
Metadata
- Authors
- Mekruksavanich, S., & Jitpattanakul, A.
- Affiliations
- University of Phayao (TH)
- King Mongkut's University of Technology North Bangkok (TH)
- Corresponding author
- Jitpattanakul, Anuchit
- Venue
- Electronics
- Publisher
- MDPI AG
- Published
- 2021-07-14
- Volume
- 10
- Issue
- 14
- Pages
- 1685
- ISSN
- 2079-9292 (ISSN-L)
- OpenAlex
- W3182319652
- Cited by
- 138
- References
- 76
- Citations by year
- 2026: 12 · 2025: 27 · 2024: 36 · 2023: 22 · 2022: 30 · 2021: 10
- FWCI
- 8.78
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Electronics (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- University of Phayao (FF64-UoE008)
- King Mongkut's University of Technology North Bangkok (KMUTNB-BasicR-64-33-2)
- Thailand Science Research and Innovation Fund (-)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Context-Aware Activity Recognition Systems
- Physical Sciences › Engineering › Biomedical Engineering › Gait Recognition and Analysis
- Physical Sciences › Computer Science › Human-Computer Interaction › Hand Gesture Recognition Systems
- Keywords
- wearable sensors, wrist-worn sensors, recurrent neural networks, long short-term memory, bidirectional LSTM, gated recurrent unit, bidirectional GRU, hybrid CNN-RNN model, CNN-BiGRU
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.