Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition
Open accessThe DOI matches its Crossref record; 48 of the 49 references it cites were found in a registry or an index; 8 of the 23 sources its claims cite were checked for support.
Abstract
Introduces DeepConvLSTM, a deep learning architecture combining convolutional and LSTM recurrent units for multimodal wearable activity recognition, outperforming non-recurrent deep networks and handcrafted feature baselines by 4% to 9%.
Overview
CoChat Summary
TL;DR
The paper proposes DeepConvLSTM, a deep framework combining four convolutional layers with two LSTM recurrent layers for wearable activity recognition from raw multimodal sensor data. It reports improved F1 scores over a baseline CNN and prior results on the OPPORTUNITY and Skoda datasets, including 0.915 F1 for OPPORTUNITY gestures with the Null class and 0.958 F1 on Skoda.
Key findings
- On OPPORTUNITY, DeepConvLSTM achieved F1 scores of 0.930 for modes of locomotion without Null, 0.895 for modes with Null, 0.866 for gestures without Null, and 0.915 for gestures with Null; the baseline CNN scored 0.912, 0.878, 0.783, and 0.883, respectively. The paper reports a 6% average improvement over the best OPPORTUNITY challenge submissions and more than 9% improvement on gesture recognition without Null.
- On Skoda, DeepConvLSTM reached 0.958 F1, compared with 0.884 for the baseline CNN, 0.893 for CNN [43], and 0.861 for CNN [23], improving the best reported result by 6%.
- Multimodal fusion improved gesture recognition on OPPORTUNITY without Null: 0.689 F1 with accelerometers only, 0.611 with gyroscopes only, 0.745 with accelerometers + gyroscopes, 0.839 with accelerometers + gyroscopes + magnetic sensors, and 0.864 with the full OPPORTUNITY sensor set.
- DeepConvLSTM had 999,122 parameters versus 7,445,458 for the baseline CNN, a 600% increase for the baseline; training took 340.3 min versus 282.2 min, and classification took 6.68 s versus 5.43 s, with DeepConvLSTM classifying almost 15 min of data per second and achieving 900× real-time on the GPU used.
Methods
- DeepConvLSTM uses four convolutional layers with 64 feature maps each, followed by two dense LSTM layers with 128 cells and a softmax output; the baseline CNN shares the convolutional layers but uses two non-recurrent dense layers with 128 units. Models were trained supervised with cross-entropy loss, mini-batch gradient descent with RMSProp, learning rate 10e-3, decay factor 0.9, mini-batches of 100 segments, and dropout p=0.5 on dense-layer inputs.
- Inputs were sliding windows of 500 ms with 250 ms step size, using raw sensor data with minimal preprocessing: linear interpolation for missing values and per-channel normalization to [0,1]. Evaluation used weighted F1 to address class imbalance.
- Experiments used the OPPORTUNITY dataset (on-body sensors, 113 channels, 30 Hz; Task A: 5-class modes of locomotion/postures; Task B: 18-class sporadic gestures) and the Skoda dataset (one subject, 10 right-arm 3D accelerometers, decimated to 30 Hz, 10 manipulative gestures, about 3 h and about 70 repetitions per gesture). Analyses included comparison to baseline CNN and published results, confusion matrices, sensor-modality fusion, and hyperparameter evaluations of sequence length and number of convolutional layers.
Limitations
- The paper states that DeepConvLSTM has not yet been implemented on a wearable device; the GPU used clearly outperforms current high-end wearable computational power, and training is best envisioned server-side. It also states that pooling was omitted because the sliding-window setup limits downsampling, though pooling could be useful without that requirement, and that training time increases with the number of layers, creating a performance–computational budget trade-off.
- Evident from the text: Skoda evaluation uses one subject and only right-arm sensors, and OPPORTUNITY testing is limited to ADL4 and ADL5 for Subjects 2 and 3, which constrains generalizability. The text does not report cross-validation or statistical significance testing.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures



![Figure 4 Placement of on-body sensors used in the OPPORTUNITYdataset (left: inertial measurements units; right: 3-axis accelerometers) [7].](https://figures.cochat.ai/paper-figures/27c0e4b39d104fe2b5315587d736fd2d/deep-convolutional-and-lstm-recurrent-neural-networks-for-multimodal-wearable-figure-4.jpg)




Verification
Claims the paper makes
5 Valid · 5 Not valid · 13 Not checked · 99 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“The choice of the number of recurrent layers is made following the results presented in [33], where the authors showed that a depth of at least two recurrent layers is beneficial when processing sequential data.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 3. Architecture > 3.1. DeepConvLSTM
What the sources say
- Supportedcites [33] Karpathy A. Johnson J. Li F.F. Visualizing and understanding recurrent networks Cornell Univ. Lab. 2015 arXiv:1506.02078 source
Section 4.1 compares recurrent networks with one, two, and three layers on two sequence datasets and states that a depth of at least two is consistently beneficial. This directly supports the attributed finding, within the paper’s tested settings. (LLM, conf=0.94)
boostedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“This seems to indicate that although deep learning techniques are often employed with large amounts of data, (e.g., millions of frames in computer vision [22]), they may actually be applicable to problem domains where acquiring annotated data is very costly, such as in supervised activity recognition.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 5. Results and Discussion > 5.4. Discussion
What the sources say
- Supportedcites [22] Pigou L. Oord A.V.D. Dieleman S. van Herreweghe M. Dambre J. Beyond Temporal Pooling: Recurrence and Temporal Convolutions for Gesture Recognition in Video arXiv Preprint 2015 arXiv:1506.01911 source
The citation is attached to the example of large-scale computer-vision data, not to the citing authors’ conclusion about costly annotation. Section 4.1 (page unavailable) reports 1,720,800 labeled video frames in the gesture-recognition dataset, supporting the cited example of data on the order of millions of frames. (LLM, conf=0.86)
hedgedstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The baseline CNN also offers better results than the OPPORTUNITY submissions in the recognition of models of locomotion.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 5. Results and Discussion > 5.1. Performance Comparison
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“From the results, we can also see how the model is able to learn from signals obtained from accelerometers, gyroscopes and magnetometers, fusing them without requiring any specific preprocessing.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 6. Conclusions
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“(iv) it is applicable even if the gestures are longer than the observation window.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 5. Results and Discussion > 5.4. Discussion
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
84 more claims
Create a free account to see all 122 claims and to verify the 2 references still open.
- It may state two separate claims as one. 14
- Probably not a substantial claim. 23
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 cited source doesn’t support the claim as written.
In the citing paper
“However, these complex network framework suffered from low ef ciency and can hardly meet real-time requirements in practice applications.”
From InnoHAR: A Deep Neural Network for Complex Human Activity Recognition (2019) · cites this paper as [19]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe cited source doesn’t support the claim as written.
In the citing paper
“However, due to the lack of filtration techniques applied to multimodal data, the system’s performance was low.”
From A Multimodal IoT-Based Locomotion Classification System Using Features Engineering and Recursive Neural Network (2023) · cites this paper as [15]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“and yield better performances than pure CNNs or LSTM networks [12]”
From Comparison of Feature Learning Methods for Human Activity Recognition Using Wearable Sensors (2018) · cites this paper as [12]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The proposed method outperforms other methods, such as those based on deep believe networks (DBNs) [10], SVM [10], and CNNs [13].”
From Deep Recurrent Neural Networks for Human Activity Recognition (2017) · cites this paper as [13]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The popular DeepConvLSTM model [7] can only afford four, moderately-sized convolution layers to learn effective data representations, and two LSTM layers for sequence modeling.”
From Complex Deep Neural Networks from Large Scale Virtual IMU Data for Effective Human Activity Recognition Using Wearables (2021) · cites this paper as [7]
Experimental — accuracy not yet measured
20 more claims about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 3. Architecture
“The length of the kernels is the same for every convolutional layer, being defined as Pl=5,∀l=2,⋯,5.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 4. Experimental Setup > 4.2. Performance Measure
“Following the OPPORTUNITY challenge experimental setup, the length of the window is 500 ms, with a step size of 250 ms.”
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
“| Total | 996,800 +(128×nc)+nc | 996,800 +(128×nc)+nc | 7,443,136+(128×nc)+nc | 7,443,136+(128×nc)+nc |”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 3. Architecture > 3.3. Model Implementation and Training
Reported 996800 · recomputed 1037888
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
“| 6 | Wai,Waf,Wac,Wao: 7232×128 | 942,592 | W: 57,856×128 | 7,405,696 |”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 3. Architecture > 3.3. Model Implementation and Training
Reported 942592 · recomputed 3702784
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
“| 7 | Wai,Waf,Wac,Wao: 128×128 | 33,280 | W: 128×128 | 16,512 |”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 3. Architecture > 3.3. Model Implementation and Training
Reported 33280 · recomputed 131072
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Going deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562 · cited 1×
“Other publications showed much deeper networks, such as “GoogLeNet”, which is a 27-layer deep neural network applied to image classification [45].”
BackgroundReducing the dimensionality of data with neural networks
Hinton, G. E., & Salakhutdinov, R. R. (2006) · Science · doi:10.1126/science.1127647 · Cited by 20,928
Speech recognition with deep recurrent neural networks
Graves, A., Mohamed, A.-R., & Hinton, G. (2013) · 2013 IEEE International Conference on Acoustics, Speech and Signal Processing · doi:10.1109/icassp.2013.6638947 · Cited by 8,893 · cited 1×
“Networks using LSTM cells have offered better performance than standard recurrent units in speech recognition, where they gave state-of-the-art results in phoneme recognition [20].”
BackgroundConvolutional networks for images, speech, and time series
LeCun, Y., & Bengio, Y. (1998) · HAL (Le Centre pour la Communication Scientifique Directe) · doi:10.5555/303568.303704 · Cited by 4,231
The class imbalance problem: A systematic study1
Japkowicz, N., & Stephen, S. (2002) · Intelligent Data Analysis · doi:10.3233/ida-2002-6504 · Cited by 3,295
DeepPose: Human Pose Estimation via Deep Neural Networks
Toshev, A., & Szegedy, C. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.214 · Cited by 3,185 · cited 1×
“Deep CNNs have had a major impact on fields, like content recommendation [24], speech recognition [25] and in computer vision [26–28], where they have become the de facto standard approach.”
BackgroundOverFeat: Integrated Recognition, Localization and Detection using Convolutional Networks
Sermanet, P., Eigen, D., Zhang, X., Mathieu, M., Fergus, R., & LeCun, Y. (2013) · arXiv · doi:10.48550/arxiv.1312.6229 · Cited by 2,680 · cited 1×
“Deep CNNs have had a major impact on fields, like content recommendation [24], speech recognition [25] and in computer vision [26–28], where they have become the de facto standard approach.”
BackgroundLee, H., Grosse, R., Ranganath, R., & Ng, A. Y. (2009) · Proceedings of the 26th Annual International Conference on Machine Learning · doi:10.1145/1553374.1553453 · Cited by 2,412 · cited 1×
“Deep learning techniques have outperformed many conventional methods in computer vision [13] and audio classification [14].”
MethodA review of wearable sensors and systems with application in rehabilitation
Patel, S., Park, H., Bonato, P., Chan, L., & Rodgers, M. (2012) · Journal of NeuroEngineering and Rehabilitation · doi:10.1186/1743-0003-9-21 · Cited by 2,396 · cited 1×
“, from simple hand gestures to complex activities, such as “cooking a meal”) and the context in which they occur from sensor data is at the core of smart assistive technologies, such as in smart homes [1], in rehabilitation [2], in health support [3,4], in skill assessment [5] or in industrial settings [6].”
Background
Research works printed but not linked: 1
- Gers F.A. Schraudolph N.N. Schmidhuber J. Learning precise timing with LSTM recurrent networks J. Mach. Learn. Res. 2003 3 115 143
Cited by
Items 2,401–2,500 of 2,807
Saam, G., & DAĞLI, C. (2026) · Procedia Computer Science · doi:10.1016/j.procs.2026.08.190
Performance Evaluation of VGG_19 Model for Brain Tumor Datasets
Reddy, G. N., Kumar, B. G., Pavithra, A., et al. (2024) · International Journal of Scientific Research in Computer Science Engineering and Information Technology · doi:10.32628/cseit241031
Pan, B., Wang, Y., & Xiang, M. (2026) · Bioengineering · doi:10.3390/bioengineering13101112
Ning, Z., Feng, Y., & Wang, J. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4871784
Encoding Physical Conditioning from Inertial Sensors for Multi-step Heart Rate Estimation
de Aguiar, D. P., & Murai, F. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-91699-2_5
Feature representation learning for human activity recognition
Khaertdinov, B. (2024) · Research Publications (Maastricht University) · doi:10.26481/dis.20240618bk · Cited by 1
Abboud, M. (2023) · theses.fr (ABES) · doi:10.70675/e7fc88d3zdaadz4db8z9759za64a65ba0e9c · Cited by 3
Spatio-temporal data analytics in the context of environmental crowdsensing
Hafyani, H. E. (2022) · theses.fr (ABES) · doi:10.70675/f05cddffzc5daz43f9za12ez7ae20d3c44d4
Li, Z., Tao, J., & Shou, D. (2026) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2026.3652182
Toward Multimodal Detection of Body-Focused Repetitive Behaviors Using Vision and Wearable Sensors
Sasikumar, N., Bhansali, J., & Tiglao, N. M. C. (2026) · International Symposium on Networks, Computers and Communications (ISNCC) · doi:10.1109/isncc70543.2026.11693919
Modeling cognitive impairment in Parkinson's patients using deep learning
Thakur, A., Sharma, R., Devi, A., Kumar, K., & Shah, T. M. (2026) · Elsevier eBooks · doi:10.1016/b978-0-443-27722-1.00013-4
Madaeni, F. (2021) · The Cryosphere · doi:10.5194/tc-2021-194-ac2
RFA: Regularized Feature Alignment Method for Cross-Subject Human Activity Recognition
Yang, Z., Fu, H., Huan, R., Qu, M., & Pan, Y. (2025) · International Journal of Pattern Recognition and Artificial Intelligence · doi:10.1142/s0218001425510085
Convolutional and Long Short-term Memory Neural Network for Earthquake Detection
(2019) · International Workshop on Computer Science and Engineering (WCSE) · doi:10.18178/wcse.2019.06.057
Li, B., Zuo, K., Li, L., & Wu, Y. (2026) · Measurement Science and Technology · doi:10.1088/1361-6501/ae6469 · Cited by 3
“Recently, deep learning methods, notably CNN and long short-term memory (LSTM) networks, have significantly advanced HAR performance due to their powerful representation learning capabilities [20–22].”
DG-IVF: A Dual-Gated IMU-Vision Fusion Method for Efficient HAR
Li, T., Li, D., Chen, Y., Wang, Y., Feng, X. T., Ma, Y., Liu, D., Wu, J. Y., & Guo, C. (2025) · Communications in computer and information science · doi:10.1007/978-981-95-4091-4_28 · Cited by 1
Octave Mix: Data augmentation using frequency decomposition for activity recognitionPreprint
Hasegawa, T. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2101.02882
“A method combining multiple sensors [27], methods inserting recurrent layer after several convolution-pooling layers [28, 29], and more advanced methods introducing Inception, Residual, and Attention modules [30, 31, 32, 33, 34] have been studied.”
MethodA multi-method explainability approach for sensor-based health activity analytics
Kurniawan, A., Caesarendra, W., Ariyanto, M., Mursalim, M. K. N., & Hakim, D. L. (2026) · Healthcare Analytics · doi:10.1016/j.health.2026.100463
Polo-Rodríguez, A., Escalera-Esteban, A., Anguita-Molina, M. Á., Valenzuela-López, I., Correa‐Rodríguez, M., Rueda‐Medina, B., & Medina-Quero, J. (2026) · Electronics · doi:10.3390/electronics15143042
“…rates for continuous MARG acquisition, and sequential deep learning architectures—in particular LSTM networks and hybrid CNN+LSTM models—have demonstrated state-of-the-art performance on MARG time series, outperforming classical methods when temporal depen-https dencies span several seconds [4].”
Du, Y. (2025) · International Conference on Computing, Robotics and System Sciences (ICRSS) · doi:10.1109/icrss67786.2025.00032
“In addition, convolutional layers are combined with long short-term memory networks (LSTM) to effectively capture temporal dependencies [3].”
Wang, W., Hua, T., & Hu, Y. (2025) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc58881.2025.11343369
“Recurrent Neural Networks (RNNs), particularly Long Short-Term Memory (LSTM) [10] and Gated Recurrent Units (GRU) [11], have become popular for their ability to model long-term dependencies in HAR [2], [12].”
Zhang, Y. (2026) · Internet Technology Letters · doi:10.1002/itl2.70199
“Recurrent neural networks (RNN) [17]: A sequence-based model for fall detection that processes sequential sensor data.”
A Method for Workflow Segmentation and Action Prediction from Video Data - AR Content
Kumar, A., Agnihotram, G., Kumar, S., Reddy, R. S., & Naik, P. (2023) · Communications in computer and information science · doi:10.1007/978-3-031-27609-5_27
Multi-Channel Deep Networks on Sequence Data for Multi-Action Recognition
Qi, Q., Zhong, S., Wang, Y., & Long, J. (2020) · International Conference on Intelligent Computation Technology and Automation · doi:10.1109/icicta51737.2020.00087
“Recently, many researchers use CNNs for wearable activity recognition [14].”
BackgroundMethodFine-Grained Gesture Recognition via Narrow-Kernel CNN and Attention-based SEMG-ACC Fusion
Zhang, H., Boudaoud, S., & Rida, I. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11460574
Deep Neural Networks with Auxiliary-Model Regulated Gating for Resilient Multi-Modal Sensor FusionPreprint
Zhao, C., Shim, M. S., Li, Y., Zhang, X., & Li, P. (2019) · arXiv (Cornell University)
Performance Evaluation of Hybrid Intelligent Models for Human Activity Recognition
Umaru, M. Y., Ziggah, Y. Y., & Nunoo, S. (2026) · Cureus Journal of Computer Science. · doi:10.7759/s44389-026-00127-6
“By capturing hidden, non-linear structures in the data, NLPCA provides a more expressive feature representation than traditional PCA, making it a powerful tool in HAR applications where human motion data can be inherently non-linear [45].”
Arsenault-Boucher, L., Boucher, É., Lhissou, R., et al. (2026) · River Research and Applications · doi:10.1002/rra.70133
LEARNING EMBEDDINGS FOR WEARABLE-BASED HUMAN ACTIVITY ANALYSIS
Sheng, T. (2020) · Cited by 1
Data augmentation for inertial sensor based human action recognition using deep learning
Kuschan, J., Filaretov, H., & Krüger, J. (2024) · AIP conference proceedings · doi:10.1063/5.0190203 · Cited by 1
Yin, Z., Pulavarthy, L. P., & Purkayastha, S. (2026) · IEEE International Conference on Healthcare Informatics (ICHI) · doi:10.1109/ichi69079.2026.00036
Paliwal, G., & Bunglowala, A. (2018) · Advances in medical diagnosis, treatment, and care (AMDTC) book series · doi:10.4018/978-1-5225-7131-5.ch007
Idrees, M., Sial, M. H., & Hassan, N. U. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4860958
Artifact: Assistant for Teacher-Student Learning: Skeleton-to-IMU Transfer for Industrial HAR
Qiao, H., Xia, Q., Rizk, H., et al. (2026) · IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops) · doi:10.1109/percomworkshops68308.2026.11585331
Zhou, T., Xie, S., Zhang, X., Wang, Y., Xu, Y., & Jiang, X. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4766842
Differentially Private Deep Learning for Smartphone-Based Human Activity Recognition
Sarkar, I., Bagchi, A. K., & Shahriar, M. S. (2025) · International Conference on Computer and Information Technology (ICCIT) · doi:10.1109/iccit68739.2025.11491348
Oishi, N., & Numao, M. (2019) · National Conference on Artificial Intelligence
Optimizing End-to-End Sensor-Based Human Activity Recognition Through Multi-Attention InteractionPreprint
Yu, Y., Wang, H., Wang, J., Yan, M., Han, X., Wu, D., Shen, Q., & Liu, H. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4777248
Yeoh, J. K. X., Kouzbary, M. A., Osman, N. A. A., Chan, C. K., Shasmin, H. N., & Tham, L. K. (2025) · International Journal of Control · doi:10.1080/00207179.2025.2537253
“For example, Ordóñez and Roggen (2016) and Ihianle et al. (2020) integrated LSTM with CNNs for activity recognition.”
Fairbairn, C. E., & Bosch, N. (2026) · Clinical Psychological Science · doi:10.1177/21677026251379441 · Cited by 2
“…and extract a range of behavioral indicators from such time-series data, including everyday activities and eating behaviors from accelerometer data (Ordóñez & Roggen, 2016; Thomaz et al., 2015), sleep duration and sleep stage from multisensor data (Boe et al., 2019), and alcohol-use episodes from…”
Deep Learning Mixed Hyper-Parameter Optimization Based on Improved Cuckoo Search Algorithm
Yu, T. O. N. G., Chen, R., & Hu, B. (2025) · Wuhan University Journal of Natural Sciences · doi:10.1051/wujns/2025302195
Efficient and Adaptive Human Activity Recognition via LLM Backbones
Bredikhin, A., Lalanda, P., & Vega, G. (2026) · IEEE International Conference on Smart Computing · doi:10.1109/smartcomp69968.2026.00022
Wavelet Neural Network-based Human Activity Recognition: A Gravity-Aware Hybrid Architecture
Puthiyaveettil, F. S., Ramanujam, E., Sahoo, A., & Tutu, P. Y. R. K. (2026) · International Conference on Intelligent Processing, Hardware, Electronics, and Radio Systems (CIPHER) · doi:10.1109/cipher70417.2026.11524015
“Similar approach in DeepConvLSTM by Ord´o˜nez and Roggen [27].”
Lu, S., Qian, Y.-Y., Yang, K.-M., et al. (2026) · International Journal of Control Automation and Systems · doi:10.1007/s12555-026-00097-4
“Typical classifi-cation methods include linear discriminant analysis (LDA) [23], hidden Markov models (HMM) [19, 24], artificial neural networks (ANN) [25, 26] and deep neural networks (DNN) [16, 27, 28].”
James, H., Rial, C., Giordano, J. O., & Erickson, D. (2026) · Frontiers in Animal Science · doi:10.3389/fanim.2026.1676504 · Cited by 2
Bouchabou, D. (2022) · theses.fr (ABES) · doi:10.70675/cdb4ca39z9dcbz4ff2z985ezfeb97c80760d · Cited by 2
Research on Ski Action Recognition Based on LSTM Method
Song, C., Cong, X., Liu, X., et al. (2025) · International Symposium on Intelligent Robotics and Systems (ISoIRS) · doi:10.1109/isoirs65690.2025.11168079
“Francisco et al. [4] proposed a method that combines deep convolutional neural networks (CNN) and LSTM recurrent neural networks for processing multimodal activity recognition tasks on wearable devices.”
Adaptive Fuzzy Residual Learning Framework for On-Skin Triboelectric Sensor Gesture Recognition
Fu, Z., Wang, H., Liao, Y., He, X., Chen, X. S., Huo, J., & Huang, J. (2026) · IEEE Transactions on Fuzzy Systems · doi:10.1109/tfuzz.2026.3695771
“To strengthen the comparison from a broader gesture-recognition and multivariate time-series perspective, we further evaluated representative deep baselines, specifically TCN-Inception [40], InceptionTime [41], DeepConvLSTM [42], TST [43], ConvTran [44], and MLSTM-FCN [45].”
Deep locomotion prediction learning over biosensors, ambient sensors, and computer vision
Javeed, M., Jalal, A., AlHammadi, D. A., & Lee, B. (2026) · PLoS ONE · doi:10.1371/journal.pone.0342793 · Cited by 1
Identificação da Doença de Parkinson através da Análise de Dados Coletados por Smartwatch
Padilha, R. A., Lobo, P. L. S., Salvini, R. L., et al. (2025) · Anais da XIII Escola Regional de Informática de Goiás (ERI-GO 2025) · doi:10.5753/erigo.2025.17114
An Advanced Attention-Enhanced Hybrid Deep Learning Model for WBSNs’ Gait Pattern RecognitionPreprint
Wu, J., Tan, Q., & Wu, X. (2022) · Research Square · doi:10.21203/rs.3.rs-1251973/v1
HT-ConvFormer: A Hierarchical Framework for Multi-scale Temporal Modeling in Table Tennis Analytics
Zhang, Y.-M., Lin, J.-H., Peng, P.-J., Chang, W.-Y., Lee, C.-C., & Fan, K. (2026) · Communications in computer and information science · doi:10.1007/978-981-95-9843-4_33
Building occupant sensing : occupancy prediction and behavior recognition
Zhu, Q. (2018) · doi:10.32657/10220/45814
“proposed a combination of convolution and Long Short-Term Memory (LSTM) recurrent neural network to extract high-level features based on different modalities of motion sensors [74].”
BackgroundAarthy, B., & Tamilpavai, G. (2026) · Signal Image and Video Processing · doi:10.1007/s11760-026-05542-9
“We compare the suggested lightweight YRYA model’s performance with three existing models: the Light Residual Model [1], the 1D CNN Residual BiLSTM Model [5] and the DeepConvLSTM Model [9].”
Adaptive Gated Multimodal Fusion for Robust and Generalized Human Activity Recognition
Gautam, S., & Shrivastava, A. (2026) · Journal of Trends in Computer Science and Smart Technology · doi:10.36548/jtcsst.2026.2.007
“All the architectural decisions such as the 2s-AGCN skeleton encoder (Shi et al., 2019) and the 1D CNN model for IMU (Ordóñez & Roggen, 2016) are chosen for their effectiveness in capturing spatiotemporal dependencies in human activity recognition.”
Pan, Y., Zhao, Z., & Xu, Y. (2026) · Sensors · doi:10.3390/s26196139
Li, Z., Miao, F., Zhang, W., & Gui, G. (2026) · Information Communication Technologies Conference (ICTC) · doi:10.1109/ictc70246.2026.11650001
Lightweight Relative Motion Stability Assessment Using Low-Sampling-Rate Smart Wristband Data
Xu, J., Bi, T., Zhang, L., Li, J., & Hu, T. (2026) · IEEE International Conference on Pattern Recognition, Machine Vision and Artificial Intelligence (PRMVAI) · doi:10.1109/prmvai70103.2026.11605629
“On the methodological side, classical hand-crafted features [19] and deep sequence models [23], [24] are the two main paradigms.”
Deep learning for stress oriented human activity recognition
Hamza, M., Uddin, N., Anjum, G., et al. (2026) · Frontiers in Digital Health · doi:10.3389/fdgth.2026.1729804
Handling Variable-Dimensional Time Series with Graph Neural NetworksPreprint
Gupta, V., Narwariya, J., Malhotra, P., Vig, L., & Shroff, G. M. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2007.00411 · Cited by 13
Détection de situations à risques pour des personnes fragiles sur des données multimodales
Yebda, T. (2022) · theses.fr (ABES) · doi:10.70675/e50bced2zd936z4369zbcf1zc0773a26d47c
Design of Ball Game Data Acquisition and Visualization Platform Supporting Mobile Terminal
Zhong, J. (2021) · International Conference on Smart Electronics and Communication (ICOSEC) · doi:10.1109/icosec51865.2021.9591624
Zhou, T., Xie, S., Zhang, X., Wang, Y., Xu, Y., & He, Y. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4978175
Bayesian Inference in Infinite Multivariate McDonald’s Beta Mixture Model
Forouzanfar, D., Manouchehri, N., & Bouguila, N. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-42508-0_29
Essafi, K., & Moussaid, L. (2026) · International Conference on Circuit, Systems and Communication (ICCSC) · doi:10.1109/iccsc71566.2026.11650057
“Third, Random Forest is widely deployed in real edge HAR systems because of its low inference cost and memory footprint [3, 10], so the present characterisation is directly relevant to production-class pipelines.”
Advanced Data Analysis Methods for Wearable Device Data
Xu, W., Zhang, W., Chen, T., Sakal, C., Yao, X., & Li, X. (2026) · Springer eBooks · doi:10.1007/978-3-032-19984-3_1
Wang, W., Wang, C., Chen, Z., & Huang, W. (2026) · Sensors · doi:10.3390/s26123681
“For instance, the DeepConvLSTM framework demonstrates high recognition precision in processing long-sequence inertial data [14,15].”
Halgamuge, M. N., Gondal, I., Jolfaei, A., Wang, H., Juang, C., & Srinivasa, N. (2026) · International Conference on Communications · doi:10.1109/icc59461.2026.11586959
Carlos, W. C., Copetti, A., Bertini, L., et al. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4818865
Zhang, B., Deng, K., Shen, J., Cai, L., Ratitch, B., Fu, H., & Guan, Y. (2021) · doi:10.2196/preprints.27934 · Cited by 5
Multi-attention mechanism based on gate recurrent unit for English text classification
Liu, H. (2018) · ICST Transactions on Scalable Information Systems · doi:10.4108/eai.27-1-2022.173166
“[16] obtained good results by using the recurrent neural network based on long and short-term memory (LSTM) to encode text.”
BackgroundContext Abstraction to Improve Decentralized Machine Learning in Structured Sensing Environments
Hamidi, M., & Osmani, A. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-26409-2_39 · Cited by 1
“– DeepConvLSTM [24]: a model encompassing 4 convolutional layers responsible for extracting features from the sensory inputs and 2 long shortterm memory (LSTM) cells used to capture their temporal dependence.”
MethodBiswas, S. (2026) · Signals and communication technology · doi:10.1007/978-3-032-09241-0_4
Wearable IMU-based System for Proactive Fall Detection via Gait Pattern Recognition
Mangai, S., Maheswari, S., Pavalakodi, S., et al. (2026) · 2026 International Conference on AI-Driven Smart Systems and Ubiquitous Computing (ICAUC) · doi:10.1109/icauc68182.2026.11441259
“On the other hand, the central hub consists of a Raspberry Pi which works with Python and acquires the data using the bleak library followed by numpy and pandas for more processing [8].”
MySecureMap: A Geolocation-Based Recommender System for Safety and Healthy Environment
Tangworakitthaworn, P., Kaewpradub, P., Hathaichot, P., et al. (2025) · IEEE Region 10 Conference (TENCON) · doi:10.1109/tencon66050.2025.11374960
Calatrava-Nicolás, F. M., Stoyanov, T., & Martinez Mozos, O. (2026) · Scientific Reports · doi:10.1038/s41598-026-64846-5
Kocal, C. I., Güngör, O., Rosing, T., & Akşanlı, B. (2026) · Electronics · doi:10.3390/electronics15153313
Activity recognition and bioinspired approaches for robotics in intelligent environments
Ranieri, C. M. (2021) · Digital Library of Theses and Dissertations (Universidade de São Paulo) · doi:10.11606/t.55.2021.tde-11082021-112227 · Cited by 1
Impact of Three-Dimensional Video Scalability on Multi-View Activity Recognition using Deep LearningPreprint
Choi, J., Cheon, M., Choi, M., et al. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1709.10206
“Our model is similar to the long-term recurrent convolutional network (LRCN) in [5] and DeepConvLSTM in [16].”
MethodDeep Learning Combined with Machine Learning for ON/OFF Home Appliances
Li, H., & Tsai, C. (2024) · 2024 2nd International Conference on Artificial Intelligence and Power Engineering (AIPE) · doi:10.1109/aipe63225.2024.00010
“Nowadays, CNN are also applied to time series data such as human physiological data and financial data [9, 10].”
STORM: Hardware-Aware Tiny Transformer Co-Design for Low-Power Inertial Human Activity Recognition
Varaldi, A., Genta, C., Manzone, A., & Vacca, M. (2026) · Electronics · doi:10.3390/electronics15091924
Improved Human Activity Recognition Model based on Sequence Mixup
Tan, R., & Yan, X. (2020) · IEEE International Symposium on Parallel and Distributed Processing with Applications (ISPA) · doi:10.1109/ispa-bdcloud-socialcom-sustaincom51426.2020.00078
“In recent years, many resear chers have applied the CNN - LSTM model to the field of HAR and proved its superiority[12][15][20][21].”
BackgroundHuang, L., & Xia, C. (2026) · International Conference on Multimedia Retrieval · doi:10.1145/3805622.3810642
Madaeni, F., Chokmani, K., Lhissou, R., et al. (2021) · The Cryosphere · doi:10.5194/tc-2021-194-cc1
Intuitive Multi‐Scale Visual Feature Fusion for Automated Classification of Human Activity
Gautam, V., Kaur, G., SINHA, A., Verma, A. K., & Nag, A. (2026) · IET Cyber-Physical Systems Theory & Applications · doi:10.1049/cps2.70054
MUSIC-CAR: A Multi-Label Complex Health-Related Activity Recognition Dataset from Smartphone Sensors
Chandrasekaran, K., Buquicchio, L., Agu, E., & Rundensteiner, E. (2026) · IEEE International Conference on Digital Health (ICDH) · doi:10.1109/icdh72779.2026.00039 · Cited by 1
Toward Elderly-Care-Oriented Smart Home HAR via Sensor-Aligned Activity Taxonomy
Tran, L. B. N., Pham, D. D., Le, H. S., Huynh, K. T., Nguyen, S., Le, T. D., & Tuyen, N. T. V. (2026) · International Conference on Applying New Technology in Green Buildings (ATiGB) · doi:10.1109/atigb70203.2026.11628469
“Deep learning approaches have achieved higher accuracy on benchmark datasets, including convolutional and recurrent models [11, 14].”
Human Activity Recognition for Packing Processes using CNN-biLSTM
Angulo, A., Beltrán, J., & Castro, L. A. (2023) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/3594739.3610735
“The structure of the CNN-biLSTM is a variation of the architecture described in [12], this architecture was proposed in [1], combining convolutional, linear, and recurrent layers.”
Benchmark Performance for the Sussex-Huawei Locomotion and Transportation Recognition Challenge 2018
Wang, L., Gjoreski, H., Ciliberto, M., et al. (2019) · Springer series in adaptive environments · doi:10.1007/978-3-030-13001-5_11
Self-citationArtificial Intelligence Algorithm for Flexible Sensors
(2022) · Wiley eBooks · doi:10.1002/9783527833115.ch11
Inertial Signal Forecasting With Foundation Model Techniques
Wieland, C., & Pankratius, V. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3603456 · Cited by 3
“While dual-view FM forecasts signals at a single sensor location (cf. Section IV-B.4), activity classifiers typically perform better with multilocation inputs [48].”
Miladinović, A., Biscontin, A., Kresevic, S., et al. (2026) · World Congress on Medical Physics and Biomedical Engineering, September 7 - 12, 2009, Munich, Germany · doi:10.1007/978-3-032-37739-5_51
Trinh, L., & Ha, B. (2022) · International Symposium on Information and Communication Technology · doi:10.1145/3568562.3568610 · Cited by 2
“We compare our model with the following state-of-the-art models: CNN [3], self-attention ConvLSTM [24], and adaptive feature fusion (AFF) [27].”
MethodInfluentialHuman Activity Recognition Based on Neural Networks and Accelerators
Hu, X. (2024) · International Conference on Intelligent Computing and Human-Computer Interaction (ICHCI) · doi:10.1109/ichci63580.2024.10807962
DNN Model Comparison for Sensor Location Robustness
Enokibori, Y., Sato, T., & Mase, K. (2024) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/3675094.3678449
A Survey on Machine Learning Techniques for Recognizing Human Activities Using Smartphone Data
Das, P., & Singh, T. R. (2025) · Communications in computer and information science · doi:10.1007/978-3-031-81342-9_23
IMU-based identification of rowing conditions through supervised machine learning
Leber, A., Siebert, T., Rapp, W., & Held, S. (2026) · BMC Sports Science Medicine and Rehabilitation · doi:10.1186/s13102-026-02082-0
“We evaluated (7) a one-dimensional convolutional neural network (CNN) [52], (8) a long short-term memory (LSTM) network [53], (9) a CNN–LSTM hybrid [54, 55], and (10) an ensemble that averaged the class probabilities of the CNN, LSTM, and hybrid models [56].”
InfluentialGarcia, A., & Huang, M. (2023) · doi:10.20944/preprints202309.1675.v1 · Cited by 9
A Systematic Review of Free-Living Human Activity Recognition (FL-HAR) Methods for Health Research
Arpitha, S., Manjaiah, D. H., & Yogesh, K. M. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-97-9523-9_39
Evaluation of Deep Learning Models for Smoking Recognition with Smartwatch and Smartphone Sensors
Akan, Y., Agac, S., & İncel, Ö. D. (2021) · Balkan Journal of Electrical and Computer Engineering · doi:10.17694/bajece.827342
Metadata
- Authors
- Ordóñez, F., & Roggen, D.
- Affiliations
- University of Sussex (GB)
- Corresponding author
- Ordóñez, Francisco
- Venue
- Sensors
- Publisher
- MDPI AG
- Published
- 2016-01-18
- Volume
- 16
- Issue
- 1
- Pages
- 115
- ISSN
- 1424-8220 (ISSN-L)
- PMC
- PMC4732148
- OpenAlex
- W2270470215
- Cited by
- 2,761
- References
- 49
- Citations by year
- 2026: 284 · 2025: 256 · 2024: 284 · 2023: 270 · 2022: 319 · 2021: 379 · 2020: 347 · 2019: 288 · 2018: 219 · 2017: 93 · 2016: 21
- FWCI
- 140.52
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Sensors (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Sussex Research Online (University of Sussex) (submittedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Figshare (submittedVersion, cc-by)
- Funders
- Google Faculty Research Award (Is deep learning useful for wearable activity recognition?)
- EPSRC (EP/N007816/1)
- 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 › Signal Processing › Time Series Analysis and Forecasting
- Keywords
- human activity recognition, wearable sensors, multimodal sensor fusion, convolutional neural networks, LSTM neural network, temporal dynamics, automatic feature extraction, hyperparameter optimization
- MeSH terms
- Machine Learning, Clothing, Humans, Human Activities, Signal Processing, Computer-Assisted, Databases, Factual, Neural Networks, Computer, Monitoring, Ambulatory
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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
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.