Efficiently Improving the Wi-Fi-Based Human Activity Recognition, Using Auditory Features, Autoencoders, and Fine-Tuning
Article
Rahdar, A., Chahoushi, M. and Ghorashi, S. 2024. Efficiently Improving the Wi-Fi-Based Human Activity Recognition, Using Auditory Features, Autoencoders, and Fine-Tuning. Computers in Biology and Medicine. 172 (Art. 108232). https://doi.org/10.1016/j.compbiomed.2024.108232
Authors | Rahdar, A., Chahoushi, M. and Ghorashi, S. |
---|---|
Abstract | Human activity recognition (HAR) based on Wi-Fi signals has attracted significant attention due to its convenience and the availability of infrastructures and sensors. Channel State Information (CSI) measures how Wi-Fi signals propagate through the environment. However, many scenarios and applications have insufficient training data due to constraints such as cost, time, or resources. This poses a challenge for achieving high accuracy levels with machine learning techniques. In this study, multiple deep learning models for HAR were employed to achieve acceptable accuracy levels with much less training data than other methods. A pre-trained encoder trained from a Multi-Input Multi-Output Autoencoder (MIMO AE) on Mel Frequency Cepstral Coefficients (MFCC) from a small subset of data samples was used for feature extraction. Then, fine-tuning was applied by adding the encoder as a fixed layer in the classifier, which was trained on a small fraction of the remaining data. The evaluation results (K-fold cross-validation and K=5) showed that using only 30% of the training and validation data (equivalent to 24% of the total data), the accuracy was improved by 17.7% compared to the case where the encoder was not used (with an accuracy of 79.3% for the designed classifier, and an accuracy of 90.3% for the classifier with the fixed encoder). While by considering more calculational cost, achieving higher accuracy using the pre-trained encoder as a trainable layer is possible (up to 2.4% improvement), this small gap demonstrated the effectiveness and efficiency of the proposed method for HAR using Wi-Fi signals. |
Keywords | Autoencoder; Channel State Information; Deep Learning; Fine-Tuning; Human Activity Recognition; Machine Learning; Mel Frequency Cepstral Coefficient |
Journal | Computers in Biology and Medicine |
Journal citation | 172 (Art. 108232) |
ISSN | 0010-4825 |
1879-0534 | |
Year | 2024 |
Publisher | Elsevier |
Publisher's version | License File Access Level Anyone |
Digital Object Identifier (DOI) | https://doi.org/10.1016/j.compbiomed.2024.108232 |
Publication dates | |
Online | 27 Feb 2024 |
Apr 2024 | |
Publication process dates | |
Accepted | 25 Feb 2024 |
Deposited | 26 Feb 2024 |
Copyright holder | © 2024, The Author |
https://repository.uel.ac.uk/item/8x595
Download files
Publisher's version
Efficiently improving the Wi-Fi-based h...pdf | ||
License: CC BY 4.0 | ||
File access level: Anyone |
67
total views113
total downloads3
views this month3
downloads this month