Evaluating BiLSTM and CNN+GRU Approaches for Human Activity Recognition Using WiFi CSI Data
This work addresses activity recognition for applications like healthcare and smart homes, but it is incremental as it compares existing methods on new data.
This paper compared BiLSTM and CNN+GRU models for Human Activity Recognition using WiFi CSI data, finding that CNN+GRU achieved 95.20% accuracy on the UT-HAR dataset while BiLSTM reached 92.05% on the NTU-Fi HAR dataset.
This paper compares the performance of BiLSTM and CNN+GRU deep learning models for Human Activity Recognition (HAR) on two WiFi-based Channel State Information (CSI) datasets: UT-HAR and NTU-Fi HAR. The findings indicate that the CNN+GRU model has a higher accuracy on the UT-HAR dataset (95.20%) thanks to its ability to extract spatial features. In contrast, the BiLSTM model performs better on the high-resolution NTU-Fi HAR dataset (92.05%) by extracting long-term temporal dependencies more effectively. The findings strongly emphasize the critical role of dataset characteristics and preprocessing techniques in model performance improvement. We also show the real-world applicability of such models in applications like healthcare and intelligent home systems, highlighting their potential for unobtrusive activity recognition.