Wi-SensiNet: Through-Wall Human Activity Recognition Based on WiFi Sensing

With the advancement of Wi-Fi sensing technology, its significant benefits in operational convenience and privacy protection have become increasingly evident, particularly in fields such as smart homes, medical monitoring, and security surveillance, where Human Activity Recognition (HAR) technology shows great promise. This study investigates a novel approach to Human Activity Recognition (HAR) using Wi-Fi Channel State Information (CSI), with a particular focus on complex conditions such as Non-Line-of-Sight (NLoS) scenarios, including Through-Wall transmissions. Traditionally, most research has focused on HAR in Line of Sight (LoS) paths, which are highly sensitive to environmental changes, while NLoS signals, especially through-wall signals, suffer from signal attenuation due to walls. To address this issue, we propose Wi-SensiNet, an innovative deep learning-based method that combines the spatial feature extraction capabilities of Convolutional Neural Networks (CNN) with the temporal sequence processing power of Bidirectional Long Short-Term Memory (BiLSTM) networks. This method also incorporates an attention mechanism to enhance the accuracy of human activity recognition in complex environments. Wi-SensiNet is specifically designed for through-wall scenarios, effectively managing the complexity of CSI data and achieving accurate through-wall human activity detection. In our experiments, we collected a through-wall CSI dataset comprising seven common activities, including running, sitting, standing, squatting, falling, punching, and walking, and verified that Wi-SensiNet achieved an average accuracy exceeding 99% on the original test set. These results demonstrate the model’s robustness and high accuracy in handling HAR tasks in complex environments.