Enhancing CSI-Based Human Activity Recognition by Edge Detection Techniques
چکیده
Human Activity Recognition (HAR) has been a popular area of research in the Internet Things (IoT) and Human–Computer Interaction (HCI) over past decade. The objective this field is to detect human activities through numeric or visual representations, its applications include smart homes buildings, action prediction, crowd counting, patient rehabilitation, elderly monitoring. Traditionally, HAR performed vision-based, sensor-based, radar-based approaches. However, vision-based sensor-based methods can be intrusive raise privacy concerns, while require special hardware, making them more expensive. WiFi-based cost-effective alternative, where WiFi access points serve as transmitters users’ smartphones receivers. method mainly using two wireless-channel metrics: Received Signal Strength Indicator (RSSI) Channel State Information (CSI). CSI provides stable comprehensive information about channel compared RSSI. In research, we used convolutional neural network (CNN) classifier applied edge-detection techniques preprocessing phase improve quality activity detection. We data converted into RGB images tested our methodology on three available datasets. results showed that proposed achieved better accuracy faster training times than simple RGB-represented data. order justify effectiveness approach, repeated experiment by applying raw long short-term memory (LSTM) Bidirectional LSTM classifiers.