Traditional human activity recognition (HAR) systems utilize vision and wearable sensors to sense and identify human movements. When applied for security and surveillance, these HAR systems can also be utilized for human identity recognition (HIR), where the sensing technologies authenticate users. Recently, Wi-Fi sensing using channel state information (CSI) has emerged as a promising alternative technology for HIR applications, wherein the users are typically identified by their gaits as the primary biometric parameter. As there are other possible biometric parameters, this study explores Wi-Fi CSI sensing in HIR where the subjects perform three physical (static) and two behavioral (dynamic) biometric parameters. A CSI dataset involving samples collected from six subjects in a controlled indoor environment was preprocessed using discrete wavelet transform (DWT) and principal component analysis (PCA). Different parameter-specific HIR models were trained and tested and the HIR model based on the ‘draw triangle’ behavioral parameter achieved the highest HIR accuracy of 98.61 %. Results show that the five investigated gestures are viable biometric parameters alternative to gait.

Conventional crowd counting (CC) systems employ cameras, LIDAR, or environmental sensors to detect the number of people in a target environment. However, these systems require line-of-sight (LOS) to recognize and estimate the crowd effectively. An alternative sensing method uses Wi-Fi channel state information (CSI) derived from the received signals. Despite Wi-Fi's capability to work in non-line-of-sight (NLOS) conditions, recent Wi-Fi CSI CC systems focus mostly on LOS scenarios, with some systems utilizing multiple pairs of TX-RX devices to improve the classification performance. Therefore, this study investigates LOS and NLOS scenarios for Wi-Fi CSI CC systems utilizing single- or multipair devices. We gathered diverse datasets and applied two preprocessing techniques: discrete wavelet transform (DWT) for noise filtering and principal component analysis (PCA) for dimension reduction. These datasets were utilized in training machine learning models to recognize three classes: 0,1, and 3 persons. Results show that adding another pair of TX-RX devices boosts the CC classification performance to reach 100% accuracy even in NLOS or obstructed conditions.

Human presence detection has various applications, with most existing methods relying on fixed-frequency RF signals like Wi-Fi, limiting transmitter-receiver distance. This study investigates the feasibility of using 4G Long Term Evolution (LTE) bands such as bands 28, 3, 1, and 41 for human presence detection in indoor environments. LTE signal data were collected during video call sessions using a Software-Defined Radio (SDR). Support Vector Machine (SVM) models were trained and evaluated on 0 vs. 1 person, 0 vs. 3 person, and 0 vs. 1 vs. 3 or multiple classifications per band. Results show that band 41, which has the highest frequency range, provides better accuracy in detecting human presence for 0 vs. 3 and multi-person classification, while Band 28, which has the lowest frequency range, performs better for detecting no-person scenarios. This highlights the impact of LTE frequency variation on human presence and crowd density detection.

Sign language recognition (SLR) systems often employ cameras and wearable sensors to detect and interpret signs and gestures. However, vision-based systems rely on LOS and may lead to privacy issues, while wearable systems require calibration and user contact. Alternatively, radio-frequency (RF) sensing, particularly Wi-Fi sensing, may be used for contactless NLOS SLR systems. While there have been some efforts in investigating different sign languages using RF systems, to our knowledge, none have explored the Filipino sign language (FSL) using Wi-Fi channel state information (CSI). In this study, we gathered CSI time series data from different participants performing FSL gestures one at a time using Wi-Fi devices. A total of 600 CSI samples were acquired and preprocessed using outlier removal, denoising and dimension reduction. Our Wi-Fi FSLR systems trained on 80% of the dataset was able to correctly classify up to 83 out of 120 test samples, that is 69.3% accuracy, showing the feasibility of using Wi-Fi CSI for FSLR.

Crowd counting systems are generally categorized as vision-based and sensor-based, wherein the former requires LOS and may cause privacy concerns, and the latter requires contact and calibration. Recent systems explored the use of radio signals, particularly from Wi-Fi, to perform crowd counting and estimation. Wi-Fi crowd counting systems use channel state information (CSI) as input to the classifier to sense the number of people in the environment. These systems often do not investigate the environment dependence of the data (i.e., the training environment is the same as the testing environment). In this study, we gathered thousands of CSI time samples for 3 crowd count classes (0-person, 3-person, and 5-person) in various indoor environments for several days in the span of 3 months. Due to the high temporal and environmental dynamics caused by the 3-month long data collection, our crowd counting system trained without a prior knowledge of the target environment did not perform as well as the system trained with partial knowledge. By partially training the crowd counting model, we can see a performance boost of around 16–20% in the classification accuracy.