CS Paper Radar

A Proof-of-Concept Study of Wi-Fi Channel State Information-Based Fall Detection Using ESP32 Devices

Amritha S. S., Anhajul S. Rahman, Mohammed Bin Saleem, Vaishnav Sankar, Jerrin Thomas Panachakel

eess.SP ·submitted 2026-11-11 14:17 UTC · v1

Abstract

This paper reports a small-scale proof-of-concept study of contactless fall detection using Wi-Fi channel state information (CSI) collected with two ESP32 devices. One device acts as the transmitter and the other as the receiver, with a separation of 3.5~m. The evaluated signal-processing pipeline uses CSI amplitude, second-level discrete wavelet transform (DWT) denoising with a Symlet-10 wavelet, z-score normalization, and fixed-length segmentation. Wavelet-scattering coefficients are used as features, and MATLAB's automated classifier-selection workflow identifies a LogitBoost ensemble with a minimum leaf size of four and a learning rate of 0.022052. The proposed classifier achieved an accuracy of 93.10%, with a precision of 100%, recall of 87.50%, and F1-score of 93.33%. These results demonstrate the potential of Wi-Fi CSI-based sensing for contactless fall detection using low-cost ESP32 hardware.

Closest papers

Nearest neighbours by meaning across the whole index.

Through-Wall Detection using Software-Defined Radio based on adaptive Principal Component Analysis

eess.SP ·2026-09-11 ·Dinuli Naotunna, Wenchao Li, Sanka Piyaratna, Phil Wandel
similarity 0.61
Abstract · 15 pages, 7 figures

Through-Wall Detection (TWD) using opportunistic WiFi signals enables non-invasive sensing for security and rescue applications; however many existing approaches rely on controlled access points or specialised hardware. This paper presents a TWD system that extracts Channel State Information (CSI) from ambient WiFi packets using a customised software-defined radio (SDR), Bluebottle, without requiring transmitter control. The key contribution is a spectral-domain scoring mechanism for adaptively selecting motion-relevant principal components from a Principal Component Analysis (PCA) decomposition of the CSI, using Welch power spectral density estimates to quantify each component's signal-to-noise ratio and spectral concentration within the frequency band associated with human motion. The selected components are then analysed using a continuous wavelet transform to robustly identify time-localised motion events. Experimental results demonstrate that the proposed adaptive component-selection method consistently reduces false detections and produces sharper time-frequency energy ridges compared to conventional fixed-component PCA.

0.26

Device Identification and Motion Classification of Passive Nonlinear RF Transponders Using Wi-Fi Channel State Information

eess.SP ·2026-10-31 ·Cory Hilton, Jeffrey A. Nanzer
similarity 0.58
Abstract

A method of joint device authentication and gesture classification using low-cost passive transponders and 802.11 signals of opportunity from commodity hardware is presented for tracking objects in cluttered environments. The approach detects nonlinear responses of diode-based radio-frequency (RF) transponders by estimating the mean time-varying phase across all subcarrier channel state information (CSI) obtained from standard preambles. These unique responses are enabled by passive transponders utilizing 2.4 GHz meander-line monopole antennas with capacitive top-hat loading and a low-barrier Schottky diode, allowing for low-power omnidirectional sensing and complex gesture recognition across multiple unique devices. The capture system uses a wireless access point to illuminate the transponder with a standard preamble. This signal is modulated by the diode's nonlinear response and transmitted to a single ESP32 receiver to calculate the channel response. Using the modulated CSI, device identification via RF fingerprinting achieves >95% accuracy in both static and dynamic conditions using simple machine learning models. For dynamically moving transponders, a time-varying velocity profile is calculated using a novel phase sanitation approach that significantly reduces the impacts of the non-cooperative network. This enables high-accuracy estimation of micro-motions on a single receiver, achieving a theoretical limit of 9.75 m/s using an amplitude-invariant binary comparator form of the sanitized waveform. The concept is experimentally validated using lightweight CNN-LSTM deep-learning models applied to a diverse dataset of five transponders and three motion profiles. This experimental validation demonstrated a classification accuracy of 98% in both gesture classification and device identification tasks.

0.35

From Activity Recognition to Shift Detection: Variational Wi-Fi Sensing with Evidential Uncertainty

cs.LG ·2026-11-18 ·Luca Cotti, Marco Cominelli, Francesco Gringoli, Lance M. Kaplan, Mani B. Srivastava +5 more
benchmark similarity 0.56
Abstract · 17 pages,8 figures, 8 tables

Human motion perturbs wireless channels, making commodity Wi-Fi devices a practical platform for device-free activity recognition. Yet Channel State Information (CSI) also depends strongly on the person, receiver position, recording session, and surrounding environment, so a classifier can continue producing activity labels after its deployment conditions have moved away from those seen during training. We propose Wi-HAR, a multi-antenna architecture in which four Variational Auto-Encoders learn compact distributions from raw Wi-Fi 6 CSI magnitudes and a delayed-fusion classifier combines their means and log-variances. On twelve activities from a public benchmark, Wi-HAR reaches 96.6% median recognition accuracy, compared with 70.8-81.5% for independently optimized single-antenna models, showing that the antennas provide complementary views of the same activity. We then use the same latent backbone to detect deployment shifts without using shifted samples for model selection or calibration. A class-conditional uncertainty score derived from Evidential Deep Learning achieves a macro-average ROC-AUC of 0.897 across eleven datasets that change receiver position, participant, recording session, or room, and detects 82.5% of shifted samples at a realized 5.2% in-distribution rejection rate. Detection performance nevertheless varies substantially across receiver positions, showing that the same physical change does not induce a uniform uncertainty response. The compact representation also supports shallow decision-tree analysis using variables from several antennas.

0.26

Temporal vs Spectral Resolution in CSI Reporting for Wi-Fi ISAC

eess.SP ·2026-10-16 ·Abolfazl Zakeri, Sumit Roy, Nhan Thanh Nguyen, Markku Juntti
similarity 0.51
Abstract · Under review for publication

CSI feedback is fundamental to proposed uses of Wi-Fi sensing such as human activity recognition (HAR); but it also constitutes significant overhead as resources must be allocated away from data communications. Wi-Fi CSI extraction can be explored along two dimensions in IEEE 802.11bf: reporting intensity, i.e., temporal resolution, and number of reported subcarriers, i.e., spectral resolution. We explore periodic and IEEE 802.11bf threshold-based reporting at matched rates, and uniform decimation and block-mean grouping at matched report sizes. We further propose reporting the mean CSI accumulated since the last report. Using the 802.11bf report size as the metric, we evaluate the accuracy-cost tradeoffs on three public datasets with our proposed model, a convolutional neural network combined with a gated recurrent unit and temporal pyramid pooling. The proposed periodic reporting with mean payload and block-mean grouping improves accuracy by up to 53 percentage points at equal CSI overhead cost. Temporal resolution is generally more cost-effective than spectral resolution. With only at most 7.7% of full-CSI feedback cost, the proposed scheme achieves 96-98% of full-CSI accuracy.

0.23

Trajectory-Guided Tokenization of Complex CSI for Wi-Fi Sensing

cs.LG ·2026-10-05 ·Ziyi Wang, Kenuo Xu, Jichu Jiang, Yumeng Yang, Zheng Chen +6 more
similarity 0.50
Abstract

Wi-Fi channel state information (CSI) enables contactless presence detection and gesture recognition. Its high-dimensional complex-valued time series require input representations that preserve informative temporal variations during compression. We propose Trajectory-Guided Tokenization (TGT), which combines complex trajectory decomposition with asymmetric attention to construct compact continuous tokens. For each antenna link and subcarrier, an orthonormal Helmert transform decomposes short, ordered temporal patches into local-center and centered-trajectory coordinates. Keys are learned from the centered-trajectory coordinates, while values retain both components. Learnable queries aggregate subcarriers into frequency slots, which are fused into temporal tokens. Trained jointly from scratch, TGT with TokenMLP achieves the highest mean accuracy of 92.83% among all evaluated frontend-backend combinations on the self-collected dataset. Experiments on EHUNAM and Widar further support the applicability of TGT to cross-domain presence detection and gesture recognition.

0.32

On-device Detection of Wearer-induced NLOS on Chest-worn UWB Tags: Machine Learning, Link Indicators, or Both?

cs.LG ·2026-10-30 ·Markus Gallacher, Vu Anh Minh Le, Matteo Trobinger, Davide Vecchia, Carlo Alberto Boano +1 more
similarity 0.49
Abstract · Submitted to ACM Transactions on Sensor Networks

Ultra-wideband (UWB) enables high-accuracy distance estimation (ranging), but performance degrades in non-line-of-sight (NLOS) conditions. When UWB tags are worn, the human body inevitably causes a wearer-induced NLOS (WNLOS), yielding significant errors; still, despite its relevance, WNLOS remains understudied. Moreover, WNLOS detection typically must occur on-device at the tag, requiring classification accuracy to be balanced against tight runtime and energy constraints often hard to meet with machine learning (ML) -- the dominating NLOS classification approach. Meanwhile, UWB radio link indicators (LI) offer a faster and lightweight alternative that has rarely been explored. We study existing on-device ML and LI classifiers when applied to WNLOS, and propose novel hybrid ones exploiting their complementarity. We focus on chest-worn tags and analyze 21 classifiers on a dataset of 1.5 million samples across channels, PHY settings, environments, subjects, and orientations. Results show that while all of them achieve good classification accuracy, relative performance, including ranging errors due to misclassification, depends heavily on channel and PHY configuration. Runtime and energy efficiency also vary dramatically; some techniques offer marginal accuracy gains while consuming up to 39.6x more energy. Therefore, we also introduce a recommender tool to identify optimal classifiers under user-defined constraints, publicly released with our dataset.

0.26

A Subcarrier-Aware Approach for Robust Respiratory Monitoring with Commodity Wi-Fi

eess.SP ·2026-08-26 ·Pei Tang, Yunpeng Ge, Ivan Wang-Hei Ho
similarity 0.49
Abstract · 13 pages, 24 figures

Wi-Fi sensing has emerged as a promising modality for contact-free respiratory monitoring in home healthcare due to its ubiquity. However, conventional approaches typically treat Channel State Information (CSI) subcarriers uniformly, neglecting their heterogeneous responses to breathing motions. To address this limitation, we propose a subcarrier-aware sensing framework that characterizes subcarrier-dependent respiratory sensitivities and exploits them for informative subcarrier selection. Based on the statistical properties of breathing-sensitive subcarriers, we further develop an unsupervised clustering-based breathing estimation method for robust respiratory monitoring. Experimental results show that the proposed method achieves 97% breathing rate estimation accuracy and reduces the mean absolute error (MAE) by 0.45 breaths per minute (bpm) compared with baseline methods. Furthermore, we extend the analysis to breathing-pause detection, which derives an amplitude-attenuation-based threshold mechanism. Under breathing pauses, the proposed method achieves an MAE of 1.6 bpm in breathing rate estimation. These results demonstrate the robustness and effectiveness of the proposed framework across different scenarios and its potential for contactless home healthcare monitoring.

0.36

Entropy-Based Wireless Technology Recognition

eess.SP ·2026-10-16 ·Mohammad Alhellani, Eli De Poorter, Mamoun Guenach, Adnan Shahid
similarity 0.48
Abstract

Wireless Technology Recognition (WTR) enables identifying wireless technologies from received radio signals and is important for spectrum monitoring and management systems. Existing approaches either rely on handcrafted signal features with limited robustness or computationally intensive Deep Learning (DL) models operating on high-dimensional In-phase and Quadrature (IQ) samples or spectrograms. In this paper, we investigate a different hypothesis: wireless technologies can be distinguished through the temporal evolution of their local time-frequency complexity. To capture this behavior, we propose a lightweight WTR framework based on two-dimensional dispersion entropy. Entropy values are extracted from successive spectrogram regions to characterize the evolution of local spectrogram complexity throughout a transmission and summarized into a compact feature representation for classification using a Multilayer Perceptron (MLP). The framework is evaluated on seven wireless technologies, namely IEEE 802.11ah, IEEE 802.15.4 SUN-FSK, IEEE 802.15.4 SUN-OFDM, LoRa, IEEE 802.11n, LTE, and 5G NR, at Signal-to-Noise Ratios (SNRs) ranging from -20 dB to 40 dB. Experimental results show that the proposed approach consistently outperforms IQ-based and Short-Time Fourier Transform (STFT)-based Convolutional Neural Network (CNN) baselines, particularly at low and moderate SNRs. Averaged across all evaluated SNR levels, it achieves 82.73% accuracy, corresponding to improvements of 23.8% and 14.6% over the IQ-based and STFT-based CNN approaches, respectively. Furthermore, the approach maintains its recognition advantage with data-augmented training while requiring substantially shorter training times. The results demonstrate that temporal evolution of local spectrogram complexity provides a robust technology-specific fingerprint and an effective, computationally efficient representation for WTR.

0.21