Radio frequency (RF) fingerprinting is considered as one of the promising techniques to enhance wireless security in the Internet of Things (IoT) applications. In this paper, a low-complexity RF fingerprinting method for classification of wireless IoT devices is proposed. The method is based on the energy spectrum of the transmitter turn-on transient signals from which unique characteristics of wireless devices are extracted. The number of spectral components to be used is determined through a proposed approach based on the estimated transient duration value. Transient duration estimation is achieved from the smoothed versions of the instantaneous amplitude characteristics of transmitter signals, which are obtained through a sliding windowaveraging method. Classification performance of the proposed spectral fingerprints is assessed using experimental data and described by a confusion matrix. The discrimination effectiveness of the spectral fingerprints is quantified by a class separability criterion and evaluated for different noise levels through Monte Carlo simulations. It is demonstrated that the proposed fingerprints outperform the classification performance of two existing fingerprints especially at low signal-to-noise ratio. Additionally, computational complexity analysis of the classifier using the proposed fingerprints is provided.

In this paper, a sequential transient detection method for radio frequency (RF) fingerprinting used in the identification of wireless devices is proposed. To the best knowledge of the authors, sequential detection of transient signals for RF fingerprinting has not been considered in the literature. The proposed method is based on an approximate implementation of the generalized likelihood ratio algorithm. The method can be implemented online in a recursive manner with low computational and memory requirements. The transients of wireless transmitters are detected by using the likelihood ratio of the observations without the requirement of any a priori knowledge about the transmitted signals. The performance of the method was evaluated using experimental data collected from 16 Wi-Fi transmitters and compared to those of two existing methods. The experimental test results showed that the proposed method can be used to detect the transient signals with a low detection delay. Our proposed method estimates transient starting points 20-times faster compared to an existing robust method, as well as providing a classification performance of a mean accuracy close to 95%.

Radio frequency (RF) fingerprinting is regarded as a promising solution to improve wireless security, especially in applications where resource-limited devices are employed. Unlike steady-state signals, such as preambles or data, the use of short-duration transient signals for RF fingerprinting offers distinct advantages for systems with low latency and low complexity requirements. One of the challenges associated with transient-based methods in practice is achieving high performance while utilizing low-cost receivers. In this study, we demonstrate for the first time that the performance of transient-based RF fingerprinting can be enhanced by designing the filter chain in a software defined radio (SDR) receiver, taking into account the relevant signal characteristics. The performance analysis is conducted using transient-based identification signals captured by the SDR receiver, focusing on the sampling rate and duration of the identification signal. In the experiments, signals collected from twenty IEEE 802.11 transmitters are used. Experimental results indicate that so long as the receiver filter parameters and the duration of the identification signal are properly determined, a high classification performance exceeding 92% can be achieved for transient-based RF fingerprinting, even at sampling rates approaching the Nyquist limit.

In this work, a method on identification of transmitter by RF signature in order to enhance security in wireless local area networks where devices operates according to IEEE 802.11b standard is presented. In this scope, RF signatures of Wi-Fi devices are examined and identities of the devices are determined by designating the features of the devices by different classifiers. Also transmitter identity detection system based on RF signature is seen to be more secure than password applications.

In recent years, advances in computer vision and deep learning have enabled automated classification and sorting systems across the food processing and agricultural industries. Automated identification of chicken meat cuts (e.g., breast, drumstick, thigh) is essential for streamlining automated packaging, inventory management, and processing lines. However, the high computational complexity and associated energy consumption of deep neural networks often restrict their deployment on resource-constrained edge platforms. This study investigates the trade-off between classification performance and energy consumption during inference among five pre-trained convolutional neural network (CNN) architectures—SqueezeNet, MobileNet-v2, EfficientNet-b0, GoogleNet, and ResNet-50—to classify five anatomical categories of chicken meat: breast, drumstick, thigh, whole-wing, and wing. Feature extraction was performed using these pre-trained networks, followed by classification with a fully connected layer. Classification performance was evaluated using accuracy, precision, sensitivity, specificity, and F1-score, while processor energy consumption during inference was measured using Intel Power Gadget. The experimental results demonstrate that EfficientNet-b0 achieves the highest overall classification accuracy (94.14%), while SqueezeNet exhibits the lowest measured energy consumption (2024.95 J) and the shortest inference latency (11.91 ms/image). Together, EfficientNet-b0 and SqueezeNet form the Pareto frontier for classification performance and energy efficiency during inference. SqueezeNet’s lightweight architecture enabled competitive accuracy while requiring significantly lower energy consumption than ResNet-50 and GoogleNet. The results indicate robustness to changes in specimen orientation and storage time under the evaluated conditions. While evaluated on a standardized processor baseline, these findings highlight the prospective viability of lightweight deep learning models for potential edge deployment in automated food sorting systems.