Modern technology has pushed us into the information age, making it easier to generate and record vast quantities of new data. Datasets can help in analyzing the situation to give a better understanding, and more importantly, decision making. Consequently, datasets, and uses to which they can be put, have become increasingly valuable commodities. This article describes the DroneRF dataset: a radio frequency (RF) based dataset of drones functioning in different modes, including off, on and connected, hovering, flying, and video recording. The dataset contains recordings of RF activities, composed of 227 recorded segments collected from 3 different drones, as well as recordings of background RF activities with no drones. The data has been collected by RF receivers that intercepts the drone's communications with the flight control module. The receivers are connected to two laptops, via PCIe cables, that runs a program responsible for fetching, processing and storing the sensed RF data in a database. An example of how this dataset can be interpreted and used can be found in the related research article "RF-based drone detection and identification using deep learning approaches: an initiative towards a large open source drone database" (Al-Sa'd et al., 2019).
15 June 2021 in IEEE Transactions on Signal Processing
This paper proposes a new performance evaluation process for time-frequency distributions (TFD) by designing a reference optimal TFD and novel accuracy and resolution measures. The motivation comes from the need for a TFD performance evaluation method that is objective, capable of quantifying the TFD accuracy and resolution, can determine the performance difference among different TFDs, and suitable for signals with an arbitrary number of components, instantaneous frequency and amplitude. We formulate the proposed optimal TFD, namely the piece-wise spline Wigner-Ville distribution (PW-WVD), by decomposing a standard non-stationary signal model using piece-wise linear frequency modulated (LFM) basis and by exploiting the Wigner-Ville distribution optimality for LFM signals. We compare the designed PW-WVD to conventional optimal TFDs and show that the former is more suitable to serve as a reference for TFD performance evaluation. Using the PW-WVD we derive TFD accuracy and resolution measures, compare them to conventional approaches, and analyze their sensitivity to form a TFD selection criteria. We evaluate the accuracy and resolution of twelve different TFDs and develop precise TFD selection strategies with or without prior information on the signal parameters. Results indicate that the compact kernel distribution is the best performing TFD given no prior information on the signal parameters and different TFDs must be selected upon the availability of prior information.
Due to the increasing number of chronic disease patients, continuous health monitoring has become the top priority for health-care providers and has posed a major stimulus for the development of scalable and energy efficient mobile health systems. Collected data in such systems are highly critical and can be affected by wireless network conditions, which in return, motivates the need for a preprocessing stage that optimizes data delivery in an adaptive manner with respect to network dynamics. We present in this paper adaptive single and multiple modality data compression schemes based on deep learning approach, which consider acquired data characteristics and network dynamics for providing energy efficient data delivery. Results indicate that: 1) the proposed adaptive single modality compression scheme outperforms conventional compression methods by 13.24% and 43.75% reductions in distortion and processing time respectively. 2) the proposed adaptive multiple modality compression further decreases the distortion by 3.71% and 72.37% when compared to the proposed single modality scheme and conventional methods through leveraging inter-modality correlations. 3) adaptive multiple modality compression demonstrates its efficiency in terms of energy consumption, computational complexity, and responding to different network states. Hence, our approach is suitable for mobile health applications (mHealth) where smart preprocessing of vital signs can enhance energy consumption, reduce storage, and cut down transmission delays to the mHealth cloud.
Social distancing is crucial to restrain the spread of diseases such as COVID-19, but complete adherence to safety guidelines is not guaranteed. Monitoring social distancing through mass surveillance is paramount to develop appropriate mitigation plans and exit strategies. Nevertheless, it is a labor-intensive task that is prone to human error and tainted with plausible breaches of privacy. This paper presents a privacy-preserving adaptive social distance estimation and crowd monitoring solution for camera surveillance systems. We develop a novel person localization strategy through pose estimation, build a privacy-preserving adaptive smoothing and tracking model to mitigate occlusions and noisy/missing measurements, compute inter-personal distances in the real-world coordinates, detect social distance infractions, and identify overcrowded regions in a scene. Performance evaluation is carried out by testing the system’s ability in person detection, localization, density estimation, anomaly recognition, and high-risk areas identification. We compare the proposed system to the latest techniques and examine the performance gain delivered by the localization and smoothing/tracking algorithms. Experimental results indicate a considerable improvement, across different metrics, when utilizing the developed system. In addition, they show its potential and functionality for applications other than social distancing.