Hadi Habibzadeh

We present a dynamic window-length classifier for steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) that does not require the user to choose a feature extraction method or channel set. Instead, the classifier uses multiple feature extraction methods and channel selections to infer the SSVEP and relies on majority voting to pick the most likely target. The classifier extends the window length dynamically if no target obtains the majority of votes. Compared with existing solutions, our classifier: (i) does not assume that any single feature extraction method will consistently outperform the others; (ii) adapts the channel selection to individual users or tasks; (iii) uses dynamic window lengths; (iv) is unsupervised (i.e., does not need training). Collectively, these characteristics make the classifier easy-to-use, especially for caregivers and others with limited technical expertise. We evaluated the performance of our classifier on a publicly available benchmark dataset from 35 healthy participants. We compared the information transfer rate (ITR) of this new classifier to those of the minimum energy combination (MEC), maximum synchronization index (MSI), and filter bank canonical correlation analysis (FBCCA). The new classifier increases average ITR to 123.5 bits-per-minute (bpm), 47.5, 51.2, and 19.5 bpm greater than the MEC, MSI, and FBCCA classifiers, respectively.

In combination with current sociological trends, the maturing development of Internet of Things devices is projected to revolutionize healthcare. A network of body-worn sensors, each with a unique ID, can collect health data, that is, orders-of-magnitude richer than what is available today from sporadic observations in clinical/hospital environments. When databased, analyzed, and compared against information from other individuals using data analytics, Healthcare Internet of Things data enables the personalization and modernization of care with radical improvements in outcomes and reductions in cost. In this article, we survey the existing and emerging technologies that can enable this vision for the future of healthcare, particularly, in the clinical practice of healthcare. Three main technology areas underlie the development of this field: 1) sensing, where there is an increased drive for miniaturization and power efficiency; 2) communications, where the enabling factors are ubiquitous connectivity, standardized protocols, and the wide availability of cloud infrastructure; and 3) data analytics and inference, where the availability of large amounts of data and computational resources is revolutionizing algorithms for individualizing inference and actions in health management. Throughout this article, we use a case study to concretely illustrate the impact of these trends. We conclude this article with a discussion of the emerging directions, open issues, and challenges.

Recent global smart city efforts resemble the establishment of electricity networks when electricity was first invented, which meant the start of a new era to sell electricity as a utility. A century later, in the smart era, the network to deliver services goes far beyond a single entity like electricity. Supplemented by a well-established Internet infrastructure that can run an endless number of applications, abundant processing and storage capabilities of clouds, resilient edge computing, and sophisticated data analysis like machine learning and deep learning, an already-booming Internet of Things movement makes this new era far more exciting. In this article, we present a multi-faceted survey of machine intelligence in modern implementations. We partition smart city infrastructure into application, sensing, communication, security, and data planes and put an emphasis on the data plane as the mainstay of computing and data storage. We investigate (i) a centralized and distributed implementation of data plane’s physical infrastructure and (ii) a complementary application of data analytics, machine learning, deep learning, and data visualization to implement robust machine intelligence in a smart city software core. We finalize our article with pointers to open issues and challenges.

Brain-computer interface (BCI)-based systems have been successfully implemented as a communication tool for people with motor deficits that obstruct normal communication. The accuracy of the algorithms used for determining user-selected commands directly impact their practicality to the user. These algorithms are divided into two principal categories, supervised and unsupervised. While the former achieves higher accuracy, the latter is useful when training is not practical for the user.In this paper, we introduce an unsupervised algorithm for steady-state visual evoked potential (SSVEP)-based BCIs, which works in three steps: (i) it selects multiple sets of EEG (electroencephalogram) channels, followed by (ii) existing feature extraction methods applied to each one of these channel sets. As its final step, (iii) it combines the extracted features from these channel sets by performing a majority vote, yielding a classification. We evaluate the information transfer rate (ITR) attained using our proposed method on a database of 35 subjects using three different (CCA, FBCCA, MSI) feature extraction methods in step (ii). We compare these results to existing methods in the literature that use a single channel set without a majority vote. The proposed method indicates an improvement in at least 7 subjects using any of the three feature extraction methods.

Today's existing smart city research involves many overtly futuristic applications such as smart transportation, in which smart roads warn drivers of bad traffic conditions ahead, smart parking, which communicates the location of unoccupied parking spaces to drivers, and smart environment, which enables fully automated homes and workplaces to adjust their temperature to conserve energy. The realization of these applications hinges on a data acquisition structure that gathers its data from a countless number of sensors, either deployed for predefined tasks (hard sensing) or built into the mobile devices of smart city residents (soft sensing). At the core of this big data infrastructure lie the 5Vs: veracity, volume, velocity, variety, and value. The soft sensing component of a smart city sensing network is particularly affected by 3Vs: veracity, volume, and velocity. To address the unique challenges of big data, recommender systems, statistical reputation systems, and context analysis are used to ensure the veracity of acquired data, machine learning algorithms are applied to handle the data volume, and data analytics algorithms are implemented to manage data velocity. Despite its seemingly insurmountable size, the acquired data is highly redundant, and systematic use of machine intelligence and data analytics can facilitate processing by extracting only the relevant information; in this article, we study the role of these algorithms through the lens of the 3Vs in facilitating soft sensing within the framework of smart city applications.

The past decade has witnessed an explosion of interest in smart cities in which a set of applications, such as smart healthcare, smart lighting, and smart transportation promise to drastically improve the quality and efficiency of these services. The skeleton of these applications is formed by a network of distributed sensors that captures data, pre-processes, and transmits it to a center for further processing. While these sensors are generally perceived to be a wireless network of sensing devices that are deployed permanently as part of an application, the emerging mobile crowd-sensing (MCS) concept prescribes a drastically different platform for sensing; a network of smartphones, owned by a volunteer crowd, can capture, pre-process, and transmit the data to the same center. We call these two forms of sensors dedicated and non-dedicated sensors in this paper. While dedicated sensors imply higher deployment and maintenance costs, the MCS concept also has known implementation challenges, such as incentivizing the crowd and ensuring the trustworthiness of the captured data, and covering a wide sensing area. Due to the pros/cons of each option, the decision as to which one is better becomes a non-trivial answer. In this paper, we conduct a thorough study of both types of sensors and draw conclusions about which one becomes a favourable option based on a given application platform.