Department Publications

This letter investigates the exploitation of an intelligent reflecting surface (IRS) to communicate securely in a two-way network consisting of an untrusted user. In particular, the transmit powers and the phase shift at each element of the IRS are optimized to maximize the sum-secrecy rate, such that the IRS-reflected and non-IRS-reflected signals are added destructively at the untrusted user. The proposed iterative algorithm converges rapidly to a feasible solution of high accuracy with a few iterations. Numerical results demonstrate sum-secrecy rate gains up to 120% compared to naive or partially optimized schemes.

Diagnosing pre-existing heart diseases early in life is important as it helps prevent complications such as pulmonary hypertension, heart rhythm problems, blood clots, heart failure and sudden cardiac arrest. To identify such diseases, phonocardiogram (PCG) and electrocardiogram (ECG) waveforms convey important information. Therefore, effectively using these two modalities of data has the potential to improve the disease screening process. We evaluate this hypothesis on a subset of the PhysioNet Challenge 2016 Dataset which contains simultaneously acquired PCG and ECG recordings. Our novel dual-convolutional neural network based approach uses transfer learning to tackle the problem of having limited amounts of simultaneous PCG and ECG data that is publicly available, while having the potential to adapt to larger datasets. In addition, we introduce two main evaluation frameworks named record-wise …

We present a MUSIC-based Direction of Arrival (DOA) estimation strategy using small antenna arrays, via employing deep learning for reconstructing the signals of a virtual large antenna array. Not only does the proposed strategy deliver significantly better performance than simply plugging the incoming signals into MUSIC, but surprisingly, the performance is also better than directly using an actual large antenna array with MUSIC for high angle ranges and low test SNR values. We further analyze the best choice for the training SNR as a function of the test SNR, and observe dramatic changes in the behavior of this function for different angle ranges.

Drones can be very useful in search and rescue missions primarily due to the aerial imaging capability. Drones can assist ground teams searching for a missing person, and law enforcement officers in crowd control. But there are instances where the observer on the ground control station misses the subtle information on the video feed coming from the drone. In fact, it is a very difficult job for the observer to be so vigilant in viewing a lot of images looking for a sign of a human in those images. This paper presents a drone-based human identification system to make search and rescue missions more effective. The drone, once detects signs of human presence in real-time aerial videos, autonomously navigates towards the suspicious location to get a better view to verify human presence. The drone has been made fully autonomous with operator override as an option. It processes images and sends selected frames for …

Convolutional Neural Networks (CNNs) have achieved state-of-the-art performance in many computer vision tasks over the years. However, this comes at the cost of heavy computation and memory intensive network designs, suggesting potential improvements in efficiency. Convolutional layers of CNNs partly account for such an inefficiency, as they are known to learn redundant features. In this work, we exploit this redundancy, observing it as the correlation between convolutional filters of a layer, and propose an alternative approach to reproduce it efficiently. The proposed `LinearConv' layer learns a set of orthogonal filters, and a set of coefficients that linearly combines them to introduce a controlled redundancy. We introduce a correlation-based regularization loss to achieve such flexibility over redundancy, and control the number of parameters in turn. This is designed as a plug-and-play layer to conveniently …

Cortical processing of arithmetic and of language rely on both shared and task-specific neural mechanisms, which should also be dissociable from the particular sensory modality used to probe them. Here, spoken arithmetical and non-mathematical statements were employed to investigate neural processing of arithmetic, compared with general language processing, in an attention-modulated cocktail party paradigm. Magnetoencephalography (MEG) data were recorded from 22 human subjects listening to audio mixtures of spoken sentences and arithmetic equations while selectively attending to one of the two speech streams. Short sentences and simple equations were presented diotically at fixed and distinct word/symbol and sentence/equation rates. Critically, this allowed neural responses to acoustics, words, and symbols to be dissociated from responses to sentences and equations. Indeed, the simultaneous …

Recent work done on lane detection has been able to detect lanes accurately in complex scenarios, yet many fail to deliver real-time performance specifically with limited computational resources. In this work, we propose SwiftLane: a simple and light-weight, end-to-end deep learning based framework, coupled with the row-wise classification formulation for fast and efficient lane detection. This framework is supplemented with a false positive suppression algorithm and a curve fitting technique to further increase the accuracy. Our method achieves an inference speed of 411 frames per second, surpassing state-of the-art in terms of speed while achieving comparable results in terms of accuracy on the popular CULane benchmark dataset. In addition, our proposed framework together with TensorRT optimization facilitates real-time lane detection on a Nvidia Jetson AGX Xavier as an embedded system while achieving …

Dynamic adjustment of signal timing based on the number of pedestrians waiting at a crossing results in safer and more efficient intersections. Traditionally, such information is obtained through push-buttons or vision-based techniques. Though counting is possible with the latter, the techniques are associated with privacy concerns, are affected by poor lighting, and equipment is costly and bulky. We present a novel alternative that uses Wi-Fi Channel State Information (CSI) processed through a Deep Neural Network (DNN) model. The amplitude and phase of 51 Orthogonal Frequency Division Multiplexing sub carriers in the IEEE802.11n standard are captured with two ESP32 devices. Of the candidate DNNs trained with the CSI data, a ID-Convolutional Neural Network is experimentally demonstrated to count up to 12 pedestrians with an accuracy of 79 % at the current level of training. This can be generalized to …

This paper investigates the communication reliability for single-input-multiple-output wireless systems with low-resolution phase quantizers. First, the maximum-likelihood detector with -bit phase quantization is derived when there are antennas at the receiver. Then, three low-complexity antenna selection strategies for data detection are proposed and their symbol error probability performance is characterized. It is shown that having 3 or more bits is sufficient to attain the full diversity order , achievable with infinite-bit quantizers, for quadrature phase shift keying modulation under Rayleigh fading. In particular, it is established that the proposed low-complexity max-distance and max-norm antenna selection strategies perform the same as the maximum-likelihood detector in terms of the asymptotic system reliability for . On the other hand, the diversity order decreases dramatically from to when is …

Blockchain is an innovative technology which enabled new applications for solving numerous problems in distributed environments such as the internet of things (IoT) equipped food supply chains (FSCs). In FSCs, the large volume of IoT data such as audio, video, images, and sensor data will be transferred to ensure the traceability of the food to its source. When blockchain technology is used in an FSC, the storage requirements in the nodes will grow with time, since blockchain only allows information adding, without deleting the already stored information. Therefore, it is evident that offchain storage offers more flexibility than onchain storage in IoT equipped supply chains. This paper proposes a scalable storage scheme using offchain storage, where the data from IoT devices in the supply chain will store offchain. To reduce the growth of offchain storage, we exploit the fact that some information regarding a …

Supporting parallel multi-priority data streams is vital for maintaining the quality of service (QoS) in vehicle-to-everything (V2X) communication. Hence, the European telecommunications standard institute (ETSI) has defined four packet types, with varying priority levels, to be used as broadcast packets in such communication. This paper studies the medium access control (MAC) layer performance of IEEE 802.11p and cellular-V2X (C-V2X) Mode 4 using discrete-time Markov chain (DTMC) based models, while considering parallel multi-priority data streams. The overall model consists of four queue models with their respective traffic generators, which are appropriately linked with the DTMCs modeling the MAC layer operations of IEEE 802.11p and C-V2X Mode 4. Closed-form solutions for the steady-state probabilities of the models are obtained, which are then utilized to derive expressions for key performance …

The acquisition of channel state information is an essential step in enabling intelligent reflecting surfaces (IRSs) for wireless communications. In this letter, we introduce a novel procedure to estimate the cascaded channel between the base station (BS), IRS, and users in a multi-user multiple-input single-output system. The common BS-IRS channel, over which all users transmit, is leveraged to decompose the cascaded channel into a series of rank-1 matrices. Low-rank matrix recovery methods are utilized to improve upon the linear minimum mean-squared error estimate of the cascaded channel. A theoretical upper bound for the mean-square error (MSE) of the proposed estimator is derived. Numerical results reveal that the proposed techniques outperform the existing counterparts in terms of the MSE and scale with the number of BS antennas.

Light fields (LFs) and light field videos (LFVs) capture both angular and spatial variation of light rays emanating from scenes. This richness of information leads to novel applications such as post-capture refocusing, depth estimation and depth-velocity filtering which are not possible with images and videos. These capabilities come, however, with a significant increase in data to be processed. In order to fully exploit opportunities provided by LFs and LFVs, low-complexity signal processing algorithms that process LF and LFV data in real-time are required. In this brief, we survey such state-of-the-art algorithms, in particular for depth filtering, refocusing and denoising of LFs and depth-velcoty filtering for LFVs, and future directions for these real-time LF an LFV processing algorithms.

An analytical method to synthesize a nonisolated multi-inductor (uncoupled) dc–dc converter having an arbitrary voltage conversion ratio is presented in this article. The proposed method provides a direct approach to synthesize the converter unlike the traditional topology synthesis methods, which involve circuit manipulations or selecting a connection matrix from a pool of possible matrices. As a result, the proposed method simplifies the topology synthesis of a converter having known properties to a four-step process. To this end, the voltage conversion ratio is decomposed, allowing to synthesize it using single inductor switching cells. The resultant single inductor switching cells can be either cascaded or differentially connected as pairs or embedded within a switching cell to realize a multi-inductor-converter. Several existing converter topologies have been synthesized to demonstrate the applicability and …

Purpose Conventional positron emission tomography can be used only preoperatively to identify lymph node metastases, and hence, these growths are difficult to locate intraoperatively. Previously, an intraoperative laparoscopic coincidence imaging system, with an external fixed detector array and a detector probe that can be moved and inserted into the stomach, was proposed to identify lymph node metastases during stomach cancer surgery. This paper proposes a prototype detector for this system. Methods GAGG:Ce (Gd 3 Al 2 Ga 3 O 12 :Ce) scintillator crystals, one to one coupled with silicon photomultipliers (SiPMs), were used to identify 511 keV photons. An optical tracking system followed the position and orientation of the movable detector. SiPM outputs were read out by time-over-threshold (TOT)-based application-specific integrated circuits, which converted the electrical charges into digital pulses. To identify …

With the advancement of technology, the digital Integrated Circuit (IC) design process has become more complex and denser. Hence, the IC testing procedure requires high-end test equipment to validate the accuracy and reliability of the manufactured components. Testers with such capabilities usually cost millions of dollars. In this paper, the authors have presented a low-cost hardware and software solution for digital IC testing. Digital ICs which operate under the 100MHz range can be easily tested in the digital domain with the FPGA-based test environment. The presented design comprises of a scalable architecture with a set of clock synchronized Altera DE0-Nano Field Programmable Gate Arrays (FPGAs) which handles the digital testing of Device Under Test (DUT) at a low cost. The digital test patterns are generated inside a computer, which transfers them to the FPGA environment and feeds them to DUT. The …

Learning a particular task from a dataset, samples in which originate from diverse contexts, is challenging, and usually addressed by deepening or widening standard neural networks. As opposed to conventional network widening, multipath architectures restrict the quadratic increment of complexity to a linear scale. However, existing multi-column/path networks or model ensembling methods do not consider any feature-dependent allocation of parallel resources, and therefore, tend to learn redundant features. Given a layer in a multi-path network, if we restrict each path to learn a context-specific set of features and introduce a mechanism to intelligently allocate incoming feature maps to such paths, each path can specialize in a certain context, reducing the redundancy and improving the quality of extracted features. This eventually leads to better-optimized usage of parallel resources. To do this, we propose …

In this paper, we investigate antenna selection strategies for multi-antenna wireless communication systems with low-resolution quantizers. We propose three sub-optimum but low-complexity antenna selection strategies and characterize their symbol error probability performance. We show that the strategy that selects the diversity branch with channel rotated constellation points being furthest away from the decision boundary is the best strategy in terms of symbol error probability. Using numerical analysis, we provide evidence to suggest that this selection strategy achieves the same diversity order as the optimum maximum likelihood (ML) detector under the same operating conditions. An extensive simulation study is performed to illustrate the accuracy of the derived results.

In four-dimensional (4-D) light field (LF) processing, 4-D linear shift-invariant filters having hyperplanar passbands are used for depth-based scene enhancement. In this paper, two low-sensitivity and low-complexity field programmable gate array (FPGA)-based digital hardware architectures for 4-D hyperplanar filters are proposed for on-chip real-time processing of LFs. Both 4-D filters are designed exploiting resonant properties of multi-dimensional passive prototype networks, and are realized as wave digital filters (WDFs). The two 4-D WDF realizations are implemented as raster-scanned processing architectures on a Xilinx Virtex 6 Sx35 FPGA with a real-time clock speed of up to 33 MHz. This corresponds to a real-time throughout of 16.8 LFs/s for an LF of size .

Real-time and online action localization in a video is a critical yet highly challenging problem. Accurate action localization requires the utilization of both temporal and spatial information. Recent attempts achieve this by using computationally intensive 3D CNN architectures or highly redundant two-stream architectures with optical flow, making them both unsuitable for real-time, online applications. To accomplish activity localization under highly challenging real-time constraints, we propose utilizing fast and efficient key-point based bounding box prediction to spatially localize actions. We then introduce a tube-linking algorithm that maintains the continuity of action tubes temporally in the presence of occlusions. Further, we eliminate the need for a two-stream architecture by combining temporal and spatial information into a cascaded input to a single network, allowing the network to learn from both types of information. Temporal information is efficiently extracted using a structural similarity index map as opposed to computationally intensive optical flow. Despite the simplicity of our approach, our lightweight end-to-end architecture achieves state-of-the-art frame-mAP of 74.7% on the challenging UCF101-24 dataset, demonstrating a performance gain of 6.4% over the previous best online methods. We also achieve state-of-the-art video-mAP results compared to both online and offline methods. Moreover, our model achieves a frame rate of 41.8 FPS, which is a 10.7% improvement over contemporary real-time methods.

This paper addresses the problem of high energy consumption in the standard take-off mode of quad-plane unmanned aerial vehicles (UAV). In the standard mode, the quad-plane takes-off vertically using the quad mode, and when the predefined altitude is reached the plane mode kicks in. However, the" bird" mode, where both quad and plane modes kick in together, the UAV could achieve altitude and airspeed while consuming lower energy. This paper presents the energy consumption and comparison in standard and bird take-off modes with actual experimental results obtained through flight tests of a small 1.35 kg quad-plane. The experiments showed that as much as 22% energy can be saved if the bird take-off mode is used.

We propose a weighted least-squares (WLS) design method for multi-dimensional (M-D) complex-coefficient finite- extent impulse response (FIR) filters. We consider the general form of M-D FIR filters having arbitrary frequency responses and low group delays. We formulate the proposed WLS design as a second-order cone programming problem. Design examples confirm that the proposed method provides the state-of-the-art M-D FIR filter designs with almost constant group delay.

Self-localization based on passive RFID-based has many potential applications. One of the main challenges it faces is the suppression of the reflected signals from unwanted objects (i.e., clutter). Typically, the clutter echoes are much stronger than the backscattered signals of the passive tag landmarks used in such scenarios. Therefore, successful tag detection can be very challenging. We consider two types of tags, namely low-Q and high-Q tags. The high-Q tag features a sparse frequency response, whereas the low-Q tag presents a broad frequency response. Further, the clutter usually showcases a short-lived response. In this work, we propose an iterative algorithm based on a low-rank plus sparse recovery approach (RPCA) to mitigate clutter and retrieve the landmark response. In addition to that, we compare the proposed approach with the well-known time-gating technique. It turns out that RPCA outperforms significantly time-gating for low-Q tags, achieving clutter suppression and tag identification when clutter encroaches on the time-gating window span, whereas it also increases the backscattered power at resonance by approximately 12 dB at 80 cm for high-Q tags. Altogether, RPCA seems a promising approach to improve the identification of passive indoor self-localization tag landmarks.

Defect detection plays an important role in product quality assurance in many industrial applications. Air-coupled ultrasound defect detection setups have the advantage that they are non-destructive and do not contaminate the investigated materials. However, this approach comes with some challenges due to the weak signaling response of the defects. Thus, sophisticated signal separation methods are required. To address this challenge, low-rank-plus-sparse recovery (LRPSR) is proposed and compared with state-of-the-art machine learning (ML) methods. The results obtained show that the LRPSR method achieves comparable results in terms of detection rate to those achieved by ML. Yet, for a small training data set, the LRPSR approach outperforms the ML algorithms. The small training data set is important for detecting defects of different production lines without having a time-consuming training process. In …

Deep neural networks (DNNs) have been successfully employed in a multitude of applications with remarkable performance. As such performance is achieved at a significant computational cost, several embedded applications demand fast and efficient hardware accelerators for DNNs. Previously proposed application specific integrated circuit (ASIC) architectures strive to utilize arrays of hundreds of processing elements (PEs) and reduce power-hungry DRAM accesses using multiple dataflows requiring complex PE architectures. These consume significant area and reduce the maximum clock frequency. This paper introduces the Kraken architecture, which optimally processes the convolutional layers, fully-connected layers, and matrix products of any DNN through a hardware-friendly uniform dataflow. This enables maximal data reuse of weights, inputs, and outputs, with a bare-bones PE design and on-the-fly dynamic reconfiguration. Kraken, implemented in 65-nm CMOS technology at 400 MHz, packs 672 PEs in 7.3 mm2, with a peak performance of 537.6 Gops. Kraken processes the convolutional layers of AlexNet, VGG-16, and ResNet-50 at 336.6, 17.5, and 64.2 frames/s, respectively, hence outperforming the state-of-the-art ASIC architectures in terms of overall performance efficiency, DRAM accesses, arithmetic intensity, and throughput, with 5.8x more Gops/mm2 and 1.6x more Gops/W.