Rich variations in gait are generated according to several attributes of the individual and environment, such as age, athleticism, terrain, speed, personal ”style”, mood, etc. The effects of these attributes can be hard to quantify explicitly, but relatively straightforward to sample. We seek to generate gait that expresses these attributes, creating synthetic gait samples that exemplify a custom mix of attributes. This is difficult to perform manually, and generally restricted to simple, human-interpretable and handcrafted rules. In this manuscript, we present neural network architectures to learn representations of hard to quantify attributes from data, and generate gait trajectories by composing multiple desirable attributes. We demonstrate this method for the two most commonly desired attribute classes: individual style and walking speed. We show that two methods, cost function design and latent space regularization, can be used individually or combined. We also show two uses of machine learning classifiers that recognize individuals and speeds. Firstly, they can be used as quantitative measures of success; if a synthetic gait fools a classifier, then it is considered to be a good example of that class. Secondly, we show that classifiers can be used in the latent space regularizations and cost functions to improve training beyond a typical squared-error cost.

This article provides a targetless extrinsic calibration method between camera and low-resolution 3D LiDAR. We carried out real-world experiments on a multi-sensor platform containing a Velodyne VLP-16 LiDAR and a camera. Ground-truth extrinsic matrix is generated by a state-of-the-art target-based method. Regarding the calibration error, the rotation and translation RMSE of our method are 0.59° and 3cm respectively. Codes are avaible at https://github.com/gitouni/Targetless-LiDAR-camera-calibration. NB: This article has been accepted for publication in IEEE Sensors Journal.  Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Citation Information: 10.1109/JSEN.2023.3263833

The COVID-19 pandemic, caused by the SARS-CoV-2 virus, continues to have a significant impact on the global population. To effectively triage patients and understand the progression of the disease, a metric-based analysis of diagnostic techniques is necessary. The objective of the present study is to identify COVID-19 from chest CT scans and determine the extent of severity, defined by a severity score that indicates the volume of infection. An unsupervised preprocessing pipeline is proposed to extract relevant clinical features and utilize this information to employ a pretrained ImageNet EfficientNetB5 model to extract discriminative features. Subsequently, a shallow feed-forward neural network is trained to classify the CT scans into three classes, namely COVID-19, Community-Acquired Pneumonia, and Normal. Through various ablation studies, we find that a domain-specific preprocessing pipeline has a significant positive impact on classification accuracy. The infection segmentation mask generated from the preprocessed pipeline performs better than state-of-the-art supervised semantic segmentation models. Further, the estimated infection severity score is observed to be well correlated with radiologists' assessments. The results confirm the importance of domain-specific preprocessing for training machine learning algorithms.

Localization of passive Radio-Frequency Identification (RFID) tags has been used to monitor landslide surface displacement since 5 years. This method, applied on slow displacements lower than 1cm per day, allows a high spatiotemporal resolution at a relatively low cost. With the feedback of the previous years, this paper proposes to summarize the various challenges encountered with the long-term outdoor RFID localization method, and presents data-processing solutions that were implemented to overcome these challenges. We propose a complex-smoothing unwrapping algorithm, a multi-frequency merging operation, as well as multi-tag and multi-antenna phase combining method. The concept of an unwrapping reference guide is presented and applied with groups of tags showing coherent displacements, or with absolute reference measurements. These approaches allow a higher data availability up to 38% for one site over multiple years, and a better phase unwrapping. Earth surface displacement monitoring with RFID proves to be a robust and accurate solution, with four equipped sites across France and Switzerland.

Most deep clustering methods despite providing complex networks to learn better from data, use a shallow clustering method. These methods have difficulty in finding good clusters due to the lack of ability to handle between local search and global search to prevent premature convergence. In other words, they do not consider different aspects of the search and it causes them to get stuck in the local optimum. In addition, the majority of existing deep clustering approaches perform clustering with the knowledge of the number of clusters, which is not practical in most real scenarios where such information is not available. To address these problems, this paper presents a novel automatic deep sparse clustering approach based on an evolutionary algorithm called Multi-Trial Vectorbased Differential Evolution (MTDE). Sparse auto-encoder is first applied to extract embedded features. Manifold learning is then adopted to obtain representation and extract the spatial structure of features. Afterward, MTDE clustering is performed without prior information on the number of clusters to find the optimal clustering solution. The proposed approach was evaluated on various datasets, including images and time-series. The results demonstrate that the proposed method improved MTDE by 18.94% on average and compared to the most recent deep clustering algorithms, is consistently among the top three in the majority of datasets. Source code is available on Github: https://github.com/parhamhadikhani/ADSMTDE_Clustering.

This article presents a detailed review and categorizing of the marker displacement method (MDM) used in vision-based tactile sensors. Vision-based tactile sensors have been proven to be a promising solution for robot tactile perception. Among such sensors, MDM is one of the most commonly used contact characterization and extraction methods. It uses visual approaches to obtain contact deformation and achieve multimodal tactile perception using physical models and post-processing algorithms. In recent years, many tactile sensors using MDM have been developed. However, the existing research does not strictly distinguish between the different types of methods but is uniformly grouped into MDM. Without differentiation, there might be a lack of systematic and comprehensive guidance in analyzing and optimizing the characteristics of MDM and selecting the most suitable method. This article is the first to classify MDM into three typical categories based on the dimensionality perspective: 2D MDM, 2.5D MDM, and 3D MDM. 2D MDM relies only on the monocular camera to acquire the marker array’s 2D displacement field. 2.5D MDM supplements 2D MDM with selected indirect features reflecting the location of the markers in the third dimension. 3D MDM employs a multi-camera system and can obtain the 3D displacement field using the stereo vision method common. Based on the latest literature, we compare the principles, characteristics, advantages and disadvantages, and applications of the three ways in detail. This work can provide a valuable reference for researchers interested in applying MDM in fields such as vision-based tactile sensors.

This paper presents a multistack switched-capacitor point-of-load (MSC-PoL) voltage regulation module (VRM) with coupled magnetics for ultrahigh-current chiplet systems. In the MSC-PoL architecture, the stacked switched-capacitor cells split the high input voltage into several intermediate voltage rails, which are loaded with the switched-inductor cells to achieve soft charging and voltage regulation. Automatic capacitor voltage balancing and inductor current sharing are realized during the soft charging process. Many inductors of the switched-inductor cells are coupled into one and operated in interleaving to reduce the inductor current ripple and boost the transient speed. A 48-to-1-V/450-A VRM containing two MSC-PoL modules is built and tested, leveraging high voltage GaN devices for the front-end and high current Silicon devices for the back-end. Two ladder-structured coupled inductor designs are developed and compared, one of which installs a leakage magnetic plate to adjust the leakage inductance for lower current ripple. Featuring 3D stacked packaging, the entire power stage, gate drivers, and bootstrap circuits of one MSC-PoL module are enclosed into a 1/16-brick/0.31-in3/6-mm-thick package. The peak and the full-load efficiencies as well as the full-load power density (including both gate loss and size) of the MSC-PoL prototype with and without using the leakage plate are 91.7% and 89.5%, 85.8% and 85.6%, and 621 W/in3 and 724 W/in3, respectively. The 6-mm-thick MSC-PoL converter can be embedded into the chiplet or CPU socket, enabling power-supply-in-package (PwrSiP) for extreme efficiency, density, and control bandwidth.

Reconfigurable intelligent surfaces (RISs) and millimeter-wave (mmWave) communications have been considered for providing wireless connectivity to mobile robots used in industrial plants and other indoor environments. However, the existing works have not sufficiently studied how the number and deployment locations of RISs should be optimized for serving a mobile robot. In this paper, we study RIS-assisted mmWave communications for a robot moving around fixed obstacles in an indoor industrial environment. For a fixed total number of reflecting elements, we formulate an optimization problem to minimize the transmission energy consumption of the access point (AP) while ensuring the robot’s received signal-to-noise ratio (SNR) above a threshold throughout its journey by jointly optimizing the number, positions and phase shifts of RISs and the beamforming vectors of the AP. To solve the formulated nonconvex optimization problem, we devise an iterative algorithm that decomposes it into two subproblems (i.e., optimizing the phase shifts of RISs and the beamforming vector of the AP, and optimizing the number and locations of RISs) and solves them alternately. Simulation results show that the proposed algorithm converges fast and can obtain the best number and locations of RISs that lead to a transmission energy consumption of the AP much lower than the benchmark schemes.

Intelligent reflecting surfaces (IRSs) have become a popular topic in recent years for their great potential for controlling the radio link environment of wireless networks. With this controlled environment, the coverage can be increased. This paper examines the coverage analysis of IRS-aided networks, considering both two-dimensional buildings and the product-distance path loss model for the first time.  Leveraging the tools from stochastic geometry, the locations of base stations (BSs), buildings, and IRSs are modeled with a homogeneous Poisson point process (PPP). A Gamma approximation for the distribution of the nearest line-of-sight (LoS)-neighbor distance is proposed, leading to a closed-form expression for the distribution of the product-distance. Feasible BSs are defined as BSs which are reachable via an IRS deployed on a specific facade of a building, and the ratio of feasible BSs is derived. Simulations are performed, which confirm the proposed analytical methods. In the numerical results, it is observed that the IRSs can introduce up to a 45\% coverage boost, and the effect of the IRS length on the coverage probability is limited beyond 1.2 meters at 60 GHz.

This paper deals with the capability of the orthogonal chirp division multiplexing (OCDM) modulation scheme to cope with multi-numerology concept in the context of next generations of communications. The OCDM-based multi-numerology transceiver is introduced, and the inter-numerology interference (INI) is analytically derived. The case of filtered-OCDM is considered as well, allowing to both reduce the INI and to shape the power spectrum of the composite signal (i.e. the signal composed of unitary signals configured with different numerologies) according to strong constraints of low out-of-band emissions. The simulations results show the inherent robustness of OCDM against INI as the interference is spread over all subcarriers, whereas it is concentrated in particular subcarriers in OFDM. It results that the (filtered)-OCDM waveform largely outperforms the orthogonal frequency division multiplexing (OFDM) in term of bit error rate (BER) performance, without need for a guard-band between numerologies.

Grasping is the most basic and important function of robots. However, the grasping performance of existing manipulators is far inferior to that of human hands. Human hands are able to evaluate the degree of incipient slip on contact surface based on tactile information during grasping process and thus regulate grasping force. As a result, humans can use the smallest possible grasping force while ensuring successful grasping, which maximizes the performance of their hands and avoids damaging the grasped object as much as possible. The existing incipient slip degree evaluation methods have certain shortcomings. So this paper proposes a novel method which can be applied to complex contact conditions where the incipient slip degree is not unidirectional and torque exists. Also, there are no restrictions on the material parameters and surface topography of the grasped object, and no need to obtain any information about it in advance. We construct a grasping force control strategy for parallel grippers based on this evaluation method, with the goal of enabling the gripper to achieve the similar grasping performance of human hands. The grasping strategy is verified in simulations and actual experiments, and the difference between the controlled force and minimum grasping force is demonstrated

Motivated by the need for ubiquitous and reliable communications in post-disaster emergency management systems (EMSs), we hereby present a novel and efficient stochastic geometry (SG) framework. This mathematical model is specifically designed to evaluate the quality of service (QoS) experienced by a typical ground user equipment (UE) residing either inside or outside a generic area affected by a calamity. In particular, we model the functioning terrestrial base stations (TBSs) as an inhomogeneous Poisson point process (IPPP), and assume that a given number of uniformly distributed unmanned aerial vehicles (UAVs) equipped with cellular transceivers is deployed in order to compensate for the damage suffered by some of the existing TBSs. The downlink (DL) coverage probability is then derived based on the maximum average received power association policy and the assumption of Nakagami-m fading conditions for all wireless links. The proposed numerical results show insightful trends in terms of coverage probability, depending on: distance of the UE from the disaster epicenter, disaster radius, quality of resilience (QoR) of the terrestrial network, and fleet of deployed ad-hoc aerial base stations (ABSs). The aim of this paper is therefore to prove the effectiveness of vertical heterogeneous networks (VHetNets) in emergency scenarios, which can both stimulate the involved authorities for their implementation and inspire researchers to further investigate related problems.

In this paper, we present the results of the MitoEM challenge on mitochondria 3D instance segmentation from electron microscopy images, organized in conjunction with the IEEE-ISBI 2021 conference. Our benchmark dataset consists of two large-scale 3D volumes, one from human and one from rat cortex tissue, which are 1,986 times larger than previously used datasets. At the time of paper submission, 257 participants had registered for the challenge, 14 teams had submitted their results, and six teams participated in the challenge workshop. Here, we present eight top-performing approaches from the challenge participants, along with our own baseline strategies. Posterior to the challenge, annotation errors in the ground truth were corrected without altering the final ranking. Additionally, we present a retrospective evaluation of the scoring system which revealed that (1) the challenge metric was permissive with the false positive predictions and (2) the size-based grouping of instances did not correctly categorize mitochondria of interest. Thus, we propose a new scoring system that better reflects the correctness of the segmentation results. Although several of the top methods are compared favorably to our own baselines, substantial errors remain unsolved for mitochondria with challenging morphologies. Thus, the challenge remains open for submission and automatic evaluation, with all volumes available for download.

As a prospective key technology for the next-generation wireless communications, reconfigurable intelligent surfaces (RISs) have gained tremendous research interest in both the academia and industry in recent years. Only limited knowledge, however, has been obtained about the channel eigenvalue characteristics and spatial degrees of freedom (DoF) of systems containing RISs, especially when mutual coupling (MC) is present between the array elements. In this paper, we focus on the small-scale spatial correlation and eigenvalue properties excluding and including MC effects, for RISs with a quasicontinuous aperture (i.e., holographic RISs). Specifically, asymptotic behaviors of far-field and nearfield eigenvalues of the spatial correlation matrix of holographic RISs without MC are first investigated, where the counter-intuitive observation of a lower DoF with more elements is explained by leveraging the power spectrum of the spatial correlation function. Second, a novel metric is proposed to quantify the inter-element correlation or coupling strength in RISs and ordinary antenna arrays. Furthermore, in-depth analysis is performed regarding the MC effects on array gain, effective spatial correlation, and eigenvalue architectures for a variety of element intervals when a holographic RIS works in the radiation and reception mode, respectively. The analysis and numerical results demonstrate that a considerable amount of the eigenvalues of the spatial correlation matrix correspond to evanescent waves that are promising for near-field communication and sensing. More importantly, holographic RISs can potentially reach an array gain conspicuously larger than conventional arrays by exploiting MC, and MC has discrepant impacts on the effective spatial correlation and eigenvalue structures at the transmitter and receiver.

We propose a tunable plasmonic semiconductor laser that exploits loss perturbation as a tuning mechanism. A metal oxide semiconductor (MOS) capacitive structure is added on top of an edge-emitting Fabry-Perot (FP) diode laser, such that a hybrid plasmonic TM mode that overlaps partly with the MOS capacitor and the semiconductor gain region is supported as the lasing mode. We also propose the use of a layer of conductive oxide, e.g., indium tin oxide (ITO), as the semiconductor of the MOS structure, because the epsilon near zero (ENZ) condition can be attained therein under accumulation, thereby producing a very large change in the effective index of the hybrid plasmonic TM mode. The change in the imaginary part of the effective index is used to tune the lasing wavelength - exploiting loss perturbation to achieve laser tuning is paradigm-shifting. The laser proposed operates at telecom wavelengths, requiring an electrical forward bias to pump the active layer, and a gate voltage to drive the MOS tuning capacitor. Simulations yield a tuning range of over 7 nm in the O-band for a 100 μm long FP laser cavity.

Researchers are interested in unmanned aerial vehicles (UAVs) because they have many uses in current 5G and future 6G networks and are safer than human-operated vehicles. Terahertz (THz)-band communications are a possible option in 6G for fast communication. THz wireless communication between UAVs is taken into consideration in this work. Different fading channels, which are significant influencing factors in THz communication channel modeling, are used to analyze the performance of UAV communications in THz networks. Consideration must be made to the structure, wireless channel parameters, and transmission characteristics when evaluating the performance of wireless technology. Nakagami-m, Rician, Weibull, and Rayleigh fading channels are all taken into consideration, along with log-normal fading. Moreover, an optimization algorithm for the THz channel is presented, which is meant to minimize transmission power by optimizing the trajectory of uplink and downlink transmissions between the UAV and users. The equations of the UAV locations and the transmission power optimization of each user are derived. Analytical formulations regarding capacity, outage probability, and bit error rate (BER) are generated when performance-influencing factors are taken into account. The analytical analysis is supported by the numerical results that are provided.

Signals of opportunity reflectometry (SoOp-R), the re-utilization of non-cooperative satellite transmissions for communication and navigation, is a promising approach to remote sensing of root-zone soil moisture (RZSM). Satellite transmissions in the frequency ranges of 137-138, 240-270, and 360-380 MHz are of interest due to the increased penetration depth. These can be combined with Global Navigation Satellite System Reflectometry (GNSS-R) in L-band (1575.42 MHz) to estimate the subsurface SM profile. The objective is to define requirements (e.g. frequency and polarization combinations, observation error, and temporal coincidence of multi-source observations) for satellite-based remote sensing of RZSM. Our approach is to use synthetic observations generated from multi-year time series of in-situ SM measurements from seven United States Climate Reference Network (USCRN) sites and dynamic vegetation structure based on a simple scaling method. A multi-frequency/polarimetric retrieval algorithm is developed and applied to these synthetic observations and used to predict retrieval errors for a range of changes in system parameters. We found that the use of both high and low frequencies improves retrieval accuracy by limiting uncertainties from vegetation and surface SM and providing sensitivity to deeper layers. Moreover, the retrieval errors were found to increase linearly with the reflectivity error and inter-frequency time delays. A bivariate model derived from this linear relationship will be useful for developing requirements on reflectivity precision based upon science requirements for SM/VWC retrievals. Although orbits of specific transmitter constellations were used to generate realistic distributions of incidence angle combinations, the method and results could be applied more generally.

Typical assessments of balance impairment are subjective or require data from cumbersome and expensive force platforms. Researchers have utilized lower back (sacrum) accelerometers to enable more accessible, objective measurement of postural sway for use in balance assessment. However, new sensor patches are broadly being deployed on the chest for cardiac monitoring, opening a need to determine if measurements from these devices can similarly inform balance assessment. Our aim in this work is to validate postural sway measurements from a chest accelerometer. To establish concurrent validity, we considered data from 16 persons with multiple sclerosis (PwMS) asked to stand on a force platform while also wearing sensor patches on the sacrum and chest. We found five (five) of 15 postural sway features derived from chest (sacrum) were significantly correlated with force platform-derived features, which is in line with prior sacrum-derived findings. Clinical significance was established using a sample of 39 PwMS who performed eyes-open, eyes-closed, and tandem standing tasks. This cohort was stratified by fall status and completed several patient-reported measures (PRM) of balance and mobility impairment. We also compared sway features derived from a single 30-second period to those derived from a one-minute period with a sliding window to create subject-specific distributions of each postural sway feature (SSD method). We find traditional computation of sway features from the chest is sensitive to changes in PRMs and task differences. Distribution characteristics from the SSD method establish additional relationships with PRMs, detect differences in more tasks, and distinguish between fall status groups. Overall, the chest was found to be a valid location to monitor postural sway and we recommend utilizing the SSD method over single-observation analyses.