Human activity recognition is a popular research field in computer vision that has already been widely studied. However, it is still an active research field since it plays an important role in many current and emerging real-world intelligent systems, like visual surveillance and human-computer interaction. Deep Reinforcement Learning (DRL) has recently been employed to address the activity recognition problem with various purposes, such as finding attention in video data or obtaining the best network structure. DRL-based human activity recognition has only been around for a short time, and it is a challenging, novel field of study. Therefore, to facilitate further research in this field, we have constructed a comprehensive survey on activity recognition methods that incorporate deep reinforcement learning. Towards the end of this survey, we summarize key challenges and open problems in this area that can be addressed by researchers in the future.

Using intelligent reflecting surfaces (IRSs) to improve the coverage and the data rate of future wireless networks is a viable option. These surfaces are constituted of a significant number of passive and nearly passive components that interact with incident signals in a smart way, such as by reflecting them, to increase the wireless system’s performance as a result of which the notion of a smart radio environment comes to fruition. In this survey we supply a study review of IRS-assisted wireless communication starting with the principles of IRS which include the hardware architecture, the control mechanisms, and the discussions of previously held views about the channel model and pathloss, then the performance analysis considering different performance parameters, analytical approaches and metrics are presented to describe the IRS-assisted wireless network performance improvements. Despite its enormous promise, IRS confronts new hurdles in integrating into wireless networks efficiently due to its passive nature. Consequently, the channel estimation for, both full and nearly passive IRS and the IRS deployments are compared under various wireless communication models and for single and multi-users. Lastly, we propose the challenges and potential future study areas for the IRS aided wireless communication systems.

A Universal Mathematical Field Theory (UMFT) is established, which states that the combination of the operations of both gradient and divergence of vector fields, such as electric field and velocity field, create the curl of an axial vector field, such as magnetic field. Utilizing UMFT, we mathematically: (1) derive the Extended-Maxwell equations and the Lorentz force from Coulomb’s law and the velocity of the source; (2) establish Maxwell-type gravitational equations and Lorentz-type gravitational force (Gravito-EM) from Newton’s law and velocity of gravitational source; (3) establish Classical-Spin-Electromagnetism (Spin-EM) from the Coulomb’s law and the spin of the spin angular velocity; (4) predicate the Spin related force. For a source moving with non-spatially-varying velocity the Extended-Maxwell equations reduce to Maxwell equations, which justifies UMFT and shows that the experiments-based Maxwell equations have mathematical origin. This derivation mathematically explains how a moving electric charge creates magnetic field, and shows that there is no magnetic monopole charge. UMFT shows that mathematical identities lead to physical dualities including duality between EM and Gravito-EM. The concepts, effects and phenomena of EM may be directly converted to that of gravity. The Gravito-EM are employed to study the accelerating universe, rotation curve and gravitation waves. The Gravito-EM can be quantized, along the line of quantizing EM, and unified with EM force. We derive, for the first time, the Spin-Lorentz-type force and Lagrangian-Lorentz-type force. If experimentally proved, the Spin-related force may be the 5th force. UMFT provides mathematical origins of physical dualities between Extended EM, Gravito-EM and Spin-EM.

The paper deals with the processing and analysis of the results of electro-optical studies of the pre-breakdown stage of an electric discharge in water using the Kerr effect. For this, the first chronogram was selected, obtained in the process of electro-optical chronographic research of water breakdown. In this work, it was possible to explain not only the features of the chronogram but also to obtain an estimate of the intensity of the appearance of the anode streamer in water E ~ 40-50 MV/cm. Using the method of computer simulation in the point-plane system near the anode point, the distributions of the electric field strength and the phase difference were constructed. The calculations were carried out both at a constant relative permittivity and its nonlinear dependence on the electric field strength. It turned out that dipole saturation is observed near the tip. As a result, the Kerr bands become denser and, at a spatial resolution of the recording equipment of 30 μm, in the pre-measuring zone, they are not recorded on the chronogram.

The performance of most array signal processing tasks relies on the presence of correlation between sensor signals. In a wireless sensor network, where sensor nodes are spread out over a relatively large area, it is useful to identify nodes observing similar sensor signals and hence common phenomenons, for example to partition the network according to the observed latent signals and corresponding correlation structure. This can be achieved via the so-called MAXVAR formulation of generalized canonical correlation analysis, which finds a low-dimensional subspace that highlights correlated signal components between multiple nodes' observed signal subspaces. The classical procedure for computing the solutions of MAXVAR consists in performing a generalized eigenvalue decomposition after collecting all the sensors' signals at a fusion center. However, this typically incurs high communication and computational costs. In this paper, we describe a low communication and computational cost distributed algorithm that computes the solutions of MAXVAR without aggregating the nodes' observations at a central location. We show how a subset of those solutions can be used locally by each node to estimate the global correlation structure across all nodes in the network, thereby allowing any node to evaluate the presence of correlated signals at any other node, even if no direct link is shared. We prove the convergence of the algorithm and validate our method for estimating the correlation structure via simulations.

With recent advances in the field of IoT (Internet of Things), the quantity of data being generated by various sensors and Internet users has increased dramatically which in turn has skyrocketed the need for efficient data compression methods. Data summarization is an efficient and effective technique for data compression that can generate a brief and succinct summary from typically larger quantities of data in an intelligent and highly useful manner, which can be done at various levels of abstraction. The impact of using such a technique in large IoT networks can be significantly advantageous in terms of reduction in the processing time, overall computation, data storage-transmission requirements, energy consumption and possible workload on IoT users. In this work, a review of existing methods for summarization techniques at various levels of abstraction of typical IoT networks are discussed. The levels of categorization that are considered are Low-level and High-level. Under each abstraction level, various techniques are further classified while briefly describing their essential characters.

Fifth-generation (5G) requires a highly accurate estimate of the channel state information (CSI) to exploit the benefits of massive multiple-input-multiple-output (MaMIMO) systems. 5G systems use pilot sequences to estimate channel behaviour using traditional methods like least squares (LS), or minimum mean square error (MMSE) estimation. However, traditional methods do not always obtain reliable estimations: LS exhibits a poor estimation when inadequate channel conditions (i.e., low-signal-to-noise ratio (SNR) region) and MMSE requires prior statistical knowledge of the channel and noise (complex to implement in practice). We present a deep learning framework based on deep neural networks (DNNs) for 5G MaMIMO channel estimation. After a first preliminary model with which we verify the good estimation capacity of our DNN-based approach, we propose two different models, which differ in the information processed by the DNN and benefit from lower computational complexity or greater flexibility for any reference signal pattern, respectively. The results show that, compared to the LS-based channel estimation, the DNN approach decreases the mean square error (MSE) and the system’s spectral efficiency (SE) increases, especially in the low-SNR region. Our approach provides results close to optimal MMSE estimation but benefits from not requiring any prior channel statistics information.

The paper evaluates the ability of the Adaptive Neuro-Fuzzy Inference System (ANFIS) Network to develop its own set of inverse kinematics (IK) equations when trained with forward kinematics data. The forward kinematics equation is developed using the Denavit-Hartenberg (D-H) approach while the inverse kinematics equations (which is used to test the ANFIS Network) is developed analytically using geometry. This evaluation is carried out on a four degree of freedom (4DOF) robotic arm.

Non-orthogonal multiple access principle applied in power domain (power-domain NOMA) enhances spectrum efficiency and connectivity of mobile communication systems through sharing between Orthogonal Multiple Access (OMA) and other multiple users. It is the almost main current trend for research and standards for 5G and beyond. Spectrum sharing generates multi-users interference (Multiple Access Interference, MAI) which destroys orthogonality in OMA systems. Meanwhile the so-called Successive Interference Cancellation (SIC) techniques are used to suppress MAI and the Cognitive Radio (CR) principle (CR-NOMA) for decoding order for multiple users inspired effective SIC algorithms. The CR-NOMA principle is applied in the present material as well, but with significant changes for SIC design as for Doubly Selective channels. Due to the above reasons, two original SIC design methods are proposed hereafter following the CR-NOMA principle. Regarding to it the OMA signal (primary user) is decoded first using a Chaos-based quasi-optimum Extended Kalman Filter (EKF). The set of “secondary users” (multiusers) are proposed to be decoded secondly by the following two different methods: - Chaos-based EKF filters or , - Sequential m hypothesis testing. The proposed approach for SIC design is rather simple, fast and accurate for practical implementation.

Designing dispersion free photonic crystal structures requires lot of computational time. Often, these structures are designed by solving the conventional Maxwell's equations. Different computational techniques are available for solving such coupled equations and obtaining the eigen frequencies of the system. However, these methods are time consuming and limit the designing of dispersion free structures. On the other side, Machine learning and deep learning are providing the assistance to the researchers in finding the solutions for various complex, time consuming and tedious problems. In the field of dispersion free structure, researchers are choosing machine learning for designing, developing and analyzing various structures. Here, in this paper, we made an attempt to adopt machine learning for predicting the dispersion values and group indices of a lattice shifted photonic crystal waveguide. Conventional methods are used to generate the eigen frequencies, group indices and wave vector points of the structure. These data sets are used for training the machine by artificial neural network. The model is trained and optimized to predict the dispersion parameters and group indices with applied lattice shift. Repeated random sub-sampling cross validation is used for testing the performance of the model. To the best of our knowledge, we made a first attempt to study a lattice shifted photonic crystal waveguide using machine learning. This work will be exciting and may lead to develop a connection between tailored waveguide structures and machine learning.

In this work, we have highlighted the recent developments on Fake news detection.

By pushing resources to far-edge servers located in the proximity of users, edge computing can greatly reduce end-to-end transmission delays. Task offloading in multi-tier networks refers to the optimization of which tasks should be offloaded from the far-edge to the edge and the cloud. Moreover, the containerization of applications can further reduce resource and time consumption and, in turn, the latency of such applications. Even though Kubernetes has become the de facto container orchestrator, not many works have considered the offloading of containerized applications in Kubernetes clusters spanning from cloud to far-edge. In this work, the problem of offloading Kubernetes tasks (or pods) in three-tier networks is formulated and optimized. First, a utility function is presented in terms of the cumulative weighted pod response time, and a utility minimization problem with central processing unit (CPU) constraints is presented. Based on the optimal theoretical solution to this problem, a three-tier offloading decision algorithm (TTODA) is developed. Vertical scaling is considered, and specific hardware capabilities of each node are taken into account by setting specific SLAs that are fed back to the algorithm. Numerical results show that TTODA outperforms a typical Kubernetes QoS model based on first-in, first-served algorithm (FIFSA) in terms of utility, average pod response time, and usage of far-edge CPU. Further, TTODA achieves an excellent trade-off between performance and computational complexity, and thus it can help achieve the requirements of latency-sensitive applications. Moreover, TTODA can easily be extended to scenarios with joint memory and CPU constraints.

We demonstrate a radio frequency (RF) phase-encoded signal generator as well as a user-defined RF arbitrary waveform generator (AWG) based on a soliton crystal micro-comb generated by an integrated MRR with a free spectral range of ~49 GHz. Owing to the soliton crystal’s robust and stable generation as well as the high intrinsic efficiency, RF phase-encoded signal generators and AWGs with simple operation and fast reconfiguration are realized. The soliton crystal micro-comb provides 60 wavelengths for RF phase-encoded signal generators, achieving a phase encoding speed of 5.95 Gb/s and a high pulse compression ratio of 29.6. Over 80 wavelengths are employed for the AWGs, achieving tunable square waveforms with a duty cycle ratio ranging from 10% to 90%, sawtooth waveforms with tunable slope ratios from 0.2 to 1, and symmetric concave quadratic chirp waveforms. Our system has great potential to achieve RF and microwave photonic signal generation and processing with low cost and footprint.

We present a decentralized advantage actor-critic algorithm that utilizes learning agents in parallel environments with synchronous gradient descent. This approach decorrelates agents' experiences, stabilizing observations and eliminating the need for a replay buffer, requires no knowledge of the other agents' internal state during training or execution, and runs on a single multi-core CPU.

FEM_2D is an open-source Finite Element Method library implemented in Rust. This MetaPaper describes some of the libraries basic functionality, as well as its unique contributions to the FEM research community. This package exhibits a powerful *hp*-refinement API with anisotropic refinements and no constraints on mesh irregularity. It also leverages Rust’s Trait-Based Genericism to provide straightforward application to many problem domains. This library has been featured in some recent research in Computational Electromagnetics (see references), depicting the validity of the underlying methodology. FEM_2D has been made available as an open-source library with the intention of providing additional explanatory value to the existing related research, as well as to facilitate the expansion of this research area

A discursive piece aimed at practitioners in Engineering Management. On the subject of Complexity Leadership and KPIs, this short article presents three models which expand on the work of Uhl-Bien and Marion (2007) by focusing on Administrative, Enabling and Adaptive Leaderships for managing complexity.

Resolving the ambiguity in both range and velocity domain is one of the essential issues for the pulse Doppler (PD) radars using medium pulse repetition frequency (PRF) waveforms. The existing solutions are mainly based on the Chinese Remainder Theorem (CRT) and its variations and extensions, but these algorithms typically have an unsatisfactory accuracy and are highly complex in computation, especially for the case of a large false-alarm rate. In this work, we proposed a new approach to tackle with ambiguity based on the partial distance matrices (PDM) and implemented it on the NVIDIA graphics processing unit (GPU) platform. The evaluations based on both Monte-Carlo simulations and real measured data present the advantages of the proposed approach in accuracy and efficiency. The results also show the proposed approach can be utilized as a real-time algorithm even for a low-threshold detection scheme.

Traditional portfolio management methods can incorporate specific investor preferences but rely on accurate forecasts of asset returns and covariances. Reinforcement learning (RL) methods do not rely on these explicit forecasts and are better suited for multi-stage decision processes. To address limitations of the evaluated research, experiments were conducted on three markets in different economies with different overall trends. By incorporating specific investor preferences into our RL models’ reward functions, a more comprehensive comparison could be made to traditional methods in risk-return space. Transaction costs were also modelled more realistically by including nonlinear changes introduced by market volatility and trading volume. The results of this study suggest that there can be an advantage to using RL methods compared to traditional convex mean-variance optimisation methods under certain market conditions. Our RL models could significantly outperform traditional single-period optimisation (SPO) and multi-period optimisation (MPO) models in upward trending markets, but only up to specific risk limits. In sideways trending markets, the performance of SPO and MPO models can be closely matched by our RL models for the majority of the excess risk range tested. The specific market conditions under which these models could outperform each other highlight the importance of a more comprehensive comparison of Pareto optimal frontiers in risk-return space. These frontiers give investors a more granular view of which models might provide better performance for their specific risk tolerance or return targets.