Search

Effective operation of smart grids relies on accurate forecasting models for renewable power generation (RPG) and power consumption. The intermittent and unpredictable nature of RPG, coupled with diverse consumption patterns, underscores the importance of robust forecasting approaches. Existing models often employ stacked layers, integrating direct features into fully connected layers, yielding suboptimal results with limited generalization capabilities. Addressing these limitations, we propose a novel two-stream architecture for RPG and power consumption forecasting. The first stream leverages dilated causal convolutional layers to capture intricate patterns, while the second stream focuses on temporal information extraction. Importantly, we fine-tune the hyperparameters of both streams using Bayesian algorithms to optimize the learning process. The outputs from these two streams are then intelligently combined and channeled into our innovative trapezoid attention module (TAM) for feature refinement, resulting in superior pattern representation. The TAM incorporates three distinct dimensions (spatial, temporal, and spatiotemporal) and enriches feature maps by integrating a skip connection from the pre-TAM features. The output post-TAM features are then employed for final forecasting. Our approach showcases remarkable performance in short-term forecasting across a spectrum of datasets, including RPG, regional, residential, and industrial power consumption. By addressing the shortcomings of existing forecasting models, our research contributes to the advancement of smart grid technologies, ensuring more reliable and efficient energy management.

We study the Kalman Filter for the linear elastic wave equation over the real line with spatially distributed partial state measurements. The dynamics of the filter are described by a spatial convolution operator with asymptotic exponential spatial decay rate. This decay rate dictates how measurements from different spatial locations must be exchanged to implement the filter: faster spatial decay implies local measurements are more relevant and the filter is more “decentralized”; slower decay implies farther measurements also become relevant and the filter is more “centralized”. Using dimensional analysis, we demonstrate that this decay rate is a function of one dimensionless group defined from system parameters, such as wave speed and noise variances. We find a critical value of such dimensionless group for which the Kalman Filter is completely decentralized.

We propose a new neck design for legged robots to achieve robust visual-inertial state estimation in dynamic locomotion. While visual-inertial state estimation is widely used in robotics, it has a problem of being disturbed by the impacts and vibration generated when legged robots move dynamically. The use of rubber dampers may be a solution, but even if the dampers are proper for some gaits, they may be excessively deformed or resonated at certain frequencies during other gait locomotion since they are not tunable. To address this problem, we develop a tunable neck system that absorbs the impacts and vibration during diverse gait locomotions. This neck system consists of two components: 1) a suspension mechanism that compensates for the weight of the head equipped with a camera and IMU (inertial measurement unit), absorbs the impacts and the head motion of high frequencies including vibration as a fixed low-pass filter; and 2) a dynamic vibration absorber (DVA) that can be reactively-adjusted to diverse gait frequencies to alleviate excessive head movements. We present a dynamics analysis of the neck system and show how to adjust the target frequency of the system. Simulation and experimental validation are performed to verify the effect of the proposed neck design, manifesting superior estimation performance and robustness across diverse gaits.

The ability to process environment maps across multiple sessions is critical for robots operating over extended periods of time. Specifically, it is desirable for autonomous agents to detect changes amongst maps of different sessions so as to gain a conflict-free understanding of the current environment. In this letter, we look into the problem of change detection based on a novel map representation, dubbed Plane Signed Distance Fields (PlaneSDF), where dense maps are represented as a collection of planes and their associated geometric components in SDF volumes. Given point clouds of the source and target scenes, we propose a three-step PlaneSDF-based change detection approach: (1) PlaneSDF volumes are instantiated within each scene and registered across scenes using plane poses; 2D height maps and object maps are extracted per volume via height projection and connected component analysis. (2) Height maps are compared and intersected with the object map to produce a 2D change location mask for changed object candidates in the source scene. (3) 3D geometric validation is performed using SDF-derived features per object candidate for change mask refinement. We evaluate our approach on both synthetic and real-world datasets and demonstrate its effectiveness via the task of changed object detection.

We developed an environmentally adaptive under‐ice navigation framework that was deployed in the Arctic Beaufort Sea during the United States Navy Ice Exercise in March 2020 (ICEX20). This navigation framework contained two subsystems developed from the ground up: (1) an on‐board hydrodynamic model‐aided navigation (HydroMAN) engine, and (2) an environmentally and acoustically adaptive integrated communication and navigation network (ICNN) that provided acoustic navigation aiding to the former. The HydroMAN synthesized measurements from an inertial navigation system (INS), ice‐tracking Doppler velocity log (DVL), ICNN and pressure sensor into its self‐calibrating vehicle flight dynamic model to compute the navigation solution. The ICNN system, which consisted of four ice buoys outfitted with acoustic modems, trilaterated the vehicle position using the one‐way‐travel‐times (OWTT) of acoustic datagrams transmitted by the autonomous underwater vehicle (AUV) and received by the ice buoy network. The ICNN digested salinity and temperature information to provide model‐assisted real‐time OWTT range conversion to deliver accurate acoustic navigation updates to the HydroMAN. To decouple the contributions from the HydroMAN and ICNN subsystems towards a stable navigation solution, this article evaluates them separately: (1) HydroMAN was compared against DVL bottom‐track aided INS during pre‐ICEX20 engineering trials where both systems provided similar accuracy; (2) ICNN was evaluated by conducting a static experiment in the Arctic where the ICNN navigation updates were compared against GPS with ICNN error within low tens of meters. The joint HydroMAN‐ICNN framework was tested during ICEX20, which provided a nondiverging high‐resolution navigation solution—with the majority of error below 15 m—that facilitated a successful AUV recovery through a small ice hole after an 11 km untethered run in the upper and mid‐water column.

In this review, recent trends in microelectronics packaging reliability are summarized. We review the technology from early packaging concepts, including wire bond and BGA, to advanced techniques used in HI schemes such as 3D stacking, interposers, fan-out packaging, and more recently developed silicon interconnect fabric integration. This review includes approaches for both design modification studies and packaged device validation. Methods are explored for compatibility in new complex packaging assemblies. Suggestions are proposed for optimizations of the testing practices to account for the challenges anticipated in upcoming HI packaging schemes.

Adding fiducial markers to a scene is a well-known strategy for making visual localization algorithms more robust. Traditionally, these marker locations are selected by humans who are familiar with visual localization techniques. This paper explores the problem of automatic marker placement within a scene. Specifically, given a predetermined set of markers and a scene model, we compute optimized marker positions within the scene that can improve accuracy in visual localization. Our main contribution is a novel framework for modeling camera localizability that incorporates both natural scene features and artificial fiducial markers added to the scene. We present optimized marker placement (OMP), a greedy algorithm that is based on the camera localizability framework. We have also designed a simulation framework for testing marker placement algorithms on 3D models and images generated from synthetic scenes. We have evaluated OMP within this testbed and demonstrate an improvement in the localization rate by up to 20 percent on four different scenes.

This paper presents normalizing flows for incremental smoothing and mapping (NF-iSAM), a novel algorithm for inferring the full posterior distribution in SLAM problems with nonlinear measurement models and non-Gaussian factors. NF-iSAM exploits the expressive power of neural networks, and trains normalizing flows to model and sample the full posterior. By leveraging the Bayes tree, NF-iSAM enables efficient incremental updates similar to iSAM2, albeit in the more challenging non-Gaussian setting. We demonstrate the advantages of NF-iSAM over state-of-the-art point and distribution estimation algorithms using range-only SLAM problems with data association ambiguity. NF-iSAM presents superior accuracy in describing the posterior beliefs of continuous variables (e.g., position) and discrete variables (e.g., data association).

Conventional discontinuous pulse width modulation (DPWM) strategies for cascaded H-bridge static compensators suffer from poor inter-phase capacitor voltage balance control during unbalanced grid voltage conditions. Specifically, the logic-based algorithm that calculates the zero-sequence voltage (ZSV) for discontinuous operation interacts with the closed-loop capacitor voltage control. This paper proposes an optimal DPWM strategy based on a finite-set optimization that chooses the best ZSV candidate that minimizes the control interactions while clamping a converter voltage. Furthermore, the effect of the different ZSV candidates on the switching loss is also considered within the optimization. Extensive experimental study is provided to demonstrate the effectiveness of the proposed approach.

A critical reduction in the total harmonic distortion in the output voltage of multilevel inverters can be obtained with an appropriate selection of the angles to switch the power semiconductors at low frequency. Among the methods used to obtain the required sequence of angles, it is possible to recognize both standard and optimized methods, like the Newton–Raphson method or the genetic algorithm, to mention a few. However, implementing the genetic algorithm implies using several configuration parameters, so it is difficult to know if the best option was chosen. This paper implements the Taguchi design of experiments, the Taguchi method, to obtain the appropriate switching angles for an multilevel inverter producing an output voltage of five and seven levels. Experimental results obtained from a laboratory prototype using the computed angles in the MATLAB GA solver demonstrate the superiority of the proposed method.

Federated Learning (FL) is a promising paradigm to empower on-device intelligence in Industrial Internet of Things(IIoT) due to its capability of training machine learning models across multiple IIoT devices, while preserving the privacy of their local data. However, the distributed architecture of FL relies on aggregating the parameter list from the remote devices, which poses potential security risks caused by malicious devices. In this paper, we propose a flexible and robust aggregation rule, called Auto-weighted Geometric Median (AutoGM), and analyze the robustness against outliers in the inputs. To obtain the value of AutoGM, we design an algorithm based on alternating optimization strategy. Using AutoGM as aggregation rule, we propose two robust FL solutions, AutoGM_FL and AutoGM_PFL. AutoGM_FL learns a shared global model using the standard FL paradigm, and AutoGM_PFL learns a personalized model for each device. We conduct extensive experiments on the FEMNIST and Bosch IIoT datasets. The experimental results show that our solutions are robust against both model poisoning and data poisoning attacks. In particular, our solutions sustain high performance even when 30% of the nodes perform model or 50% of the nodes perform data poisoning attacks.

Federated Learning (FL) is a promising paradigm to empower on-device intelligence in Industrial Internet of Things(IIoT) due to its capability of training machine learning models across multiple IIoT devices, while preserving the privacy of their local data. However, the distributed architecture of FL relies on aggregating the parameter list from the remote devices, which poses potential security risks caused by malicious devices. In this paper, we propose a flexible and robust aggregation rule, called Auto-weighted Geometric Median (AutoGM), and analyze the robustness against outliers in the inputs. To obtain the value of AutoGM, we design an algorithm based on alternating optimization strategy. Using AutoGM as aggregation rule, we propose two robust FL solutions, AutoGM_FL and AutoGM_PFL. AutoGM_FL learns a shared global model using the standard FL paradigm, and AutoGM_PFL learns a personalized model for each device. We conduct extensive experiments on the FEMNIST and Bosch IIoT datasets. The experimental results show that our solutions are robust against both model poisoning and data poisoning attacks. In particular, our solutions sustain high performance even when 30% of the nodes perform model or 50% of the nodes perform data poisoning attacks.

Federated Learning (FL) is a promising paradigm to empower on-device intelligence in Industrial Internet of Things(IIoT) due to its capability of training machine learning models across multiple IIoT devices, while preserving the privacy of their local data. However, the distributed architecture of FL relies on aggregating the parameter list from the remote devices, which poses potential security risks caused by malicious devices. In this paper, we propose a flexible and robust aggregation rule, called Auto-weighted Geometric Median (AutoGM), and analyze the robustness against outliers in the inputs. To obtain the value of AutoGM, we design an algorithm based on alternating optimization strategy. Using AutoGM as aggregation rule, we propose two robust FL solutions, AutoGM_FL and AutoGM_PFL. AutoGM_FL learns a shared global model using the standard FL paradigm, and AutoGM_PFL learns a personalized model for each device. We conduct extensive experiments on the FEMNIST and Bosch IIoT datasets. The experimental results show that our solutions are robust against both model poisoning and data poisoning attacks. In particular, our solutions sustain high performance even when 30% of the nodes perform model or 50% of the nodes perform data poisoning attacks.

In this paper, we propose an approach for computing invariant sets of discrete-time nonlinear systems by lifting the nonlinear dynamics into a higher dimensional linear model. In particular, we focus on the maximal admissible invariant set contained in some given constraint set. For special types of nonlinear systems, which can be exactly immersed into higher dimensional linear systems with state transformations, invariant sets of the original nonlinear system can be characterized using the higher dimensional linear representation. For general nonlinear systems without the immersibility property, approximate immersions are defined in a local region within some tolerance and linear approximations are computed by leveraging the fixed-point iteration technique for invariant sets. Given the bound on the mismatch between the linear approximation and the original system, we provide an invariant inner approximation of the maximal admissible invariant set by a tightening procedure.

Ground-Penetrating Radar (GPR) has been widely used in the backfill grouting detections of shield tunnel construction. The proposed approach of automatic annular thickness measurement based on GPR monitoring system and BBP-XGBoost algorithm has been demonstrated to be an effective and efficient technique for detecting the backfill grouting of shield tunnel construction. The advantages include: (1) high accuracy: the proposed model is able to achieve high coefficients of regression R2 (0.943) for training sets and comparatively low RMSE values (0.421 and 1.32 cm) for both training and testing datasets, respectively; (2) high efficiency: the proposed XGBoost algorithm outscores other models in both accuracy and efficiency; (3) real-time output: the trained models were applied to automatically measure the annular grouting thickness of the Tianjin Metro tunnel constructions (China), with consistent real-time outputs for the engineers on site. The proposed approach provides real-time and efficient guidance for the construction of shield tunnels.

In the operation of battery energy storage systems based on the cascaded H-bridge converter, it is beneficial to balance the state of charge of batteries in different submodules within the converter phase-arm. This is achieved by distributing the active power among the submodules. Although multiple methods have been proposed for this purpose, they face the challenge of rendering optimal active power distributions that maximize balancing speed while meeting power constraints in the battery energy storage system. To overcome this challenge, a model predictive control scheme is developed in this paper. The proposed method is remarkably robust against parametric uncertainties (battery voltage, capacity, etc.), as evidenced by its ability to tolerate a substantial 50% uncertainty in the parameters, resulting in a mere 0.05% steady-state error. Furthermore, because the predictive control can be executed at a low frequency, the computational burden is comparable to other existing methods.

This report presents the analysis, study, and simulation of an ultrafast charging station (UFCS) for electric vehicles (EVs) operating in steady state. The electrical architecture of the charging station uses an ac bus plus two dc buses and it is supported by a storage system based on batteries and super-capacitors. The power demand of the EVs is established taking into account the electric characteristics of their batteries and the availability of the station charging points. The analysis introduces a supervisory control based on a state machine description for different operating modes, which eventually facilitates fault detection in the electrical architecture. In addition, the study proposes different methods to handle the required energy for the charging demand and a procedure for the correct sizing of both the energy storage system and the input transformer. In laboratory experiments in a reduced-scale storage system, a SCADA supervision with CAN communication has proved successful in gathering data corresponding to modes of charge and discharge in batteries and super-capacitors, and subsequently displaying them on a computer screen.

We describe parallel and series resonant converters (PRC and SRC) via a unified set of input-dependent coordinates whose dynamics are intrinsically hybrid. We then propose hybrid feedback showing a self-oscillating behavior whose amplitude and frequency can be adjusted by a reference input ranging from zero to $\pi$ . For any reference value in that range, we give a Lyapunov function certifying the existence of a unique nontrivial hybrid limit cycle whose basin of attraction is global except for the origin. Our results are confirmed by experimental results on an SRC prototype.

The advanced modeling and estimation of overall system dynamics play a vital role in electric vehicles (EVs), as the selection of components in the powertrain and prediction of performance are the key market qualifiers. The state-space averaged model and small-signal transfer function model are useful for assessing long-term behavior in system-level analysis and for designing the controller parameters, respectively. Both models take less computation time but ignore the high-frequency switching dynamics. Therefore, these two models could be impractical for the development and testing of EV prototypes. On the other hand, the multi-domain model in available simulation tools gives in-depth information about the short-term behavior and loss analysis of power electronic devices in each subsystem, considering the switching dynamics in a long computation time. In this paper, a general mathematical framework for the dynamical analysis of complete EVs is presented using a unified, switched nonlinear model. This equation-based model runs faster than the available module-based simulation tools. Two other models, namely the time domain state-space averaged model and frequency domain small-signal transfer function model, are also developed from the switched nonlinear model for the analysis with less computation time. The design and performance of an EV with two different motors and its controllers are evaluated using the general mathematical framework.

This report presents the analysis, study, and simulation of an ultrafast charging station (UFCS) for electric vehicles (EVs) operating in steady state. The electrical architecture of the charging station uses an ac bus plus two dc buses and it is supported by a storage system based on batteries and super-capacitors. The power demand of the EVs is established taking into account the electric characteristics of their batteries and the availability of the station charging points. The analysis introduces a supervisory control based on a state machine description for different operating modes, which eventually facilitates fault detection in the electrical architecture. In addition, the study proposes different methods to handle the required energy for the charging demand and a procedure for the correct sizing of both the energy storage system and the input transformer. In laboratory experiments in a reduced-scale storage system, a SCADA supervision with CAN communication has proved successful in gathering data corresponding to modes of charge and discharge in batteries and super-capacitors, and subsequently displaying them on a computer screen.

This paper proposes a discontinuous modulation (DM) scheme for cascaded H-bridge (CHB) static compensators (StatComs) with star configuration. The proposed DM scheme guarantees zero steady-state active power in each converter phase-arm, even in the presence of severe grid imbalance conditions, thus remaining decoupled from the inter-phase capacitor voltage balancing controller. Moreover, the proposed DM scheme exploits the possibility of clamping to the zero-voltage level, thus providing a more flexible discontinuous operation than conventional DM methods. This involves lower zero-sequence-voltage injection requirements and lower switching losses. The proposed modulation scheme is experimentally corroborated and its performance under different grid voltage conditions is compared with traditional modulation schemes.

Despite the successes of deep reinforcement learning (RL), it is still challenging to obtain safe policies. Formal verifi- cation approaches ensure safety at all times, but usually overly restrict the agent’s behaviors, since they assume adversarial behavior of the environment. Instead of assuming adversarial behavior, we suggest to focus on perceived safety instead, i.e., policies that avoid undesired behaviors while having a desired level of conservativeness. To obtain policies that are perceived as safe, we propose a shield synthesis framework with two distinct loops: (1) an inner loop that trains policies with a set of actions that is constrained by shields whose conservativeness is parameterized, and (2) an outer loop that presents example rollouts of the policy to humans and collects their feedback to update the parameters of the shields in the inner loop. We demonstrate our approach on a RL benchmark of Lunar landing and a scenario in which a mobile robot navigates around humans. For the latter, we conducted two user studies to obtain policies that were perceived as safe. Our results indicate that our framework converges to policies that are perceived as safe, is robust against noisy feedback, and can query feedback for multiple policies at the same time., QC 20211215

We explore the potential to control terrain vehicles using deep reinforcement in scenarios where human operators and traditional control methods are inadequate. This letter presents a controller that perceives, plans, and successfully controls a 16-tonne forestry vehicle with two frame articulation joints, six wheels, and their actively articulated suspensions to traverse rough terrain. The carefully shaped reward signal promotes safe, environmental, and efficient driving, which leads to the emergence of unprecedented driving skills. We test learned skills in a virtual environment, including terrains reconstructed from high-density laser scans of forest sites. The controller displays the ability to handle obstructing obstacles, slopes up to 27 degrees, and a variety of natural terrains, all with limited wheel slip, smooth, and upright traversal with intelligent use of the active suspensions. The results confirm that deep reinforcement learning has the potential to enhance control of vehicles with complex dynamics and high-dimensional observation data compared to human operators or traditional control methods, especially in rough terrain.

We explore the potential to control terrain vehicles using deep reinforcement in scenarios where human operators and traditional control methods are inadequate. This letter presents a controller that perceives, plans, and successfully controls a 16-tonne forestry vehicle with two frame articulation joints, six wheels, and their actively articulated suspensions to traverse rough terrain. The carefully shaped reward signal promotes safe, environmental, and efficient driving, which leads to the emergence of unprecedented driving skills. We test learned skills in a virtual environment, including terrains reconstructed from high-density laser scans of forest sites. The controller displays the ability to handle obstructing obstacles, slopes up to 27 degrees, and a variety of natural terrains, all with limited wheel slip, smooth, and upright traversal with intelligent use of the active suspensions. The results confirm that deep reinforcement learning has the potential to enhance control of vehicles with complex dynamics and high-dimensional observation data compared to human operators or traditional control methods, especially in rough terrain.

Despite the successes of deep reinforcement learning (RL), it is still challenging to obtain safe policies. Formal verifi- cation approaches ensure safety at all times, but usually overly restrict the agent’s behaviors, since they assume adversarial behavior of the environment. Instead of assuming adversarial behavior, we suggest to focus on perceived safety instead, i.e., policies that avoid undesired behaviors while having a desired level of conservativeness. To obtain policies that are perceived as safe, we propose a shield synthesis framework with two distinct loops: (1) an inner loop that trains policies with a set of actions that is constrained by shields whose conservativeness is parameterized, and (2) an outer loop that presents example rollouts of the policy to humans and collects their feedback to update the parameters of the shields in the inner loop. We demonstrate our approach on a RL benchmark of Lunar landing and a scenario in which a mobile robot navigates around humans. For the latter, we conducted two user studies to obtain policies that were perceived as safe. Our results indicate that our framework converges to policies that are perceived as safe, is robust against noisy feedback, and can query feedback for multiple policies at the same time., QC 20211215

Despite the successes of deep reinforcement learning (RL), it is still challenging to obtain safe policies. Formal verifi- cation approaches ensure safety at all times, but usually overly restrict the agent’s behaviors, since they assume adversarial behavior of the environment. Instead of assuming adversarial behavior, we suggest to focus on perceived safety instead, i.e., policies that avoid undesired behaviors while having a desired level of conservativeness. To obtain policies that are perceived as safe, we propose a shield synthesis framework with two distinct loops: (1) an inner loop that trains policies with a set of actions that is constrained by shields whose conservativeness is parameterized, and (2) an outer loop that presents example rollouts of the policy to humans and collects their feedback to update the parameters of the shields in the inner loop. We demonstrate our approach on a RL benchmark of Lunar landing and a scenario in which a mobile robot navigates around humans. For the latter, we conducted two user studies to obtain policies that were perceived as safe. Our results indicate that our framework converges to policies that are perceived as safe, is robust against noisy feedback, and can query feedback for multiple policies at the same time., QC 20211215

We explore the potential to control terrain vehicles using deep reinforcement in scenarios where human operators and traditional control methods are inadequate. This letter presents a controller that perceives, plans, and successfully controls a 16-tonne forestry vehicle with two frame articulation joints, six wheels, and their actively articulated suspensions to traverse rough terrain. The carefully shaped reward signal promotes safe, environmental, and efficient driving, which leads to the emergence of unprecedented driving skills. We test learned skills in a virtual environment, including terrains reconstructed from high-density laser scans of forest sites. The controller displays the ability to handle obstructing obstacles, slopes up to 27 degrees, and a variety of natural terrains, all with limited wheel slip, smooth, and upright traversal with intelligent use of the active suspensions. The results confirm that deep reinforcement learning has the potential to enhance control of vehicles with complex dynamics and high-dimensional observation data compared to human operators or traditional control methods, especially in rough terrain.

Upper limb loss alters individuals’ private and professional life. Prosthetic devices are thus a solution to supply the missing upper limb segments. Nevertheless, commercial prostheses are often unaffordable, or inaccessible, to underprivileged individuals (e.g., no health insurance, low incomes, warzone). Among potential affordable alternatives, additive manufacturing, commonly “3D printing”, has been increasingly employed. This technology offers higher availability and accessibility, and can produce complex geometrical and highly customized products, which are essential features for prostheses manufacturing. Therefore, this study aims to portray an overview of reliable open-source upper limb 3D-printed prostheses currently available. We thus searched the scientific literature and online repositories hosting 3D-printable designs. We extracted data relative to mechanical and kinematic properties, 3D printing process and efficacy for each device. We found six studies implementing open-source 3DP upper limb prostheses and twenty-five open-source designs from online databases meeting selection criteria. Devices’ technical specifications were not systematically reported. In conclusion, though open-source 3D-printed upper limb prostheses can perform some functional tasks and grasps, and are widely employed to supply limb differences, further research is mandatory to validate their usage and to prove their clinical efficacy. More guidelines are required to unify contributions from private makers and non-governmental organizations with scientific groups.

In recent years, the monitoring of compositional data using control charts has been investigated in the Statistical Process Control field. In this study, we will design a Phase II Multivariate Exponentially Weighted Moving Average (MEWMA) control chart with variable sampling intervals to monitor compositional data based on isometric log-ratio transformation. The Average Time to Signal will be computed based on the Markov chain approach to investigate the performance of the proposed chart. We also propose an optimization procedure to obtain the optimal control limit, smoothing constant, and out-of-control Average Time to Signal for different shift sizes and short sampling intervals. The performance of the proposed chart in comparison with the standard MEWMA chart for monitoring compositional data is also provided. Finally, we end the paper with a conclusion and some recommendations for future research.

We study a continuous-time system that solves the optimization problem over the set of orthogonal matrices, which is also known as the Stiefel manifold. The resulting optimization flow follows a path that is not always on the manifold but eventually lands on the manifold. We introduce a generalized Stiefel manifold to which we extend the canonical metric of the Stiefel manifold. We show that the vector field of the proposed flow can be interpreted as the sum of a Riemannian gradient on a generalized Stiefel manifold and a normal vector. Moreover, we prove that the proposed flow globally converges to the set of critical points, and any local minimum and isolated critical point is asymptotically stable.

Romie is a decision support tool based on AI’s latest advances in the domain of robust scheduling. Unlike all existing systems, the tool allows to (i) visually model the operational problem and context entirely (ii) optimize to find near-optimal schedules while taking uncertainty into account and (iii) deal with a combination of various key performance indicators (KPIs). It comes with a web user interface. Part or all of the modelled activities may be associated to random variables describing their stochastic durations, in order to produce schedules that are robust w.r.t. temporal uncertainty. Hence, depending on the pursued KPIs, the schedules maximize a combination of the following terms: the probability of satisfying the problem constraints, the expected return/efficiency, the expected outcome quality, and even the operators’ wellness by minimizing its expected extra-hours. Initially developed for spatial exploration and demonstration in the context of Mars analogue missions (i.e. missions on Earth that simulate condition and aspects of Mars missions), this versatile tool is here applied to operations management in both biotechnology manufacturing and robots parametrization in a cave exploration context.

Mathematical models of social networks, i.e., communities of interacting individuals, have existed for more than 50 years and have been used extensively by sociologists, behavioral scientists, and economists. The traditional focus has been on obtaining models that capture sociological effects like interpersonal influence (tendency of individuals to be influenced by others), homophily (tendency to associate with other individuals of similar behavior, opinions, and characteristics), polarization (tendency of a community to split into opposite factions), crowd effects (tendency to follow the opinion of the majority), echo chambers (tendency of an isolated community to self-amplify their beliefs), and so on. With the advent of online social media, the breadth and scope of the research in social network theory has scaled drastically in both size and accuracy, as numbers of interacting individuals have soared, and recorded data streams have rendered the analysis of individual behaviors, preferences, and interpersonal relationships more quantitative than ever before. Alongside the graph-theoretical notions drawn from network science, such as centrality, connectivity, and resilience, tools from dynamical systems and control prove very useful when trying to capture and understand the emerging behavior of such complex systems. Indeed, the last few years have witnessed a steep increase in the number of works studying social networks from a control systems perspective. This Special Issue of IEEE Transactions on Control of Network Systems(TCNS) consolidates this trend and gathers original contributions that identify and solve some of the emerging challenges in the field. It contains 14 articles addressing a wide range of topics ranging from non-Bayesian social learning to evolutionary dynamics in population games, social influence and opinion formation processes, their coupling with epidemic processes and recommendation systems, as well as optimal targeting problems. We believe

We consider the problem of boundary feedback control of single-input-single-output one-dimensional linear hyperbolic systems when sensing and actuation are anti-located. The main issue of the output feedback stabilization is that it requires dynamic control laws that include delayed values of the output (directly or through state observers) which may not be robust to infinitesimal uncertainties on the characteristic velocities. The purpose of this paper is to highlight some features of this problem by addressing the feedback stabilization of an unstable open-loop system which is made up of two interconnected transport equations and provided with anti-located boundary sensing and actuation. The main contribution is to show that the robustness of the control against delay uncertainties is recovered as soon as an arbitrary small diffusion is present in the system. Our analysis also reveals that the effect of diffusion on stability is far from being an obvious issue by exhibiting an alternative simple example where the presence of diffusion has a destabilizing effect instead.

This article studies the tracking control problem of networked multiagent systems under both multiple networks and event-triggered mechanisms. Multiple networks are to connect multiple agents and reference systems with decentralized controllers to guarantee their information transmission, whereas the event-triggered mechanisms are to reduce the information transmission via the networks. In this article, each agent has a network to communicate with its controller and reference system, and all networks are independent and asynchronous and have local event-triggered mechanisms, which are based on local measurements and determine whether the local measurements need to be transmitted via the corresponding network. To address this scenario, we first implement the emulation-based approach to develop a novel hybrid model for the tracking control of networked multiagent systems. Next, sufficient conditions are derived and decentralized event-triggered mechanisms are designed to guarantee the desired tracking performance. Furthermore, the proposed approach is applied to derive novel results for the eventtriggered observer design problem of networkedmultiagent systems. Finally, two numerical examples are presented to illustrate the validity of the developed results.

This paper studies the stabilization and safety problems of nonlinear time-delay systems. Following both Razumikhin and Krasovskii approaches, we propose novel control Lyapunov functions/functionals for the stabilization problem and novel control barrier functions/functionals for the safety problem. The proposed control Lyapunov and barrier functions/functionals extend the existing ones from the delay-free case to the time-delay case, and allow for designing the stabilizing and safety controllers in closed-form. Since analytical solutions to time-delay optimal control problems are hard to be achieved, a sliding mode control based approach is developed to merge the proposed control Lyapunov and barrier functions/functionals. Based on the sliding surface functional, a feedback control law is established to investigate the stabilization and safety objectives simultaneously. In particular, the properties of the sliding surface functional are analyzed, and further how to construct the sliding surface functional is discussed. Finally, the proposed approaches are illustrated via two numerical examples from the connected cruise control problem of automotive systems and the synchronization problem of multi-agent systems.

Electronic chip detection is widely used in electronic industries. However, most existing detection methods cannot handle chip images with multiple classes of chips or complex backgrounds, which are common in real applications. To address these problems, a novel chip detection method that combines attentional feature fusion (AFF) and cosine nonlocal attention (CNLA), is proposed, and it consists of three parts: a feature extraction module, a region proposal module, and a detection module. The feature extraction module combines an AFF-embedded CNLA module and a pyramid feature module to extract features from chip images. The detection module enhances feature maps with a region intermediate feature map by spatial attentional block, fuses multiple feature maps with a multiscale region of the fusion block of interest, and classifies and regresses objects in images with two branches of fully connected layers. Experimental results on a medium-scale dataset comprising 367 images show that our proposed method achieved mAP0.5 = 0.98745 and outperformed the benchmark method.

National Aeronautics and Space Administration (NASA)

Despite the successes of deep reinforcement learning (RL), it is still challenging to obtain safe policies. Formal verifi- cation approaches ensure safety at all times, but usually overly restrict the agent’s behaviors, since they assume adversarial behavior of the environment. Instead of assuming adversarial behavior, we suggest to focus on perceived safety instead, i.e., policies that avoid undesired behaviors while having a desired level of conservativeness. To obtain policies that are perceived as safe, we propose a shield synthesis framework with two distinct loops: (1) an inner loop that trains policies with a set of actions that is constrained by shields whose conservativeness is parameterized, and (2) an outer loop that presents example rollouts of the policy to humans and collects their feedback to update the parameters of the shields in the inner loop. We demonstrate our approach on a RL benchmark of Lunar landing and a scenario in which a mobile robot navigates around humans. For the latter, we conducted two user studies to obtain policies that were perceived as safe. Our results indicate that our framework converges to policies that are perceived as safe, is robust against noisy feedback, and can query feedback for multiple policies at the same time., QC 20211215

We explore the potential to control terrain vehicles using deep reinforcement in scenarios where human operators and traditional control methods are inadequate. This letter presents a controller that perceives, plans, and successfully controls a 16-tonne forestry vehicle with two frame articulation joints, six wheels, and their actively articulated suspensions to traverse rough terrain. The carefully shaped reward signal promotes safe, environmental, and efficient driving, which leads to the emergence of unprecedented driving skills. We test learned skills in a virtual environment, including terrains reconstructed from high-density laser scans of forest sites. The controller displays the ability to handle obstructing obstacles, slopes up to 27 degrees, and a variety of natural terrains, all with limited wheel slip, smooth, and upright traversal with intelligent use of the active suspensions. The results confirm that deep reinforcement learning has the potential to enhance control of vehicles with complex dynamics and high-dimensional observation data compared to human operators or traditional control methods, especially in rough terrain.

We explore the potential to control terrain vehicles using deep reinforcement in scenarios where human operators and traditional control methods are inadequate. This letter presents a controller that perceives, plans, and successfully controls a 16-tonne forestry vehicle with two frame articulation joints, six wheels, and their actively articulated suspensions to traverse rough terrain. The carefully shaped reward signal promotes safe, environmental, and efficient driving, which leads to the emergence of unprecedented driving skills. We test learned skills in a virtual environment, including terrains reconstructed from high-density laser scans of forest sites. The controller displays the ability to handle obstructing obstacles, slopes up to 27 degrees, and a variety of natural terrains, all with limited wheel slip, smooth, and upright traversal with intelligent use of the active suspensions. The results confirm that deep reinforcement learning has the potential to enhance control of vehicles with complex dynamics and high-dimensional observation data compared to human operators or traditional control methods, especially in rough terrain.

Despite the successes of deep reinforcement learning (RL), it is still challenging to obtain safe policies. Formal verifi- cation approaches ensure safety at all times, but usually overly restrict the agent’s behaviors, since they assume adversarial behavior of the environment. Instead of assuming adversarial behavior, we suggest to focus on perceived safety instead, i.e., policies that avoid undesired behaviors while having a desired level of conservativeness. To obtain policies that are perceived as safe, we propose a shield synthesis framework with two distinct loops: (1) an inner loop that trains policies with a set of actions that is constrained by shields whose conservativeness is parameterized, and (2) an outer loop that presents example rollouts of the policy to humans and collects their feedback to update the parameters of the shields in the inner loop. We demonstrate our approach on a RL benchmark of Lunar landing and a scenario in which a mobile robot navigates around humans. For the latter, we conducted two user studies to obtain policies that were perceived as safe. Our results indicate that our framework converges to policies that are perceived as safe, is robust against noisy feedback, and can query feedback for multiple policies at the same time., QC 20211215

End users who cannot afford to collect and label big data to train accurate deep learning (DL) models resort to Machine Learning as a Service (MLaaS) providers, who provide paid access to accurate DL models. However, the lack of transparency in how the providers' models make predictions causes a problem of trust. A way to increase trust (and also to align with ethical regulations) is for predictions to be accompanied by explanations locally and independently generated by the end users (rather than by explanations offered by the model providers). Explanation methods using internal components of DL models (a.k.a. model-specific explanations) are more accurate and effective than those relying solely on the inputs and outputs (a.k.a. model-agnostic explanations). However, end users lack white-box access to the internal components of the providers' models. To tackle this issue, we propose a novel approach allowing an end user to locally generate model-specific explanations for a DL classification model accessed via a provider's API. First, we approximate the provider's model with a local surrogate model. We then use the surrogate model's components to locally generate model-specific explanations that approximate the explanations obtainable with white-box access to the provider's DL model. Specifically, we leverage the surrogate model's gradients to generate adversarial examples that counterfactually explain why an input example is classified into a specific class. Our approach only requires the end user to have unlabeled data of size 0.5% of the provider's training data and with a similar distribution; given the small size and unlabeled nature of these data, they can be assumed to be already available to the end user or even to be supplied by the provider to build trust in h

Deep convolutional neural networks (DCNNs) have been successfully applied to plant disease detection. Unlike most existing studies, we propose feeding a DCNN CIE Lab instead of RGB color coordinates. We modified an Inception V3 architecture to include one branch specific for achromatic data (L channel) and another branch specific for chromatic data (AB channels). This modification takes advantage of the decoupling of chromatic and achromatic information. Besides, splitting branches reduces the number of trainable parameters and computation load by up to 50% of the original figures using modified layers. We achieved a state-of-the-art classification accuracy of 99.48% on the Plant Village dataset and 76.91% on the Cropped-PlantDoc dataset.

The use of unmanned aerial vehicles (UAVs, also known as drones or RPA) in archaeology has expanded significantly over the last twenty years. Improvements in terms of the reliability, size, and manageability of these aircraft have been largely complemented by the high resolution and spectral bands provided by the sensors of the different cameras that can be incorporated into their structure. If we add to this the functionalities and improvements that photogrammetry programs have been experiencing in recent years, we can conclude that there has been a qualitative leap in the possibilities, not only of geometric documentation and in the presentation of the archaeological data, but in the incorporation of non-intrusive high-resolution analytics. The work that we present here gives a sample of the possibilities of both geometric documentation, creation of 3D models, their subsequent printing with different materials, and techniques to finally show a series of analytics from images with NGB (Nir + Green + Blue), Red Edge, and Thermographic cameras applied to various archaeological sites in which our team has been working since 2013, such as Clunia (Peñalba de Castro, Burgos), Puig Rom (Roses), Vilanera (L’Escala, Girona), and Cosa (Ansedonia, Italy). All of them correspond to different chronological periods as well as to varied geographical and morphological environments, which will lead us to propose the search for adequate solutions for each of the environments. In the discussions, we will propose the lines of research to be followed in a project of these characteristics, as well as some results that can already be viewed.

This paper deals with the control development of a wind energy conversion system (WECS) interfaced to a utility grid by using a doubly fed induction generator (DFIG), a back-to-back (B2B) converter and an RL filter for optimal power extraction. The aim was to design a sensorless controller to improve the system reliability and to simultaneously achieve the regulation of the generator speed, reactive power and DC-link voltage. The proposed global control scheme combines: (i) a high-gain observer employed to estimate the generator speed and the mechanical torque, usually regarded as accessible, (ii) a sensorless MPPT block developed to provide optimal generator speed reference, which is designed on the basis of the mechanical observer and a polynomial wind-speed estimator and (iii) a finite-time controller (FTC) applied to the B2B converter to meet the output reference's tracking objectives in a short predefined finite time by using the backstepping and Lyapunov approaches. The proposed controller performance is formally analysed, and its capabilities are verified by numerical simulations using a 2 MW DFIG wind turbine (WT) under different operating conditions.

This letter introduces a complete DC-AC conversion system fed by photovoltaic (PV) energy. The system consists of N PV panels, N DC-DC boost converters, N cascaded H-bridge inverters, a DC-link composed of N capacitors and an LCL filter. This work aims at reaching threefold control objectives: i) Extracting the available maximum power by regulating the voltages across the PV panels, ii) Ensuring a unitary power factor, iii) Regulating the DC-link voltage to a desired reference. To achieve the mentioned objectives, a multi-loop regulator is designed. The PV panels are individually controlled to track the maximum power point in order to efficiently operate at either the same or different varying climatic conditions without failures. In addition to the maximum power point tracking (MPPT) controller, two cascaded loops guaranteeing a satisfactory power factor and DC-link voltage regulation are developed. The nonlinear backstepping approach combined with Lyapunov theory are used based on the averaged model for the synthesis of the multi-loop controller. The performance of the studied system is tested via MATLAB / SimPowerSystems environment. The obtained simulation results prove that the proposed controller meets its objectives and demonstrate the efficiency of the chosen control strategy under faulty conditions.

© 2020, Springer Science+Business Media, LLC, part of Springer Nature. In this paper, we study the creation of zeros in a certain type of families of functions. The families studied are given by the difference of two basic functions with a translation made in the argument of one of these functions. The problem is motivated by applications in the 16th Hilbert problem and its ramifications. Here, we apply the results obtained to the study of bifurcations of critical periods in the Loud family of quadratic centers.

The flexibility provided by FPGAs permits the implementation of several ADCs, each one configured with the required bit resolution and sampling frequency. The paper presents the design and implementation of scalable and parametrizable analog-to-digital converters (ADC), based on a successive approximation register (SAR), on FPGAs (field programmable gate arrays). Firstly, the work develops a systematic methodology for the implementation of a parametrizable SAR-based ADC from a set of building modules, such as the pulse-width modulator (PWM), external low-pass filter (LPF) and the analog comparator. The presented method allows choosing the LPF parameters for the required performance (resolution bits and sampling frequency) of a SAR-based ADC. Secondly, the paper also presents several optimizations on the PWM module to enhance the sampling frequency of implemented ADCs, and the method to choose the LPF parameters is adapted. The PWM and SAR logic are synthesizable and parametrizable, using a low number of resources, in order to be portable for low-cost FPGA families. The methodology and PWM optimizations are tested on a Zynq-7000 device from Xilinx; however, they can be adapted to any other FPGA.

HTE 3.0 aims to support clinicians in the detection of patients with dyslipidemia, especially patients with familial hypercholesterolemia (FH), and in the recommendation of personalized lipid-lowering treatments. The core of HTE 3.0 is a clinical decision support system in which several knowledge-based systems are serialized: patient detection, therapeutic target setting, personalized treatment assessment, and treatment combination and prioritization, according to different criteria. The experimental evaluation of HTE 3.0 shows that the use of HTE 3.0 would mean increasing the capacity to detect FH by 5.7 times compared with usual clinical practice. Regarding the lipid-lowering treatment, a comparison of 18 cases among seven lipidologists shows that the differences between treatments provided by HTE 3.0 and human lipidologists are smaller than the differences between human experts.