Output Information Research Articles

Sampled‐Data Based Second‐Order Discrete‐Time Sliding Mode Control via Output Feedback

ABSTRACTThis article proposes a second‐order discrete‐time sliding mode controller (DSMC) by only using sampled‐data information of the system output. On one hand, different from the continuous‐time sliding mode control, the finite‐time stability of the closed‐loop system under DSMC is usually not kept due to the limited sampling frequency. On the other hand, in the output feedback framework, the second‐order sliding mode observer cannot independently estimate the internal state of the system if there exist unknown mismatched term and unknown time‐varying coefficient of the control input. To solve the above problems, a joint design strategy of second‐order discrete‐time sliding mode observer and sliding mode controller based on the “non‐separation principle” is proposed. The stability and robustness of the closed‐loop system is analyzed based on the Lyapunov design method. Finally, simulation example verifies the correctness of the theory.

Distributed leader-following bipartite consensus for one-sided Lipschitz multi-agent systems via dual-terminal event-triggered mechanism

This article analyses leader-following bipartite consensus for one-sided Lipschitz multi-agent systems by dual-terminal event-triggered output feedback control approach. A distributed observer is designed to estimate unknown system states by employing relative output information at triggering time instants, and then an event-triggered output feedback controller is proposed. Dual-terminal dynamic event-triggered mechanisms are proposed in sensor–observer channel and controller–actuator channel, which can save communication resources to a great extent, and the Zeno behavior is ruled out. A new generalized one-sided Lipschitz condition is proposed to handle the nonlinear term and achieve bipartite consensus. Some stability conditions are presented to guarantee leader-following bipartite consensus. Finally, one-link robot manipulator systems are introduced to demonstrate the availability of the designed scheme. The results demonstrate that the agents of the robot manipulators can track the reference trajectories bi-directionally, and effectively reduce communication resources by 61.22% and 68.04% at the sensor–observer and controller–actuator channels, respectively.

Stubborn State and Disturbance Observer Co‐Design for Nonlinear Descriptor Systems With δQC$$ \delta \mathrm{QC} $$ via a Dynamic Event‐Triggered Mechanism

ABSTRACTThis article studies the state and disturbance simultaneous estimation problem for a class of nonlinear descriptor systems with using a dynamic event‐triggered mechanism. refers to incremental quadratic constraints, which can provide a unified description of many types of common nonlinear functions. To reduce the negative impact of measurement outliers on the identification estimation, a stubborn state and disturbance observer co‐design scheme is proposed for the first time by embedding dynamic saturation output estimation errors. Meanwhile, a dynamic event‐triggered mechanism is introduced to avoid the need for continuously available output information, which can reduce the pressure on communication resources. By constructing a Lyapunov function, existence conditions of the stubborn state and disturbance observer are obtained in the form of a convex optimization problem so that the error estimation maintains an acceptable estimation performance. Finally, simulation examples illustrate the universality and stubbornness of the proposed observer.

Robust finite-time output feedback stabilisation for a class of uncertain planar systems with asymmetric output constraints

This paper considers the finite-time output feedback stabilisation (FTOFS) problem for a class of uncertain planar systems with asymmetric output constraints. The presence of unknown control coefficients, unknown measurement sensitivity and asymmetric output constraints makes the system under consideration essentially different from those in the existing related works. By constructing a new barrier Lyapunov function and applying homogeneous system theory, a simple proportional-derivative like finite-time output feedback controller is explicitly constructed, which only utilises the output information. A rigorous proof demonstrates that not only the closed-loop system is finite time stable, but also the requirement of asymmetric output constraint can be achieved. The novelty of the control approach lies in that it is capable to handle the FTOFS problem for uncertain planar systems with or without asymmetric output constraints in a unified manner, without changing the control structure. The efficiency of the proposed theoretical results are confirmed by simulation results.

Cnn-assisted multi-hop graph attention network for hyperspectral image classification

ABSTRACT Recently, the convolutional neural network (CNN) has gained widespread adoption in the hyperspectral image (HSI) classification owing to its remarkable feature extraction capability. However, the fixed acceptance domain of CNN restricts it to Euclidean image data only, making it difficult to capture complex information in hyperspectral data. To overcome this problem, much attention has been paid to the graph attention network (GAT), which can effectively model graph structure and capture complex dependencies between nodes. However, GAT usually acts on superpixel nodes, which may lead to the loss of pixel-level information. To better integrate the advantages of both, we propose a CNN-assisted multi-hop graph attention network (CMGAT) for HSI classification. Specifically, a parallel dual-branch architecture is first constructed to simultaneously capture spectral-spatial features from hyperspectral data at the superpixel and pixel levels using GAT and CNN, respectively. On this basis, the multi-hop and multi-scale mechanisms are further employed to construct a multi-hop GAT module and a multi-scale CNN module to capture diverse feature information. Secondly, an attention module is cascaded before the multi-scale CNN module to improve classification performance. Eventually, the output information from the two branches is weighted and fused to produce the classification result. We performed experiments on four benchmark HSI datasets, including Indian Pines (IP), University of Pavia (UP), Salinas Valley (SV) and WHU-Hi-LongKou (LK). The results demonstrate that the proposed method outperforms several deep learning methods, achieving overall accuracies of 95.67%, 99.04%, 99.55% and 99.51%, respectively, even with fewer training samples.

A Knowledge Distillation Method Based on Evidence Theory to Prevent Catastrophic Forgetting

Abstract Incremental learning is an emerging machine-learning approach designed to prevent catastrophic forgetting while learning a new task with knowledge retained from previous tasks. However, existing methods often must fully utilize the distributional information of the old task model’s outputs. To alleviate this, we propose a class incremental learning method grounded in evidence theory to leverage the distributional information of outputs from the old task model, which integrates uncertainty estimation into the knowledge distillation process to ensure the new task model effectively learns from the old task model’s distribution information. Specifically, we incorporate uncertainty estimation based on evidence theory to calculate knowledge distillation loss during training. We compute the uncertainty estimates of outputs from both the new and old task models, and then they are used in the knowledge distillation loss calculation. We propose a novel classification strategy that considers outputs’ probability and uncertainty estimates, which could determine the sample category without requiring an additional training phase or supplementary models. Experiments on the CIFAR-100 and ImageNet datasets demonstrate the effectiveness of the proposed method, showing that it outperforms existing methods by improving accuracy by 1%-2%. Results suggest that leveraging the distribution information of model outputs can effectively mitigate catastrophic forgetting in deep learning models within incremental learning scenarios.

Event-Triggered Model-Free Adaptive Control for Nonlinear Multiagent Systems Under Jamming Attacks.

This article addresses the security problem of tracking control for nonlinear multiagent systems against jamming attacks. It is assumed that the communication networks among agents are unreliable due to the existence of jamming attacks, and a Stackelberg game is introduced to depict the interaction process between multiagent systems and malicious jammer. First, the dynamic linearization model of the system is established by applying a pseudo-partial derivative method. Then, a novel model-free security adaptive control strategy is proposed, so that the multiagent systems can achieve bounded tracking control in the mathematical expectation sense in spite of jamming attacks. Furthermore, a fixed threshold event-triggered scheme is utilized to reduce communication cost. It is worth noting that the proposed methods only require the input and output information of the agents. Finally, the validity of the proposed methods is illustrated through two simulation examples.

Data is the fuel of generative artificial intelligence

TThe aim of the article is to present the generative technology of artificial intelligence and a general analysis of its achievements and technological limitations. In turn, the hypothesis is as follows: data is the main factor ensuring the development of generative AI. The identified research niche includes stimulants of this development, the need to clarify the concept of generative AI and the indication of the most frequently used sources and methods of obtaining data. At the beginning, the concept and genesis of artificial intelligence were explained, which was necessary to move on to further considerations regarding the analyzed generative artificial intelligence. The self-learning mechanisms used by this type of artificial intelligence are discussed, consisting mainly in the analysis of specific data sets, thanks to which computers learn things that have not been known before. The advantages of generative artificial intelligence were also indicated, including increased productivity, greater efficiency and high creativity. Part of the attention was also paid to the limitations of this technology, including inaccurate and fuzzy output information, uncritical analyzes and often biased conclusions, copyright infringement and high requirements for memory and computing power of computers. Two factors have been identified that have recently stimulated the development of generative intelligence, i.e. the growing computing power of computers and easy access to large data sets. The main subject of considerations are large data sets, dynamically and uncontrollably self-replicating in Internet servers. Today, data resources are collected very efficiently, including: in Big Data, Cloud Computing and Internet of Things technologies, and their use is almost cost-free and publicly available. It was enough to improve the technology of artificial neural networks, expand multi-layer machine learning and create deep learning for generative intelligence to gain enormous momentum and become a symbol of the beginning of the 21st century. The whole investigation was summarized with a synthetic ending.

An Automatic Vehicle Navigation System Based on Filters Integrating Inertial Navigation and Global Positioning Systems

The key technologies for advanced autonomous vehicles include those relating to perception, decision making, and execution. Path-tracking control in autonomous vehicles is heavily dependent on their positioning system. Therefore, the development of low-cost and reliable positioning systems is crucial to improving perception and decision-making technologies for autonomous vehicles. Although the accuracy of the global positioning system (GPS) is extremely high, it is vulnerable to interference. Further, despite the low positioning accuracy of inertial navigation systems (INSs), their robustness is notably high. Therefore, an integrated navigation information method based on the Adaptive Particle Filter and the Iterative Kalman Filter (APF-IKF) was developed in this study. Firstly, an integrated navigation system model was established. Then, the IKF was adopted to estimate the speed, latitude and longitude errors of the INS. Thirdly, the newest estimated error results were introduced into the APF to optimize the distribution function, and the particle quality was improved. In this process, the APF can filter non-Gaussian noise, preliminarily estimate the error, optimize the result with the IKF and correct the output information of the INS with the final estimated error. Finally, by using differential GPS positioning as the benchmark, we built a real-vehicle test platform with a low-cost and low-precision GPS and inertial units and carried out a series of real-vehicle tests. The experimental results show that compared with the traditional KF method, APF-IKF can significantly improve the positioning accuracy and robustness of the system.

Simultaneous state and fault estimation for a class of nonlinear Markov jump systems by distributed fault-tolerant observers

This paper investigates the estimation problem of a class of nonlinear Markov jump systems with actuator and sensor faults. The main goal is to design a distributed fault-tolerant observer to estimate system states and actuator faults simultaneously. Firstly, a novel distributed observer network is constructed to compensate the missing information of unobservable nodes of the system. Next, a class of new redundant sensors are set up on each distributed observer node to obtain more output measurement samples. More importantly, when some of the mentioned sensors occur faults, an index of sensor heath level is constructed to characterize the quality of the faulty sensor information. Further, a novel algorithm is designed to mask low quality output information and filter out relatively healthy one automatically. Based on the selected healthy output information, the system state and actuator fault are estimated in the case of sensor failure. Finally, an example is provided to demonstrate the effectiveness of the proposed method.

Magnetostrictive bi-perceptive flexible sensor for tracking bend and position of human and robot hand

The sensor that simultaneously perceives bending strain and magnetic field has the potential to detect the finger bending state and hand position of the human and robot. Based on unique magneto-mechanical coupling effect of magnetostrictive materials, the proposed a bi-perceptive flexible sensor, consisting of the Co–Fe film and magnetic sensing plane coils, can realize dual information perception of strain/magnetic field through the change of magnetization state. The sensor structure and interface circuit of the sensing system are designed to provide high sensitivity and fast response, based on the input–output characteristics of the simulation model. An asynchronous multi-task deep learning method is proposed, which takes the output of the position task as the partial input of the bending state task to analyze the output information of the sensor quickly and accurately. The sensing system, integrating with the proposed model, can better predict the bending state and approach distance of human or robot hand.

Exponential Boundary Control for 2-D Spatial Distributed Parameter Systems Under Boundary Collocated and Planar Linear Measurements.

For a class of 2-D spatial distributed parameter systems (DPSs) with space-dependent diffusivity, this article aims to achieve exponential realization of their desired profiles. To reduce the number of required sensors and actuators, a planar output feedback boundary control strategy is proposed with combining two nonfull-domain measurement methods, boundary collocated measurement and planar linear measurement, in which only two boundaries of the considered 2-D spatial DPSs are controlled and a little output information is measured. Moreover, by employing the Poincaré-Wirtinger inequality and variable substitution dexterously, the final exponential convergence criteria of the error system can be obtained with method of "Diverse treatment for same term." Finally, we provide a general numerical example and an application example in 2-D heat conduction systems to illustrate the effectiveness and practicability of the proposed measurement and control schemes.

Rainfall modeling with CMIP6-DCPP outputs and local characteristic information using eigenvector spatial filtering varying coefficient (ESF-VC)

Estimating rainfall at a point or region is difficult because complex factors affect rainfall. A helpful strategy is to utilize the GCM output information from CMIP6-DCPP by forming a functional relationship between GCM output data and rainfall data at a certain point or region, called statistical downscaling. However, because the resolution of the GCM output is relatively low, the model could not explain the local effects since the heterogeneity is enormous. Based on this fact, the current research proposes to add some local characteristics in the downscaling model to improve the performance to predict the rainfall levels. Further, the rainfall levels have spatial dependencies among points. Therefore, this research employed the Eigenvector Spatial Filtering-Varying Coefficient (ESF-VC) as the methodology of the modeling. The objective of this research is to perform rainfall predictive modeling with CMIP6-DCPP output and some local characteristic information as predictors using ESF-VC methodology. The approach was implemented to predict the rainfall level in the Province of Riau in Indonesia. Based on the results, the ESF-VC model provides good performance in estimating rainfall in Riau. The variables that provide local effects are altitude, equator (location), equator (distance), and wet month dummy. While the variables ENSO and vegetation (NDVI) have a significant global effect on the model.

Sustainable Resource Allocation and Base Station Optimization Using Hybrid Deep Learning Models in 6G Wireless Networks

Researchers are currently exploring the anticipated sixth-generation (6G) wireless communication network, poised to deliver minimal latency, reduced power consumption, extensive coverage, high-level security, cost-effectiveness, and sustainability. Quality of Service (QoS) improvements can be attained through effective resource management facilitated by Artificial Intelligence (AI) and Machine Learning (ML) techniques. This paper proposes two models for enhancing QoS through efficient and sustainable resource allocation and optimization of base stations. The first model, a Hybrid Quantum Deep Learning approach, incorporates Recurrent Neural Networks (RNNs) and Convolutional Neural Networks (CNNs). CNNs handle resource allocation, network reconfiguration, and slice aggregation tasks, while RNNs are employed for functions like load balancing and error detection. The second model introduces a novel neural network named the Base Station Optimizer net. This network includes various parameters as input and output information about the condition of the base station within the network. Node coverage, number of users, node count and user locations, operating frequency, etc., are different parametric inputs considered for evaluation, providing a binary decision (ON or SLEEP) for each base station. A dynamic allocation strategy aims for network lifetime maximization, ensuring sustainable operations and power consumption are minimized across the network by 2 dB. The QoS performance of the Hybrid Quantum Deep Learning model is evaluated for many devices based on slice characteristics and congestion scenarios to attain an impressive overall accuracy of 98%.

Zero-Biased Bionic Fingertip E-Skin with Multimodal Tactile Perception and Artificial Intelligence for Augmented Touch Awareness.

Electronic skins (E-Skins) are crucial for future robotics and wearable devices to interact with and perceive the real world. Prior research faces challenges in achieving comprehensive tactile perception and versatile functionality while keeping system simplicity for lack of multimodal sensing capability in a single sensor. Two kinds of tactile sensors, transient voltage artificial neuron (TVAN) and sustained potential artificial neuron (SPAN), featuring self-generated zero-biased signals are developed to realize synergistic sensing of multimodal information (vibration, material, texture, pressure, and temperature) in a single device instead of complex sensor arrays. Simultaneously, machine learning with feature fusion is applied to fully decode their output information and compensate for the inevitable instability of applied force, speed, etc, in real applications. Integrating TVAN and SPAN, the formed E-Skin achieves holistic touch awareness in only a single unit. It can thoroughly perceive an object through a simple touch without strictly controlled testing conditions, realize the capability to discern surface roughness from 0.8 to 1600µm, hardness from 6HA to 85HD, and correctly distinguish 16 objects with temperature variance from 0 to 80°C. The E-skin also features a simple and scalable fabrication process, which can be integrated into various devices for broad applications.

Automatically Extracting and Utilizing EEG Channel Importance Based on Graph Convolutional Network for Emotion Recognition.

Graph convolutional network (GCN) based on the brain network has been widely used for EEG emotion recognition. However, most studies train their models directly without considering network dimensionality reduction beforehand. In fact, some nodes and edges are invalid information or even interference information for the current task. It is necessary to reduce the network dimension and extract the core network. To address the problem of extracting and utilizing the core network, a core network extraction model (CWGCN) based on channel weighting and graph convolutional network and a graph convolutional network model (CCSR-GCN) based on channel convolution and style-based recalibration for emotion recognition have been proposed. The CWGCN model automatically extracts the core network and the channel importance parameter in a data-driven manner. The CCSR-GCN model innovatively uses the output information of the CWGCN model to identify the emotion state. The experimental results on SEED show that: 1) the core network extraction can help improve the performance of the GCN model; 2) the models of CWGCN and CCSR-GCN achieve better results than the currently popular methods. The idea and its implementation in this paper provide a novel and successful perspective for the application of GCN in brain network analysis of other specific tasks.

Fault‐tolerant control of unmanned marine vehicles with unknown parametric dynamics via integral sliding mode output feedback control

AbstractThis paper focuses on fault‐tolerant control (FTC) for unmanned marine vehicles (UMVs) subject to unknown parametric dynamics, thruster faults, and external disturbances. A novel FTC strategy based on the integral sliding mode output feedback method is proposed. Based on a high gain compensator and output information, a novel integral sliding mode fault‐tolerant controller is constructed to guarantee the dynamic positioning (DP) of UMVs. An attraction region that is related to the known positive constants with respect to dynamic uncertainties has been revealed for the first time. Finally, the closed‐loop stability can be guaranteed from the every initial time in despite of unknown parametric dynamics, thruster faults, and external disturbances. Simulation though a typical floating production ship model has verified the effectiveness of the proposed method.

Observer-based adaptive memory event-triggered consensus tracking control for high-speed train under DoS attacks

This paper focuses on the event-triggered consensus tracking of high-speed trains under denial-of-service (DoS) attacks. A novel anti-disturbance control strategy is proposed, based on an adaptive memory state observer and a disturbance observer. A memory is introduced to buffer system output information which enhancing the precision of the state observer, and the gust disturbance during the high-speed train operation is estimated by the disturbance observer. The Linear Matrix inequality technique is used to obtain the observer feedback coefficient and the controller gain, and the Lyapunov theory is employed to demonstrate the controller’s consistency tracking performance. Applying the memory to the event-triggered mechanism, a new adaptive memory event-triggered scheme is designed to better ensure system performance and effectively prevent the occurrence of Zeno behavior. To achieve precise identification of DoS attacks in the event-triggered environment, a new detection algorithm is proposed. Finally, simulation examples are used to validate the effectiveness and practicality of the proposed scheme.

Assessing fidelity in XAI post-hoc techniques: A comparative study with ground truth explanations datasets

The evaluation of the fidelity of eXplainable Artificial Intelligence (XAI) methods to their underlying models is a challenging task, primarily due to the absence of a ground truth for explanations. However, assessing fidelity is a necessary step for ensuring a correct XAI methodology. In this study, we conduct a fair and objective comparison of the current state-of-the-art XAI methods by introducing three novel image datasets with reliable ground truth for explanations. The primary objective of this comparison is to identify methods with low fidelity and eliminate them from further research, thereby promoting the development of more trustworthy and effective XAI techniques. Our results demonstrate that XAI methods based on the direct gradient calculation and the backpropagation of output information to input yield higher accuracy and reliability compared to methods relying on perturbation based or Class Activation Maps (CAM). However, these methods tend to generate more noisy saliency maps. These findings have significant implications for the advancement of XAI methods, enabling the elimination of erroneous explanations and fostering the development of more robust and reliable XAI.

Observer-based bipartite output containment control of heterogeneous multi-agent systems over structurally unbalanced signed graphs

In this paper, we propose state feedback and output feedback observer-based protocols that solve the bipartite output containment (BOC) problem for heterogeneous multi-agent systems over structurally unbalanced signed graphs. The proposed protocols rely solely on the exchange of output information among neighboring agents, resulting in a significant reduction or complete elimination of communication costs. Two design methods for determining controller parameters based on algebraic Riccati equations and linear matrix inequalities (LMIs) are proposed. Each method offers distinct advantages while ensuring that the closed-loop system poles are placed within a predefined region in the complex plane. Furthermore, an extensive solvability analysis is conducted, demonstrating that the solvability of the BOC problem using the proposed approach can be pre-guaranteed for a broad range of agent dynamics.