Time-varying Neural Networks Research Articles

Neural Network-Based Robust Guaranteed Cost Control for Image-Based Visual Servoing of Quadrotor.

In this article, a neural network (NN)-based robust guaranteed cost control design is proposed for image-based visual servoing (IBVS) control of quadrotors. According to the dynamics of three subsystems (yaw, height, and lateral subsystems) derived from the quadrotor IBVS dynamic model, the main control design is to solve the robust control problem for the time-varying lateral subsystem with angle constraints and uncertain disturbances. Considering the system dynamics, a two-loop structure is conducted. The outer loop uses the linear quadratic regulator to solve the Riccati equation for the lateral image feature system, and the inner loop adopts the optimal robust guaranteed cost control to solve the lateral velocity system. For the lateral velocity system, the optimal robust control problem is transformed to solve the modified Hamilton-Jacobi-Bellman equation of the corresponding optimal control problem utilizing adaptive dynamic programming. The implementation is accomplished with the time-varying NN and the designed estimated weight update law. In addition, the stability and effectiveness are proved by the theoretic proof and simulations.

An adaptive time-varying neural network for solving K optimal time-varying destroy locations query problem

The time-varying and destroying are two real existing factors that are generally ignored in the existing k optimal locations query study, which makes the existing study unapplicable to real-world environments. This study presents a novel adaptive time-varying neural network (ATNN) to solve the k optimal locations query problem on time-varying destroy networks. ATNN is a structure-adaptive neural network that is composed of newly designed adaptive time-varying neurons; it does not require training and can automatically construct or adjust a topological structure according to different topologies and scales of time-varying and destroying networks. Adaptive time-varying neurons consist of five layers: wave receiving layer, state verification layer, state storage layer, wave generation layer, and wave sending layer, all those five layers implements the information exchange and processing between neurons as well as the representation of the time-varying and destroying properties of the network. The proposed algorithm has been theoretically proven through time-complexity analysis, correctness analysis, and a numerical example, and its performance has been further confirmed through experiments on public road network datasets.

Adversarial self-attentive time-variant neural networks for multi-step time series forecasting

Adversarial self-attentive time-variant neural networks for multi-step time series forecasting

Multi-scale adaptive attention-based time-variant neural networks for multi-step time series forecasting

Multi-scale adaptive attention-based time-variant neural networks for multi-step time series forecasting

Chance-Constrained State Estimation for Recursive Neural Networks Under Deception Attacks and Energy Constraints: The Finite-Horizon Case.

In this article, the chance-constrained state estimation problem is investigated for a class of time-varying neural networks subject to measurements degradation and randomly occurring deception attacks. A novel energy-constrained deception attack model is proposed, in which both the occurrence of the attack and the selection of released faked packet are random and the energy of the deception attack is introduced, calculated, and analyzed quantitatively. The main purpose of the addressed problem is to design an estimator such that the prefixed probabilistic constraints of the system error dynamics are satisfied and the performance is also ensured. Subsequently, the explicit expression of the estimator gains is derived by solving a minimization problem subjected to certain recursive inequality constraints. Finally, a numerical example and a practical three-tank system are utilized to demonstrate the correctness and effectiveness of the proposed estimation scheme.

Partial-Neurons-Based $$H_{\infty }$$ State Estimation for Time-Varying Neural Networks Subject to Randomly Occurring Time Delays under Variance Constraint

Partial-Neurons-Based $$H_{\infty }$$ State Estimation for Time-Varying Neural Networks Subject to Randomly Occurring Time Delays under Variance Constraint

Variance-Constrained Resilient $$H_{\infty }$$ State Estimation for Time-Varying Neural Networks with Random Saturation Observation Under Uncertain Occurrence Probability

Variance-Constrained Resilient $$H_{\infty }$$ State Estimation for Time-Varying Neural Networks with Random Saturation Observation Under Uncertain Occurrence Probability

Finite-Time Stability of Nonlinear Impulsive Systems With Applications to Neural Networks.

This article studies the problem of finite-time stability (FTS) and finite-time contractive stability (FTCS) for nonlinear impulsive systems, where the possibility of time delay in impulses is fully considered. Some sufficient conditions for FTS/FTCS are constructed in the framework of Lyapunov function methods. A relationship between impulsive frequency and the time delay existing in impulses is established to reveal FTS/FTCS performance. As an application, we apply the theoretical results to finite-time state estimation of neural networks, including time-varying neural networks and switched neural networks. Finally, two illustrated examples are given to show the effectiveness and distinctiveness of the proposed delay-dependent impulsive schemes.

Asymptotical Stability and Exponential Stability in Mean Square of Impulsive Stochastic Time-Varying Neural Network

The effect of impulse on stability of neural network is evident not only in performance, that is, impulsive control and impulsive interference. The amount of impulse has a certain impact on stability of neural network. Unlike traditional average impulsive interval, a new strategy is applied in this paper, namely, impulsive density. Based on this strategy, by constructing Lyapunov function, we establish sufficient conditions for mean square asymptotical stability of impulsive stochastic time-varying neural network without time delay. As well as, under this strategy and uniformly asymptotically stable function, by combining trajectory based approach and improved Razumikhin method, mean square exponential stability criterion of impulsive stochastic time-varying neural network with time delay is established. Finally, to demonstrate the viability of our theoretical findings, some instances are provided.

Outer-synchronization criterions for asymmetric recurrent time-varying neural networks described by differential-algebraic system via data-sampling principles.

Asymmetric recurrent time-varying neural networks (ARTNNs) can enable realistic brain-like models to help scholars explore the mechanisms of the human brain and thus realize the applications of artificial intelligence, whose dynamical behaviors such as synchronization has attracted extensive research interest due to its superior applicability and flexibility. In this paper, we examined the outer-synchronization of ARTNNs, which are described by the differential-algebraic system (DAS). By designing appropriate centralized and decentralized data-sampling approaches which fully account for information gathering at the times t k and . Using the characteristics of integral inequalities and the theory of differential equations, several novel suitable outer-synchronization conditions were established. Those conditions facilitate the analysis and applications of dynamical behaviors of ARTNNs. The superiority of the theoretical results was then demonstrated by using a numerical example.

Self-attention-based time-variant neural networks for multi-step time series forecasting

Self-attention-based time-variant neural networks for multi-step time series forecasting

Variance-constrained robust $ H_{\infty} $ state estimation for discrete time-varying uncertain neural networks with uniform quantization

<abstract><p>In this paper, we consider the robust $ H_{\infty} $ state estimation (SE) problem for a class of discrete time-varying uncertain neural networks (DTVUNNs) with uniform quantization and time-delay under variance constraints. In order to reflect the actual situation for the dynamic system, the constant time-delay is considered. In addition, the measurement output is first quantized by a uniform quantizer and then transmitted through a communication channel. The main purpose is to design a time-varying finite-horizon state estimator such that, for both the uniform quantization and time-delay, some sufficient criteria are obtained for the estimation error (EE) system to satisfy the error variance boundedness and the $ H_{\infty} $ performance constraint. With the help of stochastic analysis technique, a new $ H_{\infty} $ SE algorithm without resorting the augmentation method is proposed for DTVUNNs with uniform quantization. Finally, a simulation example is given to illustrate the feasibility and validity of the proposed variance-constrained robust $ H_{\infty} $ SE method.</p></abstract>

Video Frame Interpolation With Learnable Uncertainty and Decomposition

Video frame interpolation can flexibly increase the temporal resolution of low frame-rate videos by generating the missing intermediate frames at any time. Existing methods generally train a renderer to predict rgb frames based on estimated clues. It tends to generate blurry outputs with unpleasing artifacts due to false clues and the neural network's bias in favor of low-frequency information. To address this problem, we propose a novel two-stage supervised framework. The inaccuracy of clues is modeled as uncertainty which can be estimated by training implicitly with a parameterized loss function in stage one. The bias is alleviated in stage two by regressing a lossless decomposition of frames instead of the raw rgbs. The decomposition can be achieved by several invertible cross-coupling layers, motivating the network to synthesize high-frequency details. Moreover, the proposed framework is equipped with a time-varying neural network that is adaptive to the timestamp of any intermediate frame, bringing benefits to multiple-frame interpolation. Both qualitative and quantitative experiments demonstrate the superiority of our proposed approach.

A wave time-varying neural network for solving the time-varying shortest path problem

A wave time-varying neural network for solving the time-varying shortest path problem

A different approach to fixed-time stability for a wide class of time-varying neural networks

We present sufficient conditions for fixed-time stability for a wide class of neural networks described by a system of differential equations with right-hand side satisfying the Carathéodory conditions. In contrast to the results given in the literature, where the settling-time function is estimated by an unknown Lyapunov function, we estimate the settling-time by a known function. In addition, the settling-time function does not depend on the initial values. We also give numerical examples, which confirm the theoretical results.

A time-varying neural network for solving minimum spanning tree problem on time-varying network

In this study, we propose a time-varying neural network (TVNN) for solving the time-varying minimum spanning tree problem with constraints (CTMST), which is a variant of the time-varying network minimum spanning tree problem (TNMSTP), a well-known NP-hard problem. Unlike traditional algorithms that use heuristic search, the proposed TVNN is based on time-varying neurons and can achieve parallel computing without any training requirements. Time-varying neurons are novel computational neurons designed in this work. They consist of six parts: input, wave receiver, neuron state, wave generator, wave sender, and output. The parallel computing strategy and self-feedback mechanism of the proposed algorithm greatly improve the response speed and solution accuracy on large-scale time-varying networks. The analysis of time complexity and experimental results on the New York City dataset show that the performance of the proposed algorithm is significantly improved compared with the existing methods.

Synchronization of recurrent neural networks with unbounded delays and time-varying coefficients via generalized differential inequalities

In this paper, we revisit the drive-response synchronization of a class of recurrent neural networks with unbounded delays and time-varying coefficients, contrary to usual in the literature about time-varying neural networks, the signs of self-feedback coefficients are permitted to be indefinite or the time-varying coefficients can be unbounded. A generalized scalar delay differential inequality considering indefinite self-feedback coefficient and unbounded delay simultaneously is established, which covers the existing result with bounded delay, the applicabilities of the sufficient conditions are discussed. Some novel criteria for network synchronization are then derived by constructing different candidate functions. These results have been improved in some aspects compared with the existing ones. Differential inequality in vector form is also derived to obtain a more refined synchronization criterion which removes some strong assumptions. Three examples are presented to verify the effectiveness and show the superiorities of our theoretical results.

Adoption of bio-image technology on rehabilitation intervention of sports injury of golf

This work aimed to explore the characteristics of surface electromyography (EMG) signal of golfers’ low back pain and the effect of rehabilitation. Based on the time-varying parameter autoregressive model and artificial neural network, ARAN algorithm was constructed, which was compared with the autoregressive moving average (ARMA) algorithm and the convolutional neural network (CNN) algorithm. Then, the established ARAN algorithm was employed to evaluate the characteristics of surface EMG signal of 106 golfers with low back pain. It was found that the accuracy, sensitivity, and specificity of the ARAN algorithm were superior to those of the CNN and ARMA algorithms. The golfer’s Roland-Morris Disability Questionnaire (RMDQ) score after treatment was less than that before treatment (P < 0.05). Moreover, there was significant negative correlation between RMDQ score and the mean values of time-varying parameters a1 and a3 (P < 0.05). The RMDQ score had a very obvious positive correlation with the mean values of a2, a4, and a6 (P < 0.001) and had a negative correlation with the mean value of a5 (P < 0.05). To sum up, the time-varying parameters of the surface EMG signal can effectively evaluate the golfer’s low back pain and the effect of treatment and rehabilitation.

Event-triggered H∞ state estimation for time-varying neural networks with variance-constraint and fading measurements

This paper addresses the event-triggered H∞ state estimation problem for a class of discrete recurrent neural networks subject to variance-constraint and fading measurements. The phenomena of fading measurements are described by introducing a set of mutually independent random variables, which reflect that each sensor has individual missing probability. In addition, for the purpose of energy saving, an event-triggered H∞ state estimation scheme is used for time-varying neural networks to determine whether the measurement output is transmitted to the estimator or not. Some sufficient conditions are obtained to guarantee that the estimation error system satisfies both estimation error variance constraint and prescribed H∞ performance requirement. Finally, the feasibility of the proposed event-triggered H∞ state estimation method is verified by a numerical example.

Global dissipativity and exponential synchronization of mixed time-varying delays neural networks with discontinuous activations

Abstract In this paper, the matters of dissipativity and synchronization for non-autonomous Hopfield neural networks with discontinuous activations are investigated. Firstly, under the framework of extending Filippov differential inclusion theory, several effective new criteria are derived. The global dissipativity of Filippov solution to neural networks is proved by using generalized Halanay inequality and matrix measure method. Secondly, the global exponential synchronization of the addressed network drive system and the response system is realized by utilizing inequality and some analysis techniques and designing the discontinuous state feedback controller. Finally, several numerical examples are given to verify the validity of the theoretical results.