Showing papers in "Journal of The Franklin Institute-engineering and Applied Mathematics in 2023"

The data-filtering based bias compensation recursive least squares identification for multi-input single-output systems with colored noises

Abstract: For the multi-input single-output (MISO) system corrupted by colored noise, we transform the original system model into a new MISO output error model with white noise through data filtering technology. Based on the newly obtained model and the bias compensation principle, a novel data filtering-based bias compensation recursive least squares (BCRLS) identification algorithm is developed for identifying the parameters of the MISO system with colored noise disturbance. Unlike the exiting BCRLS method for the MISO system (see, in Section 3), without computing the complicated noise correlation functions, still the proposed method can achieve the unbiased parameters estimation of the MISO system in the case of colored process noises. The proposed algorithm simplifies the implementation of and further expands the application scope of the existing BCRLS method. Three numerical examples clearly illustrate the validity of and the good performances of the proposed method, including its superiority over the BCRLS method and so on.

Event-Triggered formation control for time-delayed discrete-Time multi-Agent system applied to multi-UAV formation flying

Abstract: This paper studies the formation control for a time-delayed discrete-time multi-agent system (MAS). An event-triggered controller is proposed to reduce the communication load of the system. Based on the designed event-triggered condition and properties of Schur stable matrix, the stability of formation for discrete-time MAS is proved. Utilizing the virtual simulation platform integrated Robot Operating System (ROS) and Gazebo, a virtual scene with unmanned aerial vehicles (UAVs) models is built and the verification for the theoretical algorithm is completed. Finally, an experimental platform with four practical UAVs is constructed and the result shows that the expected formation is achieved and controller proposed can solve the formation control problem for time-delayed discrete-time MASs. Besides, the effectiveness of the event-triggered mechanism on reducing communication frequency is comfirmed in practical scenarios.

Quasi-projective synchronization of discrete-time fractional-order quaternion-valued neural networks

Abstract: In this article, without decomposing the quaternion-valued neural networks (QVNNs) into two complex-valued subsystems or four real-valued subsystems, quasi-projective synchronization of discrete-time fractional-order QVNNs is investigated. To this end, the sign function for quaternion number is introduced and some related properties are given. Then, two inequalities are built according to the nabla fractional difference and quaternion theory. Subsequently, a simple linear quaternion-valued controller is designed, and some synchronization conditions are given by means of our created inequalities. Finally, numerical simulations are given to prove the feasibility and correctness of the theoretical results.

General proportional integral observer (GPIO) - based disturbance compensation for minimum variance time synchronization

Abstract: Precise time synchronization is an enabling technology for mission-critical time-sensitive Industrial Internet of Things (IIoT). However, the crystal oscillator clock which is widely used in IIoT may suffer from periodic disturbances caused by repetitive motion or periodic vibration. To improve the time synchronization of distributed nodes subject to periodic disturbances, this paper proposes a novel disturbance rejection framework, General-Proportional-Integral-Observer-based Disturbance Compensation (GPIO-DC), with the proof of stability, and combined with a 2-freedom control design strategy to optimize both the disturbance rejection and clock tracking performance. And the GPIO’s unique feature of blocking zeros are fully exploited to reject the periodic disturbance at its frequencies and a zero-pole optimal design algorithm is given. With the disturbance being compensated, a disturbance-free minimum variance time synchronization protocol for a complex network is developed and optimized by using Linear Matrix Inequality (LMI) to minimize the variance of networked synchronization errors. The performance of the proposed method is devalued by intensive simulation. Comparing with recent relevant research, the proposed method achieves a better performance in disturbance rejection and minimum variance.

Cluster synchronization for directed coupled inertial reaction-diffusion neural networks with nonidentical nodes via non-reduced order method

Abstract: The cluster synchronization issues are investigated for directed coupled inertial reaction-diffusion neural networks (CIRDNNs) with nonidentical nodes by imposing two effective pinning control. A novel Lyapunov-Krasovskii functional (LKF) is established to directly analyze the dynamic behavior of CIRDNNs and deal with reaction-diffusion term, inertia term and coupling term. Moreover, based on different desired cluster synchronization states including a set of un-decoupled trajectories and the particular solutions of the decoupled node systems, two class of synchronization criteria in view of algebraic inequalities are derived under two different communication topologies, respectively. Finally, two typical examples are given to verify the theoretical results.

Leader-following consensus of multi-agent systems with connectivity-mixed attacks and actuator/sensor faults

Abstract: This study investigates the leader-following consensus issue of multi-agent systems subject to simultaneous connectivity-mixed attacks, actuator/sensor faults and disturbances. Connectivity-mixed attacks are remodeled into connectivity-maintained and connectivity-paralyzed topologies in a switched version, and actuator/sensor faults are established with unified incipient-type and abrupt-type characteristics. Then, unknown input observer-based decoupling and estimation are incorporated to achieve unknown state and fault observations with the normalized technique, and the leader-following consensus-based compensation to faults, resilience to attacks and robustness to disturbances are also realized with the neighboring output information and sensor fault estimation through the distributed framework. Criteria of achieving exponential leader-following consensus of multi-agent systems under cyber-physical threats are derived with dual attack frequency and activation rate indicators. Simulation example is conducted to exemplify the validation and merits of the proposed leader-following consensus algorithm.

Three-dimensional event-triggered fixed-time cooperative guidance law against maneuvering target with the constraint of relative impact angles

Abstract: In order to improve the flexibility and reduce the energy consumption of cooperative guidance laws considering the impact angle constraint, this paper proposes a three-dimensional event-triggered fixed-time cooperative guidance law with the constraint of relative impact angles. First, for the purpose of avoiding the precision degradation due to the estimation error of time-to-go especially facing a maneuvering target, the range-to-go and velocity along the line-of-sight (LOS) are taken as the coordination variables for achieving time-cooperative guidance. Secondly, instead of assigning specific desired impact angles for each missile, only the consensus errors of relative impact angles are utilized as the coordination variables for achieving space-cooperative guidance, which can avoid continually maneuvering for maintaining the constant desired impact angles, thus reducing the fuel consumption. Next, the guidance laws along the LOS and perpendicular to the LOS are developed, and the event-triggering mechanisms are designed to reduce the update frequency of cooperative guidance commands, thus further reducing the energy consumption. To guarantee the convergence rate, the fixed-time control theory is adopted and the stability of proposed event-triggered cooperative guidance laws are rigorously proved. In addition, it is also proved that there is no Zeno behavior when implementing the proposed event-triggered cooperative guidance laws. Finally, numerical simulations indicate that the strictly simultaneous attack is achieved and the constraint of relative impact angles is satisfied. Comparative studies demonstrate that the computation burden of cooperative guidance commands is relaxed and the fuel consumption is reduced by the proposed event-triggered cooperative guidance laws with the constraint of relative impact angles.

Non-fragile control for consensus of nonlinear multiagent systems under periodic denial of service attacks

Abstract: This paper focuses on the issues of non-fragile event-triggered consensus problem for nonlinear multiagent systems (MASs) with external disturbance subjected to randomly occurring packet losses and periodic Denial of Service (DoS) attacks. Under connected communication topology, each agent exchanges information with its neighbors. In the presence of packet losses and DoS attacks, an event-triggered mechanism is designed to decide when to communicate the current sampled-data. Randomness is solved by using Bernoulli distribution in a stochastic way and the external disturbance is handled by using H∞ performance. A sufficient condition is derived in terms of linear matrix inequalities (LMIs) by introducing time-varying Lyapunov–Krasovskii Functional (LKF), ensuring mean square consensus for the resulting closed-loop systems. Finally, to show the effectiveness and applicability of proposed control scheme, two numerical simulations are presented.

Sparsity promoting decentralized learning strategies for radio tomographic imaging using consensus based ADMM approach

Abstract: Radio tomographic imaging (RTI) has wide applications in the detection and tracking of objects that do not require any sensor to be attached to the object. Consequently, it leads to device-free localization (DFL). RTI uses received signal strength (RSS) at different sensor nodes for imaging purposes. The attenuation maps, known as spatial loss fields (SLFs), measure the power loss at each pixel in the wireless sensor network (WSN) of interest. These SLFs help us to detect obstacles and aid in the imaging of objects. The centralized RTI system requires the information of all sensor nodes available at the fusion centre (FC), which in turn increases the communication overhead. Furthermore, the failure of links may lead to improper imaging in the RTI system. Hence, a distributed approach for the RTI system resolves such problems. In this paper, a consensus-based distributed strategy is used for distributed estimation of the SLF. The major contribution of this work is to propose a fully decentralized RTI system by using a consensus-based alternating direction method of multipliers (ADMM) algorithm to alleviate the practical issues with centralized and distributed incremental strategies. We proposed distributed consensus ADMM (DCADMM-RTI) and distributed sparse consensus ADMM (DSCADMM-RTI) for the RTI system to properly localize targets in a distributed fashion. Furthermore, the effect of quantization noise is verified by using the distributed consensus algorithms while sharing the quantized data among the neighbourhoods.

Bumpless transfer control for switched systems: A time-and state-dependent multiple Lyapunov function method

Abstract: For switched systems, the time-and state-dependent multiple Lyapunov function method is set up to investigate the bumpless transfer control problem. The key point of the proposed study is to relax the decreasing assumption of traditional Lyapunov functions at switching points and the global limitation of the bumpless transfer constraint, which provides more freedom for solving the bumpless transfer control problem. First, the description of the bumpless transfer performance is proposed by using switching points directly, which makes the controller only constrain in the subinterval of the active interval rather than the whole active interval. Second, a switching controller with time-varying gain matrices and a mixed dwell time-and state-dependent switching signal are designed to reduce control bumps and avoid frequent switching. Third, based on the developed multiple Lyapunov function method, a solvability condition of the bumpless transfer control problem is developed. Finally, an example is given to show the availability of the bumpless transfer control strategy.

Dynamic event-triggering sliding mode resilient control for multi-agent systems

Abstract: The consensus problem for a multi-agent system (MAS) is investigated in this paper via a sliding mode control mechanism subject to stochastic DoS attack, which may occur on each transmission channel independently and randomly according to the Bernoulli distribution. A distributed dynamic event-triggered strategy is implemented on the communication path among agents, where dynamic parameters are introduced to adjust the threshold of event-triggered condition. After that, a distributed sliding mode controller is proposed for ensuring the stochastic consensus of the MAS. Meantime, a minimization problem is solved to obtain the correct controller gain matrix. At last, a numerical example is shown to demonstrate the presented results.

Trajectory tracking double two-loop adaptive neural network control for a Quadrotor

Abstract: In this paper, the development and experimental validation of a novel double two-loop nonlinear controller based on adaptive neural networks for a quadrotor are presented. The proposed controller has a two-loop structure: an outer loop for position control and an inner loop for attitude control. Similarly, both position and orientation controllers also have a two-loop design with an adaptive neural network in each inner loop. The output weight matrices of the neural networks are updated online through adaptation laws obtained from a rigorous error convergence analysis. Thus, a training stage is unnecessary prior to the neural network implementation. Additionally, an integral action is included in the controller to cope with constant disturbances. The error convergence analysis guarantees the achievement of the trajectory tracking task and the boundedness of the output weight matrix estimation errors. The proposed scheme is designed such that an accurate knowledge of the quadrotor parameters is not needed. A comparison against the proposed controller and two other well-known schemes is presented. The obtained results showed the functionality of the proposed controller and demonstrated robustness to parametric uncertainty.

Multilayer-neural-network observer with compensator and command-filter-based adaptive backstepping tracking control of switched nonlinear systems

Abstract: This study presents an output backstepping control architecture based on command filter via Multilayer-Neural-Network Pre-Observer with compensator to realise the reference signal tracking of an arbitrarily switching nonlinear systems with nonseperated parameter. First, a multilayer neural network pre-observer is designed before backstepping procedures to servo reconstruct the system states which can not be obtained directly. The pre-observer has superior performance in neutralizing the states abrupt chattering caused by the arbitrarily switching parameter entered in the nonlinear structure. Next, observer error compensation mechanism is designed to compensate the state estimation and shrink the approximation error domain further. Then, the backstepping controller with compensation signal based on command filter is presented to realise the stable reference signal tracking. Last, the proposed control scheme guarantees the states of the closed-loop system bounded. This mechanism makes up the shortcoming of the traditional state observer and give more flexibility in reconstructing the systems states timely, then overcomes the obstacle of the arbitrarily switching parameterized system. Furthermore, compared with the existing traditional uniform robust uncertain controller, the developed backstepping control method combining with the pre-observer not only guarantees the states servo reconstruction and servo control of the switched system, but also improves the tracking performance. Finally, a low-velocity servo turnable switched system is extensively simulated to demonstrate the effectiveness of the developed controller.

Dynamic path planning of mobile robot based on improved simulated annealing algorithm

Abstract: Dynamic path planning for mobile robots is an urgent issue that needs to be solved because of the growing use of mobile robots in daily life and industrial operations. This work focuses on avoiding moving obstacles in dynamic situations. The computational effort required by some current algorithms makes them difficult to utilize for path planning in dynamic situations whilst the computational effort required by other methods makes them simple yet prone to local minima. In this paper, an improved simulated annealing (SA) algorithm for dynamic path planning is proposed. To reduce its computational effort, the initial path selection method and deletion operation are introduced. Simulation results show the improved SA algorithm outperforms other algorithms and provides optimal solutions in static and dynamic environments.

Fixed-time stable bilateral teleoperation of underwater manipulator using prescribed performance terminal sliding surfaces

Abstract: This study proposes two novel prescribed performance terminal sliding surfaces (PPTSSs) to address the fixed time stable bilateral teleoperation issue for a class of underwater manipulators with error constraints and input saturation. A general mathematical definition of the PPTSS method is first introduced, which can predetermine the convergence rate, steady-state error, and maximum overshoot. Moreover, the system settling time would have a fixed upper bound once the PPTSS is reached. An auxiliary system for saturation compensation is utilized to overcome the difficulties caused by actuator saturation. Moreover, two control schemes based on PPTSSs are proposed to handle error constraints and ensure the bound of global settling time is fixed. Finally, numerical simulation results are presented to demonstrate the effectiveness of the developed algorithms.

A lightweight object detection network in low-light conditions based on depthwise separable pyramid network and attention mechanism on embedded platforms

Abstract: Deep learning-based object detection algorithms have been widely used and are effective in autonomous driving. But their performance degrades dramatically under low-light conditions. The existing mitigation approach for low-light object detection uses image enhancement preprocessing to improve detection performance with limited success. In addition, this preprocessing approach incurs extra computation costs, consumes more resources, and makes it more challenging to implement the algorithm on embedded platforms. This paper proposes a novel object detection network in low-light conditions for embedded platforms. The new system consists of two innovative modules, including a lightweight low-light enhancement network DS-PyLENet and an anchor-free lightweight object detector CFEDet. DS-PyLENet is a three-level pyramid network configured with depthwise separable convolution residual blocks (DSCRBs) and multiscale DSCRBs. The CFEDet is configured with an improved EfficientNet backbone network with Coordinate Attention (CA) and vanilla MBconv, i.e., CA-fused EfficientNet, an improved Path Aggregation Network (PAN) fusion module, and an anchor-free Global Focal Loss (GFL) detection head. Two evaluation metrics, relative recall ratios, are proposed to depict the effectiveness of image enhancement on object detection more intuitively. The new model is demonstrated to be superior to the assembled model of EnlightenGAN and YOLOv5n in terms of detection accuracy and speed on GPU and embedded platforms. The testing results of the new model reveal that 83.5% mAP is achieved on the Exdark dataset and 67.4% mAP on the DarkFace dataset. The running rates of GTX 1080Ti and NVIDIA Jetson Xavier NX are 22 FPS and 2.6 FPS, respectively. The recall of CFEDet with DS-PyLENet increased to 22.1% and significantly improved over 18% with Zero-DCE and 17.2% with EnlightenGAN.

Multi-channel transmission scheduling with hopping scheme under uncertain channel states

Abstract: This paper investigates the multi-channel transmission scheduling problem for remote state estimation based on a hopping scheme in cyber-physical systems. The smart sensor sends multiple subpackets over different orthogonal channels to the remote end simultaneously. Owing to the randomness and vulnerability of transmission environments, the uncertain multi-channel states are considered in this paper, which relaxes the assumption of existing deterministic models. The objective is to find an appropriate hopping scheme that minimizes the remote estimation error covariance. First, the multi-channel selection problem is modeled as a multi-arm bandits (MAB) matrix via taking the packet receiving probability as the gain. From the perspective of strategy and channel, two exponential-weight online learning algorithms are designed with the assistance of transmission energy switching policy. Then, based on Bernstein’s inequality for martingales and mini-batching loop, the upper bounds of algorithms’ regret values are analyzed under stochastic and adversarial channel states, respectively. Further, the estimator expression in iterative form and a sufficient condition for the error covariance to be bounded are derived. Finally, an example of unmanned vehicle moving demonstrates all the theoretical results.

FedHD: Communication-efficient federated learning from hybrid data

Abstract: Federated learning (FL) has attracted significant attention in the machine learning community owing to its instinct local privacy awareness. Depending on how the data are distributed over the clients, FL problems can be divided into three categories, namely the horizontal FL (HFL), the vertical FL (VFL) and the hybrid FL (HBFL). Among them, the HBFL problem is the most challenging because each client neither owns the full set of data samples nor knows the complete feature information. While many FL algorithms have been developed for the HFL and VFL problems, unfortunately they cannot handle the HBFL problem. In this paper, we propose a new FL algorithm, termed as FedHD, that can solve the challenging learning task under the HBFL setting. Since the clients cannot perform local optimization on their own under the hybrid data, a tracking variable is introduced to enable the clients to track the global gradient information and update the model based on their local data. FedHD allows the clients to perform multiple steps of local stochastic gradient descent (SGD), and hence has improved communication efficiency. Theoretical analysis is conducted to show that FedHD has a O(1/QT) convergence rate, where Q denotes the local update number and T is the total number of communication rounds. Then we further reveal the insights on how various algorithm parameters impact on the convergence performance. Experiment results show that the proposed FedHD exhibits robust performance on the hybrid data, and is largely superior to the naive local training model.

Fixed-time synchronization of multiplex networks by sliding mode control

Abstract: Multiplex networks involve different types of synchronization due to their complex spatial structure. How to control multiplex networks to achieve different types of synchronization is an interesting topic. This paper considers the fixed-time synchronization of multiplex networks under sliding mode control (SMC). Firstly, for realizing three types of synchronization of multiplex networks in a fixed time, a unified sliding mode surface (SMS) is established. After that, based on the theory of SMC, a sliding mode controller (SMCr) which is more intelligent and has a simpler form than those in the existing literature is put forward for multiplex networks. It can not only guarantee the emergence of sliding mode motion, but also can realize three different kinds of synchronization by adjusting some parameters or even one parameter of the controller. Based on some theories of fixed-time stability, this paper deduces several sufficient conditions for the trajectories of the system to reach the preset SMS in a fixed time, and derives some sufficient conditions for multiplex networks to realize three different types of fixed-time synchronization. At the same time, the settling time which can reveal what factors determine the fixed-time synchronization in multiplex networks is obtained. Finally, in numerical simulations, different chaotic systems are set for each layer of multiplex networks to represent the nodes of different layers, which can prove that the theoretical results are practical and effective.

Design of an event-triggered joint adaptive high-Gain observer for a class of nonlinear system with unknown states and parameters

Abstract: In this paper, an event-triggered joint adaptive high-gain observer design method is proposed for a class of nonlinear systems that are characterized by an unknown parameter entering the system state equations. The main difficulty in the observer design is that the event-triggered mechanism (ETM) is affected by variable delayed-sampled data and the system’s unknown parameter. To overcome this difficulty, a closed-loop output predictor is incorporated into the design of the event-triggered mechanism to compensate for the sampling and the delay affecting the system outputs. To prevent the Zeno phenomenon, and to guarantee the exponential convergence of the observer, an exponential decay factor is considered in the ETM. The effectiveness of our proposed observer is demonstrated through numerical simulations, experiments and performances comparison with previous works in the literature.

Consensus of multi-Agent systems with one-Sided lipschitz nonlinearity via nonidentical double event-Triggered control subject to deception attacks

Abstract: As for the multi-agent systems (MASs) with time-varying switching subject to deception attacks, the leader-following consensus problem is studied in this article. The one-sided Lipschitz (OSL) condition is utilized for the nonlinear functions, which makes the results more general and relaxed than those obtained by Lipschitz condition. The nonidentical double event-triggering mechanisms (ETMs) are adopted for only a fraction of agents, and each agent transmits the data according to its own necessity. Semi-Markov process modeling with time-varying switching probability is adopted for switching topology and deception attacks occurring in transmission channel are considered. By using the cumulative distribution function (CDF) and the linear matrix inequality (LMI) technology, sufficient conditions for MASs to achieve consensus in mean square are obtained. An effective algorithm is presented to obtain the event-based control gains. The merits of the proposed control scheme are demonstrated via a simulation example.

Stability of T-S fuzzy system under non-fragile sampled-data H∞ control using augmented Lyapunov-Krasovskii functional

Abstract: This article is concerned with the non-fragile sampled-data control for T-S fuzzy system with parameter uncertainties. Firstly, a novel augmented Lyapunov-Krasovskii functional with sufficient sampled-data information is constructed. And a novel h(t)-depended exponential stability criterion with H∞ performance is gotten by reciprocally convex matrix inequality. Beyond that, compared with the existing methods, the gain matrices for non-fragile sampled-data controller expected are less conservative by linear matrix inequality technique. And numerical examples are provided to support the viability and validity of the results.

Switched observer-based dynamic event-triggered consensus of multi-agent systems under DoS attacks

Abstract: This paper investigates the observer-based consensus control for high-order nonlinear multi-agent systems (MASs) under denial-of-service (DoS) attacks. When the DoS attacks appear, the communication channels are destroyed, and the blocked information may ruin the consensus of MASs. A switched state observer is designed for the followers to observe the leader’s state whether the DoS attacks occur or not. Then, a dynamic event-triggered condition is proposed to reduce the consumption of communication resources. Moreover, an observer-based and dynamic event-triggered controller is formulated to achieve leader-following consensus through the back-stepping method. Additionally, the boundedness of all closed-loop signals is obtained based on the Lyapunov stability theory. Finally, the simulation results demonstrate the effectiveness of the presented control strategy under DoS attacks.

Copyright protection of multiple CT images using Octonion Krawtchouk moments and grey Wolf optimizer

Abstract: This paper proposes a novel Octonion Krawtchouk Moments (OKMs) transform to deal with a set of images in a compact manner, and based on this transform, a local zero-watermarking scheme is proposed to protect the copyright of CT medical images. The scheme first annotates regions of interest (ROIs) on seven medical images and then uses the OKMs of these ROIs to construct a single feature image called zero-watermark. This scheme adopts the gray Wolf Optimizer (GWO) algorithm to have a blind nature and to improve robustness against common image processing manipulations and attacks (zero-watermarking requirements). In addition, our scheme uses the trained U-net (R231) model to reduce the search space for the GWO algorithm and prevent this algorithm from getting stuck in a local optimal solution. The experimental results show that the proposed method is very robust against common image processing manupilations and attacks and has superiority compared with superb other zero-watermarking methods.

Multiple UAV formations delivery task planning based on a distributed adaptive algorithm

Abstract: In this paper, six-rotor UAVs are used in the field of distribution to realize the delivery of materials arriving the demand point by UAVs. Due to the small load capacity of the six-rotor UAV, in response to actual demand, the UAV group will be used to complete the delivery task. Considered to be close to the real requirements, trajectory constraints and dynamic obstacles are established in the trajectory planning based on group perception range. In order to better deal with dynamic obstacles and related constraints, this paper designs a distributed adaptive algorithm based on individual decision-making and group decision-making. Individual decision-making is embodied in the intelligence and adjustment of UAVs, involving actor-critic methods, artificial potential field method ideas and probability finite state machines; group decision-making is embodied in the leadership mechanism and joint decision-making; self-adaptation is embodied in the adaptive adjustment of UAV level in group. In order to avoid collisions between UAV groups, a conflict resolution algorithm is designed. Through simulation analysis, the distributed adaptive algorithm proposed in this paper can not only satisfy all constraints, avoid dynamic obstacles stably, and complete tasks with small fluctuations, but also obtain the most successful decision-making UAV in the group. This article further analyzes and discusses the relevant parameters in the algorithm, and obtains the optimal parameter ranges in individual decision-making and group decision-making.