Keywords: Intelligent and AI Based Control, Smart Buildings, Nonlinear Systems and Control
Abstract: Replacing poorly performing existing controllers with smarter solutions will decrease the energy intensity of the building sector. Recently, controllers based on Deep Reinforcement Learning (DRL) have been shown to be more effective than conventional baselines. However, since the optimal solution is usually unknown, it is still unclear if DRL agents are attaining near-optimal performance in general or if there is still a large gap to bridge.

In this paper, we investigate the performance of DRL agents compared to the theoretically optimal solution. To that end, we leverage Physically Consistent Neural Networks (PCNNs) as simulation environments, for which optimal control inputs are easy to compute. Furthermore, PCNNs solely rely on data to be trained, avoiding the difficult physics-based modeling phase, while retaining physical consistency. Our results hint that DRL agents not only clearly outperform conventional rule-based controllers, they furthermore attain near-optimal performance.

Keywords: Intelligent and AI Based Control, Fault Detection and Diagnostics
Abstract: Growing penetration of renewables comes with increased cyber security threat due to inherent low inertia characteristic and sophisticated control and communication networks of power electronics. This paper proposes a data-driven cyberattack detection strategy for grid-connected photovoltaic (PV) inverters. Ideas of long short term memory (LSTM) and convolutional neural network (CNN) as the core of detection achieve time series classification to diagnose the target and mode of cyberattack. Input de-redundancy and hyperparameter selection are conducted to optimize the detection. Meanwhile, well-designed cyberattack toolboxes of false data injection (FDI), denial-of-service (DoS) and delay are applied upon the communication of both sampled signals and issued commands in a grid-connected inverter model. By observing system performance via electrical measurements, this case study evaluates the LSTM, CNN-LSTM and convolutional LSTM based detection and obtains stable high quality of classification.

Keywords: Multi-agent Systems, Modeling and Control of Complex Systems, Linear Systems
Abstract: This paper proposes a mathcal{H}_{infty} state-feedback consensus controller for linear multi-agent systems (MASs) with N number of agents. In the proposed method, the complexity of mathcal{H}_{infty} consensus problem of the multi-agent system is reduced by representing this problem by N number of linear subsystems. The proposed controller utilises the information of both the system and the controller state from neighbouring agents. The sufficient conditions of stability are derived using Lyapunov approach and the gains of the controller are determined using linear matrix inequalities (LMIs). The performance of the controller is illustrated through simulations using a numerical example. The results of the simulation demonstrate that the proposed mathcal{H}_{infty} state-feedback consensus controller could effectively achieve desired performance under the effects of external disturbances.

Keywords: Automated Guided Vehicles, Control Applications, Motion Control
Abstract: This paper presents a model-free control strategy for blending driver steering input and closed-loop path following control commands in automated lane following applications. An impedance controller is designed herein to respond to the measured driver steering torque and modify the reference steering angle to be tracked by the vehicle. A mathematical analysis is provided to prove that the vehicle trajectory remains bounded relative to the target trajectory for a given driver torque input and retains the same transient performance as the unmodified lane following controller. With this control strategy, a given steering effort translates to a steady and predictable offset relative to the target trajectory. Experimental results demonstrate that the proposed controller can deliver a desirable steering feel and is agnostic to the steering system design or the control implementation of the electronic power steering system for real world automated driving applications.

Keywords: Nonlinear Systems and Control, Optimal Control
Abstract: Fixed-time convergence control barrier functions (FxTC-CBFs) are proved to be powerful to deal with a typical signal temporal logic (STL) specification, ``reachability'', which drives the state of a dynamical system to a given set in a specific time interval and enforces the system trajectories to evolve in the safe set at all times. However, existing literature rarely discussed the effect of external disturbances on ``reachability'' specifications, and the corresponding robust FxTC-CBFs are not well studied. To address these problems, in this paper, the FxTC input-to-state safety CBFs (FxTC-ISSf-CBFs) is proposed. As a result, the original reachability specification updates to a new reachability specification. Meantime, instead of the original target set of the reachability specification, the FxTC-ISSf-CBFs guarantees forward invariance of a safety set of a larger size. Moreover, motivated by the fact that computing the derivatives of a CBF with explicit expression is often complicated and time-consuming, the implicit FxTC-ISSf-CBFs (IFxTC-ISSf-CBFs) is introduced based on an adaption of the implicit Lyapunov condition. Finally, we formulate a quadratic program (QP) to compute the control input that satisfies the conditions of the IFxTC-ISS-CBFs.

Keywords: Estimation and Identification, Linear Systems
Abstract: In this paper, the system parameters identification problem of the diagonal canonical form of discrete-time single-input and single-output (SISO) linear time-invariant (LTI) system is considered. Based on the finite input/output sample trajectories, we give the fundamental identification limit on system parameters for marginally stable SISO systems by analyzing the Fisher information matrix used in Cramer-Rao bound, under the condition of the total energy of input is finite. Furthermore, We show the identification of system parameters can be more and more difficult as the system dimension increases.

Keywords: Intelligent and AI Based Control
Abstract: With the map navigation system achieving the po- sitioning accuracy at the lane level, the lane level navigation system began to be applied in digital navigation software. The new generation of navigation system allows human drivers to accurately identify road conditions without road experience. However, in actual use, drivers still need to complete lane- changing behavior independently. The navigation system does not provide relevant real-time traffic flow information and safety prediction. Therefore, this paper innovatively combines the free space detection and lane change prediction algorithm in actual driving, and puts forward this Safety Prompt Advanced Driver-Assistance System (SPADAS). It can predict the lane- change decision and lane-change trajectory around in a short period of time in the future according to the current state and historical trajectory of the surrounding vehicles, and combine these prediction information to provide the driver or the path planning module of the self-driving vehicle with the predicted free space in the near future. The free space provided by this Advanced Driver-Assistance System (ADAS) includes three types: continuing driving according to the current state or changing lanes to the left or right. At the same time, it also provides the required recommended driving trajectory. The evaluation test of this ADAS has been done through a simulation testbed of the NGSIM dataset built with Eclipse SUMO.

Keywords: Optimal Control, Modeling and Control of Complex Systems, Factory Modeling and Automation
Abstract: This paper examines the adaptive robust control for uncertain underactuated mechanical systems. The uncertainty is time varying but bounded. An adaptive robust constraint-following control approach is developed with tunable parameters. For the seeking of optimal control parameters, a two-player cooperative game oriented optimal design problem is formulated. The problem is completely solved with the existence and uniqueness of the optimal solution proved and the analytical expression of the optimal parameters provided. Using the optimal parameters, the proposed adaptive robust control renders dual performance: guaranteed (uniform boundedness and uniform ultimate boundedness) and optimal.

Keywords: Modeling and Control of Complex Systems, Process Automation, Optimal Control
Abstract: An optimal traffic evacuation scheme is of practical importance to transport all the evacuees in potential minimized evacuation time under an emergency. In this paper, we develop a traffic dispatching control strategy under uncertain evacuees' behaviors, which consists of three parts of buses in every round as gathering, waiting, and transporting times. From the time that the disaster occurs to the time when all evacuees transport to safe shelters, the buses will be dispatched to minimize the maximal time of a bus in evacuation. A new algorithm is designed to solve the two-stage optimization process, and the simulation results demonstrate the effectiveness of the proposed method.

Keywords: Intelligent and AI Based Control
Abstract: Recent research on the prediction of driver’s lane-changing behaviour requires vehicle surrounding information, as it is believed that driver’s decision on lane changing is made consciously based on those information. However, current research has shown that the usage of such surrounding information leads to high false alarm rate of lane-changing predict system [1]. Therefore this paper contributes to developing a lane-changing prediction method which uses vehicle state information only. From the perspective of the observer’s daily experience, this paper selects vehicle’s lateral trajectory and the spectrum of its lateral trajectory as input to predict drivers’ lane-changing intention. A Direction Convolutional LSTM (DCLSTM) network has been developed to predict drivers’ lane-changing behaviours. Recent pure LSTM methods proposed by researchers provide high accuracy when predicting the generation of drivers’ lane-changing intentions, but they have relatively low accuracy in predicting drivers’ lane-changing direction. DCLSTM retains pure LSTM network’s high accuracy in the prediction of drivers’ lane-changing intentions, while its prediction of drivers’ lane-changing directions is also accurate. All the training and testing data are extracted from the NGSIM dataset.

Keywords: Signal Processing, Sensor/Data Fusion
Abstract: Map matching of vehicle trajectories is to identify the correct link in the road network for each positioning point of a trajectory. A sampling period of fewer than 10 seconds for each position point is typically regarded as a high sampling rate trajectory. Existing algorithms are mainly targeted to the low-sampling-rate trajectories, which may ignore much useful information from high-sampling-rate trajectories. Till now, no studies explore the validity of such algorithms when we feed them the high-sampling-rate trajectories, which is the target of this paper. For alleviating the positioning error influence and speeding up the matching process for high sampling rate trajectories, a batch matching strategy is studied to simultaneously match a subsequence of a trajectory. Specifically, we first estimate the mean and median speeds of a trajectory and the current speed of the GPS point. To make sure that the upcoming subsequence to be matched is on the same road segment, we utilize the speed estimations as constraints for determining the subsequence size together with the local features of a road network. Accordingly, an incremental map matching algorithm is further in this study. Experimental results on real-world datasets of high sampling rate vehicle trajectories demonstrate that the proposed algorithm outperforms the algorithm that is designed for the low-sampling-rate trajectory.

Keywords: Smart Buildings, Sensor/Data Fusion, Energy Efficiency
Abstract: Understanding and estimating the individuals' comfort level in a dynamic environment is a challenging but crucial task to optimize buildings' energy consumption. Besides, improving the occupant's comfort level affects the occupants' performance and well-being. Therefore, available personal comfort modeling methods use different information such as behavior feedback, physiological parameters, and other factors to estimate personal comfort. This study proposes a dynamical modeling approach to address this challenging task that does not require a detailed data set. We show that a dynamical system can be used to model the individual thermal conform based on occupant' adjustments (interpreted as thermal comfort response) in a ventilated space. After introducing a basic model learning heuristic, the system identification problem is formulated as an optimization problem. Then the relaxation of this optimization problem is given and tested. In addition, the proposed model can be utilized in learning-based control. Finally, the performance of the proposed method is illustrated with real test data and one simulated data set.

Keywords: Intelligent and AI Based Control, Multi-agent Systems, Nonlinear Systems and Control
Abstract: This paper considers the finite-time cooperative control problem for a multi-agent system composed of multiple heterogeneous Euler-Lagrange systems with parametric uncertainty and a dynamic leader. A distributed finite-time observer is first proposed to estimate the reference signal under a directed topology, and then a finite-time controller is constructed by transforming uncertain Euler-Lagrange systems into second-order systems through a partial design of the controller and utilizing the backstepping-like design procedure. The efficiency of the design is illustrated by simulation.

Keywords: Robotics, Adaptive Control
Abstract: In order to steer a snake robot along the desired path by detecting one edge of the path and keep a desired distance from the edge, the virtual edge guidance strategy is proposed. One edge of the path is detected by sensors. According to the detected edge, the virtual temporary target point is constructed and continuously updated. The continuously updated virtual temporary target points form the virtual edge. The locomotion direction of the robot is adjusted by the detected edge and the virtual edge. After a period of direction adjustment, the snake robot follows the centerline between the path edge and the virtual edge. The distance between the robot and the actual edge can be adjusted by the distance between the virtual edge and the actual edge. Simulations show that the snake robot can follow the desired path, and the distance between the snake robot and the path edge can meet requirements.

Keywords: Flexible Manufacturing Systems, Energy Efficiency, Control of Smart Power Delivery Systems
Abstract: This paper constructs a framework of collaborative optimization between Multi-energy system (MES) and industrial plants. The multi-energy coupling models of MES and industrial plants are established respectively. The objective of collaborative optimization problem is to minimize the total energy costs by optimizing the output power of energy equipments in MES and the production scheduling of workshops in industrial plants. In addition, considering the recovery of carbon dioxide, the power-to-gas unit is integrated into the MES. Three different cases are used to demonstrate the effectiveness and superiority of the proposed collaborative optimization method.

Keywords: Optimal Control, Nonlinear Systems and Control, Modeling and Control of Complex Systems
Abstract: In this research, an off-policy Q-learning algorithm is designed to solve the pressure tracking problem in gas pipelines by adjusting the rotating speed using only the observed data along the system trajectories. How to determine the most appropriate rotational Speed by off-policy Q-learning and enforce the pipeline pressure to follow a desired value is very challenging due to nonlinear and unknown dynamics of gas pipeline pressure systems, as well as the requirement of variation of rotational speed. To this end, an optimal control problem of gas pipeline pressure tracking control is first formulated, then an off-policy Q-function based on iterative Bellman equation which is derived, followed by an off-policy Q-learning algorithm used for searching the optimal rotational speed based on reinforcement learning technique and dynamic programming theory, such that the pressure can follow the desired value successfully. The simulation results are shown to validate the effectiveness of the proposed strategy.

Keywords: Multi-agent Systems
Abstract: This paper studies the problem of entrapping a target by mobile agents with a discrete-time system using bearing-only measurements. To locate the target without prior information of its position, a discrete-time estimator using bearing-only measurements is developed by exploiting the orthogonality property. Based on the estimation of the relative position, a controller is designed to drive the agent to entrap the target in an arbitrarily shaped orbit. We present the sufficient conditions under which the stability of the overall system is proved using LaSalle's theorem for the discrete-time system and the iterative method. An extension to the multi-agent system is also given. Simulations are given to illustrate the effectiveness of the proposed scheme.

Keywords: Multi-agent Systems, Networked Control
Abstract: This paper studies a class of distributed convex optimization problems subject to coupled equality and inequality constraints, which are affine and convex functions respectively. The objective is to minimize the sum of a strongly convex smooth function and two convex nonsmooth functions. For such a composite optimization problem with coupled constraints, we propose a distributed proximal-based primal-dual algorithm with a fixed stepsize, based on operators splitting technique and dual decomposition method, where an auxiliary variable is introduced to evict the unproximal characteristics of complex nonsmooth functions. Via Lyapunov stability theory, it is proved that global optimal solution is obtained with an O(1/t) convergence rate. Finally, the theoretical results are demonstrated in an economic dispatch problem.

Keywords: Multi-agent Systems
Abstract: In this paper, online distributed constrained optimization is investigated by employing a continuous-time multi-agent systems.The objective of the agents is to cooperatively minimize the sum of time-varying cost functions subject to a convex set at each time. Each agent can only have access to its own cost function and the convex set, and cost function in the future is not available. To address this problem, we propose a modified online distributed “projection+gradient” algorithm, which involves each agent minimizing its own cost function while exchanging local state information with others via an unbalanced digraph. Performance of the algorithm is measured by dynamic regrets. Under mild assumptions on the graph, we prove that if the rate of a minimizer’s variation is within a certain range, then regrets, as well as the violation of constraint, grow sublinearly. A simulation is presented to demonstrate the effectiveness of our theoretical results.

Keywords: Learning Systems, Nonlinear Systems and Control, Process Automation
Abstract: Aero-engine performance optimization is crucial for pursuing reliability and security during the operation of an aero-engine. However, optimization of aero-engine performance is a multiobjective, computationally expensive programming problem. Ordinarily, such a problem is assumed that analytic expressions of the objective functions are available. However, only historical data could be obtained in practice, which is impossible to apply in the existing optimization algorithm. Therefore, a data-driven multiobjective evolutionary algorithm is proposed to address these difficulties, which employs NSGA-II as the fundamental element assisted by selective ensemble learning, model management strategy, and a final solution set generation mechanism. The numerical results demonstrate that the proposed algorithm can approximate the Pareto front of the aero-engine performance optimization problem.

Keywords: Learning Systems, Learning-based Control, Control of Smart Power Delivery Systems
Abstract: This paper tries to solve the optimal power scheduling problem for microgrid system by considering the data privacy preserving. Conventionally, such problem is addressed by various numerical optimization approaches. However, these approaches usually suffer from high computational cost and result in a slow convergence speed. In order to tackle this disadvantages, a novel data-driven approach, i.e., distributed deep reinforcement learning algorithm, is developed in this paper. A deep deterministic policy gradient algorithm learning model for each distributed power source is built first. Then in a distributed framework, each local power source establishes a state space and a action space respectively. Finally, the gradient clipping strategy is used to update the local model, and the differential privacy method is introduced to protect data privacy. Simulation results show that the individual difference rate of data protected by this method reaches 40%, and the accuracy rate of aggregated data reaches more than 85%. At the same time, compared with the centralized deep reinforcement learning algorithm, the training speed is increased by 15%, and the cost function value is reduced by 60%. These results indicate that a good trade-off is achieved between accurate optimal solution and privacy preserving.

Keywords: Multi-agent Systems, Nonlinear Systems and Control, Networked Control
Abstract: This work investigates the time-varying formation tracking control (TVFTC) problem of multiple small-scale unmanned helicopters (SUHs) with disturbances. The SUHs are required to achieve a desired time-varying formation flight while guaranteeing the tracking error without violating the prescribed constraint. Based on backstepping technique, a fully distributed formation control scheme is proposed, where a barrier Lyapunov function is adopted to address the error constraint, and disturbance observers are employed to estimate the unknown disturbances. It is mathematically proved that all the closed-loop states of SUH formation are bounded under the proposed scheme while the prescribed error constraints will not be exceeded, which enhances the flight safety and applicability. Simulation results further verify the effectiveness of proposed method.

Keywords: Control Applications, Energy Efficiency, Smart Buildings
Abstract: This paper studies optimal scheduling of multiple networked microgrids using distributed coordination algorithms. Each microgrid is controlled by an individual agent using mixed-integer linear programming (MILP). The algorithm is designed towards a two-level microgrid community (MGC) structure, where the exchange power is adjusted several times through information communication between the upper-level agent and lower-level agents to minimize the operation cost. In the paper, an overview of the system model is first given, followed by a description of the distributed coordination algorithm in detail incorporating the mathematical modeling of the problem. A simulation study of the algorithm is performed based on a real MGC system. We found that the distributed coordination algorithm with multiple iterations can make the operation cost of the MGC system closer to the centralized result if appropriate parameter values are selected. Compared with other hierarchical energy management methods, our approach can achieve a lower cost.

Keywords: Control Applications, Nonlinear Systems and Control
Abstract: We propose a novel nonlinear control method for the voltage regulation of a buck converter to achieve a rapid and nonovershooting step response. The method combines feedback linearisation with a linear static state feedback controller to shape the system transient response. Experimental results involving performance comparisons with conventional linear controllers show the proposed method can deliver substantial improvement in the step reference response of a buck converter circuit, while also providing improved transient response in the rejection of both line and load disturbances.

Keywords: Optimal Control, Nonlinear Systems and Control
Abstract: In this paper an iterative algorithm is proposed to solve discrete nonlinear minimum time control problem via a transformed objective function, which changes the original problem into a tractable optimization problem. In this way, well-established numerical optimization algorithms can be applied. The fastest descent curve (Brachistochrone) is adopted as a case study to test the efficacy of this method. Gradient search algorithm is used in this framework and formulas to find the gradient are provided. Explicit method that determines the Brachistochrone curve based on initial conditions is also presented.

Keywords: Nonlinear Systems and Control
Abstract: In this paper we extend the concept of Input-Output Finite-Time Stability (IO-FTS) to the class of nonlinear quadratic systems (NLQSs), which is particularly suited to study the dynamics of robotic systems, bio-systems and of a number of other processes of interest in the field of applied sciences. So far, the research on IO-FTS has mainly concerned linear systems; this point strongly motivates the investigation of NLQSs, due to the above-mentioned wide range of applications of this class of systems. To this regard, the main result of the paper consists of a sufficient condition guaranteeing the IO-FTS of NLQSs subject to L2 input disturbances; then the analysis result is exploited to design a state-feedback controller guaranteeing the IO finite-time stabilization of a given NLQS. Such sufficient conditions are given in terms of numerically-tractable optimization problems based on Linear Matrix Inequalities (LMIs). Some numerical examples illustrate the effectiveness of the proposed approach.

Keywords: Nonlinear Systems and Control, Process Control & Instrumentation, Discrete Event Systems
Abstract: One of the key use cases for 5G cellular communication is wirelessly connected smart factory applications as part of the Industry 5.0 vision. Such applications require high throughput, low latency, and highly reliable communication for critical control applications. With limited private spectrum available for Industry 5.0 use cases, scarce wireless network resources (e.g., bandwidth, channel capacity, etc.) need to be used efficiently. In this paper, we optimize wireless channel resources, in particular the secondary control channel, by increasing channel capacity in the presence of interference through power control and self-triggered control (STC) based message transmission. We aim to maximize control channel efficiency and minimize secondary control communication. Our simulation results demonstrate that STC-based communication can achieve the same system performance as periodic communication while preserving about 85% of the control channel bandwidth.

Keywords: Networked Control, Nonlinear Systems and Control, Discrete Event Systems
Abstract: The performance of networked control systems (NCSs) is often critically dependent on communication resources such as network bandwidth, nodes energy and so on. Although in recent years, different event-triggered schemes (ETSs) have been developed to tackle these problems, the detailed comparative performance investigation of these ETSs have received few attention. The present study, therefore, carries out a detailed investigation on the performance of two recently developed ETSs, the memory-based event-triggered scheme (METS) and the discrete event-triggered scheme (DETS), considering nonlinear systems with time-varying delays and uncertainties. Further, the design procedure of the networked control system with these two ETSs are presented. Simulations are carried out considering an example of synchronization between two inertial neural networks. The results of the investigation demonstrate that the control performance of METS is better when the system is subjected to severe delays. However, the DETS performs better from the perspective of saving network resources.

Keywords: Control Applications, Robust and H infinity Control
Abstract: In this paper, a bumpless transfer control design of robust tracking control is proposed to avoid large jumps in fuel flow for turbofan aeroengine speed control system. The control design is divided into two parts: one part is the linear feedback control (LFC) which meets the requirement of bumpless indices, and the other part is the continuous sliding mode control (CSMC). First, the LFC design considering the nominal system without the disturbance and uncertainty is proposed, and the linear feedback coefficients satisfy the bumpless indices. Then, a robust integral sliding mode (RISM) design is proposed, in which the uncertain linear switched system satisfies the given H-infinity robust performance index and can resist the unmatched uncertainty. Consequently, the CSMC ensures that the RISM surface can be reached from the initial time instant. By composing the CSMC with the LFC, the control design achieves the robust trajectory tracking and the suppression of the control signal bumps during switching. Finally, the proposed bumpless transfer control design was applied to a turbofan aeroengine model, and the effectiveness was verified.

Keywords: Multi-agent Systems, Networked Control
Abstract: In this paper, we formulate a class of distributed formation control problems where the correspondence between agent labels and geometric vertices of the formation shape is not specified a prior. The new formation control problem is then decomposed into two steps including finding the correspondence between agent labels and formation's geometric vertices and the formation stabilization. A linear program that is a relaxation of an integer programing is formulated to seek the correspondence and a distributed optimization algorithm is proposed. The distributed formation stabilization control is then proposed to achieve the formation shape in a distributed way. The freedom of choosing the correspondence between agent labels and formation's geometric vertices during the group coordination could lead to potential applications where the formation is required to be adjusted in real time as a response to the environmental change.

Keywords: Multi-agent Systems, Linear Systems
Abstract: In this paper, we address the consensus problem for discrete-time multi-agent systems by proposing a novel reset consensus protocol. By introducing time-varying gains, the consensus problem is addressed by using the properties of the time-varying stochastic system matrices. It is shown that the consensus of all agents is achieved with a larger convergence rate than that of existing control protocols. A numerical example is presented to illustrate the theoretical results.

Keywords: Multi-agent Systems, Linear Systems
Abstract: This paper studies the leader-following consensus problem of linear multi-agent systems with unbounded distributed transmission delays. A novel distributed observer is first proposed to estimate the state of leader in the presence of unbounded distributed transmission delays. Then a distributed controller is further developed without using any prior knowledge of unbounded transmission delays. It is shown that not only the leader-following consensus of multi-agent systems can be achieved exponentially but also an estimate of the exponential convergence rate of consensus can be obtained. In addition, our results include those results on bounded distributed delays and constant delays as special cases. Finally, a simulation example is provided to illustrate the effectiveness of the proposed controller.

Keywords: Multi-agent Systems, Networked Control
Abstract: Recently, a new type of synchronization problem called autonomous synchronization was formulated where neither the synchronized agent dynamics nor the synchronized states are specified a priori. Through a dynamics update law together with the synchronizing control law, the heterogeneous multi-agent systems (MAS) synchronize with one another autonomously. In this paper, we propose a dynamics update scheme which can adaptively select the parameter in the dynamics update law without requiring the knowledge of the underlying communication graph or the eventual synchronized dynamics. In this way, we reduce the constraints of the design of autonomous synchronization controllers. Two conjectures are proposed on solvability of the adaptive autonomous synchronization problem in both continuous-time and discrete-time environments. Numerical simulations are provided to illustrate the solutions.

Keywords: Multi-agent Systems, Networked Control
Abstract: This paper investigates distributed control of linear physically-interconnected systems under event-triggered communication with packet losses. For each subsystem, a distributed event-triggered control law and a dynamic event-triggering mechanism are proposed such that neither continuous controller updates nor continuous communication among subsystems is needed. The resulting closed-loop system is shown to be exponentially stable. The proposed dynamic event-triggering mechanism is robust to packet losses and provides a designable positive minimum inter-event time. Simulation results on a multi-buffer DC microgrid validate the effectiveness of the proposed control strategy.

Keywords: Modeling and Control of Complex Systems
Abstract: External counterpulsation (ECP) is the mechanical therapy for ischemic cardiovascular and cerebrovascular diseases in clinical. Mean arterial pressure for patients with cerebral ischemic stroke is the main indicator assessed using traditional ECP, but improvement of the hemodynamic environment in human common carotid arteries (CCA) is critical during the long-term use of ECP, and at present there is still a lack of related quantitative hemodynamic mechanism. To overcome the above limitation, in this paper, a simplified model for regulating hemodynamic response of ECP on CCA is proposed. First, a complex high-order lumped parameter model (LPM) of human circulatory system is established to simulate hemodynamic effects with varying ECP modes on CCA. Second, counterpulsation is achieved by applying pressure on the lower limb module in the LPM. Finally, a simplified LPM is developed by fitting the input impedance of upper and lower limbs, respectively. The numerical in-silico findings show validity and feasibility of the proposed LPM compared with the clinical values. The proposed simplified LPM may provide key technical support for the future design of corresponding rehabilitation chip system.

Keywords: Modeling and Control of Complex Systems
Abstract: A kinetic model of intracellular nitric oxide (NO) response in endothelial cells (ECs) activated by dynamic shear stress is developed. Based on the hypothesis of the viscoelastic body, the Kelvin model is adopted to simulate the mechanotransduction process. The NO responses activated by different dynamic shear stress induced by constant flow, pulsatile flow, and oscillatory flow are analyzed through numerical simulations. Furthermore, the interaction between NO and reactive oxygen species (ROS) has been considered for the first time when ECs are exposed to exercise-induced shear stress, with which model prediction and experimental data are in good agreement. This work can also provide valuable insights for further study of biomechanical response in ECs.

Keywords: Modeling and Control of Complex Systems
Abstract: Objective: Cavopulmonary assist devices (CPAD) are being developed to treat failing Fontan circulation. The control algorithm to provide appropriate levels of CPAD support using cavopulmonary pressure head (CPPH) has been proposed. However, the controller estimated CPPH using Golay-Savitzky (GS) filter and extended Kalman filter (EKF), which has high computation complexity. In this study, a new model for accurately predicting CPPH for CPAD using neural network was proposed. Methods: The proposed neural network model has structure of 2-15-15-1. Inputs of the neural network model are intrinsic CPAD parameters and output is CPPH with varying physiological conditions for Fontan patients. Back-propagation algorithm with Adam training function and Tanh-Sigmoid activation function were applied in the proposed model. Efficacy and robustness of the proposed neural network model for predicting CPPH was tested with (1) CPPH without noise, (2) CPPH with 10% uniformly distributed noise, and (3) CPPH estimated with GS filter and EKF. Results and Conclusion: The proposed neural network model predicted required CPPH more accurately than using GS filter and EKF regardless of different physiologic test conditions. Significance: The proposed neural network model can quickly find the optimized solution to predict CPPH, and may be more easily applied to CPADs for providing physiological perfusion.

Keywords: Process Control & Instrumentation, Networked Control, Multi-agent Systems
Abstract: This paper presents an extension of a recently proposed hierarchical control framework applied to a cryogenic system. Under the previous assumption, each subsystem in the decomposition needed at least one control input, which induced a computational time bottleneck, especially when a tiny nonlinear entity that needs to be integrated must be attached to a large-scale linear subsystem. In this paper, this condition is removed, allowing greater flexibility in the definition of the decomposition graph, meaning that small nonlinear entities and large linear-considerable subsystems are detached, resulting in more tractable control problems. The impact of this increased flexibility on the computational time is shown using the same cryogenic station where a decomposition into four subsystems is made possible (instead of two in the previous framework).

Keywords: Nonlinear Systems and Control, Motion Control, Control Applications
Abstract: Collision-avoidance function is of great significance for autonomous vehicles. This paper proposes an artificial potential field (APF) method to plan the obstacle-avoidance path for an autonomous vehicle, and transforms the discrete path model into a continuous one for path-tracking control design. To describe the lateral dynamics and path-tracking characteristics of the vehicle, a two-degree-of-freedom (2-DOF) dynamic model and a kinematic model are established. According to the reference path information and vehicle dynamic model, a sliding mode control (SMC) strategy is presented to realize the vehicle's path-tracking behaviour, and the stability of control system is testified in Lyapunov sense. Finally, MATLAB/Simulink-Carsim co-simulation is applied to test the effectiveness of the proposed method under various road conditions.

Keywords: Nonlinear Systems and Control, Motion Control, Control Applications
Abstract: In view of the development prospect of vehicle steer-by-wire (SbW) technology, this paper proposes a recursive terminal sliding mode (RTSM) control scheme for an SbW system. Firstly, we describe the SbW system by a second-order dynamic model, where the control input is the voltage of steering motor and the system output is the steering angle of front wheels. For simplicity, the self-alignment torque and friction generated by tire-road contact are considered as external disturbances. Secondly, an RTSM controller is designed for the SbW system, which can not only deal with the parametric uncertainties, but also effectively estimate the self-alignment torque coefficient. In addition, we prove the stability of RTSM control system from the perspective of Lyapunov, and give suitable control parameters under stable conditions. Finally, the steering control is simulated along a slalom path and a circular path, respectively. Simulation results show that, compared with conventional sliding mode (CSM) control, the proposed RTSM control strategy possesses obvious advantages in terms of higher tracking accuracy, faster response speed and firmer robustness.

Keywords: Robust and H infinity Control, Estimation and Identification, Sensor/Data Fusion
Abstract: This paper investigates the secure H_infty fusion estimation (SHFE) problem for a class of cyber-physical systems (CPSs) under false data injection (FDI) attacks. The control messages received by the actuator, and the measurements sent by the local sensors are both compromised by the malicious adversary. The local intermediate estimators are designed to reconstruct states and FDI attacks, and the estimation gains can be obtained by establishing a convex optimization problem that can be solved with standard software packages. Moreover, the corresponding secure H_infty fusion structure is constructed to improve the estimation performance of system states and actuator attacks with the proposed optimal weighting fusion rules. Finally, a maneuvering vehicle example is utilized to demonstrate the effectiveness of the proposed methods.

Keywords: Control of Smart Power Delivery Systems, Energy Efficiency, Estimation and Identification
Abstract: In the process of a dynamic capacity increase of transmission lines, to prevent false data injection attacks, the stability and adjustability of transmission network capacity increase is damaged. Therefore, this study studies a risk probability forecast assessment model to improve the robustness of the model under false data injection attacks. It is important to ensure reliability and security after capacity increase operation. This novel method first probabilistically models the uncertainty of meteorological data present in risk assessment. In addition, the conductor current prediction algorithm is introduced to make the risk assessment model somewhat predictive. Finally, the conductor temperature is analyzed using Monte Carlo simulations to complete the risk assessment of instability. The simulation results show that the proposed probabilistic prediction model enables effective conductor heating risk prediction in dynamic capacity increase even in a severe cyberattacks environment.

Keywords: Networked Control, Modeling and Control of Complex Systems
Abstract: In this paper, the cluster synchronization control problem is investigated for cyber-physical systems under cyber attack. A false data injection (FDI) attack is considered in the process of data transmission through the communication network. From the defender’s point of view, the occurrence of such an attack is governed by a set of Bernoulli sequences with a given probability. The controller gains are then obtained by minimizing the upper bound of the cluster synchronization error. Finally, a numerical simulation is presented to verify the effectiveness of the proposed method.

Keywords: Control Applications, Estimation and Identification
Abstract: Based on the in-depth analysis of the mathematical model of the doubly-fed induction generator (DFIG), this paper proposes a direct voltage control (DVC) method that combines a high-order nonlinear extended state observer and super-twisting sliding mode control for the doubly-fed power generation system. Using the second-order mathematical model between the DC bus voltage and the system input, the traditional double loop control system can be replaced with a voltage loop. The advantage of this scheme is to realize direct voltage control on the basis of ensuring the variable speed constant frequency (VSCF) of the motor, which reduces the number of control variables and sensors, and simplifies the expression of control variables. The reliability of the system is improved, and the response speed of the system when dealing with large load disturbances is improved. Finally, simulation prove the effectiveness of the proposed method.

Keywords: Control of Smart Power Delivery Systems, Control of Distributed Generation Systems, Networked Control
Abstract: This paper deals with the load frequency control problem of uncertain multi-area power systems under random DoS attacks. The public communication network framework is utilized to support information transmission of power areas,which may suffer from malicious DoS attacks. When DoS attacks happen, the transmitted system states may be unavailable.Hence, a compensating strategy is adopted to alleviate the negative impact by using the previous received information. Meanwhile, the existence of parameter uncertainties can further deteriorate the robustness of multi-area power systems. A desirable sliding mode controller is proposed to suppress the load variation of power systems against time-varying uncertainty and DoS attacks. The sufficient conditions are derived to guarantee the stochastic stability with H_infinity performance of the multi-area power systems and the reachability of the specified sliding surface simultaneously. Finally, a three-area power system is utilized to show the effectiveness of the designed control scheme.

Keywords: Control Education, Linear Systems
Abstract: Interactive software tools are proven to be very useful techniques with a high impact on control education. These tools make the learning process more interactive and involve students’ participation and focus. The paper describes the features, functionalities, and applications of ‘Advanced CS’, an interactive graphical user interface (GUI) tool for time domain analysis, frequency domain analysis, and designing of the modern linear control system. It covers state-space modeling, time and frequency domain analysis, PID controller designing, state feedback controller designing by using pole-placement and quadratic regulator method, and full-order and minimum-order state observer designing. The tool, whose main features are simplicity, interactive nature, ease of use, and dynamic nature, is developed using MATLAB for the enhancement of control education. Its capability of updating performance, output parameters, and graphical elements immediately on change of system parameters makes it dynamic and makes the iterative process of analysis and designing simple and easy. Its capability to interface with the hardware setup in real-time not only adds to the purpose of control education enhancement but also validates the analysis and design.

Keywords: Intelligent and AI Based Control, Nonlinear Systems and Control, Fuzzy and Neural Systems
Abstract: The finite-time attitude tracking control for a rigid spacecraft subject to inertial uncertainties, external disturbances, actuator failures, and saturation constraints is investigated using ideas of fast nonsingular terminal-sliding mode manifold (FNTSM) and unknown nonlinear function approximation with a Chebyshev artificial neural network (CANN). First, the FNTSM is constructed, and then the unknown nonlinear function of sliding mode dynamics is approximated with a CANN. A fault-tolerant attitude control law, which ensures that all the signals in the closed-loop system with spacecraft or orbiting satellite are uniformly ultimately bounded, is then designed by combining FNTSM and CANN techniques as appropriate. An improved control algorithm is derived to achieve finite-time attitude tracking. Extensive simulation experiments were conducted to verify the effectiveness of the proposed control scheme. A selected sample of those simulation results that demonstrate an outstanding control performance is given.

Keywords: Adaptive Control, Fuzzy and Neural Systems, Nonlinear Systems and Control

Keywords: Intelligent and AI Based Control, Robotics, Signal Processing
Abstract: As the number of amputees in the world steadily increases each year, the need for fully functional prosthesis is at an all-time high. Enabling a wide variety of possible movements with fast, accurate, and fluid control is one of the main goals for a successful electric prosthesis. The use of electromyography for the detection of the intended movements allows for increased practicality of the device and easy and natural control by the user. The implementation of a fast and cheap classification method for the measured signals allows for a significantly lower price of the proposed prosthesis, thus making it available to a wider margin of the population while also maintaining the expected quality of the product. In this paper we compare multiple different machine learning algorithms for the classification of individual and combined finger flexions from eight different participants using two-channel electromyography, in order to achieve an optimal generalized classifier that allows for accurate results while providing the least complex solution to the problem. By using only two EMG sensors we can achieve a much more practical and cheaper realization of an electric prosthesis, which is the main goal of this work. The best performer out of the tested algorithms is then additionally evaluated on the same movements when they are carried out by a single participant, thus simulating a real-world case where the prosthesis would be used by a single individual. Finally, we analyze classification errors, discuss the nature of their occurrences, and propose possible solutions and future improvements.

Keywords: Real-time Systems, Learning Systems, Sensor Networks
Abstract: The Internet of Things (IoT) revolutionized maintenance technology by allowing machines to communicate among themselves over the Internet in real-time. The very focus of Industry 4.0 towards automation and data exchange calls for the application of IoT in industry and for control, supervision, and maintenance of industrial equipment based on sensor data. The goal of this paper is to test several unsupervised machine learning algorithms for detecting anomalies in unlabeled (raw) data collected from a real e-coating plant that uses IoT technology. Unsupervised learning methods are advantageous because they do not require labeled data and benefit early stage anomaly detection in industry in general. The dataset used for training and testing is composed of real-time data gathered from the plant's sensors over a period of 15 days. For the purpose of anomaly detection, we used the following unsupervised machine learning algorithms: Interquartile Range (IQR), Isolation Forest, and Elliptic Envelope. The algorithms were compared with each other by the time required for training and the total number of detected anomalies. According to the observed results, it was shown that IQR needs much less time for training and detects three times more anomalies compared to the other two algorithms.

Keywords: Intelligent and AI Based Control, Fuzzy and Neural Systems, Robotics
Abstract: In this paper we propose an algorithm for trajectory tracking of a 2-Degrees of Freedom (2-DOF) robotic arm, using fuzzy logic controllers for each joint. The idea came from the enormous benefits of the potential use of intelligent robots and robotic elements in the medical field, as this type of robotic arms can be used in heart surgeries and other procedures. The proposed fuzzy logic controller scheme uses three linguistic variables as inputs (position, velocity and acceleration). This differs from the commonly used robotic arm path tracking control techniques that usually use position and velocity as inputs. Firstly, a detailed specification and modeling of the 2-DOF robot arm is provided. For the control of each joint, we build a separate fuzzy logic controller. The inputs for the both fuzzy logic controllers are the same, but they differ in the proposed rule bases and the appropriate membership functions. The performance of the proposed control scheme is assessed by calculating the trajectory tracking error, using Root Mean Square Error, for three different types of input signals (sinusoidal, square waveform and free waveform). Because of the potential use of this type of control in robotic heart surgery manipulators, the analysis of the tracking performance of the periodic sinusoidal signal is especially emphasized. According to the obtained simulation results, it can be concluded that the algorithm shows an adequate performance in all testing scenarios. A complete simulation environment is developed through the MATLAB/Simulink software.

Keywords: Discrete Event Systems, Control Applications, Modeling and Control of Complex Systems
Abstract: To reduce the sensor utilization and improve security in data communication without undermining the control decisions, the supervisory controller dynamically adjusts which set of events needed to observe in response to its previous observations. In previous work, building upon existing methods for computing a transition-based minimal sensor activation policy, algorithms for computing the set of all minimal transition-based sensor activation policies are developed. In this paper, we extend this result for networked supervisory controls, in which delays are involved in implementing control decisions.

Keywords: Discrete Event Systems, Networked Control
Abstract: When it comes to large-scale systems with hundreds of components, it is often impossible to create a centralized supervisor due to the state size explosion. The number of states in a modular system expands exponentially as the number of system components increases. As a result, the centralized supervisory control synthesis approach has a limitation; it may only be suitable for small systems. In this paper, we study the problem of distributed control of networked discrete event systems with communication delays, for both the control channel and the observation channel. Communication channel amongst distributed supervisors is built on the finite-state automata, expanding the earlier plant automata translation method into the distributed networked control framework. A more flexible modeling method is developed by incorporating time temporal limitations. By a model transformation, we transform distributed networked supervisor synthesis problem into the (non-networked) distributed supervisor synthesis problem (for non-deterministic plants), then existing tools and algorithms can be used for synthesizing distributed networked supervisors.

Keywords: Discrete Event Systems, Robotics
Abstract: This paper proposes a mission planning algorithm for the robot to operate in an environment with interference. The robot is equipped with proximity sensor to roughly estimate the environment and can apply active perception action to detect the environment. The goal of the robots is to accomplish complex tasks, captured by co-safe Linear Temporal Logic (scLTL) formulas. With the interference created by environment, general observation-based solution will cause a huge cost even if it can complete the task. When the robot without active perception encounters a set of states that the proximity observation can not tell, it will choose a conservative but costlier action. We aim to design an active perception plan for the robot to offset the impact which can reduce the cost. Our algorithm can trade off security and cost, to find a strategy to accomplish the complex task. We illustrate our method on motion planning case studies and show that the proposed algorithm can address complex planning tasks with smaller costs through active perception action.

Keywords: Discrete Event Systems, Estimation and Identification
Abstract: In order to save storage resource, our previous paper investigated the problem of compressing strings (i.e., only keeping part of the original string) while tracking the behaviour of the given discrete event system. Recoverability was first proposed, which requires the original string can be undoubtedly reconstructed from the compressed string. If an original string itself is recoverable, then we say it is compressible. However, our previous paper can only compress a compressible string. In this paper, a new notion of k-step lossless compression is proposed. It says that the last k events in a k-step compressed string are also the last k events of the original string and all events before might be compressed in any desirable way. For any given original string, our results show that there always exists the minimal k such that the given original string is k-step compressible. We propose algorithms to determine the minimal k and calculate a minimally k-step recoverable compressed string for the given original string.

Keywords: Control Applications, Discrete Event Systems, Modeling and Control of Complex Systems
Abstract: Using traffic lights at an intersection is to resolve the conflicts among different traffic paths while retaining the capacity of passing vehicles as much as possible. The traditional methods for scheduling signals at a signalized intersection ensure the former. Correspondingly, the approach used for this purpose usually consists of engineering design followed by a verification of correctness in resolving all conflicts. This paper presents a graph model used as a framework for scheduling traffic lights that guarantees both correctness and maximum flexibility.