Achieving average consensus without disclosing the initial agents' state is critical for secure multi- agent coordination. This paper proposes a novel privacy-preserving average consensus algorithm via a matrix-weighted inter-agent coupling mechanism. Specifically, the algorithm first lifts each agent state to a higher-dimensional space, then employs a dedicatedly designed matrix-valued state coupling mechanism to conceal the initial agents' state while guaranteeing that the multi-agent network achieves average consensus. The convergence analysis is transformed into the average consensus problem on matrix-weighted switching networks with low-rank, positive semi-definite coupling matrices. We show that the average consensus can be guaranteed and discuss its performance in the presence of honest-but-curious agents and external eavesdroppers. The algorithm, involving only basic matrix operations, is computationally more efficient than cryptography-based approaches and can be implemented without relying on a centralized third party. Numerical results are provided to illustrate the effectiveness of the algorithm. (c) 2024 Published by Elsevier Ltd.

The data-driven sliding mode control (SMC) method proves to be highly effective in addressing uncertainties and enhancing system performance. In our previous work, we implemented a co-design approach based on an input-mapping data-driven technique, which effectively improves the convergence rate through historical data compensation. However, this approach increases computational complexity in multi-input and multi-output (MIMO) systems due to the dependency of the number of online optimization variables on system dimensions. To improve applicability, this paper introduces a novel input-mapping-based online learning SMC strategy with low computational complexity. First, a new sliding mode surface is established through online convex combination of pre-designed offline surfaces. Then, an input-mapping-based online learning sliding mode control (IML-SMC) strategy is designed, utilizing a reaching law with adaptively adjusted convergence and switching coefficients to minimize chattering. The input-mapping technique employs the mapping relationship between historical input and output data for predicting future system dynamics. Accordingly, an optimization problem is formulated to learn from the past dynamics of the uncertain system online, thereby enhancing system performance. The optimization problem in this paper features fewer variables and is independent of system dimension. Additionally, the stability of the proposed method is theoretically validated, and the advantages are demonstrated through a MIMO system. Note to Practitioners-The design of control strategies that reduce the impact of mismatches between practical systems and models on system performance, while also ensuring applicability, is crucial. To address this issue, this paper proposes a low-complexity IML-SMC strategy. This strategy uses historical real input-output information and the mapping relationship with future dynamics to compensate for the impact of unknown dynamics and improve the system's convergence rate. Notably, the control strategy introduced in this paper significantly reduces online computational complexity, ensuring applicability, and stability is proven. When there is a deviation between the model and the actual system, practitioners can implement the low-complexity IMC-SMC strategy proposed in this paper to more quickly achieve the control objectives in the actual system.

The coordination of traffic flow among regions is necessary for a large-scale road traffic network to avoid local congestions and improve the overall traffic efficiency. In this paper, by incorporating the random characteristic of traffic flow, we formulate the problem of perimeter traffic flow control for a multi-region traffic network as a Markov decision process with adaptive state definition. Based on stochastic macroscopic fundamental diagrams (MFD) of regions, a state transition probability model is proposed to describe the state changes of the multi-region traffic network under different perimeter control policies. With the stochastic MFD-based state transition probabilities rather than counting from the historical data, a policy iteration algorithm with perturbation analysis is introduced to get the optimal perimeter control policy in real-time without the requirement of online or offline learning. The proposed method is compared with the classic perimeter control methods by simulation, which indicates its effectiveness in mitigating the congestion and improving the network throughput, as well as the promising implementation prospect.

The research on sliding mode control strategy typically utilizes a robust approach, which may come at the cost of sacrificing some performance because large parameter space is considered. Recently, the data-driven sliding mode control method has attracted much attention and shows excellent benefits in the fact that data is introduced to compensate the controller. Nevertheless, most of the research on data-driven sliding mode control relied on identification techniques, which limits its online applications due to the special requirements of the data. In this paper, an input-mapping technique is inserted into the design framework of sliding mode control to compensate for the influence generated by the unknown dynamic of the system. The principal novelty of the proposed input-mapping sliding mode control strategy lies in that the sliding mode surface and the sliding mode controller are co-designed through online learning from historical input–output data to minimize an objective function. Then, the convergence rate of the system is improved significantly based on the method designed in this work. Finally, some simulations are provided to show the effectiveness and superiority of the proposed methods.

This paper examines the event-triggered global consensus of matrix-weighted networks subject to actuator saturation. A distributed protocol design is proposed for this category of networks to guarantee its global consensus subject to both event-triggered communication and actuator saturation. It is shown that the largest singular value of matrix-valued edge weights plays a crucial role in both protocol design and network performance, which renders the proposed framework more general than existing results that are only applicable to scalar-weighted networks. Conditions under which the global consensus can be guaranteed for leaderless matrix-weighted multi-agent networks are derived. However, the average consensus on the initial agents' states cannot be achieved due to the nonlinearities introduced in the closed-loop dynamics by the actuator saturation constraint. We further examine the scenario of leader-follower consensus for matrix-weighted multi-agent networks under the constraints of both event-triggered communication and actuator saturation. The applicability of the proposed protocol design framework to time-varying matrix-weighted networks is also examined. It is shown that the Zeno phenomenon can be excluded under the proposed interaction protocols. Simulation results in the context of the bearing-only cooperative formation of multi-vehicle systems are provided to demonstrate the effectiveness of theoretical results.

In image-based visual servoing (IBVS), parametric uncertainties tend to cause the model inaccuracy and limit the control performance. Considering these uncertainties can be embodied by the output–input data from the visual servoing system, this brief proposes an eye-in-hand visual servoing control (VSC) scheme based on the input mapping method, which directly utilizes the past output–input data to enhance the original feedback control law rather than identifying the model. The system with the input mapping method is proven to not only maintain the stability of the original VSC but also accelerate the convergent rate. The results of the experiments on a manipulator with an eye-in-hand camera demonstrate the superiority of our proposed method.

The optimization problem in model predictive control (MPC) algorithm for piecewise affine (PWA) systems often contains substantial logic variables, which requires extensive computing power for online implementation. This paper proposes a one-step distributed MPC algorithm for spatially interconnected PWA systems with both input and state constraints, where the algorithm is based on a series of r-step robustly controllable sets. Algorithms for offline computing terminal sets and a series of r-step robustly controllable sets are developed using the step-by-step backward approach. The states that lie in the robustly controllable sets are steered into the terminal set in finite steps and ultimately converge into a robustly positively invariant set. The proposed algorithm is demonstrated to significantly reduce the online computational burden. Simulation results are provided to showcase the effectiveness of the algorithm.

AbstractA robust tube‐based distributed model predictive control method is proposed for spatially interconnected systems with constraints and disturbances. The system contains multiple discrete‐time piecewise affine subsystems, which are coupled to each other through states. The predictive states of each subsystem are dependent on its states, inputs, and neighbouring states, and shrinking constraints are constructed to deal with the disturbances and the model mismatches. The differences between the predicted states at the current time instant and the optimal states at the previous time instant are constrained, which improves the accuracy of the predicted model and enhances the control performance. The conditions which ensure the recursive feasibility of optimization problems and the asymptotic stability of closed‐loop system are obtained. The effectiveness of the proposed method is verified by the simulation results.

This article proposes a synthetic robust model predictive control method (RMPC) with input mapping for the image-based visual servoing (IBVS) problem with constraints, where the novel control law is constructed by the robust control law designed offline and the online linear compensation of the past data. This proposed method can overcome the conservatism of RMPC and reduce the online computational burden. The input mapping method is suitable for the IBVS with no requirement of the slow time-varying model or time-invariant model as most adaptive control methods need. Its linear combination coefficients can be online optimized by solving a quadratic programming problem. The stability of the visual servoing system under our proposed method is proven, and its convergence speed is demonstrated to be faster than the traditional RMPC. A real-time experiment on a six-degree-of-freedom manipulator with eye-in-hand construction is designed to evaluate the proposed method. The results indicate that besides the ability to handle the constraint and the singularity problem, our proposed method provides a faster convergence rate than several classic robust control methods, and improves the computational efficiency by an order of magnitude compared with the online RMPC.

Achieving consensus via nearest neighbor rules is an important prerequisite for multiagent networks to accomplish collective tasks. A common assumption in consensus setup is that each agent interacts with all its neighbors. This article examines whether network functionality and performance can be maintained—and even enhanced—when agents interact only with a subset of their respective (available) neighbors. As shown in this article, the answer to this inquiry is affirmative. In this direction, we show that by exploring the monotonicity property of the Laplacian eigenvectors, a neighbor selection rule with guaranteed performance enhancements can be realized for consensus-type networks. For distributed implementation, a quantitative connection between entries of Laplacian eigenvectors and the “relative rate of change” in the state between neighboring agents is further established; this connection facilitates a distributed algorithm for each agent to identify “favorable” neighbors to interact with. Multiagent networks with and without external influence are examined, as well as extensions to signed networks. This article underscores the utility of Laplacian eigenvectors in the context of distributed neighbor selection, providing novel insights into distributed data-driven control of multiagent systems.

The unsupervised anomaly detection in high-dimensional and complex settings poses a formidable challenge. To tackle the challenges associated with the recognition of high-dimensional data, this paper proposes a feedback channel between the Variational Auto Encoder and the Gaussian process to enhance its data feature extraction capabilities. In order to alleviate the impact of common mislabeling issues within unsupervised contexts, the proposed model introduces a method for the selection of representative data. Subsequently, these representative data points are employed as anchors for the construction of a full-data Gaussian process model covering the entire dataset. To further disentangle the intertwined normal and abnormal data, the proposed model employs a Bayesian framework to guide the displacement of each data point within the latent space toward its position with the highest confidence. This strategic approach effectively discriminates between normal and abnormal data. This innovative method significantly enhances data consolidation within the latent space, thus preserving robustness, particularly when processing high-complexity data. Also, the numerical experiments not only validate the method’s effectiveness but also showcase the model’s robustness. These experiments highlight that even a small number of incorrect labels do not significantly impact the accuracy of anomaly detection.

This article examines the distributed nonconvex optimization problem with structured nonconvex objective functions and coupled convex inequality constraints on static networks. A distributed continuous-time primal-dual algorithm is proposed to solve the problem. We use the canonical transformation and Lagrange multiplier method to reformulate the nonconvex optimization problem as a convex–concave saddle point computation problem, which is subsequently solved by employing the projected primal-dual subgradient method. Sufficient conditions that guarantee the global optimality of the solution generated by the proposed algorithm are provided. Numerical and application examples are presented to demonstrate the proposed algorithm.

This paper examines cluster consensus for multi-agent systems on matrix-weighted switching networks. Necessary and/or sufficient conditions under which cluster consensus can be achieved are obtained, as well as quantitative characterization of the steady-state of the cluster consensus. Specifically, when the underlying network switches amongst a finite number of networks, a necessary condition for cluster consensus on matrix-weighted switching networks is derived; moreover, it is shown that the steady-state of the nodes lies at the intersection of the null spaces of the matrix-valued Laplacians of the corresponding switching networks. Furthermore, when the underlying network switches amongst an infinite number of networks, the matrix-weighted integral network is employed to provide sufficient conditions for cluster consensus; analogous to the previous case, the quantitative characterization of the corresponding steady-state of the nodes pertains to the null space analysis of matrix-valued Laplacian of the “integral” network. Lastly, conditions for bipartite consensus of matrix-weighted switching networks are provided. Simulation examples demonstrate the presented theoretical results.

The data-driven model predictive control (MPC) approach has been an effective tool for unknown constrained systems. However, most of the existing designs rely on the prior collected data sequence through offline trials, which may be affected by time-varying disturbances, or the online estimated system model with the non-trivial computation cost. This limits the applications of these designs. To alleviate these restrictions, in this paper, an input-mapping data-driven scheme is developed. This scheme online directly maps the future control policy and the predicted state to the past online noisy input/state data which are updated once new data come. This overcomes the limitations of previous designs. To ensure the system constraints, this scheme is combined with a tube MPC method and the proposed input-mapping data-driven tube MPC can guarantee the recursive feasibility and the input-to state stability. Two examples illustrate the advantages of the proposed method.

The paper proposes an adaptive control algorithm for the uncalibrated position-based visual servoing to complete the tracking task of robot system. The desired and current poses of the end-effector are obtained by the camera and the system error is described in the camera coordination, which can avoid the work of hand-eye calibration. The connection matrix between the robot and the end-effector as well as the camera measurement bias are estimated together by the input/output data in the online control process, which can avoid the calibration between the robot terminal and the end-effector and reduce the influence of the measurement bias. By our proposed algorithm, the controller can update the model. Moreover, the control law and the estimation updated law are designed by the Lyapunov theory, which is proved that the system error and the parameter estimation error can be closed to zero. To verify the performance of the proposed controller, a numerical simulation experiment on a 6 degrees-offreedom robot manipulator equipped with the end-effector and the eye-to-hand camera under our proposed control method is designed and presented.

The increasing prevalence of connected and autonomous vehicles (CAVs) offers new opportunities for the traffic regulation of the mixed traffic flow consisting of CAVs and regular vehicles. Understanding traffic dynamics of mixed traffic flow with the speed drops of micro CAVs is necessary for the new mode of traffic management. This paper presents a regulation model to calculate the macroscopic traffic flow parameters for the multi-lane mixed traffic flow in the existence of one slowdown CAV from the wave propagation perspective. The propagation of kinetic waves goes through three stages generated by speed difference and random lane changing behavior due to local deceleration. The wave interference principle is proposed to model the superposition of different types of waves in the scene of multiple speed drops. The combination of the regulation model and the wave interference principle describes the macroscopic flow propagation process in lane level under the speed drops of micro CAVs with different distributions. The accuracy of the model output was over 90%, and the maximum error of the average speed was 2.4 km/h with experiments performed in SUMO_CACC traffic simulator. The proposed model gives a relatively accurate dynamic process, making it suitable for further applications of traffic regulation.

Achieving average consensus without disclosing sensitive information can be a critical concern for multi-agent coordination. This paper examines privacy-preserving average consensus (PPAC) for vector-valued multi-agent networks. In particular, a set of agents with vector-valued states aim to collaboratively reach an exact average consensus of their initial states, while each agent's initial state cannot be disclosed to other agents. We show that the vector-valued PPAC problem can be solved via associated matrix-weighted networks with the higher-dimensional agent state. Specifically, a novel distributed vector-valued PPAC algorithm is proposed by lifting the agent-state to higher-dimensional space and designing the associated matrix-weighted network with dynamic, low-rank, positive semi-definite coupling matrices to both conceal the vector-valued agent state and guarantee that the multi-agent network asymptotically converges to the average consensus. Essentially, the convergence analysis can be transformed into the average consensus problem on switching matrix-weighted networks. We show that the exact average consensus can be guaranteed and the initial agents' states can be kept private if each agent has at least one "legitimate" neighbor. The algorithm, involving only basic matrix operations, is computationally more efficient than cryptography-based approaches and can be implemented in a fully distributed manner without relying on a third party. Numerical simulation is provided to illustrate the effectiveness of the proposed algorithm.

In this work, we develop an event-triggered distributed robust model predictive control algorithm to stabilize the origin of a class of interconnected systems with nonlinear coupling terms and additive bounded disturbances. Each subsystem is assumed to be able to exchange information with its neighboring subsystems. The transmitted information between neighbors is used to estimate their future behaviors. Then, a robust distributed model predictive control algorithm for each subsystem is developed by integrating the estimations of its neighbors and each subsystem executes a local model predictive control law after solving its optimization problem. Furthermore, a distributed event-triggered mechanism is designed to trigger a series of asynchronous computations of the optimization problems, which achieves a trade-off between communication resource usage and control performance. Theoretical conditions on ensuring feasibility and closed-loop stability are provided. Finally, a practical example of the multi-machine power system with governor controllers is provided to show the effectiveness of the proposed algorithm.

The hydrogen gas turbine can be the key technology for carbon-neutral in the future. The flexibility of gas turbine under various operation conditions is an important issue. A state-space model of hydrogen gas turbine with different hydrogen volume ratio mixed fuel is constructed based on Rowen model, and a full-dimensional state observer is designed to obtain the states of state-space model. The model predictive control method based on amplitude decaying aggregation strategy is presented to stabilize the rotational speed of hydrogen gas turbine under various constraints, which are to ensure the safety of operation. The number of variables contained in the optimization problem to be solved online is reduced by aggregation strategy. The simulation results confirmed that the presented method works well, and the effects of the hydrogen volume ratio for the performance of hydrogen gas turbine are discussed.

A critical prerequisite for controlling complex networks is to find a driver node set with a structural controllability guarantee. This paper introduces the control capacity for a driver node set and solves the problem of finding a complete minimum driver node set that not only guarantees network structural controllability but achieves the desired level of control capacity. A novel algorithmic framework is proposed which is based on the concept of equivalent set and approximate matching replacement technique. The proposed algorithmic framework is shown to outperform the state-of-the-art approaches in the literature. The validity of the proposed algorithm is analyzed and the performance is evaluated by experiments on artificial and real-world complex networks.