Control Strategy For Multi-agent Systems Research Articles

Optimal leader-following consensus under switching topologies based on event-triggered NN-based observer

This paper proposes a novel control strategy for multi-agent systems (MASs) to achieve optimal leader-following consensus under jointly connected topologies. The devised approach incorporates two key components: an event-triggered neural network (NN)-based leader observer and an NN-based optimal controller. The proposed leader observer allows followers to locally reconstruct the leader’s dynamics, even in scenarios where the communication topology is jointly connected. The implementation of an event-triggered communication protocol significantly reduces the communication burden, enhancing the overall efficiency of information exchange. As the observer’s state converges to the actual leader’s state, the simplification of achieving leader-follower consensus becomes evident by directly setting the state of the leader observer as the tracking goal. An NN-based optimal controller is utilized to approximate the solution to the Hamilton–Jacobi–Bellman (HJB) equation, thus enabling optimal tracking of the leader by followers. Finally, the performance of the proposed control strategy is verified by simulations.

A Deep Reinforcement Learning Approach Combined With Model-Based Paradigms for Multiagent Formation Control With Collision Avoidance

Generating collision-free formation control strategy for multiagent systems faces huge challenges in collaborative navigation tasks, especially in a highly dynamic and uncertain environment. Two typical methodologies for solving this problem are the conventional model-based paradigm and the data-driven paradigm, particularly the widely used deep reinforcement learning (DRL) method. However, both the model-based and data-driven paradigms encounter inherent drawbacks. In this paper, we present two novel general schemes that combine these two paradigms together in an online mode. Specifically, the two paradigms are combined in a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">parallel</i> and a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">serial</i> structure in these two schemes, respectively. In the parallel scheme, the outputs of the model-based and DRL-based controllers are lumped together. In the serial scheme, the output of the model-based controller is fed as an input of the DRL-based controller. The interpretation of the two combined schemes is suggested from a control-oriented perspective, where the parallel DRL controller is viewed as a complementary uncertainty compensator and the serial DRL controller is taken as an inverse dynamics estimator. Finally, comprehensive simulations are conducted to demonstrate the superiority of the proposed schemes, and the effectiveness is further verified by deploying our schemes to a physical experiment platform based on a set of three-wheeled omnidirectional robots.

Reinforcement Learning-Based Control Strategy for Multi-Agent Systems Subjected to Actuator Cyberattacks During Affine Formation Maneuvers

Reinforcement Learning-Based Control Strategy for Multi-Agent Systems Subjected to Actuator Cyberattacks During Affine Formation Maneuvers

Safety control using barrier certificates for multiagent systems with input saturation and formation constraints

Maintaining the given operational area is critical in guaranteeing the safety of nonlinear second-order multiple autonomous agents. The properties of multiagent systems and several physical constraints, including bounded modeling error and actuator saturation, dramatically affect the maneuverability of multiagent systems inside the specified operational area. Moreover, the existing safety control algorithms heavily rely on the boundaries of the operational area. To overcome this issue, by constructing a novel scalable control technique, the safety area for multiagent systems can be transformed into input-constrained control barriers along each coordinate of motion for agents. It is shown that the safety of each agent and the global asymptotic stability are guaranteed under the proposed distributed control algorithm. The asymmetrical closed-form scheme for the agent's safety rule is built by applying the adjustable low and high bounds of the control signals associated with the actual control inputs, which are repeatedly computed by using new local measurements as the agents move, and the saturated control inputs with asymmetrical constraints are ensured. The absolute values of the modeling errors and external disturbances can be tracked by the proposed safety controller. Super-twisting control (STC) is employed to address the formation constraint problem of multiagent systems, where the effect that arises from uncertain nonlinear complexity of the agents and external disturbances is eliminated. Moreover, finite-time convergence, a desirable robust behavior of multiagent systems, and the formation constraint are simultaneously achieved. Furthermore, the stability of the proposed integrated control strategy for multiagent systems is analyzed, which reveals that the proposed distributed safety control can seamlessly integrate with the robust control protocol with minimum modification under the directed information interaction topology. Safety formation control calibration and tuning are carried out, and comparative simulation results are provided to illustrate the effective performance of the obtained theoretical results.

Adaptive Multi-Agent Coverage Control With Obstacle Avoidance

This letter presents an adaptive coverage control strategy for multi-agent systems with obstacle avoidance in the presence of actuator faults and time-varying uncertainties. The strategy is based on a leader-follower approach. Assuming that the motion of the leader is given, one distributes the followers within the leader's obstacle-free sensing range so that collisions with obstacles can be avoided. An optimized distribution is achieved through the Centroidal Voronoi Tessellation (CVT) and a function approximation technique based immersion and invariance (FATII) coverage controller is constructed to realize the CVT. The stability of the FATII coverage controller is established and its validity is tested by simulations.

A New Decentralized Control Strategy of Microgrids in the Internet of Energy Paradigm

The Energy Internet paradigm is the evolution of the Internet of Things concept in the power system. Microgrids (MGs), as the essential element in an Energy Internet, are expected to be controlled in a corporative and flexible manner. This paper proposes a novel decentralized robust control strategy for multi-agent systems (MASs) governed MGs in future Energy Internet. The proposed controller is based on a consensus algorithm applied with the connected distributed generators (DGs) in the MGs in the energy internet paradigm. The proposed controller’s objectives are the frequency/voltage regulation and proportional reactive/active power-sharing for the hybrid DGs connected MGs. A proposed two-level communication system is implemented to explain the data exchange between the MG system and the cloud server. The local communication level utilizes the transmission control protocol (TCP)/ internet protocol (IP) and the message queuing telemetry transport (MQTT) is used as the protocol for the global communication level. The proposed control strategy has been verified using a hypothetical hybrid DGs connected MG such as photovoltaic or wind turbines in MATLAB Simulink environment. Several scenarios based on the system load types are implemented using residential buildings and small commercial outlets. The simulation results have verified the feasibility and effectiveness of the introduced strategy for the MGs’ various operating conditions.

A Robust Formation Control Strategy for Multi-Agent Systems with Uncertainties via Adaptive Gain Robust Controllers

This paper deals with a design problem of an adaptive gain robust controller which achieves consensus for multi-agent system (MAS) with uncertainties. In the proposed controller design approach, the relative position between the leader and followers are considered explicitly, and the proposed adaptive gain robust controller consisting of fixed gains and variable ones tuned by time-varying adjustable parameters can reduce the effect of uncertainties. In this paper, we show that sufficient conditions for the existence of the proposed adaptive gain robust controller are reduced to solvability of linear matrix inequalities (LMIs). Finally, the effectiveness of the proposed robust formation control system is verified by simple numerical simulations. A main result of this study is that the proposed adaptive gain robust controller can achieve consensus and formation control giving consideration to relative distance in spite of uncertainties.

Hybrid Event-Triggered and Impulsive Control Strategy for Multiagent Systems With Switching Topologies.

This article investigates the hybrid event-triggered and impulsive consensus problems for leaderless and leader-following multiagent systems (MASs) with switching topologies. Based on the state information of neighboring agents at event-triggered moments and impulsive instants, a hybrid event-triggered and impulsive control strategy (HETICS) is designed to reduce the communication frequency between neighboring agents and to ensure consensus of leaderless and leader-following MASs. By utilizing the Lyapunov direct method, some consensus criteria are obtained for leaderless and leader-following MASs with switching topologies. It is shown that the HETICS excludes the Zeno behavior. Several numerical examples and simulations are given to illustrate the effectiveness of the proposed consensus strategy and a comparison with previous consensus control methods is given.

Predefined containment control for general linear multiagent systems with time‐varying delays and switching topologies

AbstractContainment control problems for general linear multiagent systems with switching interaction topologies and time‐varying delays are studied. Firstly, state observers are constructed to estimate the multiple leaders under the influence of switching graphs and time‐varying delays. Secondly, a predefined containment protocol is presented based on the distributed state observer, where the expected convex combination of multiple leaders is predefined by several given weights. Moreover, an algorithm with two steps to confirm the gain matrices of the containment protocol and state observer is given. Based on linear matrix inequality technique and common Lyapunov‐Krasovskii stability theory, the convergence of this control strategy for multiagent systems with switching topologies and time‐varying delays to achieve state containment is proved. Finally, a numerical simulation is given to verify the validity of the theoretical results.

Event-triggered Finite-time Consensus under Directed Graphs

Abstract This paper focuses on deal with the finite-time consensus with event-triggered control strategy for multi-agent systems (MASs). An event-triggered protocol for finite-time consensus is designed using relative measurements. The coordination measurement error is utilized in the triggering condition design for the purpose of removing the prerequisite of topology graph knowledge. Under strongly connected graph assumptions, by utilizing the proposed consensus protocol, all agents can complete consensus and Zeno behaviour will not happen in a settling time. Next, by decomposing the Laplacian matrix in Frobenius norm form, the results are extended to the more general graphs containing a directed spanning tree. At last, a numerical example demonstrates the validity of the algorithm results.

ICS-Zooids - An Experimental Testbed for Cooperative Control Strategies

Abstract In this paper, we present an experimental test bench to implement various cooperative control strategies for multi-agent systems, and illustrate its use with experimental results for a source-seeking problem, where a group of small wheeled robots termed as Zooids should locate a source of a given spatial scalar field. This algorithm is implemented as a validation to demonstrate the capabilities of the test bench. We propose to achieve this by utilising an internal target-based position controller, under the assumptions of convexity of the scalar, continuous/discrete field and availability of local measurements of the field, so that agents can calculate its gradient and its Hessian. We then show in experiments, that using estimated gradients and Hessians (with data communicated from neighbours) in the presence of noisy measurements of the field strength provides satisfactory results for convex fields, under various algorithms such as Steepest Descent, Gauss-Newton, Levenberg Marquardt. These algorithms are analysed, and experimental results are discussed.

Feedback control strategies for multi-agent systems under a fragment of signal temporal logic tasks

Multi-agent systems under temporal logic tasks have great potential due to their ability to deal with complex tasks. The control of these systems, however, poses many challenges and the majority of existing approaches result in large computational burdens. We instead propose computationally-efficient and robust feedback control strategies for a class of systems that are, in a sense, feedback equivalent to single integrator systems, but where the dynamics are partially unknown for the control design. A bottom-up scenario is considered in which each agent is subject to a local task from a limited signal temporal logic fragment. Notably, the satisfaction of a local task may also depend on the behavior of other agents. We provide local continuous-time feedback control laws that, under some sufficient conditions, guarantee satisfaction of the local tasks. Otherwise, a local detection & repair scheme is proposed in combination with the previously derived feedback control laws to deal with infeasibilities, such as when local tasks are conflicting. The efficacy of the proposed method is demonstrated in simulations.

Model-Free Cooperative Adaptive Sliding-Mode-Constrained-Control for Multiple Linear Induction Traction Systems.

In order to deal with the speed cooperative control problem in the multiple linear induction traction systems consists of multiple linear induction motors, a model-free cooperative adaptive sliding-mode-constrained-control strategy is proposed considering the input magnitude and rate constraints which may cause the problem of actuator and integral saturation. First, the equivalent circuit topology of the single motor in the system is investigated. Besides, the system is considered as the multiagent system with fixed communication topology due to the interaction between adjacent motors. Then, the output observer is presented to estimate the output and the estimation algorithm of pseudo-partial derivative parameter and uncertainties is proposed. Based on the above, the proposed control scheme is presented by designing an integral sliding-mode surface containing the systematic error and an anti-windup compensator is added to eliminate the saturation. Finally, the simulations of the proposed control strategy for multiagent systems are carried out to demonstrate the effectiveness and superiority of the proposed control strategy.

Hybrid Control of Multiagent Systems With Contingent Temporal Tasks and Prescribed Formation Constraints

In this paper, we present a distributed hybrid control strategy for multiagent systems with contingent temporal tasks and prescribed formation constraints. Each agent is assigned a local task given as a linear temporal logic formula. In addition, two commonly seen kinds of cooperative robotic tasks, namely, service and formation, are requested and exchanged among the agents in real time. The service request is a short-term task provided by one agent to another. On the other hand, the formation request is a relative deployment requirement with predefined transient response imposed by an associated performance function. The proposed hybrid control strategy consists of four major components: 1) the contingent requests handling module; 2) the real-time events monitoring module; 3) the local discrete plan synthesis module; and 4) the continuous control switching module, and it is shown that all local tasks and contingent service/formation requests are fulfilled. Finally, a simulated paradigm demonstrates the proposed control strategy.

A novel finite-time average consensus protocol based on event-triggered nonlinear control strategy for multiagent systems

We present a novel finite-time average consensus protocol based on event-triggered control strategy for multiagent systems. The system stability is proved. The lower bound of the interevent time is obtained to guarantee that there is no Zeno behavior. Moreover, the upper bound of the convergence time is obtained. The relationship between the convergence time and protocol parameter with initial state is analyzed. Lastly, simulations are conducted to verify the effectiveness of the results.

Extremum Seeking Based Fault-Tolerant Cooperative Control for Multiagent Systems

We propose a novel fault-tolerant cooperative control strategy for multiagent systems. A set of unknown input observers for each agent are constructed for fault detection. Then a real-time adaptive extremum seeking algorithm is utilized for adaptive approximation of fault parameter. We prove that the consensus can be still reached by regulating the interconnection weights and changing the connection topology of the fault agent. A numerical simulation example is given to illustrate the feasibility and effectiveness of the proposed method.

Combined Top-down and Bottom-up Approach to Cooperative Distributed Multi-agent Control with Connectivity Constraints

We propose a novel framework to synthesize coordination and control strategies for multi-agent systems from team tasks that are in the form of regular languages specifying service requests. The framework incorporates top-down design approaches with bottom-up control schemes: on one hand, a formal approach is applied to distribute the local tasks based on service capabilities of each agent and to synthesize local supervisors associated with each agent to fulfill the local tasks; on the other hand, by modeling the physical motion dynamics of an agent as a single integrator system, a connectivity-preserving controller with only position measurements is designed such that motion behaviors of cooperating agents are subject to communication constraints. The performance of the underlying framework is guaranteed by using a technique that is inspired by compositional verification while the communication constraints can always be obeyed. A motion and mission co-planning example of a robotic system is presented.

Optimal and event-based networked control of physically interconnected systems and multi-agent systems

Many interconnected systems like vehicle platoons or energy networks consist of similar or identical subsystems. The subsystem interconnections are either caused by the physical relations among the subsystems or have to be introduced by the controller to cope with cooperative control goals. This paper proposes strategies to reduce the complexity of the controller design problem (offline information reduction) and to reduce the amount of the system information, which is necessary for the implementation of the designed controller (online information reduction). It consists of two parts. The first part deals with the linear quadratic regulator (LQR) design problem for interconnected systems. A decomposition based on a state transformation is introduced, which allows to design the optimal controller for the interconnected system by considering modified subsystems separately. The proposed decomposition approach can be uniformly applied to multi-agent systems and physically interconnected systems.The second part of the paper introduces an event-based control strategy for multi-agent systems. The event-based control is a means to reduce the communication effort by invoking an information exchange among the subsystems only when the deviation between the estimated and current subsystem state exceeds an event threshold. An event-based controller is proposed, which mimics the continuous state-feedback controller with a desired precision. The relation between the event threshold and the approximation error is analysed.

Special Issue on Advances in Agent-mediated Automated Negotiations

Complex Automated Negotiations have been studied very widely and becoming an emerging area in the field of Autonomous Agents and Multi-Agent Systems. In general, a negotiation could be complex because many factors need to be considered to specify such a negotiation. These factors may include a number of negotiation issues, dependency of issues, representation of utilities, negotiation protocols, negotiation forms (bilateral or multi-party), time constraints, etc. In complex situations, agents are not always selfish but rather cooperative in order to achieve realistic and win-win agreements. In a complex negotiation, there could be multiple issues that are interdependent. Thus, agent’s utilities will become more complex than simple utility functions. Furthermore, negotiation forms and protocols could be different between bilateral situations and multi-party situations. To realize such a complex automated negotiation, we have to incorporate strong agent technologies including search, CSP, graphical utility models, Bays nets, auctions, utility graphs, prediction and learning methods. This special issue brings together a collection of papers addressing issues of designing both negotiation mechanisms and scheduling and control strategies for multi-agent systems into the same forum. This special issue includes 7 papers. Each paper was reviewed by 3 experts in the area. Some of the issues addressed in these 7 papers include:

Swarm Tracking Using Artificial Potentials and Sliding Mode Control

In this paper, we present a stable and decentralized control strategy for multiagent systems (swarms) to capture a moving target in a specific formation. The coordination framework uses artificial potentials to take care of both tracking and formation tasks. First, a basic controller is designed based on a kinematic model. After that, sliding mode control technique is used to force the agents with general vehicle dynamics to obey the required motion. Finally, specific potential functions are discussed and corresponding simulation results are given.