Multi-agent Research Articles

Feedforward compensator based fuzzy adaptive asymptotic consensus for nonlinear multiagent systems

Feedforward compensator based fuzzy adaptive asymptotic consensus for nonlinear multiagent systems

Distributed formation tracking with connectivity preservation and orientation alignment

This paper investigates the formation tracking problem incorporating connectivity preservation and orientation alignment for multi-agent systems. Consider the body-fixed frame of each agent in the multi-agent system is initialized with different orientations. To achieve formation controller design by relative position in the body-fixed frame, the orientation of the frames should be aligned first. A distributed passive-based orientation controller is proposed to ensure convergence of all body-fixed frames toward a unified desired orientation. Once the frames are aligned, a barrier Lyapunov function (BLF)-based formation tracking controller is proposed utilizing the displacement of the agent to ensure the distances between agents always satisfy predefined distance constraints. This strategy guarantees inter-agent collision avoidance and connectivity preservation throughout the formation. Numerical simulations are conducted to validate the efficacy of the proposed approach.

Enhancing Flexibility and Reliability in Smart Distribution Networks: A Self-Healing Approach

This paper presents a pioneering approach that integrates a Self-Healing System through a Smart Distribution Network, which employs a two-step implementation of a comprehensive Energy Management algorithm and a multi-agent system with an autonomous distributed regeneration method. The novel method improves network operation through optimized management of local microgrids, which include diverse components comprising renewable energy sources, electric vehicle charging infrastructure, and energy storage systems. The first phase concentrates on minimizing network operation costs by adhering to system utilization restrictions and optimal load distribution, while the second phase addresses the optimization of regeneration processes, which decreases discrepancies between recoverable priority loads and switching operations. The evaluation introduces a stochastic programming approach to model and manage uncertainties, and utilizes the Kantorovich method and the roulette wheel mechanism to improve reliability. ...

R-DOCO: Resilient Distributed Online Convex Optimization Against Adversarial Attacks

This paper addresses the problem of distributed constrained optimization in a multi-agent system where some agents may deviate from the prescribed update rules due to failures or malicious adversarial attacks. The objective is to minimize the collective cost of the unattacked agents while respecting the constraint limitations. To tackle this, we propose a resilient distributed projected gradient descent algorithm for online optimization that achieves sublinear individual regret, defined as the difference between the online and offline solutions. Additionally, we extend the cost function from convex combinations to more general distributed optimization scenarios. The proposed algorithm demonstrates resilience under adversarial conditions, allowing it to handle an unknown number of adversarial nodes while maintaining performance. Compared to existing methods, this approach offers a robust solution to adversarial attacks in constrained distributed optimization problems.

Consensus analysis for heterogeneous multiagent systems subject to input saturation

AbstractThis paper investigates the consensus problem of heterogeneous multiagent systems (HMASs) consisting of first‐order and second‐order agents, where second‐order linear and second‐order nonlinear agents are included in the second‐order agents. Firstly, the consensus problem of HMASs without input saturation is studied in the first part. However, in practical applications, due to the physical limitations of actuators, saturation often occurs. Therefore, we focus on the consensus problem of HMASs with input saturation in the second part. Then based on the knowledge of Lyapunov stability theory and graph theory, it gives sufficient conditions for these HMASs to reach consensus, which ensures that each agent eventually reaches consensus. Eventually, the effectiveness of the theory is verified by performing simulations.

Securing Consensus in Fractional-Order Multi-Agent Systems: Algebraic Approaches Against Byzantine Attacks

This paper investigates the behavior of fractional-order nonlinear multi-agent systems subjected to Byzantine assaults, specifically focusing on the manipulations of both sensors and actuators. We employ weighted graphs, both directed and undirected, to illustrate the system's topology. Our methodology combines algebraic graph theory with fractional-order Lyapunov techniques to develop algebraic requirements for leader-following consensus, providing a robust framework for analyzing consensus dynamics in these complex systems. We present quantitative results demonstrating the effectiveness of our approach, including two numerical examples that validate the proposed requirements for consensus evaluation. Notably, our work highlights the novelty of using fractional-order systems to enhance resilience against adversarial conditions, contributing significantly to the field of multi-agent systems. By clarifying key terms and streamlining our language, we ensure accessibility for a broader audience while emphasizing the implications of our findings for real-world applications.

Adaptive finite-time consensus tracking for nonlinear second-order multi-agent systems based on integral sliding mode

This paper delves into the investigation of finite-time consensus for second-order nonlinear multi-agent systems (MASs) with external disturbances under directed topology. The MASs considered in this study consist of n followers and a leader, all of which are subject to bounded disturbances. First, a continuous nonsingular integral terminal sliding mode is designed, which can effectively eliminate the singularity and chattering. Then, a finite-time disturbance observer is introduced to estimate and compensate for disturbances. Subsequently, an integral sliding mode adaptive controller is designed to enhance system’s robustness, improve response speed, and increase tracking accuracy. Furthermore, Lyapunov theory is utilized to demonstrate that the system achieves finite-time consensus under a directed connected topology. Finally, we apply a simulation to verify the efficacy of the proposed consensus control protocol.

A multi-agent system to dynamically devise an LCA framework weighting system taking into account socio-technical and environmental consideration

This research article explores the need to adapt the weighting system of a life cycle assessment (LCA) framework to a wide range of socio-technical and environmental considerations, which are complex and sensitive to external stressors. The study demonstrates the potential of agent-based systems to adapt a weighting system to dynamic conditions, and suggests integrating an agent-based model to handle uncertainties of weighting in environmental impact assessment. Generative adversarial networks (GANs) and multi-agent systems (MAS) are utilized to address data limitations and simulate diverse scenarios. We confirm the potential of agent-based systems to analyse the effective management of uncertainties and customization of weighting systems in environmental impact assessments to improve decision making. Furthermore, the study emphasizes the importance of continuous adaptation and recalibration to ensure the system remains relevant in dynamic environments. The results confirm that MAS is a powerful tool for managing uncertainty, customizing weighting systems and improving decision making in environmental assessments. Moreover, the study acknowledges challenges and sets the groundwork for future research.

An integrated decision-execution framework of cooperative control for multi-agent systems via reinforcement learning

Cooperative control is both a crucial and hot research topic for multi-agent systems (MASs). However, most existing cooperative control strategies guarantee tracking stability under various non-ideal conditions, while the path decision capability is often ignored. In this paper, the integrated decision-execution (IDE) framework is newly presented for cooperative control of multi-agent systems (MASs) to accomplish the integrated task of path decision and cooperative execution. This framework includes a decision layer and a control layer. The decision layer generates a continuous trajectory for the virtual leader to reach the target from its initial position in an unknown environment. To achieve the goal of this layer, (1) the Step-based Adaptive Search Q-learning (SASQ-learning) algorithm is proposed based on reinforcement learning to efficiently find the discrete path, (2) an Axis-based Trajectory Fitting (ATF) method is developed to convert the discrete path into a continuous trajectory for mobile agents. In the control layer, this trajectory is used to regulate the following MASs to achieve cooperative tracking control with the presence of input saturation. Simulation experiments are presented to demonstrate the effectiveness of this framework.

Reinforcement Learning for Synchronization of Heterogeneous Multiagent Systems by Improved Q-Functions.

This article dedicates to investigating a methodology for enhancing adaptability to environmental changes of reinforcement learning (RL) techniques with data efficiency, by which a joint control protocol is learned using only data for multiagent systems (MASs). Thus, all followers are able to synchronize themselves with the leader and minimize their individual performance. To this end, an optimal synchronization problem of heterogeneous MASs is first formulated, and then an arbitration RL mechanism is developed for well addressing key challenges faced by the current RL techniques, that is, insufficient data and environmental changes. In the developed mechanism, an improved Q-function with an arbitration factor is designed for accommodating the fact that control protocols tend to be made by historic experiences and instinctive decision-making, such that the degree of control over agents' behaviors can be adaptively allocated by on-policy and off-policy RL techniques for the optimal multiagent synchronization problem. Finally, an arbitration RL algorithm with critic-only neural networks is proposed, and theoretical analysis and proofs of synchronization and performance optimality are provided. Simulation results verify the effectiveness of the proposed method.

Analysis of Factory Automated Guided Vehicles Systems by using Multi-agent System and Contract Net Protocol

Analysis of Factory Automated Guided Vehicles Systems by using Multi-agent System and Contract Net Protocol

Event-triggered iterative learning containment control for one-sided Lipschitz nonlinear singular switched multi-agent systems with multiple leaders and error quantization

In this work, under a repeatable control environment, the iterative learning containment control issue for multi-agent systems with one-sided Lipschitz nonlinear singular switched dynamics is addressed. It focuses on the output containment for one-sided Lipschitz nonlinear singular switched multi-agent systems (SWMASs) with multiple leaders, for which an event-triggered containment quantized iterative learning control (QILC) scheme is designed. Due to the limited system resources, SWMASs are often restricted in their communication capabilities and the frequency of controller updates in many applications, the proposed event-triggering mechanism can greatly restrict the use of communication and computation resources throughout the iterative learning process. When the triggering requirement is met, the event-triggered QILC input is repeatedly updated; alternatively, the control input stays unchanged. To conquer the nonidentical initial states of followers, an initial state learning with error quantization and event-triggering is also utilized. By virtue of some appropriate learning gains and sufficient convergent conditions, the containment error convergence can be accomplished for one-sided Lipschitz SWMASs. Finally, some numerical examples via Matlab are provided to demonstrate the efficacy of the event-triggered QILC scheme, it indicates that followers’ outputs can converge to the convex hull formed by leaders and the containment errors gradually converge to the 10−3 - vicinity of the origin along the iteration axis, where the communication resources can save more than 50% in the transmission channel.

Nodal Consensus in Multiagent Systems With Edge Communications

Nodal Consensus in Multiagent Systems With Edge Communications

A Model-Based Reinforcement Learning Algorithm for Multi-Agent Cooperation Nash Equilibrium With Unstable Communication

Solving Nash equilibrium is important to multi-agent systems, in which the communication is an important factor. However, many proposed reinforcement learning (RL) based algorithms take into account the communication factors by assuming stable communication conditions, which does not hold in the real environment. In this paper, we analyze the effect of a typical RL algorithm in the cases of unstable communication and communication failure, which causes information loss between agents, leading to isolation of agents and affecting algorithm convergence. Then, we propose a model-based RL algorithm to solve Nash equilibrium for multi-agent systems when agents are isolated, and prove its convergence and rationality through mathematical proofs. The simulations results show the effectiveness of the proposed algorithm.

Multi-agent reinforcement learning with synchronized and decomposed reward automaton synthesized from reactive temporal logic

Multi-agent systems (MAS) consist of multiple autonomous agents interacting to achieve collective objectives. Multi-agent reinforcement learning (MARL) enhances these systems by enabling agents to learn optimal behaviors through interaction, thus improving their coordination in dynamic environments. However, MARL faces significant challenges in adapting to complex dependencies on past states and actions, which are not adequately represented by the current state alone in reactive systems. This paper addresses these challenges by considering MAS operating under task specifications formulated as Generalized Reactivity of rank 1 (GR(1)). These synthesized strategies are used as a priori knowledge to guide the learning. To tackle the difficulties of handling non-Markovian tasks in reactive systems, we propose a novel synchronized decentralized training paradigm that guides agents to learn within the MARL framework using a reward structure constructed from decomposed synthesized strategies of GR(1). We initially formalize the synthesis of GR(1) strategies as a reachability problem of winning states of the system. Subsequently, we develop a decomposition mechanism that constructs individual reward structures for decentralized MARL, incorporating potential values calculated through value iteration. Theoretical proofs are provided to verify that the safety and liveness properties are preserved. We evaluate our approach against other state-of-the-art methods under various GR(1) specifications and scenario maps, demonstrating superior learning efficacy and optimal rewards per episode. Additionally, we show that the decentralized training paradigm outperforms the centralized training paradigm. The value iteration strategy used to calculate potential values for the reward structure is compared against two other strategies, showcasing its advantages.

Fixed-time prescribed performance control for nonlinear multi-agent systems with novel uncertainties

This paper addresses an approach to fixed-time prescribed performance control design for multi-agent systems with novel uncertainties, i.e., the uncertainties from system input powers, and a new neural backstepping control design is proposed such that the prescribed performance can be guaranteed within a fixed-time. Unlike existing adaptive results, the upper boundary of stability time is decided only by the design parameters and also is independent of the initial states of the system. By leveraging Lyapunov stability theory and graph theory, it is demonstrated that every follower agent can synchronize to the leader agent, and tracking errors converge to the origin’s neighborhood. Finally, the simulation results verify the effectiveness of the control strategy.

Time-varying formation tracking of multi-agent systems with multiple leaders based on aperiodic time-triggered intermittent control

In this article, the distributed aperiodic time-triggered intermittent control protocol (DATICP) is presented for the time-varying formation tracking of multi-agent systems (MASs) with multiple leaders. Firstly, we present the mathematical description of DATICP. The characteristic of DATICP lies in that for each agent, its control signal is updated at aperiodic sampling instants on work time intervals and disappears on rest time intervals. Secondly, based on a mixed time-dependent Lyapunov functional (MTLF), the exponential stability theorem of the derived system is established. Thirdly, according to the theorem, sufficient conditions are given to achieve time-varying formation tracking for MASs with multiple leaders by DATICP. As a result, a corollary is obtained which compromises conservativeness and complexity. Finally, two simulation examples are offered for illustration.

Robust Collaborative Visual-Inertial SLAM for Mobile Augmented Reality.

Achieving precise real-time localization and ensuring robustness are critical challenges in multi-user mobile AR applications. Leveraging collaborative information to augment tracking accuracy on lightweight devices and fortify overall system robustness emerges as a crucial necessity. In this paper, we propose a robust centralized collaborative rnulti-agent VI-SLAM system for mobile AR interaction and server-side efficient consistent mapping. The system deploys a lightweight VIO frontend on mobile devices for real-time tracking, and a backend running on a remote server to update multiple submaps. When overlapping areas between submaps across agents are detected, the system performs submap fusion to establish a globally consistent map. Additionally, we propose a map registration and fusion strategy based on covisibility areas for online registration and fusion in multi-agent scenarios. To improve the tracking accuracy of the frontend on agent, we introduce a strategy for updating the global map to the local map at a moderate frequency between the camera-rate pose estimation of the frontend VIO and the low-frequency global map optimization, using a tightly coupled strategy to achieve consistency of the multi-agent frontend poses estimation in the global map. The effectiveness of the proposed method is further confirmed by executing backend mapping on the server and deploying VIO frontends on multiple mobile devices for AR demostration. Additionally, we discuss the scalability of the proposed system by analyzing network traffic, synchronization frequency, and other factors at both the agent and server ends.

Double adaptive event-triggered strategy for leader-following group consensus of heterogeneous multi-agent systems in cooperative–competitive networks

This article investigates adaptive event-triggered control for leader-following group consensus of hybrid heterogeneous multi-agent systems (HMASs) composed of single- and double-integrator agents under cooperative–competitive relationships. By using fully distributed control protocol and adaptive protocol, the leader-following group consensus can be achieved without global information. Additionally, a novel adaptive double dynamic event-triggered mechanism (DDETM) is proposed, in which two independent adaptive event-triggered conditions are devised to maintain events of communication and updates of controllers. Furthermore, adaptive parameters and internal dynamic variables are employed to reduce the number of triggers and exclude Zeno phenomenon. By utilizing Lyapunov stability theory, some sufficient conditions are acquired for the leader-following group consensus of the HMASs. Finally, simulation examples are given to prove the effectiveness of the proposed method.

Adaptive fault-tolerant containment control for heterogeneous uncertain nonlinear multi-agent systems under actuator faults

As modern control systems involve a growing number of actuators, the consequent number of faults occurrence, which may deteriorate the closed-loop system performance, is increasing. Focusing on the containment control problem of generic nonlinear high-order heterogeneous uncertain Multi-Agent Systems, this study proposes a novel distributed adaptive fault-tolerant Proportional-Integral-Derivative control protocol to improve the reliability and the resilience of the whole networked systems against actuator failures. The Lyapunov theory, combined with the Barbalat lemma, provides the design of suitable adaptive mechanisms that adjust the PID control actions to ensure the asymptotic convergence of the closed-loop containment errors for the overall network, hence implying the convergence of each agent state toward the convex region shaped by the multiple leaders, despite the occurrence of actuators faults. The fulfillment of the fault-tolerant containment objective is ensured by requiring a lower computational burden with respect to alternative solutions, hence resulting in more attractive for wider practical engineering applications. Numerical simulation results, also including a comparison analysis with a state-of-the-art controller, corroborate the efficiency and the advantages of the proposed distributed PID controller.