Distributed anti-windup approach for consensus tracking of second-order multi-agent systems with input saturation☆
Introduction
In recent years, intensive research attention has been devoted to the coordination control of multi-agent systems due to its broad applications in areas such as mobile sensor network, unmanned aerial vehicles and satellite attitude alignment [1]. In distributed control of multi-agent systems, the goal is to design distributed control laws which only use local information such that some global control objectives can be achieved. Common coordination tasks include consensus, consensus tracking, flocking, swarming and so on. In the consensus tracking problem, there exists some leader agent which determines the final desired trajectory of the network and the follower agents need to track the trajectory of the leader asymptotically. Various results have been obtained for this problem regarding different agent dynamics and communication topologies [2], [3], [4], [5], [6].
When implementing the coordination controllers to real systems, practical limitations of the agent dynamics have to be considered. Input saturation is one ubiquitous control limitation imposed by physical systems resulted from finite actuation power of the agents. It may lead to serious performance degradation or even instability if not properly handled. In this sense, designing coordination controllers subject to input saturation is of great importance. In [7], [8], saturated consensus of first-order integrators with directed communication graphs was studied. Consensus algorithms for double-integrator dynamics with bounded inputs were proposed in [9]. In [10], consensus strategies accounting for actuator saturations and the lack of velocity measurements were designed for double-integrator dynamics based on auxiliary systems. [11] investigated the saturated global consensus tracking problem for neutrally stable systems and double integrator systems under detail-balanced directed communication graphs. A globally bounded nonlinear controller was proposed in [12] for second-order multi-agent systems under detail-balanced directed graphs from the input constrained optimal consensus perspective. Semi-global and global containment control were studied for second-order multi-agent systems with input saturation in [13]. In [14], [15], [16], [17], semi-global coordination of linear multi-agent systems with input saturation was studied with low-gain and low-and-high gain control approaches. Using similar techniques, semi-global consensus of linear multi-agent systems with relative output feedback and input saturation under directed switching networks was investigated in [18]. In [19], a multi-hop relay based distributed controller was proposed for global leader-following consensus of linear multi-agent systems with input saturation. Global saturated consensus problem for discrete-time neutrally stable and double integrator multi-agent systems was studied in [20]. Model predictive control methods were employed in [21], [22] to handle input saturation constraints in consensus control.
Note that for global consensus tracking of second-order multi-agent systems with input saturation, existing results are mostly only applicable for undirected or detail-balanced directed communication graphs [9], [10], [11], [13], [19]. For general directed communication graphs, the asymmetry of the graph Laplacian matrix makes it much harder to analyze the effect of saturation nonlinearity on the stability of the closed-loop systems. In this work, we overcome such difficulty by generalizing the traditional anti-windup control ideas to distributed control of multi-agent systems. Anti-windup control is an effective method to counteract the influence of input saturation [23], [24], [25] and has recently been applied to the cooperative control of multi-agent systems. In [26], the global consensus of multiple identical Hurwitz linear continuous-time systems within a fully connected network in the presence of saturation was studied. A decentralized dynamic output feedback control law was designed with an anti-windup compensation term. The results were generalized to the regional synchronization of general linear multi-agent systems under input saturation in [27]. In this work, we consider the global input saturated consensus tracking problem for second-order multi-agent systems under general directed communication graphs. The controllers are obtained by adapting linear consensus tracking controllers designed for the unconstrained case and the design process is both simple and effective. Furthermore, only relative output measurements are needed for the controller implementation.
The rest of the paper is organized as follows. In Section 2, some preliminaries and the problem formulation are given. In Section 3, the distributed anti-windup approach is proposed and analyzed. In Section 4, some simulation examples are provided to illustrate the proposed control strategy. Finally, some concluding remarks are given in Section 5.
Notations For a vector , is a diagonal matrix with the diagonal elements , is the Euclidean norm, and is the signum function. For a matrix , is the induced -norm and denotes the smallest eigenvalue when all the eigenvalues are real. For a signal defined on , its -norm is defined as . A signal is said to belong to , i.e., , if .
Section snippets
Graph theory
The communication relation among the agents in the leader–follower network can be represented by a directed graph , consisting of a finite set of vertices and a finite set of edges . Each agent is represented by a vertex in and an edge is an ordered pair which represents the information flow from agent to agent . The set of neighbors of node is denoted by . A path in is a sequence where
Distributed anti-windup controller design
In this section, we consider the input constrained consensus tracking problem under an anti-windup control framework. The philosophy behind anti-windup control is to first design the controller without considering the input constraint. Then, when input saturation is presented, find a modification of the pre-designed controller such that the closed-loop trajectory remains unchanged when no saturation is encountered and remains to be satisfactory during saturation. This is an attractive design
Simulation
To illustrate the performance of the proposed controller, consider a leader–follower network composed of four followers and one leader modeled by (1), (2) respectively. The communication graph of the network is shown in Fig. 1 which satisfies Assumption 1. The elements of the Laplacian matrix are taken as 0s and 1s.
The controller (12) is first designed for the unconstrained case with the controller parameters , and determined from Algorithm 1. With the initial
Conclusion
In this paper, global consensus tracking for second-order multi-agent systems with input saturation under general directed communication graphs is studied. A distributed anti-windup controller design approach is proposed. The design method can adapt linear consensus tracking controller to handle input saturation which offers great flexibility in designing saturated global consensus tracking controllers with different local performances. Future directions include generalizing the results to
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (33)
- et al.
Tracking control for multi-agent consensus with an active leader and variable topology
Automatica
(2006) - et al.
Observer-based finite-time coordinated tracking for general linear multi-agent systems
Automatica
(2016) - et al.
On consensus algorithms for double-integrator dynamics without velocity measurement and with input constraints
Systems Control Lett.
(2010) - et al.
On global leader-following consensus of identical linear dynamic systems subject to actuator saturation
Systems Control Lett.
(2013) - et al.
Global optimal consensus for multi-agent systems with bounded controls
Systems Control Lett.
(2017) - et al.
Synchronization of identical linear dynamic systems subject to input saturation
Systems Control Lett.
(2014) - et al.
Global leader-following consensus of a group of general linear systems using bounded controls
Automatica
(2016) - et al.
Global consensus for discrete-time multi-agent systems with input saturation constraints
Automatica
(2014) - et al.
Constructive nonlinear anti-windup design for exponentially unstable linear plants
Systems Control Lett.
(2007) - et al.
Anti-windup design for input-coupled double integrator systems with application to quadrotor UAV’s
Eur. J. Control
(2017)
On constructing Lyapunov functions for multi-agent systems
Automatica
Synchronization in networks of minimum-phase, non-introspective agents without exchange of controller states: homogeneous, heterogeneous, and nonlinear
Automatica
Consensus problems in networks of agents with switching topology and time-delays
IEEE Trans. Automat. Control
A theoretical study of different leader roles in networks
IEEE Trans. Automat. Control
Controllability of multi-agent systems from a graph-theoretic perspective
SIAM J. Control Optim.
Distributed tracking control for linear multiagent systems with a leader of bounded unknown input
IEEE Trans. Automat. Control
Cited by (0)
- ☆
This work was supported by the Natural Science Foundation of Jiangsu Province under Grant BK20170695, the National Natural Science Foundation of China under Grant 61703094, National Priority Research Project NPRP 8-274-2-107 funded by Qatar National Research Fund .