Output feedback consensus control for fractional-order nonlinear multi-agent systems with directed topologies

https://doi.org/10.1016/j.jfranklin.2019.10.032Get rights and content

Abstract

This paper is devoted to the output feedback consensus control problem for a class of nonlinear fractional-order multi-agent systems (MASs) with general directed topologies. It is worth noting that the considered fractional-order MASs including the second-order MASs as special cases. By introducing a distributed filter for each agent, a control algorithm uses only relative position measurements is proposed to guarantee the global leaderless consensus can be achieved. Also the derived results are further extended to consensus tracking problem with a leader whose input is unknown and bounded. Finally, two simulation examples are provided to verify the performance of the control design.

Introduction

Consensus control of multi-agent systems (MASs) is to design appropriate controller for each agent only using local information between neighbors such that the states of all agents reach general agreement. From the viewpoint of existing number of leaders in MASs, existing consensus problem of MASs can be classified into two categories: leaderless consensus problem [1], [2], [3] and consensus tracking (or leader-follower consensus) problem with a leader [4], [5], [6], [7]. During the last decade, consensus of MASs has received considerable attention and there are many results available in the literature, see the recent survey papers [8], [9] and references therein.

Early work often focuses on consensus of integer-order MASs, i.e., first-order MASs [1], [4], [5], second-order MASs [2], [3], [6], and high-order MASs [7], [10], [11]. However, several phenomena can be explained naturally by the collective group behavior of agents with fractional-order dynamics rather than classical integer-order dynamics [12], [13]. For example, the synchronization motion of multiple agents in fractional circumstances, i.e., viscoelastic materials, macromolecule fluids and porous media. Up to now, some researchers have tackled the consensus control of MASs with fractional-order dynamics. Leaderless consensus problem of nonlinear fractional-order double integrator MASs is studied in [14], [15]. Both leaderless consensus problem and consensus tracking problem are considered in [16] for fractional-order MASs with input time delay. Consensus tracking control problem is also presented for fractional-order single integrator MASs with undirected topology [17] or directed topology [18], [19], and for fractional-order double integrator MASs with undirected topology [20] or directed topology [14], [21]. It should be emphasized that fractional-order systems are an extension to the traditional integer-order ones, which have properties of infinity memory and hereditary due to the existence of a memory term in the model [22].

Note that state feedback-based controllers are mainly based on a restrictive assumption that the state variables of each agent can be measured directly. However, in many real applications, full-state measurements are unavailable due to economical concerns or physical constraints. Especially when there exist multiple states in second-order or higher-order MASs, it is unrealistic to obtain the information of multiple states accurately. Thus, state feedback-based control in those cases should be replaced by output feedback-based control. Some results on output feedback-based consensus control problem are presented in MASs with general linear dynamics [23], [24], with second-order agent dynamics [25], [26] and with high-order agent dynamics [27]. So far, no author has studied the output feedback-based consensus problem for MASs with nonlinear fractional-order dynamics.

In reality, the agents might be affected by the interaction among neighboring agents, but also by its own intrinsic nonlinear dynamic. So the MASs with intrinsic nonlinear dynamics are considered recently in [2], [3], [5], [14], [18]. Since the limited view field or nonuniform sensing ranges of sensors, one agent may be able sense another agent, but not vice versa. The communication topology among the agents, in general is directed. Taking into consideration these practical cases, in this paper, we consider the consensus problem of fractional-order double integrator MASs with intrinsic nonlinear dynamics and general directed topologies using only relative output information. Due to the well-known Leibniz rule for fractional derivatives is invalid [28], how to construct a suitable Lyapunov function for analysing the stability of nonlinear fractional-order MASs is very challenging. The output feedback based consensus control of double integrator MASs in the presence of nonlinear fractional-order dynamics is even more challenging as the communication topology among the agents is not only directed but also local.

The main contributions of this paper are summarized as follows. Firstly, for the leaderless consensus problem of fractional-order MASs with intrinsic nonlinear dynamics and directed graph, a novel distributed algorithm combined with a filter is derived. By generalizing an important nonlinear fractional-order inequality, it is shown that all the agents can achieve global consensus if the interaction graph is strongly connected and the control parameters are chosen properly. Secondly, for the tracking problem with a leader whose input is bounded and unknown to any follower, a newly distributed algorithm combined with a similar filter is further developed to guarantee the tracking error and control input are uniformly ultimately bounded (UUB). Of particular interest is all the algorithms designed in this paper can be implemented only using relative out measurements between neighbors.

The rest of this paper is organized as follows. Some preliminaries are briefly presented in Section 2. Output feedback based leaderless consensus problem is first studied in Section 3. In Section 4, the results are then extended to the output feedback based tracking problem when there exists an unknown leader. Simulation examples and conclusions are outlined in Sections 5 and 6, respectively.

Let Rm×n and Rm be the sets of m × n real matrices and m-dimensional Euclidean space, respectively. Let 1nRn(0nRn) stand for the n × 1 column vector of all ones (zeros) and In (On) be the n × n identity (zero) matrix. Denote by diag(d1,,dn)Rn×n a diagonal matrix with diagonal entries d1 to dn. We use ⊗ to represent the Kronecker product. For a vector x=[x1,,xn]TRn, x=xTx denotes the Euclidean norm. Let λ¯(A)(λ̲(A)) denote the maximal (minimum) eigenvalue of a positive definite matrix ARn×n. Let |b| and σ(B) be the absolute value of a real number bR and the maximal singular value of a matrix BRm×n, respectively.

Section snippets

Graph theory

Let G=(V,E,A) denote a weighted directed graph of order N,  where V={v1,,vN} and EV×V are, respectively, the set of nodes and the set of directed edges, and A=[aij]N×N is a weighted adjacency matrix with weights aij > 0 if (vj,vi)E and aij=0 otherwise. Moreover, it is assumed that  aii=0, iI={1,,N}. The edge (vj,vi)E means that the node vi can access information from vj,  then node vj is a neighbor of node vi. The set of neighbors of node vi is denoted as Ni={jV:(vj,vi)E}. The

Output feedback based leaderless consensus control

We first consider the leaderless consensus problem for a group of N agents distributed on a communication graph G. The dynamics of each agent is described by{Dqxi=wi,Dqwi=f(t,xi,wi)+ui,yi=xi,iI={1,2,,N},where 0 < q ≤ 1 xiRn, wiRn, uiRn and yiRn denote, respectively, the position, velocity, control input and output of agent i, and f(t,xi,wi)Rn is the intrinsic nonlinear dynamics of agent i.

Assumption 1

Suppose that there exist two constants ρ1 and ρ2 such that, x,v,y,zRn and t ≥ 0f(t,x,v)f(t,y,z)

Output feedback based tracking control

In this section, we focus on the output consensus tracking problem for a fractional-order MAS consisting of N followers and one leader. Let the leader indexed by 0 and all followers indexed by 1,,N. The dynamics of each agent is governed by{Dqxi=wi,Dqwi=f(t,xi,wi)+ui,yi=xi,iI{0},where we consider a more general case that the leader’s input u0 is possibly time-varying and nonzero. Moreover, the leader’s input information without being effected by those of the followers or unknown to any

Simulation examples

To illustrate the effectiveness of our proposed algorithms, we provide the following two simulation examples.

Example 1 (Leaderless consensus problem)

Consider the leaderless consensus problem for a group of six agents modeled by fractional-order MASs (4). The topology structure G among the six agents is shown in Fig. 2, where each nonzero edge weight is assumed to be 1. The nonlinear function is given by f(t,xi,wi)=0.1isinxi0.5wi. Choose q=0.95, a=b=5 and c=10. The initial states are chosen as xi(0)=0.02i and wi(0)=0.01i, i=1,2,,6.

Conclusions

This paper focus on designing output feedback based consensus protocols for nonlinear fractional-order MASs with double-integrator dynamics under general directed topologies. The leaderless consensus problem and consensus tracking problem are addressed in Sections 3 and 4, respectively, by proposing a distributed algorithm only using relative positive measurements between neighbors. By using the fractional Lyapunov direct method, some sufficient conditions are derived to guarantee the global

References (34)

  • M.A. Duarte-Mermoud et al.

    Using general quadratic lyapunov functions to prove lyapunov uniform stability for fractional order systems

    Commun. Nonlinear Sci. Numer. Simul.

    (2015)
  • R. Olfati-Saber et al.

    Consensus problems in networks of agents with switching topology and time-delays

    IEEE Trans. Autom. Control.

    (2004)
  • J. Mei et al.

    Distributed consensus of second-order multi-agent systems with heterogeneous unknown inertias and control gains under a directed graph

    IEEE Trans. Autom. Control.

    (2016)
  • G. Wen et al.

    Consensus tracking of multi-agent systems with lipschitz-type node dynamics and switching topologies

    IEEE Trans. Circuits Syst. I, Reg. Papers.

    (2014)
  • L. Ding et al.

    An overview of recent advances in event-triggered consensus of multiagent systems

    IEEE Trans. Cybern.

    (2018)
  • Z. Zuo et al.

    An overview of recent advances in fixed-time cooperative control of multi-agent systems

    IEEE Trans. Ind. Electron

    (2018)
  • Y. Cao et al.

    Distributed coordination of networked fractional-order systems

    IEEE Trans. Syst., Man, Cybern., B, Cybern.

    (2010)
  • Cited by (0)

    View full text