Elsevier

ISA Transactions

Volume 106, November 2020, Pages 181-199
ISA Transactions

Research article
Distributed adaptive fault-tolerant close formation flight control of multiple trailing fixed-wing UAVs

https://doi.org/10.1016/j.isatra.2020.07.005Get rights and content

Highlights

  • The fault-tolerant close formation flight control of multiple UAVs is considered.

  • The nonlinear 6-DOF fixed-wing UAV model is used in the control design.

  • A composite learning algorithm is proposed to enhance the estimation capability.

Abstract

This paper considers the reliable control problem for multiple trailing fixed-wing unmanned aerial vehicles (UAVs) against actuator faults and wake vortices. A distributed adaptive fault-tolerant control (FTC) scheme is proposed by using a distributed sliding-mode estimator, dynamic surface control architecture, neural networks, and disturbance observers. The proposed control scheme can make all trailing fixed-wing UAVs converge to the leading UAV with pre-defined time-varying relative positions even when all trailing UAVs encounter the wake vortices generated by the leading UAV and a portion of trailing UAVs is subjected to the actuator faults. It is shown that under the proposed distributed FTC scheme, the tracking errors of all trailing UAVs with respect to their desired positions are bounded. Comparative simulation results are provided to illustrate the effectiveness of the proposed control scheme.

Introduction

Recently, formation control of multi-agent system (MAS) has received a great deal of interest and numerous control methods have been obtained [1], [2], [3], [4]. Due to the high efficiency and robustness, the formation control of multiple unmanned aerial vehicles (multi-UAVs) has been widely investigated [5], [6], [7], [8]. Nevertheless, the formation control of multiple fixed-wing UAVs is very challenging if the outer-loop position and inner-loop attitude are simultaneously considered, since both of them are highly coupled and nonlinear. Currently, massive existing results focus on developing formation control methods based on the outer-loop position model [9], [10], [11]. Regarding these methods, the outer-loop position model is first transformed into a double-integrator system, and then a virtual formation control signal is constructed for each UAV, followed by an extraction algorithm for the overall control signal. However, the aerodynamic uncertainties and actuator faults may not be included in the control methods for the transformed outer-loop double-integrator system, which may significantly limit the implementations of these aforementioned formation control methods. Therefore, it is necessary to develop effective formation control protocols for multiple fixed-wing UAVs with simultaneous considerations of outer-loop position and inner-loop attitude models by explicitly handling unknown aerodynamic parameters and actuator faults.

When multiple fixed-wing UAVs are flying in a close formation, the wake vortices generated by the leading UAV impose induced velocities on the trailing UAVs, extending the trailing UAVs’ ranges and endurances. However, catastrophes may be encountered by the trailing UAVs and leading UAV if the adverse effects of wake vortices on the station-keeping and tracking performances of trailing UAVs are not well attenuated. Autonomous aerial refueling (AAR), as a utilization of close formation, can significantly increase the endurances and extend the flight radii of receiver UAVs (trailing UAVs) [12], [13]. In AAR, the receiver UAVs keep the close formation flight with a tanker UAV (leading UAV) and receive the fuels from the tanker UAV via an aerial refueling hose if the probe–drogue method is adopted, or a flying boom if the boom refueling method is adopted. Regarding AAR, the vortices generated by the tanker UAV may degrade the position tracking and keeping performances of receiver UAVs if the effects of wake vortices are not evidently considered. The authors in [14] proposed a proportional–integral (PI) method for the trailing aircraft to track the leading aircraft. In [15], a visual guidance method was studied for the tight formation of two UAVs. By using uncertainty and disturbance estimator, a robust formation control algorithm was investigated within a backstepping architecture [16]. To further increase the flight safety and avoid the collisions among UAVs, the wake vortices should be explicitly considered for trailing UAVs.

As a key method to increase operational safety, fault-tolerant control (FTC) has been widely investigated for various plants [17], [18], [19], [20], [21], [22], [23], [24]. To ensure safe formation flight, the distributed FTC schemes of multiple agents have been widely investigated. In [25], an FTC scheme was proposed for multiple nonlinear uncertain agent systems by using a fault diagnosis module. By designing adaptive laws to compensate for the uncertain items, a distributed sliding-mode FTC scheme has been developed in [26] for heterogeneous multiple agent systems. In [27], the distributed FTC scheme was investigated to synchronize the attitudes of UAVs with robust adaptive mechanism. Recently, several FTC methods have been developed to tolerate the actuator faults encountered by a portion of UAVs. In [28], an FTC protocol was studied for multi-UAVs in a leader–follower communication network. With a similar communication network, a finite-time FTC method was further developed in [29]. In [30], [31], the authors investigated the FTC for the longitudinal motions with distributed communications. Different from the leader–follower communication network, distributed communication topology has better robustness against communication link failures. More recently, the distributed finite-time FTC method was studied in [32] for the attitude synchronization of multi-UAVs. In [33], the fuzzy neural networks and the fractional-order calculus are integrated to improve the FTC performance. To further reduce the performance degradations induced by the actuator faults and wind effects simultaneously, the fractional-order backstepping FTC scheme was proposed for multi-UAVs [34]. Moreover, it should be noted that the aforementioned existing results are only concerned with the FTC for UAVs without the consideration of wake vortices. In [35], [36], the FTC problem was investigated for the lead-wing close formation flight involving two UAVs. In [37], the authors proposed an FTC scheme to further take the wake vortices into the controller design. However, all aerodynamic parameters needed to be accurately known in the constructed FTC scheme, which significantly increased the cost of obtaining aerodynamic parameters. Therefore, fault-tolerant close formation flight control methods should be further investigated especially when multiple trailing UAVs with unknown aerodynamic parameters are involved.

Motivated by the analysis mentioned, this paper attempts to develop a distributed fault-tolerant close formation flight control scheme for trailing UAVs against actuator faults, wake vortices, and aerodynamic uncertainties. The distributed sliding-mode estimator (DSME) is designed for estimating the leading UAV’s position. Adding the pre-defined time-varying relative positions to the estimated leading UAV’s position signals, the FTC scheme is then developed for the nonlinear trailing UAVs based on the newly constructed position references. Moreover, to overthrow the controller development difficulty due to the partially unknown aerodynamic parameters, neural networks (NNs) are utilized in the FTC scheme. Furthermore, the disturbance observer (DO) is used to estimate the unknown disturbances containing actuator faults and wake vortices. Prediction errors are incorporated to improve the composite learning capability of NNs and DOs. The main contributions of this paper are stated as follows.

  • (1)

    Different from the previous work [37], which investigated the FTC method for a single trailing UAV with exactly known aerodynamic parameters, this paper further considers the fault-tolerant close formation flight control of multiple trailing fixed-wing UAVs with little knowledge on the aerodynamic parameters in a distributed communication network. Actuator faults, wake vortices, and partially unknown aerodynamic parameters are explicitly considered.

  • (2)

    In contrast to [9], [10], [38] about the formation control of multi-UAVs with three-degree-of-freedom (3-DOF) outer-loop position models, the close formation flight control scheme is developed on the nonlinear 6-DOF UAV model. Therefore, the developed control method is more practical.

  • (3)

    To improve the control performance degraded by thelumped uncertainties including unknown aerodynamic parameters, actuator faults, and wake vortices, a composite learning algorithm is proposed to enhance the estimation capability.

The remaining parts of this paper are organized as follows. Section 2 gives the nonlinear trailing UAV model and some useful preliminaries. Section 3 presents the distributed fault-tolerant close formation flight control protocol. Comparative simulation results and concluding remarks are given in Sections 4 Simulation results, 5 Conclusions and future works, respectively.

Section snippets

Trailing UAV model

In this paper, as illustrated in Fig. 1, a close formation team consisting of N trailing UAVs and one leading UAV is investigated. The nonlinear model of the ith trailing UAV is expressed as [37], [39] ẋi=Vicosβicosαicosθicosψi+Visinβi(cosϕisinψi+sinϕisinθicosψi)+Vicosβisinαi(sinϕisinψi+cosϕisinθicosψi)+wixcosθicosψi+wiy(cosϕisinψi+sinϕisinθicosψi)+wiz(sinϕisinψi+cosϕisinθicosψi)ẏi=Vicosβicosαicosθisinψi+Visinβi(cosϕicosψi+sinϕisinθisinψi)+Vicosβisinαi(sinϕicosψi+cosϕisinθisinψi)+wixcosθi

Distributed adaptive fault-tolerant close formation controller design

In this section, the developed FTC scheme consists of three parts: (1) a DSME is first developed to estimate the leader’s position for each trailing UAV with the information from its neighboring UAVs and a newly constructed position signal can be obtained for each trailing UAV by adding the pre-defined time-varying relative position to the estimated leader’s position signal; (2) Based on the newly constructed longitudinal position signal, an FTC is developed for each trailing UAV to achieve the

Simulation results

Considering the typical implementation of close formation in AAR, the proposed adaptive FTC scheme is verified on the AAR, which consists of one leading UAV (tanker UAV), and three trailing UAVs (receiver UAVs). Regarding the formation team, if the UAV i and j can communicate with each other, the corresponding element aij of the adjacency matrix A is set as 1, otherwise, the element is chosen as 0. Moreover, if the ith trailing UAV has access to the leading UAV, the corresponding element bi of

Conclusions and future works

In this paper, a distributed adaptive fault-tolerant close formation flight control scheme has been developed for multiple trailing fixed-wing UAVs against actuator faults and wake vortices. A DSME is first designed to estimate the leader’s position for each trailing UAV. Then, by dividing the trailing UAV model into the longitudinal dynamics containing the forward and vertical subsystems, and the lateral-directional dynamics containing the side-distance and sideslip angle subsystems, two FTCs

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 (46)

  • ZhangY.M. et al.

    Bibliographical review on reconfigurable fault-tolerant control systems

    Annu. Rev. Control

    (2008)
  • AoW. et al.

    Adaptive robust fault tolerant control design for a class of nonlinear uncertain MIMO systems with quantization

    ISA Trans

    (2017)
  • LiZ. et al.

    A data-driven fault-tolerant control design of linear multivariable systems with performance optimization

    ISA Trans

    (2017)
  • XieC.-H. et al.

    Data-based fault-tolerant control for affine nonlinear systems with actuator faults

    ISA Trans

    (2016)
  • DongH.Y. et al.

    Dual-quaternion based fault-tolerant control for spacecraft formation flying with finite-time convergence

    ISA Trans

    (2016)
  • JiangY. et al.

    Adaptive backstepping fault-tolerant control for flexible spacecraft with unknown bounded disturbances and actuator failures

    ISA Trans

    (2010)
  • KhaliliM. et al.

    Distributed adaptive fault-tolerant control of uncertain multi-agent systems

    Automatica

    (2018)
  • YangH.L. et al.

    Synchronization of multiple 3-DOF helicopters under actuator faults and saturations with prescribed performance

    ISA Trans

    (2018)
  • YuZ.Q. et al.

    Safe control of trailing UAV in close formation flight against actuator fault and wake vortex effect

    Aerosp Sci Technol

    (2018)
  • XiangX. et al.

    On decentralized adaptive full-order sliding mode control of multiple UAVs

    ISA Trans

    (2017)
  • HongY. et al.

    Tracking control for multi-agent consensus with an active leader and variable topology

    Automatica

    (2006)
  • PengZ.H. et al.

    Path-guided time-varying formation control with collision avoidance and connectivity preservation of under-actuated autonomous surface vehicles subject to unknown input gains

    Ocean Eng

    (2019)
  • Liao F, Wang JL. Formation and reconfiguration control for nonholonomic UAVs with velocity constraints. In: AIAA...
  • Cited by (37)

    • Cross-dimensional formation control of second-order heterogeneous multi-agent systems

      2022, ISA Transactions
      Citation Excerpt :

      Among the various cooperative control objectives of MASs, formation control aims to regulate the involved agents to move with a pre-specified configuration in terms of the agents’ states or outputs [10,15]. In the past two decades, there is a surge of interest in the analysis and design for formation control problems of some typical MASs, such as UAVs, autonomous surface vehicles, mobile robots, and dynamical agents in generally linear state–space forms [14–26]. Some interesting formation control policies, including behavior-based ones, virtual structure-based ones, and leader–follower-based ones, were proposed [6,27,28].

    • Robust bipartite tracking consensus of multi-agent systems via neural network combined with extended high-gain observer

      2022, ISA Transactions
      Citation Excerpt :

      Over the past two decades, significant research efforts have been devoted to investigating the cooperative control problem of multi-agent systems [1–3]. As the fundamental and important research topic of distributed cooperative control, the consensus problem means the agents finally approach the same value of states, which has received much attention due to its wide applications [4]. In a general way, the consensus problem can be divided into the leaderless consensus [5] and the tracking consensus [6], depending on whether there exists a leader.

    • A review on fault-tolerant cooperative control of multiple unmanned aerial vehicles

      2022, Chinese Journal of Aeronautics
      Citation Excerpt :

      In the proposed control architecture, the fixed-wing UAV model was first divided into the longitudinal and lateral-directional dynamics, and then disturbance observer-based FTC methods were developed for the longitudinal and lateral-directional subsystems, respectively. Recently, by using the longitudinal and lateral-directional control architecture similar to Ref.156, an intelligent distributed FTCC scheme was presented in Ref.157 for multiple follower fixed-wing UAVs, which was developed by integrating a distributed SMO, dynamic surface control, and composite learning algorithms. The distributed SMO was first used for estimating the desired position references for each follower UAV.

    • Standoff tracking of a ground target based on coordinated turning guidance law

      2022, ISA Transactions
      Citation Excerpt :

      Today, there have been rapid developments in unmanned aerial vehicles (UAVs) [1–4].

    View all citing articles on Scopus

    This work was supported in part by National Natural Science Foundation of China (No. 61833013), State Key Laboratory of Synthetical Automation for Process Industries, Northeastern University, Shenyang, China (No. 2019-KF-23-05), 111 Project (No. B20007), and Natural Sciences and Engineering Research Council of Canada .

    View full text