electronics

Article
Fault-Tolerant Cooperative Control of Multiple Uncertain
Euler - Lagrange Systems with an Uncertain Leader
Rongrong Gu 1 and Xudong Sun 2,3, *

1 School of Electronic and Information Engineering, Shanghai Research Institute for Intelligent Autonomous
Systems, Tongji University, Shanghai 201804, China; rongronggu@tongji.edu.cn
2 College of Mechanical and Vehicle Engineering, Chongqing University, Chongqing 400044, China
3 China Society of Automotive Engineers, Beijing 100176, China
* Correspondence: 20220701098g@stu.cqu.edu.cn

Abstract: This paper explores the fault-tolerant cooperative control of multiagent systems, which
are modeled via an uncertain leader system and multiple uncertain Euler–Lagrange systems with
actuator faults. A self-adjusting observer is initially proposed to estimate the signal of the uncertain
leader for different followers and compute the observer gain in real time. An adaptive fault-tolerant
controller is designed based on the above observer and nonsingular fast terminal sliding mode surface.
This controller estimates lumped uncertainty and ensures that tracking errors are ultimately bounded.
The controller designed in this paper has the following three advantages. Firstly, the observer can
estimate and transmit the leader’s state to each follower even without precise knowledge of the
leader’s system matrix. Secondly, the controller is robust to actuator faults, uncertain parameters
and external disturbances, the upper bounds of which can be arbitrarily large and unidentified.
Thirdly, the controller has a simple structure and is also suitable for situations where the actuator is
healthy. Lastly, simulations are provided to demonstrate the effectiveness of both the observer and
the controller with or without actuator fault.

Keywords: fault-tolerant cooperative control; self-adjusting observer for uncertain leader system;
uncertain Euler–Lagrange multiagent systems; actuator fault

Citation: Gu, R.; Sun, X.

Fault-Tolerant Cooperative Control of
Multiple Uncertain Euler–Lagrange
1. Introduction
Systems with an Uncertain Leader. The Euler–Lagrange system serves as a paradigmatic model for unraveling complex
Electronics 2024, 1, 0. https:// dynamics and finds extensive application in various engineering domains, including
doi.org/ robotic manipulators [1,2], spacecrafts [3,4] and unmanned aerial vehicle systems [5,6].
Over the past few years, there has been a notable focus on cooperative control in multiagent
Academic Editors: He Cai and
Maobin Lv
systems, which involve multiple Euler–Lagrange systems and a leader system. The main
difficulty in multiagent cooperative control lies in devising a distributed control law for
Received: 22 April 2024 each agent to accomplish particular global goals, such as consensus [7], synchronization [8]
Revised: 19 May 2024 and formation [9].
Accepted: 24 May 2024 Various efforts have been devoted to addressing this issue [10]. Networked coop-
Published:
erative systems offer numerous advantages over single subsystems, including enhanced
efficiency and capabilities. Over recent years, there has been significant attention toward
the cooperative control of multiple Euler–Lagrange systems. The fundamental challenge is
Copyright: © 2024 by the authors.
to devise a distributed control scheme for each entity so that particular overarching goals
Licensee MDPI, Basel, Switzerland. (e.g., consensus [1,7], tracking [11,12], synchronization [8,13] and formation [9,14]) can be
This article is an open access article accomplished via local maneuvers. The challenge arises from the incomplete knowledge
distributed under the terms and of the parameters associated with multiple Euler–Lagrange systems, and the existence
conditions of the Creative Commons of uncertainty will diminish the controllability performance of the system. Currently,
Attribution (CC BY) license (https:// the adaptive method [11,13], neural networks method [12,14] and backstepping design
creativecommons.org/licenses/by/ procedure [8] have become effective methods for addressing the uncertainties associated
4.0/). with multiple Euler–Lagrange systems. More specifically, [11] investigated the distributed

Electronics 2024, 1, 0. https://doi.org/10.3390/electronics1010000 https://www.mdpi.com/journal/electronics

Electronics 2024, 1, 0 2 of 20

adaptive control of a group of underactuated flexible spacecraft operating under a leader–

follower framework, relying solely on the measurements of the rigid bodies. Ref. [13]
developed strategies to facilitate synchronization among networked uncertain Lagrange
systems through adaptive coordination control protocols. In [12], a distributed finite-time
tracking control algorithm was formulated, leveraging neural networks for estimating
model uncertainties. In [14], a design of an adaptive neural network multilayer formation
controller was proposed to handle model uncertainties. In [8], a fixed-time robust controller
was developed by converting uncertain Euler–Lagrange systems into second-order systems
and introduced the backstepping design methodology. However, these systems also exhibit
increased complexity and a higher frequency of faults, attributed to their extensive array
of sensors, controllers and communication equipment. At times, a fault within a system
can lead to a decline in performance or even result in catastrophic accidents. Therefore,
ensuring reliability is a crucial objective in engineering system design. It is imperative to
maintain stable and appropriate performance in systems regardless of whether components
are functioning properly or have failed. Fault-tolerant control is seen as one of the most
promising control technologies because it maintains specified performance of dynamic
systems even if component faults occur. Fault tolerance control is divided into passive
fault tolerance and active fault tolerance according to design methods. As the complexity
of the system deepens, the fixed control gain designed by passive fault tolerance and the
limited conservatism of fault tolerance are amplified. Active fault tolerance can adjust the
control gain or reconstruct the controller online, thereby compensating for the limitations
of passive fault tolerance and having stronger fault tolerance. However, it can also lead to
time delays caused by fault detection, isolation and controller reconstruction, as well as
computational resource waste caused by time estimation of fault information [15].
When dealing with Euler–Lagrange systems, controlling and executing tasks becomes
essential, particularly in the face of environmental disruptions, model uncertainties and
actuator malfunctions. This issue has garnered significant attention in recent years due to
mission failures caused by these uncertainties. To achieve the desired control performance,
various studies have been carried out and put into practice, e.g., sliding mode control [16],
back-stepping control [17], adaptive control [18,19], etc. In current research, the extensive
use of sliding mode control for fault-tolerant control demonstrates strong robustness to
parameter uncertainties and external disturbances. Nonetheless, it is vital to deal with the
inherent chattering issues in sliding mode control when utilizing this control approach.
In [20], a resilient fault-tolerant controller based on sliding mode was introduced to lessen
chattering. Moreover, a quasi-continuous second-order sliding mode control law was
developed to address system malfunctions and alleviate chattering. Terminal sliding mode
has delivered satisfactory outcomes in fault-tolerant control. In [21], a decentralized fault-
tolerant controller for robot manipulators was proposed by combining terminal sliding
modes and a super-twist algorithm to minimize chattering and expedite convergence.
In [22], a sliding mode surface for integral terminal was carefully developed, resulting in
the development of a model-free finite-time resilient adaptive fault-tolerant control for
robot manipulators, demonstrating advantages in global finite-time convergence. Fault-
tolerant cooperative control stands out as a leading control technology ensuring consistent
and optimal performance of dynamic systems despite component faults [18,19,23]. Ref. [18]
devised an adaptive fault-tolerant cooperative controller to attain the coordinated tracking
error to be uniformly ultimately bounded for networked uncertain Lagrange systems.
Ref. [23] proposed a distributed coordinated control scheme to make the leader-following
tracking error uniformly ultimately bounded. Ref. [19] introduced an adaptive neural
networks-based fixed-time fault-tolerant controller to make leader-following tracking
errors converge into compact sets.
In the domain of output regulation, as described in [24], it is common to reference
an exosystem that produces both the reference signal and external disturbances. Further-
more, in the realm of cooperative output regulation, as discussed in [25], the exosystem is
regarded as the leader system, distinct from the follower system. To design a distributed
Electronics 2024, 1, 0 3 of 20

control law, a distributed observer was proposed as a dynamic compensator following

the certainty equivalence principle [26]. This observer possessed the capability to assess
the leader’s state and transmit it to all followers via the communication network. The
observer presented in [27] initially addressed the cooperative output regulation issue in
linear multiagent systems and was subsequently extended to multiple Euler–Lagrange
systems [28]. Nevertheless, it was contingent on the assumption of possessing knowledge
regarding the system matrix of the leader system. To loosen the above assumption, an
adaptive distributed observer was introduced in [28,29], capable of estimating both the
state and matrix of the leader system. A self-tuning distributed observer was designed
in [30] to address the challenge of calculating the observer gain online when dealing with
numerous agents. These observers assume that certain followers have knowledge of the
leader system’s dynamics. Nonetheless, in practice, the dynamics of the leader system
might remain unknown to any follower, meaning that none of the followers possesses
precise knowledge of the leader system’s dynamics. In light of this, a distributed dynamic
compensator was introduced in [10].
Through the above investigation, we have learned how to save resources and achieve
fault-tolerant performance in multiagent collaborative control without using fault detec-
tion. The designed controller can handle actuator faults, system uncertainties and external
disturbances while tracking uncertain leader signals, which is a challenging research topic.
In this paper, we synthesize a controller based on an adaptive distributed observer and non-
singular fast terminal sliding mode surface to solve the fault-tolerant cooperative control
problem of multiple uncertain Euler–Lagrange systems, which are subjected to actuator
faults. In existing research, designing a distributed observer requires assuming that some
followers know the dynamics of the leader system [27,28]. However, in many practical
applications, followers may not have precise knowledge of the leader system’s dynam-
ics. The main contributions of this work can be stated as follows. Initially, an adaptive
distributed observer is devised for a linear leader system with neutral stability, featuring
an uncertain system matrix. This observer can estimate and transmit the leader’s state to
each follower via the system’s communication network, even without precise knowledge
of the leader’s system matrix. Secondly, an adaptive approach is applied to estimate the
aggregate uncertainty, removing the need to compute its upper limit. This method exhibits
resilience to actuator faults, uncertain parameters and external disturbances, the upper
bounds of which can be arbitrarily large and unidentified. Finally, the controller in this
paper has a simple structure without fault diagnosis, and it is also suitable for situations
where the actuator is healthy. The complicated and costly fault detection, diagnosis and
identification are not required.
This paper is structured as follows. Section 2 presents the formulation of the fault-
tolerant cooperative control problem. Section 3 describes the development of a self-
adjusting observer for dynamic leader systems with unknown parameters. In Section
4, we design the adaptive fault-tolerant controller. In Section 5, an example is presented.
Finally, we conclude this paper.
In what follows, the following notation will be adopted. For xi ∈ Rni , i = 1, . . . , m,
col(x1, . . . , xm ) = x1T , . . . , xm T T . ||a|| denotes the Euclidean norm of a vector a. For b ∈
 

Rn , c ∈ Rn and b ⊙ c is defined as b ⊙ c = [b1 c1, b2 c2, . . . , bn cn ]T . For d ∈ Rn , |d|r =

T
|d1 | , |d2 |r , . . . , |dn |r , sgn(d) = [sign(d1 ), sign(d2 ), . . . , sign(dn )]T . For e = col(e1 , . . . , e2n ) ∈
 r

− e2 e1 . . .
 
0 0
R2n , let ϕ : R2n 7→ Rn×2n be such that ϕ(e) =  ... .. . . .. ..
.
 
. . . .
0 0 ... −e2n e2n−1

2. Problem Formulation
Consider the following Euler–Lagrange equations with unknown dynamic uncertainties:

Mi (qi )q̈i + Ci (qi , q̇i )q̇i + Gi (qi ) = τi + di (1)

Electronics 2024, 1, 0 4 of 20

where i = 1, · · · , N, qi ∈ Rn and q̇i ∈ Rn are the joint variable vectors of generalized

position and velocity of the ith agent, Mi (qi ) ∈ Rn×n is the symmetric positive definite
inertia matrix, Ci (qi , q̇i ) ∈ Rn×n is the Coriolis and Centripetal forces, Gi (qi ) ∈ Rn is
the gravity term, Mi = Mi (qi ) + ∆Mi (qi ), Ci (qi , q̇i ) = Ci (qi , q̇i ) + ∆Ci (qi , q̇i ), Gi (qi ) =
Gi (qi ) + ∆Gi (qi ), Mi (qi ), Ci (qi , q̇i ) and Gi (qi ) are nominal values of inertia, Coriolis and
centripetal forces and gravity, respectively, ∆Mi (qi ), ∆Ci (qi , q̇i ) and ∆Gi (qi ) are unknown
dynamic uncertainties of inertia , Coriolis and centripetal forces and gravity, respectively,
τi ∈ Rn is the control torque and also represents the actuator’s output and di (t) ∈ Rn
signifies uncertain external disturbances.
The following exosystem produces the reference signal:

v̇ = Π( φ)v
(2)
q0 = Cv

where v ∈ Rm is the state, Π( φ) ∈ Rm×m is the system matrix, φ ∈ Rl is an unspecified

parameter vector, C ∈ Rn×m is a known constant matrix and q0 ∈ Rn is the desired
trajectory to track.
Inspired by [31], this paper considers the following actuator faults:

τi = τAi + [( Ei (t) − In )τAi + τ̄i ] = τAi + τFi (3)

where τAi is nominal actuator output, τFi is actuator fault, τ̄i is bounded bias fault, Ei (t) =
diag{ Ei1 , Ei2 , . . . , Ein } ∈ Rn×n , Ein ∈ (0, 1] is actuator effectiveness, In is the identity matrix.
The multiagent system consists of N followers as described in (1) and an uncertain
leader as defined in (2). The communication graph is represented by a directed graph
G = (V , E ), where V = {0, 1, · · · , N } is the node set and E ⊆ V × V is the edge set.
The edge set E is defined such that, for i = 1, · · · , N, j = 0, 1, · · · , N, i ̸= j, ( j, i ) ∈ E
if and only
 if (the node i can receive information of node j. The weighted adjacency
A = aij ∈ R N + 1 )×( N +1) , i, j ∈ V , where a = 0; a > 0, if ( j, i ) ∈ E ; otherwise a = 0.
ii ij ij
 N
Define the Laplacian matrix associated with G as H = hij i,j=1 , where hii = ∑ N j=0 aij
and hij = − aij for i ̸= j, i, j = 1, · · · , N. Define the subgraph G = (V , E ) of G where
V = {1, · · · , N } and E ⊆ V × V is obtained from E by eliminating all edges between the
node 0 and the nodes in V .
Fault-tolerant Cooperative Control Problem: Consider a multiagent system com-
posed of multiple uncertain Euler–Lagrange systems (1) subject to actuator faults (3) and a
leader system (2), a fixed graph G ; find an adaptive control law such that for any initial con-
ditions qi (0), q̇i (0) and v(0), the solution of the closed-loop system exists and is bounded
for all t ≥ 0 and the leader-following cooperative tracking error ei (t) = qi (t) − q0 (t) is
ultimately bounded, i = 1, · · · , N.

Remark 1. The fault-tolerant cooperative control problem is also addressed by [32], ensuring
uniform bounds on tracking errors.

To solve the above problem, we need the following assumptions.

Assumption 1. The actuator fault τFi is bounded by ∥τFi ∥ ≤ k τi , where the constant k τi > 0.

Assumption 2. The graph G contains a directed spanning tree with the leader as the root and the
subgraph G is undirected.

Assumption 3. The matrix Π( φ) is neutrally stable.

Remark 2. Assumption 1 is the standard assumption. The fault τFi in this article considers
partial loss failure fault and bounded bias fault of the actuator; thus, it can be assumed that τFi
Electronics 2024, 1, 0 5 of 20

is bounded [31]. Assumption 2 is standard in a communication topology graph for multiagent

systems and has been widely found in existing results [26,33]. Assumption 3 is also assumed
in [10,28] and it is not conservative. Under Assumption
 3, without loss of generality, we can
0 1
presume that Π( φ) = diag( φ) ⊗ a, where a = , diag(·) represents a diagonal matrix,
−1 0
φ = col( φ01 , · · · , φ0l ) ∈ Rl , m = 2l.

3. A Self-Adjusting Observer Design for the Uncertain Leader System

By delving deeper into prior research, we initially recollect the notion of observer as
introduced in [7]:

η̇i = S(ωi )ηi + µ1 ∑ η j − ηi

j∈N i
(4)
ω̇i = µ2 ϕ( ∑ (ηj − ηi ))ηi , i = 1, . . . , N
j∈N i

where η0 = v, for i = 1, . . . , N, ηi ∈ Rm , S(ωi ) is system matrix, ωi is unknown parameter

vector, S(ωi ) = diag(ωi ) ⊗ a, and µ1 and µ2 are positive constants.
The limitation of observer (4) lies in the requirement for a sufficiently large observer
gain µ1 to ensure the problem’s solvability, making it challenging to derive a suitable
value of µ1 for large values of N [34]. Therefore, in this section, we design the following
self-adjusting observer:

η̇i = Π( φi )ηi − κi ∑


µij ηi − η j
j∈N i

T
(5)
µ̇ij = k ij aij κi ηi − η j ηi − η j
φ̇i = −γϕ(ηi∗ )ηi

where i = 1, . . . , N, j = 0, 1, . . . , N; let ηvi = ∑ N T

j=0 aij ηi − η j , κi = κi ηvi ηvi is a suffi-
ciently smooth function, k ij = k ji > 0, i, j = 1, . . . , N, k i0 > 0, i = 1, . . . , N, µij = µ ji for
i, j = 1, . . . , N,

ηi − η0 , agent i is the leader ′ s child


ηi∗ = (6)
ηi − ηk + ηk + · · · − η p + η p − η0 , otherwise.

Let η̄i = ηi − η0 , φ̄i = φi − φ, for i = 1, . . . , N, j = 0, 1, . . . , N; let µ̄ij = µij − µ, with

some unknown constant µ > 0. Then, (5) can be written as:

η̄˙ i = Π( φ)η̄i + Π( φ̄i )ηi − κi ∑ µij η̄i − η̄ j

j∈N i

T (7)
µ̄˙ ij = k ij aij κi η̄i − η̄ j


η̄i − η̄ j
φ̄˙ i = −γϕ(ηi∗ )ηi , i = 1, . . . , N

Define F1 as follows:
 N
F1 = f ij i,j=1 (8)

where f ii = ∑ N j=0 µij aij , f ij = − µij aij when i ̸ = j, i, j = 1, · · · , N. Define F2 = diag ( µi0 ai0 , · · · ,
µ N0 a N0 ). Define κ = diag(κ1 , · · · , κ N ). Let η = col(η1 , η2 , . . . , η N ), η̄ = col(η̄1 , η̄2 , . . . , η̄ N ),
Electronics 2024, 1, 0 6 of 20

φ̄ = col( φ̄1 , φ̄2 , . . . , φ̄ N ), ηv = col(ηv1 , . . . , ηvN ), Sd ( φ̄) = block diag (S( φ̄1 ), . . . , S( φ̄ N )).
Then, the compact form for (7) is:

η̄˙ = ( IN ⊗ Π( φ))η̄ + Πd ( φ̄)η − ( F3 ⊗ Im )η + ( F4 ⊗ Im )(1 N ⊗ v), (9a)

φ̄˙ = −γϕd (η ∗ )η. (9b)

where F3 = κF1 , F4 = κF2 , ϕd (η̄ ) =
block diag (ϕ
(η̄1 ), . . . , ϕ(η̄ N )), for i = 1, . . . , N, j =
0, 1, . . . , N, let ηvi = ∑ N a η
j=0 ij i − η j , κ i = κ η Tη
i vi vi is a sufficiently smooth nondecreasing
function and κi (·) ≥ 1, k ij = k ji > 0, i, j = 1, . . . , N, k i0 > 0, i = 1, . . . , N, µij = µ ji for i, j =
1, . . . , N. Define κ = diag(κ1 , · · · , κ N ). Let µ̄ = col(µ10 , . . . , µ N0 , µ12 , . . . , µ1N , . . . , µ( N −1)1 ,
. . . , µ ( N −1) N ).

Lemma 1. Given Systems (2) and (5), under Assumptions 2 and 3, there exist µij > 0, k ij > 0
and a smooth nondecreasing function κi (·) ≥ 1 , i = 1, . . . , N, such that for any η0 ∈ Rm , ηi (t)
exists and is bounded and we have: (1) limt→∞ η̄i (t) = 0 asymptotically; (2) limt→∞ φ̄˙ (t) =
0 asymptotically; (3) limt→∞ Π( φ̄i )ηi (t) = 0 asymptotically; (4) µ̄ij (t), i = 1, . . . , N, j =
0, 1, . . . , N, are bounded for all t ≥ 0.

Proof. The energy function for system (7) is presented as follows:

W (η̄, φ̄, µ̄, t) = W1 (η̄, t) + W2 ( φ̄, t) + W3 (µ̄, t) (10)

where

1 N T
2 i∑
W1 (η̄, t) = η̄i η̄i
=1
1 N −1 T
2 i∑
W2 ( φ̄, t) = γ φ̄i φ̄i (11)
=1
 
1 N µ̄2ij µ̄2i0
W3 (µ̄, t) = ∑ ∑ 2kij + ∑ ki0 
2 i=1 j∈N (t) 0∈ N̄ (t)
i i

The derivatives of the functions W1 ,W2 and W3 along system (7) are as follows:

N
∑ η̄iT (Π( φ)η̄i + Π( φ̄i )ηi − κi ∑


Ẇ1 (η̄, t) = µij η̄i − η̄ j )
i =1 j∈N i
(12)
N N
∑ η̄iT (Π( φ)η̄i + Π( φ̄i )ηi ) − ∑ η̄iT κi ∑


= µij η̄i − η̄ j
i =1 i =1 j∈N i

Since Π( φ) is the skew symmetric, we have

N N
∑ η̄iT Π( φ̄i )ηi − ∑ η̄iT κi ∑


Ẇ1 (η̄, t) = µij η̄i − η̄ j (13)
i =1 i =1 j∈N i
Electronics 2024, 1, 0 7 of 20

By proposition in [7], for i = 1, · · · , N, we have, η̄iT Π( φ̄i )ηi = φ̄iT ϕ(η̄i )ηi , then

N N
∑ φ̄iT ϕ(η̄i )ηi − ∑ η̄iT κi ∑


Ẇ1 (η̄, t) = µij η̄i − η̄ j
i =1 i =1 j∈N i
N N
∑ φ̄iT ϕ(η̄i )ηi − ∑ η̄iT κi ∑


= µij η̄i − η̄ j (14)
i =1 i =1 j∈N i
N N
∑ φ̄iT ϕ(η̄i )ηi − ∑ ∑ κi µij η̄iT η̄i − η̄ j


=
i =1 i =1 j∈N i

The derivative of W2 along (7) is :

1 N −1 T N
Ẇ2 ( φ̄, t) = ∑
2 i =1
γ φ̄i φ̄˙ i = − ∑ φ̄iT ϕ( ηi∗ )ηi (15)
i =1

The derivative of W3 along (7) is :

!
N µ̄ij µ̄˙ ij µ̄ µ̄˙
Ẇ3 (µ̄, t) = ∑ ∑ 2k ij
+ ai0 i0 i0
k i0
i =1 j∈Ni
N N
µ̄ij

µ̄i0
= ∑ ∑ 2kij ij ij i i j i j
k a κ ( η̄ − η̄ ) T
( η̄ − η̄ ) + a i0 k κ η̄
k i0 i0 i i i
T
η̄ (16)
i =1 j =1
!
N N N N
1
= ∑ κi ∑ aij µ̄ij η̄i η̄i − η̄ j + ∑ aij µ̄ij η̄ j η̄ j − η̄i
T T
+ ∑ κi ai0 µ̄i0 η̄iT η̄i



i =1
2 j =1 j =1 i =1

Since aij µ̄ij = a ji µ̄ ji , i, j = 1, . . . , N, we have

N N N N
∑ ∑ aij µ̄ij η̄jT ∑ ∑ aij µ̄ij η̄iT



η̄ j − η̄i = η̄i − η̄ j (17)
i =1 j =1 i =1 j =1

Thus,

N
∑ κi ( ∑ µ̄ij η̄iT η̄i − η̄ j + ∑ µ̄i0 η̄iT η̄i )


Ẇ3 (µ̄, t) =
i =1 j∈Ni 0∈ N̄i (t)
(18)
N
∑∑ κi µ̄ij η̄iT


= η̄i − η̄ j
i =1 j∈ N̄i

Combining (14), (15) and (18), we can obtain

N N N
∑ φ̄iT ϕ(η̄i )ηi − ∑ ∑ κi µij η̄iT η̄i − η̄ j + ∑ ∑ κi µ̄ij η̄iT η̄i − η̄ j



Ẇ (η̄, φ̄, µ̄, t) =
i =1 i =1 j∈N i i =1 j∈ N̄i
N
− ∑ φ̄iT ϕ( ηi∗ )ηi (19)
i =1
N N N
= ∑ φ̄iT ϕ(η̄i )ηi − ∑ ∑ µκi η̄iT (η̄i − η̄ j ) − ∑ φ̄iT ϕ( ηi∗ )ηi
i =1 i =1 j∈ N̄i i =1
Electronics 2024, 1, 0 8 of 20

By the definition of ηi∗ in (6), we can obtain ∑iN=1 φ̄iT ϕ(η̄i )ηi − ∑iN=1 φ̄iT ϕ( ηi∗ )ηi = 0; thus,

N
Ẇ (η̄, φ̄, µ̄, t) = − ∑ ∑ µκi η̄iT (η̄i − η̄ j ) (20)
i =1 j∈ N̄i

By Lemma 4 in [35], under Assumption 2, H is positive definite and symmetric; thus, the
eigenvalue of H is a positive real number. We define κ = diag(κ1 , κ2 , · · · , κ N ) above, then
we have

Ẇ (η̄, φ̄, µ̄, t) = − µη̄ T (κH ⊗ Im )η̄

(21)
≤ − µλmin (κH )∥η̄ ∥2 ≤ 0

Thus, limt→∞ W (t) exists, which implies that η̄, φ̄ and µ̄ are all bounded. v is bounded by
Assumption 3; thus, η = η̄ + v is bounded. µ̄ij and µij are bounded.
Rt
Next, define W4 (t) = 0 µλmin (κH )∥η̄ (τ )∥2 dτ. If one can show

lim Ẇ4 (t) = lim µλmin (κH )∥η̄ (t)∥2 = 0, (22)

t→∞ t→∞

then, one can conclude limt→∞ η̄ (t) = 0. Integrating both sides of (21) gives
Z t
W4 (t) ≤ −Ẇ (r )dr = W (0) − W (t) ≤ W (0). (23)
0

Since Ẇ4 (t) = µλmin (κH )∥η̄ (t)∥2 ≥ 0, then limt→∞ W4 (t) exists. Next, we analyze whether
Ẅ4 (t) is bounded.

Ẅ4 (t) = 2µλmin (κH )η̄ (t)η̄˙ (t). (24)

Since η̄ and η have already been proven to be bounded as shown above, combining
Equation (9a) leads to the conclusion that η̄˙ is also bounded, thereby allowing us to
conclude that Ẅ4 (t) is bounded as well. Thus, applying Barbalat’s lemma, we can show
limt→∞ Ẇ4 (t) = 0; thus, we can obtain

lim η̄ (t) = 0 (25)

t→∞

which together with (9b) yields limt→∞ φ̄˙ (t) = 0 asymptotically. Differentiating η̄˙ i gives

η̄¨ i = Π( φ̇)η̄i + Π( φ)η̄˙ i + Π( φ̄˙ i )ηi + Π( φ̄i )η̇i − κ̇i ∑

µij η̄i − η̄ j
j∈N i
(26)
∑ η̄˙ i − η̄˙ j


− κi µij
j∈N i

We have shown η̄i , ηi , φ̄, µ̄ij , κi (·), µij , η̄˙ i , φ̄˙ , µ̄˙ ij , κ̇i (·) are all bounded on t ≥ 0; thus, φ̄i
and φ̄˙ i are all bounded on t ≥ 0; by (5), we can obtain that η̇i is bounded on t ≥ 0; thus,
η̄¨ i is bounded. Then, by Lemma 8.2 in [36], we obtain limt→∞ η̄˙ i (t) = 0, which together
with η̄i = Π( φ)η̄i + Π( φ̄i )ηi − κi ∑ j∈N µij η̄i − η̄ j in (7) yields limt→∞ Π( φ̄i )ηi (t) = 0
˙


i
asymptotically.

4. Adaptive Fault-Tolerant Cooperative Controller Design

For the solvability of the fault-tolerant cooperative control problem, we need three
properties for system (1) [8,37].

Property 1. There are two positive constants k mi and k m̄i , such that k mi Im ≤ Mi (qi ) ≤ k m̄i Im .
Electronics 2024, 1, 0 9 of 20

Property 2. There are two positive constants k ci and k gi , such that ∥Ci (qi , q̇i )∥ ≤ k ci ∥q̇i ∥ and
∥Gi (qi )∥ ≤ k gi , respectively..

Property 3. The matrix Ṁi (qi ) − 2Ci (qi , q̇i ) is skew-symmetric.

 T T T T
, x2i = qiT , q̇iT ; the dynamics model of (1) subject to actuator faults

Let xi = x1i
(3) is equivalently rearranged as the following form:

ẋ1i = x2i
(27)
ẋ2i = Mi−1 ( x1i )(τAi + Fi ( xi ) + Di ( xi , t))

where Fi ( xi ) = −(Ci ( xi ) x2i + Gi ( x1i )) and the lumped uncertainty Di ( xi , t) = di + τFi −

∆Mi ( x1i ) ẋ2i − ∆Ci ( xi ) x2i − ∆Gi ( x1i ).
Using the self-adjusting observer ηi for the leader signal v in (2) which is designed in
(5), we further design an adaptive controller to solve the fault-tolerant cooperative control
problem for the multiple uncertain Euler–Lagrange systems. The definition of the tracking
error z1i is as follows:

z1i = x1i − Cηi (28)

The second error z2i is defined as follows:

z2i = x2i − αi (29)

where αi is the virtual control, αi = −Ki z1i + C η̇i , Ki = ai In , ai is a positive constant and In
is an n-dimensional identity matrix. Differentiating z1i , we obtain

ż1i = ẋ1i − C η̇i = −Ki z1i + z2i (30)

The time derivative of z2i is

ż2i = Mi−1 ( xi )(τAi + Fi ( xi ) + Di ( xi , t)) − α̇i (31)

By Theorem 1 in [38], we can choose a nonsingular fast terminal sliding mode surface

λ p /qi
si = z1i + k1i z1ii + k2i z2ii (32)

where k1i = diag(k1i1 , k1i2 , · · · , k1in ) and k2i = diag(k2i1 , k2i2 , · · · , k2in ) are positive ma-
trices, pi and qi are positive odd numbers satisfying the relations 1 < pi /qi < 2 and
λi > pi /qi . If si = 0, the convergence time T is defined as T = max { Ti }, i = 1, · · · , N,
where
q /pi
Z |z (0)|
1i k2ii
Ti =  qi /pi dz1i
0 λ
z1i (t) + k1i z1ii
(33)
pi 1−qi /pi
qi | z1i (0)| pi /qi − 1 pi /qi − 1
= × Fi (qi /pi , ;1+ ; −k1i |z1i (0)|λi −1 )
k1i ( pi /qi − 1) (λi − 1) pi /qi (λi − 1) pi /qi

where z1i (0) is the initial value of z1i (t) and Fi (·) is Gauss’ hypergeometric function.
As the quantity of follower agents expands, the task of computing the upper limit
becomes more cumbersome and involved. Therefore, we design an adaptive method to
T
estimate the lumped uncertainty Di , denoted as Di = Di1 , · · · , Din ∈ Rn ; we establish


the lumped uncertainty estimation as D̂i = [ D̂i1 , · · · , D̂in ] T ∈ Rn . Subsequently, the

estimation error is expressed as D̃i = Di − D̂i = [ D̃i1 , · · · , D̃in ] T ∈ Rn . Afterwards, we
offer the subsequent assumption.
Electronics 2024, 1, 0 10 of 20

Assumption 4. It is assumed that the lumped uncertainty Di ( xi , t) is upper bounded [31] and
the rate of change is also upper bounded [39]; i.e., there exist unknown constants D̄im > 0
and Ξim , m = 1, · · · , n, such that ∀t > 0 and || Dim (t)|| ≤ D̄im , such that ∀t > 0 and
||(d/dt) Dim (t)|| ≤ Ξim .

Remark 3. In [39], the norm of partial derivatives for the unknown component is bounded, which
is composed of system state variables, disturbances and actuator faults. In this paper, our lumped
uncertainty Di ( xi , t) is also composed of system state variables, disturbances and actuator faults;
thus, it is reasonable to make Assumption 4 in the following theoretical analysis.

The controller τAi is designed as

τAi = τ1i + τ2i + τ3i (34)

where
q i −1
τ1i = − k ( Mi ż1i + Ci si + Mi k1i λi ż1i ⊙ |z1i |λi −1 + σi |si | ⊙ sgn(si )) ⊙ |z2i |1− pi /qi (35)
pi 2i
p
q i −1 T + T 1− i
τ2i = − k2i ((si ) z1i z2i ) ⊙ |z2i | qi (36)
pi
τ3i = Mi α̇i − Fi − D̂i (37)

and the adaption law is designed as

p
D̂˙ im = sim k2im i |z2im |( pi /qi )−1 − δim D̂im (38)
qi

where i = 1, · · · , N, m = 1, · · · , n, (siT )+ is the Moore–Penrose pseudoinverse of siT , which

0, if siT = 0

T T

+
satisfies si si = , σi and δim are positive constants.
1, others

Theorem 1. Consider a multiagent system consisting of (27) and (2), a fixed graph G and assume
that Assumptions 1–4 hold. Then, take the adaptive fault-tolerant controller (34), such that for any
initial conditions qi (0), q̇i (0), ηi (0), v(0), i = 1, · · · , N, the solution of the closed-loop system
exists and is bounded for all t ≥ 0 and the tracking errors z1i , si and the lumped uncertainty
estimation error D̃i are ultimately bounded.

Proof. The Lyapunov function candidate is proposed

N
1 T 1 N T 1 N n
V= ∑ 2 1i 1i 2 ∑ i i i 2 ∑ ∑ D̃im
z z + s M s + 2
. (39)
i =1 i =1 i =1 m =1
Electronics 2024, 1, 0 11 of 20

Differentiating (39), we have

N
V̇ = ∑ (−z1i
T T
Ki z1i + z1i z2i + siT Ci si + siT ( Mi ż1i + Mi k1i λi ż1i ⊙ |z1i |λi −1 )
i =1
N n
pi
+ siT k2i ((τAi + Fi + Di − Mi α̇i ) ⊙ |z2i |( pi /qi )−1 )) + ∑ ∑ D̃im ( Ḋim − D̂˙ im )
qi i =1 m =1
N
≤ ∑ (−z1i
T T
Ki z1i + z1i z2i + siT Ci si + siT ( Mi ż1i + Mi k1i λi ż1i ⊙ |z1i |λi −1 ) (40)
i =1
N n
pi
+ siT k2i ((τAi + Fi + Di − Mi α̇i ) ⊙ |z2i |( pi /qi )−1 )) + ∑ ∑ D̃im Ξim
qi i =1 m =1
N n
−∑ ∑ D̃im D̂˙ im
i =1 m =1

Substituting (34) into (40), we have

N n N n
p
V̇ ≤ ∑ (−z1i
T
Ki z1i − σi ∥si ∥2 + ∑ D̃im Ξim ) − ∑ ( ∑ D̃im D̂˙ im + siT k2i i (( Di − D̂i )
i =1 m =1 i =1 m =1
qi
⊙ |z2i |( pi /qi )−1 ))
N n n
≤ ∑ (−z1i
T
Ki z1i − σi ∥si ∥2 + ∑ D̃im Ξim + ∑ D̃im δim D̂im )
i =1 m =1 m =1
N N n
≤ ∑ (−z1i
T
Ki z1i − σi ∥si ∥2 ) + ∑ ∑ D̃im (Ξim + δim Dim − δim D̃im ) (41)
i =1 i =1 m =1
N N n N n N n
≤ ∑ (−z1i
T
Ki z1i − σi ∥si ∥2 ) − ∑ ∑ 2
δim D̃im +∑ ∑ D̃im Ξim + ∑ ∑ D̃im δim Dim
i =1 i =1 m =1 i =1 m =1 i =1 m =1
N N n nN
1 1
≤ ∑ (−z1i
T
Ki z1i − σi ∥si ∥2 ) − ∑ ∑ 2
δim D̃im + ∑ ∑ ( D̃im 2 2
+ Ξ2im − δim D̃im
i =1
2 i =1 m =1 2 i =1 m =1
+ 2 D̃im δim Dim )

where

1 N n
2 i∑ ∑ (−δim D̃im
2
+ 2 D̃im δim Dim )
=1 m =1
1 N n N n
≤− ∑ ∑
2 i =1 m =1
2
δim ( Dim 2
− 2Dim D̂im + D̂im ) + ∑ ∑ ( Dim − D̂im )δim Dim
i =1 m =1

1 N n
2 i∑ ∑ δim ( Dim
2 2
≤ − D̂im ) (42)
=1 m =1
1 N n
2 i∑ ∑ δim Dim
2
≤
=1 m =1
1 N n
2 i∑ ∑ δim D̄im
2
≤
=1 m =1
Electronics 2024, 1, 0 12 of 20

Hense, substituting (42) into (41), we can have

N
1 N n 1 N n
V̇ ≤ ∑ (−z1i
T
Ki z1i − σi ∥si ∥2 ) − ∑ ∑ 2
δim D̃im + ∑ ∑ ( D̃im 2 2
+ Ξ2im + δim D̄im )
i =1
2 i =1 m =1 2 i =1 m =1
N
1 N n 1 N n (43)
≤ ∑ (−z1i
T
Ki z1i − σi ∥si ∥2 ) − ∑ ∑ 2
(δim − 1) D̃im + ∑ ∑ (Ξ2im + δim D̄im
2
)
i =1
2 i =1 m =1 2 i =1 m =1
≤ − ρi V + ci

where ρi , ci are two positive constants provided as

 
2 min{σi }
ρi = min 2λmin (Ki ), , min{δim } − 1 (44)
λmax ( Mi ( x1i ))
1 N n
2 i∑ ∑ (Ξ2im + δim D̄im
2
ci = ), (45)
=1 m =1

where i = 1, · · · , N and m = 1, · · · , n. In order to guarantee that ρi > 0, the gains δim are
chosen to satisfy

min{δim } − 1 > 0. (46)

By multiplying both sides with eρi t in Equation (43), we obtain

V (t) ≤ (V (0) − ci /ρi )e−ρi t + ci /ρi

(47)
≤ V (0) + ci /ρi

Define

Pi = 2(V (0) + ci /ρi ) (48)

we have
T
z1i z1i ≤ Pi
siT Mi ( x1i )si ≤ Pi
(49)
n
∑ 2
D̃im 2
= ∥ D̃i ∥ ≤ Pi .
m =1

Thus, we obtain the compact sets

n p o
Υz1i = z1i ∈ Rn | ∥z1i ∥ ≤ Pi
 q 
Υsi = si ∈ Rn | ∥si ∥ ≤ Pi /λmin ( Mi ) (50)
n p o
Υ D̃i = D̃i ∈ Rn | ∥ D̃i ∥ ≤ Pi

where Pi = 2(V (0) + ci /ρi ) with ci , ρi given in (45) and (44), respectively. Consequently,
we can obtain that z1i and si and D̃i are ultimately bounded. By (25), we can obtain
limt→∞ (Cηi (t) − Cη0 (t)) = 0 and combine the above z1i = x1i − Cηi , which is ultimately
bounded; thus, we can obtain that ei = x1i − Cη0 is ultimately bounded; thus, the fault-
tolerant cooperative problem can be solved and the tracking error ei can be made as small
as possible.

Remark 4. By (48) and (49), we can see that ρi and ci determine the size of ∥z1i ∥, ∥si ∥ and ∥ D̃i ∥;
that is, the smaller the tracking errors, the bigger ρi and the smaller ci should be. By (44), we can
see that the bigger ρi is, the bigger the control parameters ai , σi and δim should be. By (45), we can
Electronics 2024, 1, 0 13 of 20

see that the smaller ci is, the smaller the parameter δim should be. Combining (44), (45) and (48),
the influence of δim becoming smaller in (45) is greater than the influence of δim becoming larger in
(44) for Pi , i.e., δim becoming smaller can make Pi smaller. Thus, the right sides of (50) can be set as
small as possible, i.e., the tracking error ei is guaranteed to be small enough by choosing the small
parameter δim and the big parameters ai , σi . Compared with the semi-globally bounded tracking
error in [40], the tracking error in this paper is globally bounded and the bound can be sufficiently
small from the above analysis.

5. Simulation Studies
In this section, we consider a group of six robotic manipulators given by (1), where
 T
qi = q1i , q2i , i = 1, · · · , 6 and
   
M11i M12i C11i C12i  T
Mi ( q i ) = , C (qi , q̇i ) = , G ( qi ) = G1i G2i (51)
M21i M22i C21i C22i

where
2 2
M11i = (m1i + m2i )r1i + m2i r2i + 2m2i r1i r2i cos(q2i ) + J1i ,
2
M12i = m2i r2i + m2i r1i r2i cos(q2i ),
2
M21i = m2i r2i + m2i r1i r2i cos(q2i ),
2
M22i = m2i r2i + J2i ,
C11i = −m2i r1i r2i sin(q2i )q̇2i ,
(52)
C12i = −m2i r1i r2i sin(q2i )(q̇1i + q̇2i ),
C21i = m2i r1i r2i sin(q2i )q̇1i,
C22i = 0,
G1i = (m1i + m2i )r1i g cos(q1i ) + m2i r2i g cos(q1i + q2i ),
G2i = m2i r2i g cos(q1i + q2i )

where m1i and m2i are the mass of two joints, r1i and r2i are the length of two joints, J1i and
J2i are the moment inertia and g is the gravity acceleration constant. The parameters of the
system are selected as m1i = 0.5 kg, m2i = 1.5 kg, r1i = 1 m, r2i = 0.8 m, J1i = 5 kg · m2 ,
J2i = 5 kg · m2 , and g = 9.81 m/s2 . The parameter uncertainties of the system are assumed
as ∆m1i = 0.3 ∗ rand and ∆m2i = 0.3 ∗ rand, where “rand” denotes the random value
of interval [0, 1]. The disturbance is set as di (t) = [0.5sin(2t) + 0.3sin(q1i ), 0.3cos(2t) +
0.4sin(q2i )] T .
The agents’ communication network is depicted in Figure 1, thus fulfilling
Assumption 2.
Electronics 2024, 1, 0 14 of 20

Figure 1. Communication network Ḡ .

The leader’s signal is

q0 = col( A1 sin( φ01 t + ϕ1 ), A2 sin( φ02 t + ϕ2 )) (53)

where the values of A1 , A2 , φ01 and φ02 can be any unknown positive real number, while
ϕ1 and ϕ2 can be any arbitrary unknown real number.The leader’s  signal is produced by

0 1 1 0 0 0
(2) with v = col (v11 , v12 , v21 , v22 ), Π( φ) = diag( φ) ⊗ , C= ,
−1 0 0 0 1 0
 T
where diag(·) is a diagonal matrix and φ = col( φ01 , φ02 ) = 4 2 . Thus, Assumption 3
is also satisfied. The software used for simulation is Matlab2021b.

5.1. Simulation for Observer

Initially, we furnish the simulation outcomes for the self-adjusting observer. According
to Lemma 1, we have the ability to formulate the observer for the leader (2) in the following
manner:

η̇i = Π( φi )ηi − κi ∑ µij ηi − η j

j∈N i

T
(54)
µ̇ij = k ij aij κi ηi − η j ηi − η j
φ̇i = −γϕ(ηi∗ )ηi , i = 1, · · · , N

2
where i = 1, · · · , 6, j = 0, 1, · · · , 6, let ηvi = ∑6j=0 aij ηi − η j , κi = 5 + ηvi
Tη


vi , k ij = k ji =
1, i, j = 1, . . . , 6, k i0 = 1, i = 1, . . . , 6, γ = 40 and φi ∈ R2 and randomly generated initial
conditions. The observer’s performance is depicted in Figure 2, which shows that the
estimation errors asymptotically converge to zero. Figure 3 shows that the dynamic gain of
the observer has the ability to adaptively converge towards the desired constant. Figure 4
shows that the estimation error of the unknown leader system’s parameter φ̄i can converge
to some constant, which means that the limt→∞ φ̄˙ i (t) = 0 can be achieved.
The controller designed in this article is suitable for both actuator health and the
situation of actuator failure. Two sets of simulation experiments will be conducted to verify
these two situations.
Electronics 2024, 1, 0 15 of 20

1 3.5
-v 112
-v12
111 11
0.8 3 -v12
211
-v11 212

-v11 312
-v12
0.6 311
-v11 2.5 412
-v12
411
0.4 -v11 512
-v12
511
-v11 2 -v
611 612 12
0.2

-v12
-v11
0 1.5

i12
i11
-0.2
1
-0.4
0.5
-0.6
0
-0.8

-1 -0.5
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Time t (sec) Time t (sec)
4 3
-v -v22
121 21 122
3.5 221
-v21 -v22
2.5 222
-v -v22
321 21 322
3 421
-v21 -v22
422
-v21 2
521 -v22
522
2.5 621
-v21 -v22
622
1.5
-v21

-v22
2
i21

i22
1
1.5

0.5
1

0.5 0

0 -0.5
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Time t (sec) Time t (sec)

Figure 2. The error output of adaptive observer ηi11 − v11 and ηi12 − v12 , ηi21 − v21 and ηi22 − v22 ,
i = 1, · · · , 6.

4.5
10
4
14

24
3.5
34

3 45

56
2.5
ij

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3
Time t (sec)

Figure 3. Dynamic gain µij , i = 1, · · · , 6 , j ∈ N̄i .

Electronics 2024, 1, 0 16 of 20

-1 -0.8

-1
-2
-1.2

-1.4
-3
-1.6

-4 -1.8

-2
-5
-2.2

-2.4
-6
-2.6

-7 -2.8
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Time t (sec) Time t (sec)

Figure 4. Trajectory of φ̄i1 and φ̄i1 , i = 1, · · · , 6.

5.2. Simulation for Adaptive Fault-Tolerant Controller with Healthy Actuator

The actuators are healthy with

Ei1 = 1, Ei2 = 1
(55)
τ̄i1 = 0, τ̄i2 = 0

Based on the observer, we can design the adaptive fault-tolerant controller, which is given
in (34), where i = 1, · · · , 6, j = 1, 2, the designed parameters δij = 2, σi = 100, Ki = 100I2×2 ,
k1i = 300I2×2 , k2i = 30I2×2 , pi = 13, qi = 11, λi = 3. The initial values of joint angle
position and angle velocity are chosen as qi (0) = [1, 2] T and q̇i (0) = [0.3, 0.1] T , respectively,
and there are randomly generated initial conditions. The performance of Controller 1 can
be found in Figure 5. It can be observed that the tracking errors can converge as close to
zero as possible. Figure 6 shows that the adaptive parameters D̂i1 and D̂i2 can adaptively
converge to the required constant.

3 2.5
q11-q01 q -q
12 02
2.5 q21-q01 q22-q02
2
q31-q01 q -q
32 02
2 q41-q01 q42-q02
q51-q01 1.5
q52-q02
1.5 q61-q01 q -q
62 02
1
qi1-q 01

qi2-q 02

1
0.5
0.5

0
0

-0.5 -0.5

-1 -1
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Time t (sec) Time t (sec)

100 60

40
50
20

0
0

-50 -20

-40
-100
-60

-150
-80

-200 -100
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Time t (sec) Time t (sec)

Figure 5. The tracking error qi1 − q01 and qi2 − q02 , q̇i1 − q̇01 and q̇i2 − q̇02 , i = 1, · · · , 6.
Electronics 2024, 1, 0 17 of 20

104 104
6 2.5

5
2

1.5
3

2 1

1
0.5

0
-1

-2 -0.5
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Time t (sec) Time t (sec)

Figure 6. The adaptive parameters D̂i1 and D̂i2 , i = 1, · · · , 6.

5.3. Simulation for Adaptive Fault-Tolerant Controller with Actuator Fault

The actuators suffer from partial loss failure fault and bias fault, which are established
in the following manner
 
1.0 if t < 1 s 1.0 if t < 2 s
Ei1 = , Ei2 =
0.5 otherwise 0.4 otherwise (56)
τ̄i1 = 0.2, τ̄i2 = −0.2

Obviously, the fault values above satisfy Assumption 1. Based on the observer, we can
design the adaptive fault-tolerant controller, which is given in (34), where i = 1, · · · , 6, j =
1, 2, the designed parameters δij = 2, σi = 100, Ki = 100I2×2 , k1i = 300I2×2 , k2i = 30I2×2 ,
pi = 13, qi = 11, λi = 3. The initial values of joint angle position and angle velocity are
chosen as qi (0) = [2, 0.8] T and q̇i (0) = [0.2, 0.4] T , respectively, and there are randomly
generated initial conditions. Based on the bounded disturbances, faults and uncertain
parameters above, it can be determined that Assumption 4 is satisfied. The performance
of Controller 1 can be found in Figure 7. It can be observed that the tracking errors can
converge as close to zero as possible. Figure 8 shows that the adaptive parameters D̂i1 and
D̂i2 can adaptively converge to the required constant.
Electronics 2024, 1, 0 18 of 20

2 3
q11-q01 q -q
12 02

1.5 q21-q01 q -q
2.5 22 02
q31-q01 q -q
32 02

1 q41-q01 q42-q02
2
q51-q01 q52-q02

0.5 q61-q01 q62-q02

1.5

qi2-q 02
qi1-q 01
0
1
-0.5

0.5
-1

0
-1.5

-2 -0.5
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Time t (sec) Time t (sec)
60 20

40 0

-20
20

-40
0

-60
-20
-80

-40
-100

-60
-120

-80 -140
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Time t (sec) Time t (sec)

Figure 7. The tracking error qi1 − q01 and qi2 − q02 , q̇i1 − q̇01 and q̇i2 − q̇02 , i = 1, · · · , 6.

104 104
3 3

2 2.5

1 2

0 1.5

-1 1

-2 0.5

-3 0

-4 -0.5
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Time t (sec) Time t (sec)

Figure 8. The adaptive parameters D̂i1 and D̂i2 , i = 1, · · · , 6.

6. Conclusions
This paper investigates the fault-tolerant cooperative control problem in multiagent
systems, which are characterized by an uncertain leader system and multiple uncertain
Euler–Lagrange systems with actuator faults. A self-adjusting observer is initially designed
to estimate the uncertain leader’s signal for different followers and calculate the observer
gain in real time. We further synthesize the controller based on the above observer and
a nonsingular fast terminal sliding mode surface to address the control problem. The
adaptive approach is applied to estimate the aggregate uncertainty, removing the necessity
to compute its upper limit. This method is robust to actuator faults, uncertain parameters
and external disturbances. Future research will consider fault-tolerant cooperative control
of strongly nonlinear multiagent systems in complex communication topologies.
Electronics 2024, 1, 0 19 of 20

Author Contributions: Methodology, R.G.; software, R.G.; validation, X.S.; writing—original draft
preparation, R.G.; writing—review and editing, R.G. All authors have read and agreed to the pub-
lished version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: Data are contained within the article.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Zhao, X.; Zheng, X.; Ma, C.; Li, R. Distributed consensus of multiple Euler–Lagrange systems networked by sampled-data
information with transmission delays and data packet dropouts. IEEE Trans. Autom. Sci. Eng. 2015, 14, 1440–1450.
2. Chen, Q.; Jin, Y.; Song, Y. Fault-tolerant adaptive tracking control of Euler–Lagrange systems–An echo state network approach
driven by reinforcement learning. Neurocomputing 2022, 484, 109–116.
3. Hu, Y.; Geng, Y.; Wu, B.; Wang, D. Model-free prescribed performance control for spacecraft attitude tracking. IEEE Trans. Control
Syst. Technol. 2020, 29, 165–179.
4. Xiao, B.; Hu, Q.; Zhang, Y. Adaptive sliding mode fault tolerant attitude tracking control for flexible spacecraft under actuator
saturation. IEEE Trans. Control Syst. Technol. 2011, 20, 1605–1612.
5. Han, Q.; Liu, Z.; Su, H.; Liu, X. Filter-Based Disturbance Observer and Adaptive Control for Euler–Lagrange Systems With
Application to a Quadrotor UAV. IEEE Trans. Ind. Electron. 2022, 70, 8437–8445.
6. Bernardes, E.; Boyer, F.; Viollet, S. Modelling, control and simulation of a single rotor UAV with swashplateless torque modulation.
Aerosp. Sci. Technol. 2023, 140, 108433.
7. Wang, S.; Huang, J. Adaptive leader-following consensus for multiple Euler–Lagrange systems with an uncertain leader system.
IEEE Trans. Neural Netw. Learn. Syst. 2018, 30, 2188–2196.
8. Dong, Y.; Chen, Z. Fixed-time synchronization of networked uncertain Euler–Lagrange systems. Automatica 2022, 146, 110571.
9. Naderolasli, A.; Shojaei, K.; Chatraei, A. Finite-time velocity-free adaptive neural constrained cooperative control of Euler–
Lagrange systems. Trans. Inst. Meas. Control 2023, https://doi.org/10.1177/014233122311546.
10. Cai, H.; Su, Y.; Huang, J. Cooperative Control of Multi-Agent Systems: Distributed-Observer and Distributed-Internal-Model Approaches;
Springer Nature: Cham, Switzerland, 2022.
11. Chen, T.; Shan, J.; Wen, H. Distributed adaptive attitude control for networked underactuated flexible spacecraft. IEEE Trans.
Aerosp. Electron. Syst. 2018, 55, 215–225.
12. Chen, L.; Li, C.; Sun, Y.; Ma, G. Distributed finite-time tracking control for multiple uncertain Euler–Lagrange systems with error
constraints. Int. J. Control 2021, 94, 698–710.
13. Chen, G. Cooperative controller design for synchronization of networked uncertain Euler–Lagrange systems. Int. J. Robust
Nonlinear Control 2015, 25, 1721–1738.
14. Li, D.; Ge, S.S.; He, W.; Li, C.; Ma, G. Distributed formation control of multiple Euler–Lagrange systems: A multilayer framework.
IEEE Trans. Cybern. 2020, 52, 3325–3332.
15. Amin, A.A.; Hasan, K.M. A review of Fault Tolerant Control Systems: Advancements and applications. Measurement 2019,
143, 58–68.
16. Van, M.; Ceglarek, D. Robust fault tolerant control of robot manipulators with global fixed-time convergence. J. Frankl. Inst. 2021,
358, 699–722.
17. Van, M.; Mavrovouniotis, M.; Ge, S.S. An adaptive backstepping nonsingular fast terminal sliding mode control for robust fault
tolerant control of robot manipulators. IEEE Trans. Syst. Man Cybern. Syst. 2018, 49, 1448–1458.
18. Chen, G.; Song, Y.; Lewis, F.L. Distributed fault-tolerant control of networked uncertain Euler–Lagrange systems under actuator
faults. IEEE Trans. Cybern. 2016, 47, 1706–1718.
19. Li, X.; Qin, H.; Li, L.; Sun, Y. Adaptive fault-tolerant control for multiple Euler–Lagrange systems considering time delays and
output constraints. Asian J. Control 2023, 25, 2822–2837.
20. Meng, Q.; Zhang, T.; Gao, X.; Song, J. Adaptive sliding mode fault-tolerant control of the uncertain stewart platform based on
offline multibody dynamics. IEEE/Asme Trans. Mechatron. 2013, 19, 882–894.
21. Lu, Z.; Li, Y.; Fan, X.; Li, Y. Decentralized fault tolerant control for modular robot manipulators via integral terminal sliding mode
and disturbance observer. Int. J. Control Autom. Syst. 2022, 20, 3274–3284.
22. Chen, Z.; Wang, X.s.; Cheng, Y. Model free based finite time fault-tolerant control of robot manipulators subject to disturbances
and input saturation. Int. J. Robust Nonlinear Control 2022, 32, 5281–5303.
23. Li, H.; Liu, C.L.; Zhang, Y.; Chen, Y.Y. Adaptive neural networks-based fixed-time fault-tolerant consensus tracking for uncertain
multiple Euler–Lagrange systems. ISA Trans. 2022, 129, 102–113.
24. Huang, J. Nonlinear Output Regulation: Theory and Applications; SIAM—Society for Industrial and Applied Mathematics: Philadel-
phia, PA, USA, 2004.
25. Dong, Y.; Gu, R. Adaptive coordinated control of networked non-affine nonlinear systems with a non-autonomous nonlinear
leader. Nonlinear Dyn. 2023, 111, 13111–13124.
Electronics 2024, 1, 0 20 of 20

26. Liu, T.; Huang, J. A distributed observer for a class of nonlinear systems and its application to a leader-following consensus
problem. IEEE Trans. Autom. Control 2018, 64, 1221–1227.
27. Su, Y.; Huang, J. Cooperative output regulation of linear multi-agent systems. IEEE Trans. Autom. Control 2011, 57, 1062–1066.
28. Cai, H.; Huang, J. Leader-following consensus of multiple uncertain Euler–Lagrange systems under switching network topology.
Int. J. Gen. Syst. 2014, 43, 294–304.
29. Cai, H.; Lewis, F.L.; Hu, G.; Huang, J. The adaptive distributed observer approach to the cooperative output regulation of linear
multi-agent systems. Automatica 2017, 75, 299–305.
30. Dong, Y.; Chen, J.; Huang, J. A self-tuning adaptive distributed observer approach to the cooperative output regulation problem
for networked multi-agent systems. Int. J. Control 2019, 92, 1796–1804.
31. Gao, M.; Jin, X.; Ding, L. A Novel Nonsingular Fixed-Time Sliding Mode Control of Uncertain Euler–Lagrange Systems. IEEE
Syst. J. 2022, 17, 467–478.
32. Chen, G.; Song, Y.D. Robust fault-tolerant cooperative control of multi-agent systems: A constructive design method. J. Frankl.
Inst. 2015, 352, 4045–4066.
33. Wang, C.; Wen, C.; Guo, L. Adaptive Consensus Control for Nonlinear Multi-Agent Systems With Unknown Control Directions
and Time-Varying Actuator Faults. IEEE Trans. Autom. Control 2020, 66, 4222–4229.
34. Dong, Y.; Huang, J. The leader-following rendezvous with connectivity preservation via a self-tuning adaptive distributed
observer. Int. J. Control 2017, 90, 1518–1527.
35. Hu, J.; Hong, Y. Leader-following coordination of multi-agent systems with coupling time delays. Phys. A Stat. Mech. Appl. 2007,
374, 853–863.
36. Khalil, H.K. Nonlinear Systems, 3rd ed.; Patience Hall: Hoboken, NJ, USA, 2002.
37. Roy, S.; Roy, S.B.; Kar, I.N. Adaptive–robust control of Euler–Lagrange systems with linearly parametrizable uncertainty bound.
IEEE Trans. Control Syst. Technol. 2017, 26, 1842–1850.
38. Yang, L.; Yang, J. Nonsingular fast terminal sliding-mode control for nonlinear dynamical systems. Int. J. Robust Nonlinear Control
2011, 21, 1865–1879.
39. Van, M. An enhanced robust fault tolerant control based on an adaptive fuzzy PID-nonsingular fast terminal sliding mode control
for uncertain nonlinear systems. IEEE/ASME Trans. Mechatron. 2018, 23, 1362–1371.
40. Zhang, S.; Yang, P.; Kong, L.; Chen, W.; Fu, Q.; Peng, K. Neural networks-based fault tolerant control of a robot via fast terminal
sliding mode. IEEE Trans. Syst. Man Cybern. Syst. 2019, 51, 4091–4101.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.