1 s2.0 S0016003223001084 Main

Available online at www.sciencedirect.
com
Journal of the Franklin Institute 360 (2023) 3914–3966

www.elsevier.com/locate/jfranklin
Three-dimensional event-triggered fixed-time

cooperative guidance law against maneuvering target
with the constraint of relative impact angles
Zhongyuan Chen, Xiaoming Liu, Wanchun Chen∗
School of Astronautics, Beihang University, Beijing 100191, PR China
Received 23 May 2022; received in revised form 4 January 2023; accepted 14 February 2023
Available online 22 February 2023
Abstract
In order to improve the flexibility and reduce the energy consumption of cooperative guidance laws
considering the impact angle constraint, this paper proposes a three-dimensional event-triggered fixed-
time cooperative guidance law with the constraint of relative impact angles. First, for the purpose of
avoiding the precision degradation due to the estimation error of time-to-go especially facing a maneu-
vering target, the range-to-go and velocity along the line-of-sight (LOS) are taken as the coordination
variables for achieving time-cooperative guidance. Secondly, instead of assigning specific desired impact
angles for each missile, only the consensus errors of relative impact angles are utilized as the coordina-
tion variables for achieving space-cooperative guidance, which can avoid continually maneuvering for
maintaining the constant desired impact angles, thus reducing the fuel consumption. Next, the guidance
laws along the LOS and perpendicular to the LOS are developed, and the event-triggering mechanisms
are designed to reduce the update frequency of cooperative guidance commands, thus further reducing
the energy consumption. To guarantee the convergence rate, the fixed-time control theory is adopted
and the stability of proposed event-triggered cooperative guidance laws are rigorously proved. In addi-
tion, it is also proved that there is no Zeno behavior when implementing the proposed event-triggered
cooperative guidance laws. Finally, numerical simulations indicate that the strictly simultaneous attack
is achieved and the constraint of relative impact angles is satisfied. Comparative studies demonstrate
that the computation burden of cooperative guidance commands is relaxed and the fuel consumption is
∗ Corresponding author.
E-mail addresses: zhongyuan@buaa.edu.cn (Z. Chen), chenwanchun_buaa@163.com (W. Chen).
https://doi.org/10.1016/j.jfranklin.2023.02.027
0016-0032/© 2023 The Franklin Institute. Published by Elsevier Inc. All rights reserved.
Z. Chen, X. Liu and W. Chen Journal of the Franklin Institute 360 (2023) 3914–3966
reduced by the proposed event-triggered cooperative guidance laws with the constraint of relative impact
angles.
© 2023 The Franklin Institute. Published by Elsevier Inc. All rights reserved.
1. Introduction
Recently, the penetration difficulty of a target by utilizing traditional single missile is

increasing, because the multi-layered defense system for missile is being perfect, and the
target maneuverability is enhanced [1–3]. In such a situation, the concept of cooperative
guidance is proposed as an efficient countermeasure to reinforce the penetration capacity of
missiles [4–6]. Besides achieving small or even zero miss distance, cooperative guidance laws
are also required to satisfy the constraints in both time and space dimensions to improve the
lethality of multiple missiles.
Time-cooperative guidance means that multiple missiles attack the target simultaneously,
which can be achieved through two methods. The first approach is called impact time control
guidance (ITCG), which requires all missiles to attack the target at a pre-specified common
impact time [7–9]. However, it is difficult to assign an appropriate common impact time in
advance for multiple missiles with different initial conditions. In the second approach, multiple
missiles coordinate their time-to-go to be consensus for achieving simultaneous attack through
a communication topology [10–12]. It should be noted that the explicit expression of time-to-
go estimation is required for time-cooperative guidance in [10–12]. Nevertheless, the precise
estimation of time-to-go is challenging due to large initial heading errors [13], unknown
target maneuver [14], and external disturbances [15], which will degrade the performance
of cooperative guidance. In the present paper, a new time-cooperative guidance strategy is
designed, which directly utilizes the consensus errors of range-to-go and the radial relative
velocity as the coordination variables. As a result, the decrease of guidance precision due to
the estimation error of time-to-go can be avoided.
Space-cooperative guidance means that multiple missiles attack a target in different impact
angles to improve the lethality. Conventional cooperative guidance laws with the constraint
of impact angle are achieved by specifying desired impact angles for each missile before the
engagement [16–18], so multiple missiles may frequently adjust their guidance commands for
reaching and maintaining the desired impact angles under some uncertainties such as unknown
target maneuver and disturbances. Accordingly, the guidance laws with specific desired impact
angles may cause unnecessary fuel consumption. The present paper will pre-specify desired
relative impact angles between missiles for achieving space-cooperative guidance Because the
relative impact angles to the target will rarely change greatly under some disturbances (such
as unknown target maneuver and model uncertainties), it is unnecessary for the multi-missile
system to continually maneuver for maintaining the constant desired impact angles. As a result,
the cooperative guidance laws will be more flexible and the fuel consumption can be reduced.
The research on space-cooperative guidance laws that utilize the constraint of relative impact
angles are rare [19–21]. The cooperative guidance laws in [19,20] are developed to attack
the target with assigned relative flight-path angles, but the time-cooperative guidance is out
of consideration and the dynamic models are linearized based on the small angle assumption.
The nonlinear cooperative guidance law proposed in [21] can achieve simultaneous attack and
3915
satisfy the constraint of relative impact angles. However, the target in [21] is stationary, not
a maneuvering one, which limits the scope of applications.
Important to note is that the missiles are assumed to fly in a plane in [16–21]. Practically,
the missiles fly in the nonlinear three-dimensional (3-D) environment in which the line-of-
sight (LOS) angles and angular rates may be very large. In such a situation, the precision of
guidance laws that are based on the small angle assumption [19,20] will degrade dramatically.
Therefore, it is vital to design a robust nonlinear cooperative guidance law in 3-D environment
that can satisfy the constraints in both time and space dimensions. The 3-D cooperative
guidance laws proposed in [22,23] can achieve simultaneous attack and desired impact angles,
yet the estimation of time-to-go is utilized as the consensus variable. As mentioned before,
the estimation error of time-to-go will significantly influence the accuracy of time-cooperative
guidance. The cooperative guidance law proposed in [21,24] utilizes the consensus of range-
to-go and radial relative velocity to achieve simultaneous attack, but the targets are stationary,
not maneuvering ones.
More importantly, the above cooperative guidance laws are required to continuously up-
date the cooperative guidance commands with periodically transmitted input signals, which is
called time-triggered guidance [25–27], which may demand a heavy transmission load in the
controller-to-actuator channel [28–30]. It should be noted that high regulation frequency of
guidance commands even when there is no significant change in the coordination variables
may waste the limited fuel that a missile carries [31–33]. In order to save the limited re-
sources of the multi-missile system while guaranteeing the desirable guidance performance,
an event-triggered cooperative guidance scheme is proposed in the present paper. The coop-
erative guidance commands will not be updated and hold constant until trigger conditions are
satisfied, which can greatly reduce the update frequency of guidance commands, thus reducing
the resources consumption.
Event-triggered control scheme has attracted wide attention in recent years. An event-
driven-based fixed-time adaptive controller is constructed in [34] to reduce the communication
burden while excluding the Zeno-behavior. The designed controller not only guarantees all
the signals of the closed-loop system (CLS) are practically fixed-time bounded, but also
the tracking error can be regulated to the predefined boundary. A self-triggered scheme is
formulated in [35] to predict sampling instants automatically and to reduce the computational
burden of the on-line solving of model predictive control. The co-design of the self-triggered
scheme and the MPC approach can design the control input when keeping the state trajectories
within a pre-specified bound over a given time interval. Nevertheless, the approaches in [34,35]
are designed for single agent, which are not suitable for the cooperative guidance law design
of multi-missile system.
There are still few research related to the distributed event-triggered cooperative guidance
[36–38]. An event-triggered optimal cooperative guidance law for simultaneous attacks with
impact angle constraints is proposed in [36], yet only a stationary target is considered. In [37],
an adaptive event-triggered distributed optimal guidance law is developed by which multiple
missiles can simultaneously attack a maneuvering target with less update numbers. Never-
theless, the engagement geometries in [36,37] are planer, not the practical three-dimensional
ones. An event-triggered 3-D nonlinear cooperative guidance law for simultaneous attack is
proposed in [38], which can optimize the usage of available resources. However, only the
target stationary is considered. In [39], the distributed optimal event-triggered cooperative
guidance law for nonlinear multi-missile guidance systems with partially unknown dynamics
3916
is proposed, which can deal with maneuvering targets. However, the constraint of impact
angle is out of consideration in [37–39], which limits the lethality of multiple missiles.
Furthermore, the flight time of terminal guidance is very short, so the high convergence rate
of guidance errors is a vital demand. Nevertheless, the proposed event-triggered cooperative
guidance scheme aims to reduce the update frequency of cooperative guidance commands,
which may in turn decrease the convergence rate. Some guidance laws have been developed
based on the theory of finite-time control to achieve a faster convergence rate than the asymp-
totic ones [40–42]. However, the convergence time of finite-time cooperative guidance laws is
dependent on the initial conditions, and the convergence time may become infinity with huge
initial errors [43–45]. As an extension of finite-time control, the bound of convergence time
for the fixed-time control is independency to initial conditions, which can further improve
the convergence rate [46–48]. A three-dimensional adaptive fixed-time cooperative guidance
law with impact time and angle constraints in [49], yet the target is stationary. Robust fixed-
time three-dimensional cooperative guidance laws with the constraint of impact angle are
proposed in [50,51], which can adapt to unknown target maneuvers and model uncertainties.
Nevertheless, the performance of the guidance laws in [50,51] will suffer from the estimation
error of time-to-go. In addition, the cooperative guidance laws proposed in [48–51] require
the guidance commands to be updated continuously with time-periodical input signals, which
may cause unnecessary consumption of resources.
Aiming at the above observations and discussions, we present 3-D event-triggered fixed-
time robust cooperative guidance scheme with the constraint of relative impact angles, which
can achieve accurate simultaneous attack and fast convergence with less resource consumption.
The main contributions or advantages are summarized as follows:
(1) Although the 3-D cooperative guidance laws developed in [22,23,50,51] can achieve
simultaneous attack and satisfy the impact angle constraint, the estimation of time-to-go
is utilized as the consensus variable, and the estimation error will reduce the guidance
precision. In this study, the consensus errors of range-to-go and relative velocity along
the LOS are regarded as the coordination variables, which can avoid the precision
degradation due to the estimation error of time-to-go. The guidance laws proposed in
[21,24] utilized consensus errors of range-to-go and radial relative velocity for achieving
time-cooperative guidance, yet the targets are stationary, and the impact angle constraint
is out of consideration.
(2) Instead of assigning certain desired impact angles for each missile in [21-24,36,50],
only the desired relative impact angles are pre-specified to achieve space-cooperative
guidance in the present paper. The relative impact angles will rarely change greatly
under some disturbances, so the requirement of maintaining constant desired impact
angles can be relaxed. As a result, the regulation frequency of cooperative guidance
commands can be decreased and the fuel consumption can be reduced. Therefore, the
proposed cooperative guidance laws that utilize the constraint of desired relative impact
angles will be more flexible.
(3) Contrary to the time-triggered cooperative guidance laws that continuously update the
cooperative guidance commands [1,27], an event-triggered cooperative guidance law is
developed. By designing appropriate triggering mechanisms, the cooperative guidance
commands are not updated and hold constant until triggering conditions are satisfied,
which can further reduce the resource consumption as well as ensure the guidance sys-
tem’s performance. Distributed event-triggered cooperative guidance laws are proposed
3917
in [37–39], but the constraint of impact angle is out of consideration in both works,
which limits the lethality of multiple missiles.
(4) Note that the proposed event-triggered cooperative guidance scheme aims to reduce the
update frequency of cooperative guidance commands, which may in turn decrease the
convergence rate of the coordination variables. A fixed-time event-triggered cooperative
guidance is designed in the present paper, whose convergence rate is independent of
initial conditions. Hence, the proposed guidance law will ensure a fast convergence rate
under a lower update frequency.
The rest of the paper is organized as follows. Section 2 provides some necessary prelimi-
naries including graph theory, definition of fixed-time stability, and the problem formulation
of simultaneous attack with desired relative impact angles. In Section 3, the 3-D fixed-time
event-triggered cooperative guidance laws with the constraint of relative impact angle is pro-
posed, and the corresponding stability proof is also given in detail. In order to demonstrate
the effectiveness of the proposed cooperative guidance laws, some comparative case studies
are discussed in Section 4. Section 5 concludes the paper.
2. Preliminaries and problem formulation of cooperative attack
2.1. Graph theory

Consider a group of N missiles, whose communication topology graph can be mathemat-
ically described as G = (V, E ), where V = {1, 2, · · · , N } is the set of vertexes related to the
missile members, E ⊆ V × V is the set of edges related to the communication links [52]. An
edge between vertex Vi and vertex V j is denoted by (V j , Vi ) ∈ E , which means that the ith
missile can obtain the information of the jth missile. DefineA = [ai j ] as the adjacency matrix
of graph G, where ai j = 1 if and only if (V j , Vi ) ∈ E , and ai j = 0 if (V j , Vi ) ∈
/ E . In addition,
the neighborhood set of the ith missile is defined as Ni = {V j : (V j , Vi ) ∈ E }. A graph is called
connected graph if there is a path consisting of edges in E to reach from the ith missile to
the jth missile for any pair of vertices (Vi , V j ). In this study, the graph is assumed undirected,
which means that the missile in a pair of neighbors can be detected by each other, namely,
ai j = a ji [53].
The degree of a vertex Vi is denoted as di , which is defined as the number of neighbors of a
missile Vi . The degree matrix of graph G is expressed as D = diag {di } (i ∈ [1, 2, . . . , N ]). The
Laplacian matrix L of G is then defined as L = D − A. For a connected and undirected graph,
the Laplace matrix L is symmetric and semi-positive definite. The minimum eigenvalue of L
is zero with corresponding eigenvector 1N = [1, 1, . . . , 1]T . In addition, the second smallest
eigenvalue λ2 of L is greater than zero, i.e., λ2 > 0, which can be regarded as an indicator
for evaluating the connectivity of graph G [54].
Lemma 1. [55]: For an undirected connected graph G, the eigenvalues of L are 0, λ2 , . . . , λN
and satisfy 0 < λ2 ≤ . . . ≤ λN . Moreover, if 1TN x = 0 with x = [x1 , x2 , . . . , xN ]T , then x T Lx ≥
λ2 x T x.
2.2. Preliminaries
Consider the following nonlinear system,

x˙(t ) = f (t , x(t ))
(1)
x(0) = x0
3918
where x ∈ Rn and f : Rn → Rn is a nonlinear function such that f (0) = 0, i.e., x = 0 is an

equilibrium point of system Eq. (1). Some standard definitions and lemmas are introduced as
follows.
Definition 1. (Fixed-Time Stability [46,56]): The equilibrium point of the system Eq. (1) is
fixed-time convergent if the settling time is bounded and independent of initial conditions,
i.e., ∃Tmax > 0 : ∀x0 ∈ Rn and T (x0 ) ≤ Tmax .
Lemma 2. (Practical fixed-time stable [57]): If there exists a Lyapunov function V (x(t ))
satisfying
V˙ (x(t )) ≤ −c1V p (x(t )) − c2V q (x(t )) + ι (2)
with the constants c1 > 0, c2 > 0, ι > 0, p ∈ (0, 1), and q ∈ (1, ∞ ), the origin of system
Eq. (1) is practical fixed-time stable. Moreover, the residual set of the solution is
1p q1
− 1p ι − q1 ι
lim x(t )|V (x(t )) ≤ min c1 , c2 (3)
t→T 1−κ 1−κ
where κ ∈ (0, 1) is a constant. The settling time T satisfies
1 1
T ≤ Tmax := + (4)
c1 κ (1 − p) c2 κ (q − 1)
In addition, if ι = 0, the origin of system Eq. (1) is fixed-time stable, which is defined in
Definition 1, and Tmax will be (1/[c1 (1 − p)] ) + (1/[c2 (q − 1)] ).
Lemma 3. [58]: For any real numbers ξ1 , ξ2 , . . . , ξN > 0, the following inequality holds
⎧ N p
⎪
⎪
N
1−p
⎪
⎨ ξ i
P
≥ N ξi , if p > 1
i=1 i=1
N p (5)
⎪
⎪ N
⎪
⎩ ξ i
P
≥ ξ i , if 0 < p ≤ 1
i=1 i=1
Lemma 4. [59]: For any real variable y, we have

0 ≤ |y| − y tanh (μy) ≤ ζ /μ (6)
where μ 1 and ζ > 0 whose minimum value ζ ∗ satisfies ζ ∗ = x ∗ (1 − tanh x ∗ ), and x ∗

∗
satisfies the equation e−2x + 1 − 2x ∗ = 0.
2.3. Problem formulation of cooperative guidance in 3-D environment

Fig. 1 illustrates the 3-D engagement geometry of N missiles against a maneuvering target,
where ri is the relative distance between the ith missile and the target. The elevation and
azimuth LOS angles are denoted by qεi and qβi , respectively [60]. A reference frame Mi xI yI zI
is located at the ith missile, where Mi denotes the ith missile. Moreover, Mi xL yL zL is the LOS
frame.
The 3-D engagement dynamics of the ith missile expressed in the Mi xL yL zL frame can be
expressed as [61]:
r̈i − ri q˙εi
2
− ri q˙βi
2
cos2 qεi = aT ri − aMri (7)
3919
Fig. 1. Three-dimensional engagement geometry of the ith missile against a maneuvering target.
ri q̈εi + 2r˙i q˙εi + ri q˙βi

2
sin qεi cos qβi = aT εi − aMεi (8)
−ri q̈βi cos qεi − 2r˙i q˙βi cos qεi + 2ri q˙εi q˙βi sin qεi = aT βi − aM βi (9)
T T
where [aMri aMεi aM βi ] and [aT ri aT εi aT βi ] are the acceleration of the i−th missile
and the target expressed in Mi xL yL zL , respectively.
In this study, the maneuverers of target are regarded as disturbances, then the engagement
dynamics can be rewritten as
r̈i = ri q˙εi
2
+ ri q˙βi
2
cos2 qεi − aMri + dri = fride f + fridis − aMri (10)
−2r˙i q˙εi − ri q˙βi

2
sin qεi cos qβi − aMεi aMεi
q̈εi = +dεi = fεide f + fεidis − (11)
ri ri
−2r˙i q˙βi cos qεi + 2ri q˙εi q˙βi sin qεi + aM βi aM βi
q̈βi = + dβi = fβide f + fβidis + (12)
ri cos qεi ri cos qεi
where
⎡ de f ⎤ ⎡ ⎤ ⎡ dis ⎤ ⎡ ⎤
fri ri q˙εi2
+ ri q˙βi 2
cos2 qεi fri aT ri
⎢ de f ⎥ ⎢ −2r˙i q˙εi −ri q˙βi2 sin qεi cos qβi ⎥
⎢f ⎥ = ⎢ ⎥, and⎢ dis ⎥ ⎢ a
⎣ fεi ⎦ = ⎣ rTiεi ⎦.
⎥
⎣ εi ⎦ ⎣ ri ⎦ −a
−2r˙i q˙βi cos qεi +2ri q˙εi q˙βi sin qεi T βi
fβide f fβidis ri cos qεi
r cos q
i εi
Time-cooperative guidance (i.e., simultaneous attack) means that multiple missiles synchro-
nize their time-to-go to hit the target at the same time. Here, in order to avoid the precision
degradation of simultaneous attack due to the estimation error of time-to-go, the range-to-go
and relative velocity in the LOS direction (radial relative velocity) are utilized as the coor-
dination variables. Thus, the main objective is to design a cooperative guidance law that can
bring the range-to-go r and radial relative velocity r˙ to the consensus. In practice, the miss
3920
distances of multiple missiles are not strict zero at the impact point, so the consensus errors
of coordination variables are unnecessary to be strict zero. Therefore, given different initial
r(0) and r˙(0), the objective of time-cooperative guidance can be expressed as

ri (t ) − r j (t ) ≤ δr1 , ∀t ≥ Tr
(13)
vri (t ) − vr j (t ) ≤ δr2 , ∀t ≥ Tr
where δr1 and δr2 are sufficiently small positive constants, so the guidance objective described
by Eq.(13) is called “practical time-cooperative guidance”. Moreover, the convergence time
Tr satisfies ∃Tmax > 0, s.t., Tr ≤ Tmax .
Space-cooperative guidance means that multiple missiles attack a target in different impact
angles to improve the lethality. Impact angle is defined as the intersect angle between the
velocity vectors of the missile and the target at the impact time. According to [40,62,63], the
constraint of impact angle can be converted to the constraint of LOS angle. Denote q˜εi (t )
and q˜βi (t ) are the errors between the current LOS angles and desired LOS angles, that is,

q˜εi (t ) = qεi (t ) − qεi
d
(14)
q˜βi (t ) = qβi (t ) − qβi
d
d d
where qεi and qβi represent the desired terminal LOS angles in the elevation and azimuth
directions, respectively. Therefore, given different initial qε (0) and qβ (0), the objective of
space-cooperative guidance can be described by

q˜εi (t ) − q˜ε j (t ) ≤ δε1 , ∀t ≥ Tε
(15)
q˙εi (t ) ≤ δε2 , ∀t ≥ Tε
and

q˜βi (t ) − q˜β j (t ) ≤ δβ1 , ∀t ≥ Tβ
(16)
q˙βi (t ) ≤ δβ2 , ∀t ≥ Tβ
where δε1 , δε2 , δβ1 , and δβ2 are sufficiently small positive constants, so the guidance objec-
tives described by Eqs. (15) and (16) are called “practical space-cooperative guidance”. The
convergence time Tε and Tβ satisfy ∃Tmax > 0, s.t., Tε ≤ Tmax and Tβ ≤ Tmax .
d d
Remark 1. In this study, instead of specifying certain values for qεi and qβi in [21-24,50],
only the desired relative LOS angles qεi − qε j and qβi − qβ j are utilized to achieve space-
d d d d
cooperative guidance, which means that the requirement of maintaining constant desired im-
pact angles can be relaxed. During the engagement, the relative impact angles to the target
will rarely change greatly. As a result, less regulation of guidance commands is required and
the fuel consumption can be reduced by utilizing the proposed cooperative guidance laws.
3. Design of event-triggered cooperative guidance laws
As shown in Fig. 2, the guidance command of traditional time-triggered guidance is updated

at periodic sampling instants. The sampling period of the guidance system is assumed as a
constant T . In such a situation, the time-triggered guidance commands may be constantly
updated even though the changes of engagement states are negligible or have no impact on
the guidance performance. As a result, the traditional time-triggered guidance may generate
too much redundant regulation, thus causing the waste of limited fuel.
3921
Fig. 2. Diagram of the time-triggered guidance scheme.
Fig. 3. Diagram of the event-triggered guidance scheme.
In order to save the limited resources of the multi-missile system while guaranteeing the
desirable guidance performance, an event-triggered robust cooperative guidance scheme is
proposed in the present paper. In contrast to time-triggered control, the cooperative guidance
commands are not updated until the triggering conditions are violated, as shown in Fig. 3.
When the trigger conditions are not satisfied, the zero-order holder (ZOH) is applied to keep
continuity of guidance commands. Hence, the update frequency of guidance commands and
resources consumption can be reduced.
Remark 2. Just as shown in Fig. 2, event-triggered controllers require the constant monitoring
of a triggering condition. For some applications this is a reasonable assumption, e.g., when
we can use dedicated hardware for this purpose. Taking the team of multiple missiles as an
example, missile can use the inter-missile data link to communicate continuously.
Definition 2. (Zeno Behavior [64]): Zeno behavior implies that an infinite number of events
are triggered under a finite time interval. To be specific, a multi-missile system exhibits the
∞
Zeno behavior if lim tki = (tk+1
i
− tki ) = t∞
i
for the i−th missile, where tki denotes the
k→∞ k=0
triggered time of the k −th event, and t∞
i
is a finite time.
Remark 3. In Fig. 3, the minimum triggering interval of event-triggered guidance is greater
than or equal to the sampling period T , so Zeno behavior is naturally avoided. In the following
design progress of guidance law, we will still try to demonstrate that Zeno behavior will not
3922
appear in the continuous guidance. The reason is that if Zeno behavior appears, event-triggered
guidance will require faster and faster update [25]. Therefore, the guidance laws should be
designed to ensure that no ZENO behavior occurs. Accordingly, compared with continuous
updating guidance laws, less computing effort is required by the event-triggered guidance,
thus reducing the computation cost.
The framework of the event-triggered cooperative guidance scheme is illustrated in Fig. 4.
As can be seen, the triggering conditions are set in the sensor-to-guidance channel. The
cooperative guidance laws and triggering conditions will be designed in the following sections.
3.1. Design of guidance law for simultaneous attack

In this section, a virtual consensus radial relative velocity vri∗ (t ) is first constructed by
utilizing the consensus error of range-to-go. Secondly, the event-triggered guidance law aMri (t )
along the LOS direction is developed to lead the range-to-go ri (t ) and radial relative velocity
vri (t ) of multiple missiles to be consensus within fixed time. Finally, it is proved that the
multiple-missile system can achieve the objective of time-cooperative guidance described by
Eq.(13).
First, the virtual consensus radial relative velocity vri∗ (t ) is designed as
br /cr
vri∗ (t ) = −αr1 i (t ) − αr2 tanh (μr i (t ) ) + vˆr , (17)
where αr1 , αr2 , and μr are positive constants. In addition, br and cr are two positive odd
numbers that satisfy br /cr > 2. Furthermore, vˆr is a constant and can be chosen as the average
value of the initial radial relative velocities. In the following contents, we will demonstrate
that the radial relative velocity vri (t ) can entry a small neighborhood of vˆr . The term i (t )
is the consensus error of range-to-go, that is,

N

i (t ) = ai j ri (t ) − r j (t ) , (18)
j=1
Next, the tracking error of the radial relative velocity is defined as

br /cr
evri (t ) = vri (t ) − vri∗ (t ) = vri (t ) + αr1 i (t ) + αr2 tanh (μr i (t ) ) − vˆr . (19)
For leading the current radial relative velocity vri (t ) to reach the virtual consensus radial
relative velocity vri∗ (t ) within fixed time, the event-triggered cooperative guidance law along
the LOS is designed as
br br /cr −1 i i
aMri (t ) = ri (t )q˙εi
2
(t ) + ri (t )q˙βi
2
(t )cos2 qεi (t ) + αr1 tk i tk
cr i

pr /gr i
+ αr2 μr 1 − tanh2 μr i tki i tk + αr3 evri
i
tk
i i i
+ αr4 tanh μr evri tk , t ∈ tk , tk+1 (20)
where αr3 and αr4 are positive constants, pr and gr are two positive odd numbers that satisfy
pr /gr > 1, and tki (tk+1
i
) is the latest (next) triggering time of the i-th missile. The term i (t )
is the consensus error of radial relative velocities, i.e.,

N

i (t ) = ai j vri (t ) − vr j (t ) . (21)
j=1
3923
Fig. 4. Event-triggered cooperative guidance scheme for the ith missile.
3924
Fig. 5. Communication topology of the multi-missile system.
Finally, the triggering mechanism is designed as

pr /gr
i
tk+1 = inf t > tki |ri (t )| ≥ ηr αr3 evri (t ) + ηr αr4 , (22)
where ηr is a constant that satisfies 0 < ηr < 1. The term ri (t ) is the triggering error between
the latest triggered states and the present states, which is defined as
br br /cr −1 i i i
ri (t ) = αr1 i tk i tk + αr2 μr 1 − tanh2 μr i tki i tk
cr
pr /gr i
+ αr3 evri tk + αr4 tanh μr evri tki
br br /cr −1
− αr1 i (t ) i (t ) − αr2 μr 1 − tanh2 (μr i (t ) ) i (t )
cr
p /g
− αr3 evrir r (t ) − αr4 tanh (μr evri (t ) ). (23)
Through the triggering mechanism Eq. (22), the guidance command aMri (t ) for the i-th
missile will be updated at the triggering time sequence t0i , t1i , . . ..
Assumption 1. The nonlinear terms and uncertain disturbances along the LOS of the ith
missile are bounded, that is,

de f
fri + fridis ≤ f¯r (24)
where f¯r is non-negative constants.

Theorem 1. For the multi-missile system described by Eq.(10) with model uncertainties and
unknown target maneuverer, the objective of the time-cooperative guidance expressed as
Eq.(13) can be achieved through the time-cooperative guidance law Eq. (20) and trigger-
ing mechanism Eq. (22), if the following conditions hold:

αr2 > vˆr + τr
αr4 (1 − ηr ) > f¯r (25)
with
p g+r g
N αr4 ζr 1 αr4 ζr r r
τr = min , N 2
¯
μr (1 − κr ) αr4 (1 − ηr ) − fr αr3 μr (1 − κr )(1 − ηr )
3925
Fig. 6. Trajectories of multiple missiles that attack a maneuvering target.
where κr ∈ (0, 1) is a constant. The constant ζr satisfies ζr > 0 whose minimum value ζ ∗ is
∗
ζ ∗ = x ∗ (1 − tanh x ∗ ), where x ∗ satisfies the equation e−2x + 1 − 2x ∗ = 0. In addition, there
is no Zeno behavior when employing the event-triggered time-cooperative guidance law.
Proof. First, we will prove that the current radial relative velocity vri (t ) can reach the virtual
consensus radial relative velocity vri∗ (t ) within fixed time. Based on Eqs. (10) and (17)-(21),
the first-order derivative of the velocity tracking error evri (t ) with respect to time can be
obtained, that is,
e˙vri (t ) = ri (t )q˙εi
2
(t ) + ri (t )q˙βi
2
(t )cos2 qεi (t ) − aMri (t ) + fride f (t ) + fridis (t )
br br /cr −1
+ αr1 (t ) i (t ) + αr2 μr 1 − tanh2 (μr i (t ) ) i (t )
cr i
br br /cr −1 i i i
= −αr1 i tk i tk − αr2 μr 1 − tanh2 μr i tki i tk
cr
br br /cr −1
+ αr1 (t ) i (t ) + αr2 μr 1 − tanh2 (μr i (t ) ) i (t )
cr i
pr /gr i
− αr3 evri tk − αr4 tanh μr evri tki + fride f (t ) + fridis (t )

pr /gr
= − ri (t ) + αr3 evri (t ) + αr4 tanh (μr evri (t ) ) + fride f (t ) + fridis (t ) (26)
3926
Fig. 7. Consensus errors of the coordination variables and time-to-go under the proposed event-triggered guidance
laws.
Then the following Lyapunov candidate function is considered,
1 2
N
Vr1 (t ) = e (t ) (27)
2 i=1 vri
3927
Fig. 8. Results under the proposed event-triggered guidance laws.
According to the triggering mechanism Eq. (22), as well as Lemmas 3 and 4, the derivative
of Eq. (27) can be written as

N
V˙r1 (t ) = evri (t )e˙vri (t )
i=1

N
N pr/ +1
pr /gr g
≤ |evri (t )| ηr αr3 evri (t ) + ηr αr4 − αr3 evri r (t )
i=1 i=1
3928

N
N
N αr4 ζr
− αr4 |evri (t )| + f¯r |evri (t )| +
i=1 i=1
μr

N
pr +gr N
N αr4 ζr
≤ −αr3 (1 − ηr ) |evri (t )| gr − αr4 (1 − ηr ) − f¯r |evri (t )| +
i=1 i=1
μr
gr −pr pr +gr N αr4 ζr
− αr4 (1 − ηr ) − f¯r (2Vr1 (t ) ) 2 +
1
≤ −αr3 (1 − ηr )N 2gr (2Vr1 (t ) ) 2gr (28)
μr
According to Lemma 2, the tracking error of radial relative velocities can converge into a
small region around zero in fixed time, i.e.,

N αr4 ζr
|evri (t )| ≤ evr (t ) ≤ τr = min ,
μr (1 − κr ) αr4 (1 − ηr ) − f¯r
p g+r g
1 αr4 ζr r r
N2 (29)
αr3 μr (1 − κr )(1 − ηr )
and the bound of convergence time Tr1 is provided by

√
2 2
Tr1 ≤ + (30)
κr αr4 (1 − ηr ) − f¯r
gr −pr pr +gr
p
αr3 κr (1 − ηr )( grr − 1)N 2gr 2 2gr
Next, we will prove that the consensus error of range-to-go i (t ) can coverage within fixed
time. Consider the following Lyapunov candidate function,
1 T
Vr2 (t ) = r (t )Lr(t )
2
where r(t ) = [r1 (t ) r2 (t ) , . . . ri (t ) . . . , rN (t )]T . Based on Lemmas 2 and 3, it can be ob-
tained that

N
V˙r2 (t ) = r T (t )Lr˙ (t ) = i (t )vri (t )
i=1

N br +cr
N
N N
N αr2 ζr
≤ −αr1 cr
(t ) − αr2 | i (t )| + vˆr | i (t )| + | i (t )||evri (t )| +
i=1
i
i=1 i=1 i=1
μr
cr −br br +cr N αr2 ζr
− αr2 − vˆr − τr (2λ2Vr2 (t ) ) 2 +
1
≤ −αr1 N 2cr (2λ2Vr2 (t ) ) 2cr (31)
μr
where λ2 is the second smallest eigenvalue of Laplacian matrix L. According to Lemma 2,
one can obtain the following residual set,
⎧ ⎧ 2 ⎫⎫
⎨ ⎨ b 2c+rc ⎬⎬
−1 Nαr2 ζr N αr2 ζr r r
lim (t )|Vr2 (t ) ≤ min (2λ2 ) , (32)
⎩t→Tr1 +Tr2 ⎩ μr (1 − κr ) αr2 − vˆr − τr 2λ2 αr1 μr (1 − κr ) ⎭⎭
where Tr2 is provided by

√
2 2
Tr2 ≤ √ + (33)

αr2 − vˆr − τr κr λ2
cr −br br +cr
αr1 ( cr − 1)N 2cr κr (2λ2 ) 2cr
b r
3929
Consequently, the practical consensus of range-to-go can be achieved within fixed time,
and the bound of total convergence time satisfies
Tr = Tr1 + Tr2
√
2 2
≤ +
κr αr4 (1 − ηr ) − f¯r
gr −pr pr +gr
αr3 κr (1 − ηr )( gprr − 1)N 2gr 2 2gr
√
2 2
+ √ + (34)

αr2 − vˆr − τr κr λ2
cr −br
αr1 ( r − 1)N 2cr κr (2λ2 ) 2cr
b br +cr
cr
Moreover, it can be obtained that vri (t ) = vri∗ (t ) + evri (t ) ≤ |−αr1 ibr /cr (t )
−αr2 tanh (μr i (t ) )| + vˆr + τr , which means that vri (t ) can entry a small neighborhood
of vˆr .
Thus, based on inequalities Eqs. (29), (30), (32) and (34), it can be concluded that the
objective of time-cooperative guidance described by Eq.(13) can be achieved in fixed time by
utilizing guidance law Eq. (20) and triggering mechanism Eq. (22).
In the followings, it will be proved that there is no Zeno behavior. According to the
definition of range-to-go consensus error i (t ) and Lemma 1, we can get

N
λN r T (t )Lr(t ) ≥ i (t )
2
= r T (t )L2 r(t ) ≥ λ2 r T (t )Lr(t ). (35)
i=1
The norm of i (t ) satisfies

N N
N

| i (t )| ≤
ai j evri (t ) − evr j (t ) + 2αr2 ai j + αr1 ai j − br /cr
(t ) + bj r /cr (t )
i
j=1 j=1 j=1
br /cr
≤ evr (t ) 1 + (lii − 1 ) evr (t ) 2 + 2αr1 lii (t ) 2 + 2αr2 lii ≤ ςi1 (36)
with ςi1 (t ) = (lii − 1 + N (1/2) )(2Vr1 (0) )(1/2) + 2αr2 lii + 2αr1 lii (2λN Vr2 (0) )(br /2cr ) , and λN de-
notes the largest eigenvalue of L. According to the definition of triggering error ri (t ) ex-
pressed by Eq.(23), it can be obtained that

D+ (|ri (t )| ) ≤ ˙ ri (t )

pr prg−gr r

≤ − αr3 evri (t ) + αr4 μr 1 − tanh (μr evri (t ) ) e˙vri (t )
2
gr

br brc−cr r
− αr1 (t ) + αr2 μr 1 − tanh (μr i (t ) ) ˙ i (t )
2
cr i
b −2c
br br r r
˙
− i (t ) i (t ) αr1 −1 i cr
(t )−2αr2 μr tanh (μr i (t ) ) × 1−tanh (μr i (t ) )
2 2
cr cr

≤ ς2 |aMri (t )| + f¯r + ς2 ς3 | i (t )| + ς4 2i (t )
⎛ ⎞

N
+ ς3 ⎝ ai j −aMri (t ) + aMr j (t ) + 2lii f¯r ⎠
j=1
⎛ ⎞

i N
≤ ς2 aMri tk + f¯r + ςi1 ς2 ς3 + ςi12 ς4 + ς3 ⎝ li j aMr j tkj + 2lii f¯r ⎠
j=1
3930

≤ ϕ tki , tkj (37)
where

pr pr −gr
ς2 = αr3 (2Vr1 (0) ) 2gr + αr4 μr ,
g r

br br −cr

ς3 = αr1 (2λN Vr2 (0) ) 2c r + αr2 μr ,
c r
br −2cr
ς4 = |(αr1 bcrr ( bcrr −1)( (2λN Vr2 (0) ) 2cr + 2αr2 μ2r )|, and
ϕ(t i , t j ) = ς2 (|aMri (t i )| + f¯r ) + ς3 (| N li j aMr j (t j )| + 2lii f¯r ) + ςi1 ς2 ς3 + ς 2 ς4 .
k k k j=1 k i1 More-
over, D+ () represents the right derivative. According to Eq.(23), we can obtain ri (tki ) = 0.
Through inequality Eq. (37), we have
% t % t

|ri (t )| ≤ ˙ ri (s)d s ≤ ϕ tki , tkj d s (38)
tki tki
Based on the inequality Eq. (38) and the triggering mechanism Eq. (22), one can get
% tk+1
i
i
ri t = ηr αr3 e pr /gr t i + ηr αr4 ≤ ϕ t i

, j
k+1 vri k+1 k k ds
t (39)
tki
which yields
ηr αr4 ηr αr4
i
tk+1 − tki ≥ ≥ >0 (40)
ϕ tk , tk
i j ϕ̄
where ϕ̄ is the maximum value of ϕ(tki , tkj ). Inequality Eq. (40) indicates that there is a strictly
positive bound on the inter-event intervals. Thus, there is no Zeno behavior. This completes
the proof.
3.2. Design of guidance laws for achieving desired relative impact angles
In this section, the event-triggered guidance laws along the elevation and azimuth directions
of the LOS are designed, which can achieve desired relative impact angles (i.e., the objectives
of space-cooperative guidance described by Eqs.(15) and (16)) in fixed time.
First, the virtual consensus angular rate for qεi (t ) is designed as
∗
q˙εi (t ) = −αε1 bi ε /cε (t ) − αε2 tanh (με i (t ) ), (41)
where i (t ) is the consensus error of the relative LOS angles in the elevation direction, that
is,

N

i (t ) = ai j q˜εi (t ) − q˜ε j (t ) , (42)
j=1
where αε1 , αε2 , and με are positive constants. In addition, bε and cε are two positive odd
numbers that satisfy bε /cε > 2.
Next, the tracking error of angular rate in the elevation direction of LOS is defined as
∗
evεi (t ) = q˙εi (t ) − q˙εi (t ) = q˙εi (t ) + αε1 bi ε /cε (t ) + αε2 tanh (με i (t ) ) (43)
3931
Fig. 9. Triggering instants of the guidance commands under the event-triggered cooperative guidance scheme.
∗
For leading the current angular rate q˙εi (t ) to reach the virtual consensus angular rate q˙εi (t )
within fixed time, the guidance command in the elevation direction of LOS is designed to be
of form
aMεi (t ) = −2r˙i (t )q˙εi (t ) − ri (t )q˙βi

2
(t ) sin qεi (t ) cos qβi (t )
&
bε
+ ri (t ) αε1 bi ε /cε −1 tki i tki + αε2 με 1 − tanh2 με i tki i tki
cε
3932
Fig. 10. Comparisons of triggering numbers. (TT: Time-triggered; ET: Event-triggered.).
pε /gε i '
+ αε3 evεi tk + αε4 tanh με evεi tki , t ∈ tki , tk+1
i
(44)
where i (t ) is the consensus error of angular rate, i.e.,

N

i (t ) = ai j q˙εi (t ) − q˙ε j (t ) (45)
j=1
in which αε3 and αε4 are positive constants, pε and gε are two positive odd numbers that
satisfy pε /gε > 1, and tki (tk+1
i
) is the latest (next) triggering time of the i-th missile.
To achieve event-triggered guidance, the triggering mechanism for .. is designed as

pε /gε
i
tk+1 = inf t > tki |εi (t )| ≥ ηε αε3 evεi (t ) + ηε αε4 (46)
3933
where 0 < ηε < 1, and the triggering error εi (t ) is defined as

bε bε /cε −1 i i
εi (t ) = αε1 i tk i tk + αε2 με 1 − tanh2 με i tki i tki
cε
pε /gε i
+ αε3 evεi tk + αε4 tanh με evεi tki
bε
− αε1 bi ε /cε −1 (t )i (t ) − αε2 με 1 − tanh2 (με i (t ) ) i (t )
cε
pε /gε
− αε3 evεi (t ) − αε4 tanh (με evεi (t ) ) (47)
Assumption 2. The nonlinear terms and uncertain disturbances in the elevation and azimuth
directions of the LOS are bounded, that is,
⎧
⎨ fεide f + fεidis ≤ f¯ε
(48)
⎩ f de f + f dis ≤ f¯β
βi βi
where f¯ε and f¯β are non-negative constants.

Remark 4. It should be noted that if ri = 0, | fεide f + fεidis | and | fβide f + fβidis | will become
unbounded. In practice, before ri = 0, the missile will have already attacked the target
body. Therefore, the terminal value of ri will be a small value, not zero, which means that
| fεide f + fεidis | and | fβide f + fβidis | are bounded. Thus, Assumption 2 is reasonable.
Theorem 2. For the multi-missile system Eq. (11) with model uncertainties and unknown
target maneuverer, if the following inequalities hold:
αε2 > τε
αε4 (1 − ηε ) > f¯ε (49)
with
p g+εg
N αε4 ζε 1 αε4 ζε ε ε
τε = min , N 2
με (1 − κε ) αε4 (1 − ηε ) − f¯ε αε3 με (1 − κε )(1 − ηε )
where κε ∈ (0, 1) is a constant. The constant ζε satisfies ζε > 0 whose minimum value ζ ∗
∗
is ζ ∗ = x ∗ (1 − tanh x ∗ ), where x ∗ satisfies the equation e−2x + 1 − 2x ∗ = 0. The objective
of space-cooperative guidance described by Eq.(15) can be achieved through guidance law
Eq. (44) and triggering mechanism Eq. (46). In addition, there is no Zeno behavior.
Proof. Based on Eqs. (11) and (41)-(45), we can get the first derivative of the angular rate
tracking error evεi (t ) with respect to time, that is

e˙vεi (t ) = −2r˙i (t )q˙εi (t ) − ri (t )q˙βi
2
(t ) sin qεi (t ) cos qβi (t ) − aMεi (t ) /ri (t )+ fεide f + fεidis
bε
+ αε1 bi ε /cε −1 (t )i (t ) + αε2 με 1 − tanh2 (με i (t ) ) i (t )
cε
bε pε /gε i
= −αε1 bi ε /cε −1 tki i tki − αε2 με 1 − tanh2 με i tki i tki − αε3 evεi tk
cε
bε b /c −1
− αε4 tanh με evεi tki + αε1 i ε ε (t )i (t ) + αε2 με 1 − tanh2 (με i (t ) ) i (t )
cε
3934
Fig. 11. Triggering intervals of guidance command aMri (t ).
+ fεide f + fεidis

pε /gε
= − εi (t ) + αε3 evεi (t ) + αε4 tanh (με evεi (t ) ) + fεide f + fεidis (50)
Consider the following Lyapunov candidate function
1 2
N
Vε1 (t ) = e (t ) (51)
2 i=1 vεi
3935
Fig. 12. Trajectories and coordination variables under the cooperative guidance law in [22].
According to the triggering mechanism Eq. (46), as well as Lemmas 3 and 4, the derivative
of Eq. (51) is

N
V˙ε1 (t ) = evεi (t )e˙vεi (t )
i=1

N
N
pε /gε pε +gε
≤ |evεi (t )| ηε αε3 evεi (t ) + ηε α4 − αε3 evεigε (t )

i=1 i=1

N
N
N αε4 ζε
− αε4 |evεi (t )| + f¯ε |evεi (t )| +
i=1 i=1
με
3936
gε −pε pε +gε N αε4 ζε

− αε4 (1 − ηε ) − f¯ε (2Vε1 (t ) ) 2 +
1
≤ −αε3 (1 − ηε )N 2gε (2Vε1 (t ) ) 2gε (52)
με
Based on Lemma 2, the tracking error of angular rate in the elevation direction of LOS
can converge into a small region around zero in fixed time, i.e.,

N αε4 ζε
|evεi (t )| ≤ evεi (t ) ≤ τε = min ,
με (1 − κε ) αε4 (1 − ηε ) − f¯ε
pεg+gε
1 αε4 ζε ε
N2 (53)
α3 με (1 − κε )(1 − ηε )
and the bound of convergence time Tε1 is provided by

√
2 2
Tε1 ≤ + (54)
¯
κε αε4 (1 − ηε ) − fε p gε −pε pε +gε
α κ (1 − η )( − 1)N 2gε 2 2gε
ε
3 ε ε gε
Then, we will prove that the consensus error of the relative LOS angles in the elevation
direction of LOS can coverage within fixed time. The following Lyapunov candidate function
is considered
1
Vε2 (t ) = q˜εT (t )Lq˜ε (t )
2
where q˜ε (t ) = [q˜ε1 (t ) q˜ε2 (t ) . . . q˜εN (t )]T . Based on Lemmas 2 and 3, we can obtain

N
V˙ε2 (t ) = q˜εT (t )Lq˙˜ε (t ) = i (t )vεi (t )
i=1

N bε +cε
N
N
N αε2 ζε
≤ −αε1 i cε
(t ) − αε2 |i (t )| + |i (t )||evεi (t )| +
i=1 i=1 i=1
με
cε −bε
2cε bε +cε 1 N αε2 ζε
≤ −αε1 N (2λ2Vε2 (t ) ) 2cε − (αε2 − τε )(2λ2Vε2 (t ) ) 2 + (55)
με
where λ2 denotes the second smallest eigenvalue of L. According to Lemma 2, we can obtain
the following residual set
2 b 2c+εc
−1 N αε2 ζε N αε2 ζε ε ε
lim (t )|Vε2 (t ) ≤ min (2λ2 ) ,
t→Tε1 +Tε2 με (1 − κε )(αε2 −τ ) 2λ2 αε1 με (1 − κε )
(56)
where Tε2 can be provided by
√
2 2
Tε2 ≤ √ + cε −bε bε +cε
(57)
(αε2 − τε )κε λ2 αε1 ( cε − 1)N 2cε κε (2λ2 ) 2cε
b ε
Therefore, the consensus of the relative LOS angles in the elevation direction of LOS can
be achieved in fixed time, and the bound of total convergence time satisfies
Tε = Tε1 + Tε2
3937
Fig. 13. Results under the cooperative guidance law in [22].
√
2 2
≤ +
¯
κε αε4 (1 − ηε ) − fε
gε −pε pε +gε
αε3 κε (1 − ηε )( pε /gε − 1)N 2gε 2 2gε
√
2 2
+ √ + cε −bε bε +cε
(58)
(αε2 − τε )κε λ2 αε1 (bε /cε − 1)N 2cε κε (2λ2 ) 2cε
3938
Fig. 14. Trajectories, consensus errors and time-to-go under the cooperative guidance law in [48].
Next, it will be proved that there is no Zeno behavior. From the definition of i (t ) and
Lemma 1, we can get

N
λN q˜εT (t )Lq˜ε (t ) ≥ 2i (t ) ≥ λ2 q˜εT (t )Lq˜ε (t ). (59)
i=1
According to the definition of i (t ), it can be obtained that

N N
N

|i (t )| ≤
ai j evεi (t ) − evε j (t ) + 2αε2
ai j + αε1 ai j −i (t ) + j (t )
bε /cε bε /cε
j=1 j=1 j=1

bε /cε
≤ evε (t ) 1 + (lii − 1 ) evε (t ) 2 + 2αε1 lii (t ) 2 + 2αε2 lii ≤ i1 (60)
3939
Fig. 15. Results under the cooperative guidance law in [48].
with i1 (t ) = (lii − 1 + N (1/2) )(2Vε1 (0) )(1/2) + 2αε2 lii + 2αε1 lii (2λN Vε2 (0) )(bε /2cε ) , and λN
denotes the largest eigenvalue of L. Based on Eq. (47), we get

D+ |εi (t )| ≤ ˙ εi (t )

pε pεg−gε ε

≤ − αε3 evεi (t ) + αε4 με 1 − tanh (με evεi (t ) ) e˙vεi (t )
2
gε

bε bεc−cε ε
− αε1 i (t ) + αε2 με 1 − tanh (με i (t ) )
2 ˙ i (t )
cε
bε −2cε
bε bε
˙
−i (t )i (t ) αε1 −1 i cε
(t )−2αε2 με tanh (με i (t ) ) × 1−tanh (με i (t ) )
2 2
cε cε
3940

aMεi (t )
≤ 2 + f¯ε + 2 3 |i (t )| + 4 2 (t )

ri (t ) i
⎛ ⎞

N
+ 3 ⎝ ai j −aMεi (t ) + aMε j (t ) + 2lii f¯ε ⎠
j=1
⎛ ⎞
a t i N

Mεi k
≤ 2 i + f¯ε + i1 2 3 + i12 4 + 3 ⎝ li j aMε j tk + 2lii f¯ε ⎠
j
ri tk j=1

≤ φ tki , tkj (61)
where

pε pε −gε
2 = αε3 (2Vε1 (0) ) + αε4 με ,
2gε
gε

bε bε −cε

3 = αε1 (2λN Vε2 (0) ) 2c ε + αε2 με ,
cε
bε −2cε
4 = |(αε1 bcεε ( bcεε −1)( (2λN Vε2 (0) ) + 2αε2 μ2ε )|, and
2cε
⎛ ⎞
N

φ tki , tkj =2 aMεi tki /ri tki + f¯ε +3 ⎝ li j aMε j tkj +2lii f¯ε ⎠ + i1 2 3 +i12 4 .
j=1
The maximum value of φ(tki , tkj ) is φ̄. Since εi (tki ) = 0, through Eq. (61), we have
% t % t

|εi (t )| ≤ ˙ εi (s)d s ≤ φ tki , tkj d s (62)
tki tki
Based on the triggering mechanism Eqs. (46) and (62), one can get
% tk+1
i
i
εi t = ηε αε3 e pε /gε t i + ηε αε4 ≤ φ tki , tkj ds (63)

k+1 vεi k+1
tki
which yields
( '
i
tk+1 − tki ≥ [ηε αε4 ]/ φ tki , tkj ≥ ηε αε4 /φ̄ > 0 (64)
Inequality Eq. (64) indicates that there is a strictly positive bound on the inter-event inter-
vals. Thus, there is no Zeno behavior. This completes the proof.
Similar to the design progress of the event-triggered cooperative guidance law along the
elevation direction of LOS, the design of the event-triggered cooperative guidance law along
the azimuth direction of LOS is as the followings.
First, the virtual consensus angular rate for qβi (t ) is designed as
∗ b /c
q˙βi (t ) = −αβ1 Yi β β (t ) − αβ2 tanh μβ Yi (t ) , (65)
where Yi (t ) is the consensus error of the relative LOS angles in the azimuth direction, that
is,

N

Yi (t ) = ai j q˜βi (t ) − q˜β j (t ) , (66)
j=1
3941
Fig. 16. Energy consumption under different guidance schemes.
where αβ1 , αβ2 , and μβ are positive constants. In addition, bβ and cβ are two positive odd
numbers that satisfy bβ /cβ > 2.
Next, we define the tracking error of angular rate as
∗ b /c
evβi (t ) = q˙βi (t ) − q˙βi (t ) = q˙βi (t ) + αβ1 Yi β β (t ) + αβ2 tanh μβ Yi (t ) (67)
∗
For leading the true angular rate q˙βi (t ) to reach the virtual consensus angular rate q˙βi (t )
within fixed time, the guidance command in the azimuth direction of LOS is proposed to be
of form
aM βi (t ) = 2r˙i (t )q˙βi (t ) cos qεi (t ) − 2ri (t )q˙εi (t )q˙βi (t ) sin qεi (t )

&
bβ b /c −1
− ri (t ) cos qεi (t ) αβ1 Yi β β tki i tki + αβ2 μβ 1 − tanh2 μβ Yi tki i tki
cβ
pβ /gβ i '
+αβ3 evβi tk + αβ4 tanh μβ evβi tki , t ∈ tki , tk+1 i
(68)
where i (t ) is the consensus error of angular rate, i.e.,

N

i (t ) = ai j q˙βi (t ) − q˙β j (t ) (69)
j=1
in which αβ3 and αβ4 are positive constants. Moreover, pβ and gβ are two positive odd
numbers that satisfy pβ /gβ > 1.
Finally, the triggering mechanism is designed as

pβ /gβ
i
tk+1 = inf t > tki βi (t ) ≥ ηβ αβ3 evβi (t )ηβ αβ4 (70)
3942
Fig. 17. Consensus errors of the coordination variables and time-to-go under different values of vˆr .
where 0 < ηβ < 1, and the triggering error βi (t ) is
bβ bβ /cβ −1 i i
βi (t ) = αβ1 Yi tk i tk + αβ2 μβ 1 − tanh2 μβ Yi tki i tki
cβ
pβ /gβ i
+ αβ3 evβi tk + αβ4 tanh μβ evβi tki
bβ b /c −1
− αβ1 Yi β β (t )i (t ) − αβ2 μβ 1 − tanh2 μβ Yi (t ) i (t )
cβ

− αβ3 eγvβi (t ) − αβ4 tanh μβ evβi (t ) (71)
3943
Table 1
Initial engagement conditions.
Missile no. r(0) r˙(0) qε (0) qβ (0) q˙ε (0) q˙β (0)
Missile 1 4620 m −310 m/s 20 deg 60 deg −1.03 deg/s −1.03 deg/s
Missile 2 4640 m −320 m/s −10 deg 20 deg 0.86 deg/s 0.86 deg/s
Missile 3 4660 m −330 m/s 10 deg 30 deg −0.43 deg/s −0.43 deg/s
Missile 4 4680 m −340 m/s −20 deg 10 deg 0.43 deg/s 0.43 deg/s
Theorem 3. For the multi-missile system Eq. (12) with model uncertainties and unknown
target maneuverer, if the following inequalities hold:
αβ2 > τβ
αβ4 (1 − ηβ ) > f¯β (72)
with
p g+βg
N αβ4 ζβ 1 αβ4 ζβ β β
τβ = min , N 2
¯
μβ (1 − κβ ) αβ4 (1 − ηβ ) − fβi αβ3 μβ (1 − κβ )(1 − ηβ )
where κβ ∈ (0, 1) is a constant. The constant ζβ satisfies ζβ > 0 whose minimum value ζ ∗
∗
is ζ ∗ = x ∗ (1 − tanh x ∗ ), where x ∗ satisfies the equation e−2x + 1 − 2x ∗ = 0. The objective
of space-cooperative guidance described by Eq.(16) can be achieved through guidance law
Eq. (68) and triggering mechanism Eq. (70). In addition, there is no Zeno behavior.
Proof. As the proof of Theorem 3 is similar with that of Theorem 2, it is omitted here.
4. Simulations and comparative results
To verify the effectiveness of the proposed event-triggered cooperative guidance laws for
simultaneous attack with the constraint of desired relative impact angles, we conduct various
simulations that consider four missiles from different positions in the 3-D space to attack a
maneuvering target.
First, the initial conditions and the parameters of guidance laws are listed; secondly, sim-
ulation curves and results of the proposed guidance laws are illustrated; finally, comparison
studies to the finite-time cooperative guidance law in [22] and the fixed-time cooperative
guidance law in [48] that predefines specific desired impact angles are conducted to further
illustrate the advantages of the proposed robust event-triggered cooperative guidance law with
the constraint of desired relative impact angles.
4.1. Initial conditions

The initial engagement conditions and desired relative impact angles are listed in Ta-
bles 1 and 2, respectively. The maximum value of the available missile acceleration is as-
sumed to be 200 m/s2 . The initial position and velocity of target in the inertial frame are
(5000, 5000, 5000) m and (−100, 0, 0) m/s, respectively. The acceleration of the target in
the inertial frame is given as
aT y = 60 sin (0.35t + π4 ) m/s2 , and aT z = 60 cos(0.35t ) m/s2 , where aT y and aT z are the
components along the axes of inertial frame.
3944
Table 2.
Desired relative impact angles.
Direction of LOS M1 -M4 M2 -M1 M3 -M2 M4 -M3

Elevation (qε ) 37 deg −17 deg 20 deg −40 deg
Azimuth (qβ ) 40 deg −30 deg 1 deg −11 deg
The communication topology of the multi-missile team is illustrated in Fig. 5, whose

Laplacian matrix L is
⎡ ⎤
2 −1 −1 0
⎢−1 2 −1 0⎥
L=⎢ ⎣−1 −1
⎥. (73)
2 0⎦
0 0 −1 1
The guidance parameters for the time-cooperative guidance law Eq. (20) are chosen as vˆr =
−325,αr1 = 0.45, αr2 = 406, αr3 = 20, αr4 = 600,μr = 60, br = 11, cr = 5, pr = 23, gr =
11, ηr = 0.6. The guidance parameters for the space-cooperative guidance laws are chosen
as αε1 = αβ1 = 1, αε2 = αβ2 = 1, αε3 = αβ3 = 1, αε4 = αβ4 = 0.5, με = 80, μβ = 35, bε =
bβ = 11, cε = cβ = 5, pε = pβ = 23, gε = gβ = 11, ηε = ηβ = 0.7.
4.2. Simulation results

As illustrated in Fig. 6, four missiles with different initial conditions can attack a maneu-
vering target from different directions under the proposed event-triggered cooperative guidance
laws. The consensus of rang-to-go and radial relative velocity can be achieved within three
seconds (the total fight time is about 14 s), as can be seen in Fig. 7(a) and (b). The consensus
errors of radial relative velocities are no more than 0.02 m/s, which can be nearly neglected
compared with the norm of consensus radial relative velocity (about 325 m/s). As depicted
in Fig. 7(c), the time-to-go of each missile becomes nearly the same after three seconds.
The consensus of coordination variables is maintained until multiple missiles attack the tar-
get. In the end, the miss distance and impact time of multiple missiles are the same, which
are 0.3083 m and 14.267 s, respectively. Therefore, the objective of time-cooperative guidance
described by Eq. (13) is reached, i.e., multiple missiles with different initial conditions can
attack a maneuvering target simultaneously.
The errors of relative LOS angles and the LOS angles are shown in Fig. 8(a) and (b),
respectively. It can be seen that the errors between the current relative LOS angles and the
desired relative LOS angles are zero at the end of the engagement, which means that the
objectives express by Eqs. (15) and (16) are achieved. Note that the constraints of relative
desired LOS angles are utilized to achieve space-cooperative guidance in this study, instead of
assigning specific desired LOS angles, so the LOS angles in both the elevation and azimuth
directions are not necessary to maintain constant values, as illustrated in Fig. 8(b). As depicted
in Fig. 8(b) and (c), the angular rates of the LOS angles in the elevation and azimuth directions
reach consensus after about eight seconds, and the differences of LOS angles between missiles
are nearly constant. The guidance commands of multiple missiles are shown in Fig. 8(d),
which are rarely saturated, benefiting from only utilizing the constraint of relative impact
angles.
The triggering instants (the updates of cooperative guidance commands) are displayed in
Fig. 9. The inter-event times, i.e., triggering intervals between two consecutive update instants
3945
Fig. 18. Triggering numbers of aMr , aMε , and aMβ under different values of ηr .
of guidance laws, are dynamically updated through the proposed triggering mechanisms de-
scribed by Eqs. (22), (46) and (70). The events may be not triggered totally in a period,
during which the guidance errors are so small that it is unnecessary to regulate the guidance
commands. As a result, compared with conventional time-triggered guidance scheme with
constant update period, the update frequency of cooperative guidance commands under the
proposed event-triggered cooperative guidance laws appear to be lower, thus further reducing
the unnecessary energy consumption.
The specific numbers of triggering instants are illustrated in Fig. 10, where “TT” represents
the time-triggered guidance scheme and “ET” represents the proposed event-triggered guidance
laws. Since the flight time of missile in this scenario is 14.267 s and the simulation step is
3946
Fig. 19. Energy consumption of aMr , aMε , and aMβ under different values of ηr .
chosen as 0.001 s, so the update number for time-triggered guidance is 14,267. By utilizing
the proposed event-triggered guidance scheme, the triggering number can be reduced by more
than 53.52% to 6632 (e.g., the triggering number of aMr for missile 1), even can be reduced
to only 1190, about 8.34% of the update number for time-triggered guidance scheme (e.g.,
the triggering number of aMε for missile 1). Thus, the event-triggered cooperative guidance
scheme can greatly reduce the update frequency of the guidance commands.
3947
Fig. 20. Convergence time of vr under different value of ηr .
Table 3
The triggering numbers and intervals of the time-cooperative guidance commands aMri (t ).
Statistics Missile 1 Missile 2 Missile 3 Missile 4

Triggering number 6632 3469 5523 2844
Min interval (s) 0.001 0.001 0.001 0.001
Mean interval (s) 0.002 0.004 0.003 0.005
Max interval (s) 0.500 0.465 1.586 1.410
Reduction rate (%) 53.52 75.69 61.29 80.07
Moreover, taking the guidance command aMri (t ) as an example, the triggering intervals
(time intervals between two consecutive triggered instants) of multiple missiles are shown in
Fig. 11. As can be seen, the triggering intervals are dynamically tuned through the proposed
triggering mechanisms. When the guidance errors are large, frequently updating the guidance
commands are required and the triggering intervals are small. When the guidance errors
are samll enough, cooperative guidance instructions are updated sparsely and the triggering
intervals become large, even the updates of guidance commands are not triggered totally in
a period. The quantitative comparison with the time-triggered guidance scheme are listed in
Tables 3 and 4. As can be seen, the reduction rate of triggering numbers are from 53.52% to
91.66. Combined with the results illustrated in Figs. 6–8, it can be inferred that the update
requirements are drastically reduced to save computational resources without compromising
the performance of cooperative guidance by utilizing the proposed even-triggered cooperative
guidance.
3948
Fig. 21. Guidance performance of aMε under different qε and qβ .
Table 4
The triggering numbers and intervals of the space-cooperative guidance commands aMεi (t ) and aM βi (t ).
Missile1 Missile 2 Missile 3 Missile 4
Statistics aMε aMβ aMε aMβ aMε aMβ aMε aMβ

Triggering number 1190 2008 2521 1604 2233 1789 2751 1864
Min interval (s) 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001
Mean interval (s) 0.004 0.003 0.005 0.005 0.004 0.004 0.007 0.007
Max interval (s) 6.493 3.833 2.571 4.004 3.177 5.141 3.246 4.772
Reduction rate (%) 91.66 85.93 82.33 88.76 84.35 87.46 80.72 86.94
4.3. Comparative study

To further verify the effectiveness of the proposed guidance scheme, the comparative sim-
ulations to the cooperative guidance law in [22] that utilized the estimation of time-to-go to
achieve simultaneous attack and the fixed-time cooperative guidance law in [48] that prede-
fines specific desired impact angles are conducted. The scenario and the initial conditions are
the same as those in Section 4.2.
3949
Fig. 22. Guidance performance of aMβ under different qε and qβ .
First, as shown in Fig. 12, multiple missiles can attack the maneuvering target from differ-
ent directions under the cooperative guidance law proposed in [22]. Note that the coordination
variable for achieving simultaneous attack in [22] is directly chosen as the explicit estimation
of time-to-go, i.e.,
r
tˆgo = − . (74)
r˙
Therefore, the estimation error of time-to-go will significantly influence the accuracy of
cooperative guidance. Compared the results in Fig. 12(b) and (d) with those in Fig. 7(a) and
(c), it can be seen that the terminal relative distances and time-to-go are not strict consensus
under the guidance law in [22], so the simultaneous attack is not precisely achieved. In this
study, the consensus errors of range-to-go and relative velocity along the LOS are utilized as
the coordination variables, so the relative distances and time-to-go can reach strict consensus,
which means that simultaneous attack can be precisely achieved, as shown in Fig. 7.
As presented in Fig. 14, under the fixed-time cooperative guidance law proposed in [48],
multiple missiles can attack the maneuvering target simultaneously and the time-to-go achieves
strict consensus. As can be seen in Fig. 15, the errors between the current LOS angles and
3950
Table 5
The computational burden of the proposed method and other methods (s).
Proposed cooperative Fixed-time cooperative Finite-time cooperative Argument Proportional

guidance guidance [48] guidance [22] Navigation Guidance 0
M1 Mean 3.7722 × 10−5 2.7498 × 10−5 1.2311 × 10−5 2.9645 × 10−7
SD 2.4035 × 10−5 1.8924 × 10−5 0.2811 × 10−5 5.4246 × 10−7
M2 Mean 3.6650 × 10−5 2.6941 × 10−5 1.0156 × 10−5 2.0622 × 10−7
SD 1.7264 × 10−5 0.9358 × 10−5 0.3469 × 10−5 6.1555 × 10−7
M3 Mean 3.7016 × 10−5 2.6831 × 10−5 0.9835 × 10−5 1.9193 × 10−7
SD 1.0489 × 10−5 0.4320 × 10−5 0.3511 × 10−5 4.9462 × 10−7
M4 Mean 3.6314 × 10−5 2.6797 × 10−5 0.9810 × 10−5 1.8906 × 10−7
SD 0.8013 × 10−5 0.4595 × 10−5 0.3306 × 10−5 4.9723 × 10−7
the desired LOS angles both in the elevation and azimuth directions of LOS become zero
after about three seconds. However, compared the results in Fig. 15(d) with those in Fig. 8(d),
the guidance commands under the cooperative guidance law in [48] sink into saturation for
a longer time than the guidance commands under the cooperative guidance laws proposed
in the present paper. This is because that the guidance laws in [48] assign specific desired
impact angles for each missile, and multiple missiles are required to conduct maneuvers
for maintaining the constant impact angles. In contrast, the proposed cooperative guidance
laws with the constraint of relative impact angles are not required to maintain the constant
desired impact angles. As a result, the proposed cooperative guidance laws will not require
too much regulation and the norm of guidance commands are generally smaller than those of
the guidance laws in [48].
Compared the results in Figs. 7, 8 and 14, 15, one can see that the convergence time of
the proposed cooperative guidance laws is not longer than that of the fixed-time cooperative
guidance law in [48]. Therefore, the proposed fixed-time event-triggered cooperative guidance
laws can guarantee the convergence rate with fewer updates of guidance commands.
The detailed comparisons of the energy consumption under different guidance laws are
illustrated in Fig. 16, where “FxT” represents the fixed-time cooperative guidance law in [48],
“ET” represents the event-triggered cooperative guidance laws proposed in the present paper,
and “AT” represents the situation in which the proposed event-triggered cooperative guidance
laws are set to be always triggered. It can be derived from Fig. 16 that the proposed fixed-time
event-triggered cooperative guidance laws with the constraint of relative impact angles can
reduce the consumption of energy greatly. In order to take a deep look at the contribution of
“event-triggered guidance scheme” and “utilizing the constraint of relative impact angles” to
the reduction in energy consumption, we set the triggering conditions described by Eqs. (22),
(46) and (70) are always satisfied, and corresponding results are labeled “AT” in Fig. 16. As
can be seen, the energy consumption can be reduced by only pre-specifying desired relative
impact angles, instead of assigning specific desired impact angles for each missile, and the
“event-triggered guidance scheme” can further reduce the energy consumption. Taking the
No.4 missile as an example, the total energy consumption can be reduced by more than 80%
through the proposed guidance laws.
Finally, the statistical computation time of generating guidance commands for each sim-
ulation step are listed in Table 5. As can be seen, the computation time of the proposed
cooperative guidance, fixed-time cooperative guidance [48], and finite-time cooperative guid-
ance [22] are basically on the same order of magnitude, about several 10−5 s. Since the
3951
Fig. 23. Tracking error of radial velocity under different value of μr .
classical Argument Proportional Navigation Guidance [65] has a simple form, so the com-
putation burden is very light, only about several 10−7 s. Generally, the guidance cycle of the
practical online application is no more than 1 ms (i.e., 10−3 s), so the computation time of the
proposed guidance law is small enough for on-board flight applications.
4.4. Analysis of parameter sensitivities

As shown in the guidance laws and triggering conditions described by Eqs. (20), (22),
(44), (46), (68), and (70), the parameters that should be determined before utilizing the pro-
posed event-triggered cooperative guidance law consist of vˆr , ηi , μi , αi1 , αi2 , αi3 , αi4 , bi /ci ,
and pi /gi (i = r, ε, β). In the following, the sensitivities of these parameters on the guidance
performance will be analysed.
First, the parameter vˆr is a constant and can be chosen as the average value of the initial
radial relative velocities. Moreover, it can be demonstrated that the radial relative velocity
vri (t ) can entry a small neighborhood of vˆr . In order to verify the adaptability of the proposed
cooperative guidance law to different values of vˆr , we conduct numerical simulations where
the value of vˆr is set from −375 m/s to −275 m/s.
As can be seen in Fig. 17, under different values of vˆr , the range-to-go, radial relative
velocities, and time-to-go can achieve consensus. As a result, multiple missiles can attack the
maneuvering target simultaneously. In addition, if the magnitude of vˆr is set as a larger one,
the flight time of missiles will be shorter. In this paper, the magnitude of vˆr is just taken as
the average of the initial relative velocities.
Secondly, the parameter ηr in the triggering mechanism (22) will influence the triggering
number and fuel consumption. As shown in Fig. 18, when the value of ηr increases, the
triggering number of aMr decreases. Moreover, it can be seen that the change of ηr has no
influence on the triggering numbers of aMε and aMβ . Similarly, as illustrated in Fig. 19, when
3952
Fig. 24. Triggering numbers and energy consumption of aMr , aMε , and aMβ under different values of μr .
the value of ηr increases, the energy consumption of aMr decreases. In addition, as can be
seen, the change of ηr has no influence on the energy consumption of aMε and aMβ . In fact, as
shown in the triggering mechanism expressed by Eq. (22), if the value of ηr becomes smaller,
pr /gr
the threshold value ηr αr3 |evri (t )| + ηr αr4 will become smaller, so the triggering number and
energy consumption will become greater and vice versa.
On the other hand, as presented in Fig. 20, if the value of ηr increases, the convergence
time of vr will become longer. Since the engagement time of terminal guidance is quite short,
about tens of seconds, the consensus error of vr should converge as fast as possible. Therefore,
a trade-off needs to be considered between less triggering number/energy consumption and fast
3953
Fig. 25. Energy consumption of aMr and convergence time of vr under different values of αr1 , αr2 , αr3 , and αr4 .
convergence rate. Here, the value of ηr is taken as 0.6 for the smaller triggering number/energy
consumption and the faster convergence rate.
As illustrated in Figs. 21 and 22, similar to ηr , when the values of ηε and ηβ increase,
the triggering number and energy consumption decrease, while the consensus error of qε
and qβ increase. As a result, a trade-off is required to be considered between less triggering
number/energy consumption and excellent guidance precision. In addition, as shown in Fig.
20, smaller ηε and ηβ will enlarge the convergence time of vr , because missiles should con-
duct much regulation to achieve both the objectives of time-cooperative and space-cooperative
guidance. It should be noted that the change of ηε rarely has influence on the performance of
aMβ , and the change of ηβ rarely has influence on the performance of aMε . Therefore, com-
3954
bined the results in Figs. 20–22, ηε and ηβ are both taken as 0.7 for the trade-off between less
triggering number/energy consumption, faster convergence rate, and higher guidance precision.
Thirdly, according to Lemma 2, the tracking error of radial relative velocities can converge
into a small region around zero in fixed time, i.e.,

N αr4 ζr
|evri (t )| ≤ evr (t ) ≤ τr = min ,
μr (1 − κr ) αr4 (1 − ηr ) − f¯r
p g+r g
1 αr4 ζr r r
N2 . (75)
αr3 μr (1 − κr )(1 − ηr )
It can be seen in Eq. (29) that the value of μr will influence the velocity tracking error, and
smaller μr will cause a larger upper bound of |evri (t )|. As shown in Fig. 23, the increasement
of μr will decrease the velocity tracking error. When the value of μr is larger than about 60,
the velocity tracking error will be smaller than 0.01 m/s. However, as illustrated in Fig. 24(a)
and (c), when the values of μr is larger than about 65, the triggering number and energy
consumption of aMr will increase dramatically. Therefore, a trade-off between less triggering
number/energy consumption and higher guidance precision is required. Similar to ηr , as can
be seen in Fig. 24(b) and (c), the change of μr nearly has no influence on the triggering
numbers and energy consumption of aMε and aMβ . Here, the value of μr is taken as 60. The
value of με and μβ are taken as 80 and 35, respectively.
The influence of αr1 , αr2 , αr3 , and αr4 on the energy consumption of aMr and convergence
time of vr are shown in Fig. 25. The convergence time of vr denotes the time that radial relative
velocities of multiple missiles achieve consensus. As illustrated in Fig. 26, the convergence
time of vr is 4.94 s when αr1 equals to 0.2, and the convergence time of vr is 11.62 s when
αr1 equals to 0.5. The detailed results are shown in Fig. 25(a). As the value of αr1 increases,
the convergence time of vr becomes longer and the energy consumption of aMr decreases. For
the balance between the convergence time and energy consumption, the value of αr1 is taken
as 0.45. As can be seen in Fig. 25(c) and (d), when the values of αr2 , αr3 , and αr4 increase,
the convergence time of vr will drop sharply first and then reach a small range of variation.
On the other hand, with the increasement of αr2 , αr3 , and αr4 , the energy consumption of aMr
becomes larger. Therefore, a trade-off should be considered between less energy consumption
and fast convergence of the radial velocity consensus error. Here, αr2 , αr3 , and αr4 are taken
as 406, 20 and 600, respectively. Similarly, αε1 , αε2 , αε3 , and αε4 are set as 1, 1, 1 and 0.5,
respectively. Moreover, αβ1 , αβ2 , αβ3 , and αβ4 are set as 1, 1, 1 and 0.5, respectively.
Finally, as depicted in Fig. 27(a), taking pr /gr as 23/11 can achieve less energy con-
sumption of aMr and fast convergence rate of vr . It can be seen in Fig. 27(b) that when br /cr
increases, the convergence time of vr becomes larger and the energy consumption of aMr be-
comes lower. Thus, considering a balance between the energy consumption and convergence
rate, br /cr is taken as 11/5. Similarly, pε /gε and pβ /gβ are both taken as 23/11. In addition,
bε /cε and bβ /cβ are both set as 11/5.
4.5. Monte Carlo simulations

In order to future verify the robustness of the proposed event-triggered cooperative guidance
laws against initial errors and measurement noises and inquiry the sensitivities of parameters
in the guidance laws, 1000 Monte Carlo simulations are carried out, and the simulation
conditions are set as follows:
3955
Fig. 26. Radial relative velocities vr under different values of αr1 .
(a) The dispersions of uncertainties in the initial LOS angles are assumed to follow uniform
distributions with zero means, and standard deviations of 5 degrees.
(b) The uncertainty in the initial relative range follows a uniform distribution with zero
mean and standard deviation of 100 m.
(c) A standard deviation of 0.1 deg is assumed for the LOS angle measurement noise with
zero mean.
(d) A standard deviation of 10 m is assumed for the relative range measurement noise with
zero mean.
(e) The autopilot lag is considered as 0.1 s.
(f) The information of the maneuvering target is unknown.
The corresponding results for the miss distance, consensus error of impact time, and con-
sensus errors of impact angle are presented in Figs. 28–32. As shown in Fig. 28, the range-to-
go, radial relative velocity, and time-to-go can achieve consensus, so the simultaneous attack
can be realized. The detailed Monte Carlo results of the miss distance, maximum error of
impact time, maximum error of impact angle, and consensus error of impact angle are pre-
sented in Figs. 29–32 and Tables 6, 7. The maximum impact time error denotes the maximum
difference of the impact time between each two missiles of the multi-missile team, which can
be utilized to evaluate the performance of time-cooperative guidance. Furthermore, the max-
imum error of impact angle denotes the maximum difference of the terminal relative impact
angle and the desired relative impact angle between each two missiles of the multi-missile
team, which can be utilized to evaluate the performance of space-cooperative guidance. The
3956
Z. Chen, X. Liu and W. Chen
3957
Journal of the Franklin Institute 360 (2023) 3914–3966

Fig. 27. Energy consumption of aMr and convergence time of vr under different values of pr /gr and br /cr .
Fig. 28. Results of Monte Carlo simulations.
Table 6
The statistical results of miss distance.
Statistical results M1 M2 M3 M4
Miss Mean 0.23456 0.42369 0.27294 0.39673
Dis- Standard Deviation 0.14053 0.16050 0.14976 0.16433
tance
(m)
consensus error of impact angle is the difference between the terminal relative impact angle
and the desired relative impact angle.
The 1000-run Monte Carlo results of the miss distance are given in Fig. 29 and Table 6. As
can be seen, over 60% of the miss distances are less than 0.5 m. The maximum miss distance
3958
Fig. 29. Monte Carlo results of miss distance.
3959
Fig. 30. Monte Carlo results of maximum impact time error.
3960
Fig. 31. Monte Carlo results of maximum impact angle error.
is about 1.2 m, and the mean value is no more than 0.424 m. The Fig. 30 depicts the maximum
impact angle error of 1000 Monte Carlo simulations, the maximum is about 0.0034 s, and the
proportion of the maximum impact angle error that is less than 0.003 s is more than 97%. As
listed in Table 7, the mean of the maximum impact time error is 0.00221 s, and the standard
deviation is 0.00039 s.
As illustrated in Fig. 31(a) and (c), the maximum impact angle errors in the elevation
direction and that in the azimuth direction are no more than 0.12 deg and 0.2 deg, respectively.
As shown in Fig. 31(a) and Fig. 31(c), the maximum impact angle error in the elevation
direction that is less than 0.1 deg is more than 99%, and the maximum impact angle error in
the azimuth direction is less than 0.1 deg is more than 89%. As listed in Table 7, the mean
and the standard deviation of the maximum impact angle error in the elevation direction are
3961
Fig. 32. Consensus error of impact angle.
Table 7
The statistical results of impact time and impact angle.
Statistical results Mean Standard deviation

Maximum error of impact time (s) 0.00221 0.00039
Maximum error of impact angle in the elevation direction (deg) 0.03963 0.02041
Maximum error of impact angle in the azimuth direction (deg) 0.05360 0.03281
0.03963 deg and 0.02041 deg, respectively. In addition, the mean and the standard deviation
of the maximum impact angle error in the azimuth direction are 0.05360 deg and 0.03281 deg,
respectively.
As depicted in Fig. 32, the 1000-run Monte Carlo results of the impact angle consensus
error with different perturbed conditions can converge to near zero. The detailed statistical
results are shown in Fig. 31 and Table 7, which has been analysed above.
In conclusion, the presented Monte Carlo simulation results further demonstrate the ro-
bustness of the proposed event-triggered cooperative guidance laws against the maneuvering
target with multiple disturbances. Additionally, the Monte Carlo results indicate that the se-
lected parameters of the proposed guidance laws can adaptive to variable initial conditions
and measurement noises.
5. Conclusion
In this paper, a 3-D event-triggered fixed-time robust cooperative guidance scheme is pro-
posed, which can satisfy the constraint of desired relative impact angles. First, the consensus
errors of range-to-go and relative velocity along the LOS are chosen as the coordination vari-
ables, so the guidance error suffering from the estimation error of time-to-go can be avoided.
Secondly, the desired relative impact angles are pre-specified for multiple missiles to achieve
space-cooperative guidance, instead of assigning specific desired impact angles for each mis-
3962
sile before launch. As a result, the commands of the proposed guidance laws will require less
regulation to maintain constant impact angles, thus reducing the fuel consumption. Thirdly, by
designing an appropriate event-triggering mechanism, the cooperative guidance commands are
not updated and hold constant until trigger conditions are satisfied, which can significantly
reduce the update frequency of the cooperative guidance commands as well as ensure the
guidance system’s stability. Finally, in order to improve the convergence rate of guidance er-
rors when the update frequency of guidance commands decrease, a fixed-time event-triggered
cooperative guidance is developed, whose convergence rate is independent of initial condi-
tions. Therefore, the proposed event-triggered cooperative guidance laws can guarantee the
performance of cooperative guidance while decreasing the utilization of limited resources on
the missile. In the further, the constraint of the LOS rate will be further considered to avoid
the degradation of seeker. In addition, since there are too many parameters in the cooperative
guidance laws, some intelligent technology can be considered to select appropriate parame-
ters for achieving better guidance performance or lower energy consumption such as including
data-driven method [66], deep learning method [67], and reinforcement learning method [68].
Declaration of Competing Interest
Authors declare that they have no conflict of interest.
Acknowledgments
Project funded by China Postdoctoral Science Foundation (Grant no 2021M700321).
References
[1] J. Zhou, X. Wu, Y. Lv, X. Li, Z. Liu, Recent progress on the study of multi-vehicle coordination in cooperative
attack and defense: an overview, Asian J. Control 24 (2) (2021) 794–809.
[2] S.R. Kumar, T. Shima, Cooperative nonlinear guidance strategies for aircraft defense, J. Guid. Control Dyn. 40
(1) (2017) 124–138.
[3] H. Yu, K. Dai, H. Li, Y. Zou, X. Ma, S. Ma, H. Zhang, Distributed cooperative guidance law for multiple
missiles with input delay and topology switching, J. Frankl. Inst. 358 (17) (2021) 9061–9085.
[4] I.S. Jeon, J.I. Lee, M.J. Tahk, Homing guidance law for cooperative attack of multiple missiles, J. Guid. Control
Dyn. 33 (1) (2010) 275–280.
[5] J. Zhao, R. Zhou, Unified approach to cooperative guidance laws against stationary and maneuvering targets,
Nonlinear Dyn. 81 (4) (2015) 1635–1647.
[6] W. Dong, C. Wang, J. Wang, M. Xin, Three-dimensional nonsingular cooperative guidance law with different
field-of-view constraints, J. Guid. Control Dyn. 44 (11) (2021) 2001–2015.
[7] I.-S. Jeon, J.-I. Lee, M.-J. Tahk, Impact-time-control guidance law for anti-ship missiles, IEEE Trans. Control
Syst. Technol. 14 (2) (2006) 260–266.
[8] N. Harl, S. Balakrishnan, Impact time and angle guidance with sliding mode control, IEEE Trans. Control Syst.
Technol. 20 (6) (2012) 1436–1449.
[9] Y. Chen, J. Wang, C. Wang, J. Shan, M. Xin, A modified cooperative proportional navigation guidance law, J.
Frankl. Inst. 356 (11) (2019) 5692–5705.
[10] Z. Li, Z. Ding, Robust cooperative guidance law for simultaneous arrival, IEEE Trans. Control Syst. Technol.
27 (3) (2018) 1360–1367.
[11] J. Zhao, S. Yang, Integrated cooperative guidance framework and cooperative guidance law for multi-missile,
Chin. J. Aeronaut. 31 (3) (2018) 546–555.
[12] G. Li, J. Lü, G. Zhu, K. Liu, Distributed observer-based cooperative guidance with appointed impact time and
collision avoidance, J. Frankl. Inst. 358 (14) (2021) 6976–6993.
[13] N. Dhananjay, D. Ghose, Accurate time-to-go estimation for proportional navigation guidance, J. Guid. Control
Dyn. 37 (4) (2014) 1378–1383.
3963
[14] C. Wang, X. Ding, J. Wang, J. Shan, A robust three-dimensional cooperative guidance law against maneuvering
target, J. Frankl. Inst. 357 (10) (2020) 5735–5752.
[15] X. Wang, C. Tan, 3-D impact angle constrained distributed cooperative guidance for maneuvering targets without
angular-rate measurements, Control Eng. Pract. 78 (2018) 142–159.
[16] Y. Zhang, X. Wang, H. Wu, A distributed cooperative guidance law for salvo attack of multiple anti-ship
missiles, Chin. J. Aeronaut. 28 (5) (2015) 1438–1450.
[17] B. Li, D. Lin, H. Wang, Finite time convergence cooperative guidance law based on graph theory, Optik 127
(21) (2016) 10180–10188.
[18] Z. Hou, Y. Yang, L. Liu, Y. Wang, Terminal sliding mode control based impact time and angle constrained
guidance, Aerosp. Sci. Technol. 93 (2019) 105142.
[19] V. Shaferman, T. Shima, Cooperative optimal guidance laws for imposing a relative intercept angle, J. Guid.
Control Dyn. 38 (8) (2015) 1395–1408.
[20] V. Shaferman, T. Shima, Cooperative differential games guidance laws for imposing a relative intercept angle,
J. Guid. Control Dyn. 40 (10) (2017) 2465–2480.
[21] T. Lyu, Y. Guo, C. Li, G. Ma, H. Zhang, Multiple missiles cooperative guidance with simultaneous attack
requirement under directed topologies, Aerosp. Sci. Technol. 89 (2019) 100–110.
[22] J.H. Song, S.M. Song, S.L. Xu, Three-dimensional cooperative guidance law for multiple missiles with finite–
time convergence, Aerosp. Sci. Technol. 67 (2017) 193–205.
[23] X. Wang, X. Lu, Three-dimensional impact angle constrained distributed guidance law design for cooperative
attacks, ISA Trans 73 (2018) 79–90.
[24] T. Lyu, C. Li, Y. Guo, G. Ma, Three-dimensional finite-time cooperative guidance for multiple missiles without
radial velocity measurements, Chin. J. Aeronaut. 32 (5) (2019) 1294–1304.
[25] W.P. Heemels, K.H. Johansson, P. Tabuada, An introduction to event-triggered and self-triggered control, in:
The 51st IEEE conference on decision and control (CDC), IEEE, 2012, pp. 3270–3285.
[26] X. Ge, Q. Han, X. Zhang, L. Ding, F. Yang, Distributed event-triggered estimation over sensor networks: a
survey, IEEE Trans. Cybern. 50 (3) (2020) 1306–1320.
[27] Z. Peng, J. Wang, D. Wang, Q.-L. Han, An overview of recent advances in coordinated control of multiple
autonomous surface vehicles, IEEE Trans. Ind. Inform. 17 (2) (2020) 732–745.
[28] Q. Hu, Y. Shi, Event-based coordinated control of spacecraft formation flying under limited communication,
Nonlinear Dyn. 99 (3) (2020) 2139–2159.
[29] D. Duan, C. Liu, Event-based optimal guidance laws design for missile-target interception systems using fuzzy
dynamic programming approach, ISA Trans. (2021) published online 11 Nov, doi:10.1016/j.isatra.2021.10.037.
[30] M. Li, C. Guo, H. Yu, Y. Yuan, Event-triggered containment control of networked underactuated unmanned
surface vehicles with finite-time convergence, Ocean Eng. 246 (2022) 110548.
[31] C. Zhang, G. Zhang, X. Zhang, DVSL guidance-based composite neural path following control for underactuated
cable-laying vessels using event-triggered inputs, Ocean Eng. 238 (2021) 109713.
[32] A.K. Behera, B. Bandyopadhyay, M. Cucuzzella, A. Ferrara, X. Yu, A survey on event-triggered sliding mode
control, IEEE J. Emerg. Sel. Top. Power Electron. 2 (3) (2021) 206–217.
[33] L. Dou, S. Cai, X. Zhang, X. Su, R. Zhang, Event-triggered-based adaptive dynamic programming for distributed
formation control of multi-UAV, J. Frankl. Inst. 359 (8) (2022) 3671–3691.
[34] X. Song, P. Sun, S. Song, V. Stojanovic, Event-driven NN adaptive fixed-time control for nonlinear systems
with guaranteed performance, J. Frankl. Inst. 359 (9) (2022) 4138–4159.
[35] P. He, J. Wen, V. Stojanovic, F. Liu, X. Luan, Finite-time control of discrete-time semi-Markov jump linear
systems: a self-triggered MPC approach, J. Frankl. Inst. 359 (13) (2022) 6939–6957.
[36] J. Liu, J. Yang, An event-triggered optimal cooperative guidance law for simultaneous attacks with impact angle
constraints, Optim. Control Appl. Methods (2022) published online 20 Apr, doi:10.1002/oca.2894.
[37] T. Long, Y. Cao, J. Sun, G. Xu, Adaptive event-triggered distributed optimal guidance design via adaptive
dynamic programming, Chin. J. Aeronaut. (2021) published online 16 Sep, doi:10.1016/j.cja.2021.08.005.
[38] A. Sinha, S.R. Kumar, D. Mukherjee, Three-dimensional nonlinear cooperative salvo using event-triggered
strategy, J. Guid. Control Dyn. 44 (2) (2021) 328–342.
[39] Y. Gao, C. Liu, D. Duan, S. Zhang, Distributed optimal event-triggered cooperative control for nonlinear multi-
missile guidance systems with partially unknown dynamics, Int. J. Robust Nonlinear Control (2022) published
online 18 Jul, doi:10.1002/rnc.6285.
[40] B. Biswas, A. Maity, S.R. Kumar, Finite-time convergent three-dimensional nonlinear intercept angle guidance,
J. Guid. Control Dyn. 43 (1) (2020) 146–153.
3964
[41] D. Duan, C. Liu, J. Sun, Adaptive periodic event-triggered control for missile-target interception system with
finite-horizon convergence, Trans. Inst. Meas. Control 42 (10) (2020) 1808–1822.
[42] X. Yang, S. Song, Three-dimensional consensus algorithm for nonsingular distributed cooperative guidance
strategy, Aerosp. Sci. Technol. 118 (2021) 106958.
[43] Z. Xu, X. Liu, J. Cao, M. Song, Fixed-time bipartite consensus of nonlinear multi-agent systems under directed
signed graphs with disturbances, J. Frankl. Inst. 359 (6) (2022) 2693–2709.
[44] Y. Liu, H. Li, Z. Zuo, X. Li, R. Lu, An overview of finite/fixed-time control and its application in engineering
systems, IEEE-CAA J. Autom. Sin. (2022) published online Jan, doi:10.1109/JAS.2022.105413.
[45] M.A. Jamal, R. Kumar, S. Mukhopadhyay, S. Das, Fixed-time stability of dynamical systems with impulsive
effects, J. Frankl. Inst. 359 (7) (2022) 3164–3182.
[46] A. Polyakov, Nonlinear feedback design for fixed-time stabilization of linear control systems, IEEE Trans.
Autom. Control 57 (8) (2011) 2106–2110.
[47] R. Wei, J. Cao, A. Alsaedi, Finite-time and fixed-time synchronization analysis of inertial memristive neural
networks with time-varying delays, Cogn. Neurodyn. 12 (1) (2018) 121–134.
[48] Z. Chen, W. Chen, X. Liu, J. Cheng, Three-dimensional fixed-time robust cooperative guidance law for simul-
taneous attack with impact angle constraint, Aerosp. Sci. Technol. 110 (2021) 106523.
[49] H. Yu, K. Dai, H. Li, Y. Zou, X. Ma, S. Ma, H. Zhang, Three-dimensional adaptive fixed-time cooperative
guidance law with impact time and angle constraints, Aerosp. Sci. Technol. 123 (2022) 107450.
[50] W. Dong, C. Wang, J. Wang, Z. Zuo, J. Shan, Fixed-time terminal angle constrained cooperative guidance law
against maneuvering target, IEEE Trans. Aerosp. Electron. Syst. 58 (2) (2021) 1352–1366.
[51] H. Li, H. Li, Y. Cai, Three-dimensional cooperative guidance law to control impact time and angle
with fixed-time convergence, Proc. Inst. Mech. Eng., Part G (2022) published online 4 Feb, doi:10.1177/
09544100211043093.
[52] J. Yu, X. Dong, Q. Li, J. Lü, Z. Ren, Task coupling based layered cooperative guidance: theories and applications,
Control Eng. Pract. 121 (2022) 105050.
[53] M. Zhuang, L. Tan, K. Li, S. Song, Fixed-time formation control for spacecraft with prescribed performance
guarantee under input saturation, Aerosp. Sci. Technol. 119 (2021) 107176.
[54] Z. Chen, M.R. Emami, W. Chen, Connectivity preservation and obstacle avoidance in small multi-spacecraft
formation with distributed adaptive tracking control, J. Intell. Robot. Syst. 101 (1) (2021) 1–23.
[55] R. Olfati-Saber, R.M. Murray, Consensus problems in networks of agents with switching topology and time-de-
lays, IEEE Trans. Autom. Control 49 (9) (2004) 1520–1533.
[56] A. Polyakov, D. Efimov, W. Perruquetti, Finite-time and fixed-time stabilization: implicit Lyapunov function
approach, Automatica 51 (2015) 332–340.
[57] J. Liu, Y. Yu, H. He, C. Sun, Team-triggered practical fixed-time consensus of double-integrator agents with
uncertain disturbance, IEEE Trans. Cybern. 51 (6) (2020) 3263–3272.
[58] Z. Zuo, Nonsingular fixed-time consensus tracking for second-order multi-agent networks, Automatica 54 (2015)
305–309.
[59] R.J. Wallsgrove, M.R. Akella, Globally stabilizing saturated attitude control in the presence of bounded unknown
disturbances, J. Guid. Control Dyn. 28 (5) (2005) 957–963.
[60] G. Li, Y. Wu, P. Xu, Fixed-time cooperative guidance law with input delay for simultaneous arrival, Int. J.
Control 94 (6) (2019) 1664–1673.
[61] C.-D. Yang, C.-C. Yang, Analytical solution of three-dimensional realistic true proportional navigation, J. Guid.
Control Dyn. 19 (3) (1996) 569–577.
[62] B.S. Kim, J.G. Lee, H.S. Han, Biased PNG law for impact with angular constraint, IEEE Trans. Aerosp.
Electron. Syst. 34 (1) (1998) 277–288.
[63] S.R. Kumar, S. Rao, D. Ghose, Nonsingular terminal sliding mode guidance with impact angle constraints, J.
Guid. Control Dyn. 37 (4) (2014) 1114–1130.
[64] T. Xu, G. Lv, Z. Duan, Z. Sun, J. Yu, Distributed fixed-time triggering-based containment control for networked
nonlinear agents under directed graphs, IEEE Trans. Circuits Syst. I 67 (10) (2020) 3541–3552.
[65] P. Zarchan, Tactical and Strategic Missile Guidance, American Institute of Aeronautics and Astronautics, Inc.,
2012.
[66] V. Djordjevic, V. Stojanovic, H. Tao, X. Song, S. He, W. Gao, Data-driven control of hydraulic servo actuator
based on adaptive dynamic programming, Discrete Contin. Dyn. Syst. 15 (7) (2022) 1633–1650.
3965
[67] N. Rodríguez-Barroso, A.R. Moya, J.A. Fernández, E. Romero, E. Martínez-Cámara, F. Herrera, Deep learn-
ing hyper-parameter tuning for sentiment analysis in twitter based on evolutionary algorithms, in: 2019 IEEE
Federated Conference on Computer Science and Information Systems (FedCSIS), IEEE, 2019, pp. 255–264.
[68] N. Wang, X. Wang, N. Cui, Y. Li, B. Liu, Deep reinforcement learning-based impact time control guidance law
with constraints on the field-of-view, Aerosp. Sci. Technol. 128 (2022) 107765.
3966

1 s2.0 S0016003223001084 Main

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

1 s2.0 S0016003223001084 Main

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S0016003223001084 Main

Uploaded by

Copyright:

Available Formats

Available online at www.sciencedirect.

Journal of the Franklin Institute 360 (2023) 3914–3966

Three-dimensional event-triggered fixed-time

Recently, the penetration difficulty of a target by utilizing traditional single missile is

2. Preliminaries and problem formulation of cooperative attack

2.1. Graph theory

where x ∈ Rn and f : Rn → Rn is a nonlinear function such that f (0) = 0, i.e., x = 0 is an

Lemma 4. [59]: For any real variable y, we have

where μ 1 and ζ > 0 whose minimum value ζ ∗ satisfies ζ ∗ = x ∗ (1 − tanh x ∗ ), and x ∗

2.3. Problem formulation of cooperative guidance in 3-D environment

ri q̈εi + 2r˙i q˙εi + ri q˙βi

−2r˙i q˙εi − ri q˙βi

3. Design of event-triggered cooperative guidance laws

As shown in Fig. 2, the guidance command of traditional time-triggered guidance is updated

Fig. 2. Diagram of the time-triggered guidance scheme.

Fig. 3. Diagram of the event-triggered guidance scheme.

3.1. Design of guidance law for simultaneous attack

Next, the tracking error of the radial relative velocity is defined as

Fig. 5. Communication topology of the multi-missile system.

Finally, the triggering mechanism is designed as

where f¯r is non-negative constants.

Fig. 6. Trajectories of multiple missiles that attack a maneuvering target.

Then the following Lyapunov candidate function is considered,

Fig. 8. Results under the proposed event-triggered guidance laws.

and the bound of convergence time Tr1 is provided by

where Tr2 is provided by

ri t  = ηr αr3 e pr /gr t i  + ηr αr4 ≤ ϕ t i

aMεi (t ) = −2r˙i (t )q˙εi (t ) − ri (t )q˙βi

Fig. 10. Comparisons of triggering numbers. (TT: Time-triggered; ET: Event-triggered.).

where 0 < ηε < 1, and the triggering error εi (t ) is defined as

where f¯ε and f¯β are non-negative constants.

Fig. 11. Triggering intervals of guidance command aMri (t ).

Consider the following Lyapunov candidate function

≤ |evεi (t )| ηε αε3 evεi (t ) + ηε α4 − αε3 evεigε (t )

gε −pε pε +gε   N αε4 ζε

and the bound of convergence time Tε1 is provided by

Fig. 13. Results under the cooperative guidance law in [22].

According to the definition of i (t ), it can be obtained that

 j=1  j=1  j=1 

Fig. 15. Results under the cooperative guidance law in [48].

εi t  = ηε αε3 e pε /gε t i  + ηε αε4 ≤ φ tki , tkj ds (63)

Fig. 16. Energy consumption under different guidance schemes.

aM βi (t ) = 2r˙i (t )q˙βi (t ) cos qεi (t ) − 2ri (t )q˙εi (t )q˙βi (t ) sin qεi (t )

where i (t ) is the consensus error of angular rate, i.e.,

where 0 < ηβ < 1, and the triggering error βi (t ) is

4. Simulations and comparative results

4.1. Initial conditions

Direction of LOS M1 -M4 M2 -M1 M3 -M2 M4 -M3

The communication topology of the multi-missile team is illustrated in Fig. 5, whose

4.2. Simulation results

Fig. 20. Convergence time of vr under different value of ηr .

Statistics Missile 1 Missile 2 Missile 3 Missile 4

Fig. 21. Guidance performance of aMε under different qε and qβ .

Missile1 Missile 2 Missile 3 Missile 4

Statistics aMε aMβ aMε aMβ aMε aMβ aMε aMβ

4.3. Comparative study

Fig. 22. Guidance performance of aMβ under different qε and qβ .

Proposed cooperative Fixed-time cooperative Finite-time cooperative Argument Proportional

ri t = ηr αr3 e pr /gr t i + ηr αr4 ≤ ϕ t i

where 0 < ηε < 1, and the triggering error εi (t ) is defined as

≤ |evεi (t )| ηε αε3 evεi (t ) + ηε α4 − αε3 evεigε (t )

gε −pε pε +gε N αε4 ζε

According to the definition of i (t ), it can be obtained that

j=1 j=1 j=1

εi t = ηε αε3 e pε /gε t i + ηε αε4 ≤ φ tki , tkj ds (63)

where i (t ) is the consensus error of angular rate, i.e.,

where 0 < ηβ < 1, and the triggering error βi (t ) is