machines
Article
GPS Path Tracking Control of Military Unmanned Vehicle
Based on Preview Variable Universe Fuzzy Sliding
Mode Control
Houzhong Zhang 1 , Xiangtian Yang 1 , Jiasheng Liang 1, *, Xing Xu 1 and Xiaoqiang Sun 1,2






Citation: Zhang, H.; Yang, X.; Liang,
J.; Xu, X.; Sun, X. GPS Path Tracking
Control of Military Unmanned
Vehicle Based on Preview Variable
Universe Fuzzy Sliding Mode
Control. Machines 2021, 9, 304.
https://doi.org/10.3390/
machines9120304
Academic Editor: Dario Richiedei
Received: 25 September 2021
Accepted: 12 November 2021
Published: 23 November 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Machines 2021, 9, 304. https://doi.org/10.3390/machines9120304
1 Automobile Engineering Research Institute, Jiangsu University, Zhenjiang 212013, China;
1000004025@ujs.edu.cn (H.Z.); 2221904110@stmail.ujs.edu.cn (X.Y.); xuxing@ujs.edu.cn (X.X.);
sxq@ujs.edu.cn (X.S.)
2 Yangzhou Aoterui Automobile Electronic Technology Co., Ltd., Yangzhou 225008, China
* Correspondence: 2211804024@stmail.ujs.edu.cn
Abstract: In the process of the continuous development and improvement of modern military
systems, military unmanned vehicles play an important role in field reconnaissance and strategic
deployment. In this paper, the precise tracking algorithm of a military unmanned vehicle, based
on GPS navigation, is studied. Firstly, the optimal preview point is obtained according to the data
points of a differential GPS signal. Secondly, the pure tracking algorithm is used to calculate the
demand steering angle, and a variable universe fuzzy sliding mode controller is designed to control
the lateral motion of the vehicle, which is verified by the joint simulation platform of Simulink
and CarSim, under multiple working conditions. Finally, the actual vehicle is verified by using the
Autobox platform. The results show that the lateral motion control of path tracking designed in
this paper can achieve an accurate and effective control effect, and has real-time performance for
engineering applications.
Keywords: global positioning system; unmanned vehicle; path tracking control; lateral stability;
variable universe fuzzy sliding mode control
1. Introduction
In recent years, with the continuous evolution of modern military war, the demand
for intelligent and unmanned military vehicles, such as military reconnaissance vehicles
and unmanned combat vehicles, is increasing. China is also paying increasing attention
to the unmanned technology research of military vehicles, to ensure a high scientific and
technological level, and modern combat strength of its army [1,2]. Among them, the
unmanned reconnaissance off-road vehicle for field operations is an important military
vehicle, which can complete path exploration and tactical reconnaissance tasks in dangerous
field environments (e.g., swamps, minefields, etc.). A military unmanned vehicle realizes
autonomous path tracking mainly by using GPS coordinates and its own GPS positioning
as the input of the desired path. Based on the above background, this paper carries out
GPS path tracking control of a military unmanned vehicle. Research on the path following
control of an unmanned vehicle can be defined as work in which the vehicle is regarded as
a point (the control point) to track a geometric curve (expected path), which has nothing to
do with time, and only represents the position information, that is, the control point can
continuously track the target path preview point on the desired path at a given speed [3].
Usually, a lateral motion controller is designed on the assumption that the longitudinal
speed is constant, so that the lateral displacement error between the control point and the
desired path converges gradually to zero. Therefore, the control problem of path tracking
can be divided into the following two parts: preview point selection and lateral motion
control. The pilot’s preview model was first proposed by the academician Guo Konghui.
The pilot’s preview model was established to simulate the driver’s consciousness, and,
https://www.mdpi.com/journal/machines
Machines 2021, 9, 304 2 of 19

finally, an in-loop controller considering the driver’s input was formed. At present, the
calculation of preview points is mainly divided into single-point preview and multipoint
preview. Ding Nenggen and other scholars established a vehicle driver model based on
the optimal curvature of the single-point preview, and designed a method to obtain the
optimal curvature after selecting a preview point according to the vehicle speed, to find the
optimal path to approach the preview point [4]. Based on the traditional optimal preview
driver model, Dong Ting modified the steering wheel angle by considering the acceleration
feedback link, to obtain the improved preview driver model [5]. Based on the single-point
preview method, Ren Dianbo and other scholars established the dynamic models of vehicle
lateral position error and yaw angle error [6]. Most studies, including the above studies,
use the method based on vehicle speed and preview distance to calculate the preview point,
but do not explain, in detail, how to find the reasonable preview point on the specified
path, but only use the preview point for subsequent motion control. Single-point preview
has certain limitations. When the current vehicle attitude angle deviation is too large, or the
path curvature is too large, it is difficult to find the appropriate preview point. At present,
there is little research on multi-point preview. Zhao Kai et al. selected multiple preview
points with equal intervals, after selecting the first one. Xu et al. [7], of the University of
Michigan, proposed a path tracking algorithm based on the optimal preview control, which
introduced the multipoint preview road curvature within the finite time window into the
augmented state vector, and reconstructed the nonlinear optimal control problem into the
augmented LQR problem, but did not introduce the calculation of the preview point in
detail. There are three kinds of lateral control schemes for path following, which are the
geometry model, kinematics model, and dynamics model. The geometric model depends
on the relationship between vehicle attitude and position, and Ackermann steering. The
kinematic model further considers the motion equation of the vehicle, while it does not
involve the physical properties of the vehicle and the force on the vehicle. In a word, the
path following control method based on the geometric and/or kinematic model ignores the
dynamic characteristics of the vehicle system, and its application range is limited. Therefore,
to obtain a more accurate tracking control effect, especially under high-speed and high
curvature conditions, it is necessary to consider the vehicle system dynamics characteristics
in the path tracking control design. At present, the algorithms based on the dynamic model
mainly include PID control [8–10], fuzzy control [11,12], sliding mode control [12–16],
and model predictive control [17–23]. All kinds of algorithms have unique advantages
and are widely used, especially model predictive control. However, model predictive
control belongs to optimal control, and its real-time performance is inevitably reduced.
Although sliding mode control does not need to establish an accurate dynamic model,
and uses the nonlinear theory to deal with multidimensional complex coupling problems
effectively, the chattering phenomenon greatly hinders its further application. Aiming
at the shortcoming of the buffeting phenomenon, many scholars use neural networks to
adaptively approximate the uncertainty and optimize it [16,24]. However, many deep
learning algorithms, including neural networks, need a large number of accurate and
reasonable training data, and the acquisition of data sets depends on other algorithms. The
scholars Chen and Zhang et al. combined fuzzy control with sliding mode control and
used fuzzy rules to adjust the gain coefficient of the switching term, and achieved a good
control effect [25,26].
To sum up, after considering the characteristics of single-point preview and multipoint
preview, this paper proposes a single-point preview algorithm that can be applied to a large
curvature path, which can guide the vehicle to track the curve with large curvature, under
the premise of ensuring the real-time advantage of single-point preview. At the same time,
on the basis of fuzzy sliding mode control (FSMC), the research on variable universe FSMC
is carried out to achieve lateral control with faster control and more accurate responses,
under the premise of eliminating the chattering phenomenon. Finally, the effectiveness of
the proposed algorithm is verified by the joint simulation platform and real vehicle test.
The layout of Preview FSMC system are shown as Figure 1.
Machines 2021, 9, x FOR PEER REVIEW 3 of 20

Machines 2021, 9, 304 3 of 19

effectiveness of the proposed algorithm is verified by the joint simulation platform and
real vehicle test. The layout of Preview FSMC system are shown as Figure 1.

δf
Preview point γ
Maximum distance point αf
vy
δ
β

Closest αr 0 vx
Fxr
point Fyr Fxf
Fyf

lr lf
Target
path Vehicle l

δd

Figure
Figure 1. Preview
1. Preview FSMC
FSMC system
system layout.
layout.

2. Acquisition of Path and Preview Information

2. Acquisition of Path and Preview Information
For vehicles that rely on GPS to complete autonomous tracking, it is necessary to
For vehicles that rely on GPS to complete autonomous tracking, it is necessary to
determine the path information of human planning first, and then find the appropriate
determine the path information of human planning first, and then find the appropriate
preview point on the path in front of the vehicle, according to the path information of the
preview point on the path in front of the vehicle, according to the path information of the
moving destination of the vehicle at the next moment. Generally speaking, the man-made
moving destination of the vehicle at the next moment. Generally speaking, the man-made
path information is composed of multiple GPS discrete points. It is necessary to convert
path information is composed of multiple GPS discrete points. It is necessary to convert
the longitude and latitude information into the path information in the vehicle coordinate
the longitude and latitude information into the path information in the vehicle coordinate
system, and then obtain the preview information.
system, and then obtain the preview information.
2.1. GPS Information Processing
3. GPS Information Processing
Whether it is manually input or collected in advance, the final actual path information
is aWhether
series of itGPS is manually
longitude inputand latitudeor collected
coordinate in advance, the final actual
points. Therefore, path infor-
these longitude and
mation
latitude is acoordinates
series of GPS longitude
(lon, lat) need and to latitude
be converted coordinate points.
first, and thenTherefore,
the discrete these longi-
coordinate
tude and(X,
points latitude
Y) in the coordinates (lon, lat) need
geodetic coordinate systemto be areconverted
obtained.first, and then the discrete
coordinate2 points (X,Y) in the geodetic coordinate system are obtained.
2 + 9η 2 + 4η 4 ) cos2 ( lat ) l 2 + 1 (61 − 58t2 + 4t4 ) cos4 ( lat ) l 4 ]

 X = ζ + Nt cos2 lat) l 2 [0.5 + 24 1
( 5 − t 2
p p 720 p4
 l2 1 l2 1 l4 (1)
X = ζ + Nt cos 2
( lat )
 Y = N cos(lat) l [1 + 1 (1 − t2 + ηp2 2) cos2 (24 [0.5 + (5 − 2t 2
+ 9η 2
+ 4η 4
) cos 2
( lat ) + (61 − 58 t 2
+ 4 t 4
) cos 4
( lat ) 4 ]
p 6 lat) pl 2 + 120 1
(5 − 18t2 + t4p 2+ 14η 720 2 − 58η 2 t2 ) cos4 ( lat )pl44 ]
p
l 1 l2 1 l4
(1)
Y = N cos( lat ) [1 + (1 − t 2 + η 2 ) cos 2 ( lat ) + (5 − 18t 2 + t 4 + 14η 2 − 58η 2 t 2 ) cos 4 ( lat ) 4 ]
 p 26 2 2
where η 2 = e cos (lat), t = tan(plat)120 , and N is the radius of curvature of
prime unitary p
the
circle; the value of ζ can be obtained by the following:
where η 2 = e 2 c o s 2 ( la t ) , t = tan( lat ) , and N is the radius of curvature of the prime unitary 
circle;
ξ = 111133.0047lat− sin (lat)the
cosvalue
(lat)(of
32009 ζ can be obtained
.8575 + 133.9602 the2 (following:
bysin lat) + 0.6976 sin4 (lat) + 0.0039 sin6 (lat) (2)
ξ = 1 1 1 1 3 3 .0 0 4 7 la t -s in ( l a t ) c o s ( la t ) ( 3 2 0 0 9 .8 5 7 5 + 1 3 3 .9 6 0 2 s in 2 ( l a t ) + 0 .6 9 7 6 s in 4 ( l a t ) + 0 .0 0 3 9 s i n 6 ( la t ) )
(2)
Through the Gauss projection coordinate system, the positive direction of the X-axis
Through the
corresponds Gauss
to the projection
north, and thecoordinate system,ofthe
positive direction thepositive direction of the
Y-axis corresponds X-axis
to the south.
corresponds to the north,
Then, the geodetic and the system
coordinate positiveisdirection of theinto
transformed Y-axis
thecorresponds to the south.
vehicle coordinate system
(x, y).
Then, the geodetic coordinate system  is transformed into the vehicle coordinate system
(x, y). x = Y sin θ + X cos θ
(3)
y = Y cos θ − X sin θ
 x = Y sin θ + X cos θ
The vehicle heading angle yis=θ. The
Y cos θ −preview
X sin θ point can be selected after coordinate
(3)

transformation.
The vehicle heading angle is θ. The preview point can be selected after coordinate
2.2. Calculation of Preview Point
transformation.
In this paper, a single-point preview method, considering curvature, is used to calcu-
4. Calculation of Preview
late the preview point, asPoint
shown in Figure 2.
(1)In According
this paper,to
a the
single-point previewofmethod,
GPS information the path,considering curvature,
the nearest point of theisvehicle
used to cal-
position
culate the preview
at the point,
current as shown
moment in Figure 2.after coordinate transformation. According
is determined
to Formula (1), the following formula is obtained after simplification and operation
processing, and the nearest point is found by traversing the following formula:
 Machines 2021, 9, 304 4 of 19


DisY = (lat p − latv ) ∗ 111000
(4)
DisX = lon p − lonv ) ∗ 111000 ∗ cos(latv ∗ 0.01745)

Machines 2021, 9, x FOR PEER REVIEW

where latp and lonp are the latitude and longitude coordinates of the path points; 4latofv and
20
lonv are the latitude
√ and longitude coordinates of the self-driving vehicle. After obtaining
DisX and DisY, DisX 2 + DisY 2 is used to calculate the distance between the discrete
points on the expected path and the self-workshop.

Figure
Figure 2. Schematic
2. Schematic diagram
diagram of of preview
preview point
point screening.
screening.

(1)(2) According
The preview to the distance is determined
GPS information according
of the tonearest
path, the the following
point offormula:
the vehicle posi-
tion at the current moment is determined after coordinate transformation. According
(
k ∗ Vt, k ∗ Vt > ld_min
to Formula (1), the following ld =
formula is obtained after simplification and operation(5)
l , k ∗ Vt ≤ ld_min
processing, and the nearest point isd_min found by traversing the following formula:
Vt is the vehicle speed, k is the proportional coefficient, and ldmin is the minimum
preview distance.  DisY =(lat p − latv ) *111000
 (4)
 DisX = (lon
(3) Find the coordinates p − lon
of the v )*111000*cos(
nearest latv *0.01745)
point according to the method described in step 1,
calculate the distance between the point on the path and the vehicle from the nearest
where latp and lonp are the latitude and longitude coordinates of the path points; latv and
point, take the path point closest to the preview distance as the preview point, return
lonv are the latitude and longitude coordinates of the self-driving vehicle. After obtaining
the coordinates of that point, and then obtain the pre-selected preview point.
DisX
(4) andAfterDisY, D i s X 2 + D i s Y 2 is used to calculate the distance between the discrete
the preview point is determined, the line between the current vehicle position
points on (Cxr , Cxr ) and thepath
the expected and the
preview self-workshop.
point is calculated.
(2)(5) The preview distance is determined
Traverse all the points in the nearest according to the
point and following
preview point,formula:
calculate the distance
between these points and the line in step 4, and then determine the maximum dis-
𝑘 ∗𝑉， 𝑘 ∗𝑉 > 𝑙 _
tance point. 𝑙 = (5)
𝑙 _ , 𝑘∗𝑉 ≤𝑙 _
(6) Determine whether the distance maxdi between the maximum distance point and the
line
Vt is theinvehicle
step 4 exceeds
speed, kthe is thethreshold,
proportional and then judge whether
coefficient, and ldminthe preselected
is the minimum preview
pre-
point
view distance. is reasonable. The curvature value does not need to be calculated specifically,
(3) Find and the
the curvature of
coordinates canthe benearest
reflected by judging
point accordingthetodistance between
the method the maximum
described in step
distance point
1, calculate and thebetween
the distance line in step 1.
the point on the path and the vehicle from the near-
(7) estIfpoint,
step 3takedetermines
the path that pointthe threshold
closest to thevalue is exceeded,
preview distance asthethecurrent maximum
preview point,
distance point is selected as the preselected preview point, and steps 1–3 are repeated
return the coordinates of that point, and then obtain the pre-selected preview point.
until the requirements are met, and the preselected preview point is output as the
(4) After the preview
preview point. point is determined, the line between the current vehicle position
(Cxr, Cxr) and the preview point is calculated.
The pseudo code of one of the above procedures is shown in the Figure 3:
(5) Traverse all the points in the nearest point and preview point, calculate the distance
The minimum value of x and its location is represented by fmin (x), ffit ((x1,y1),(x2, y2))
between these points and the line in step 4, and then determine the maximum dis-
is used to fit a straight line according to two points, and fdis is the distance from the
tance point.
point to the line. The final output (x(maxdisIndex), y(maxdisIndex)) is the preview point
(6)coordinates
Determine whether
after the distance
the above processmaxd between the
cycle ijudgment maximum
is repeated distance
many point and the
times.
line in step 4 exceeds the threshold, and then judge whether the preselected preview
point is reasonable. The curvature value does not need to be calculated specifically,
and the curvature can be reflected by judging the distance between the maximum
distance point and the line in step 1.
(7) If step 3 determines that the threshold value is exceeded, the current maximum dis-
tance point is selected as the preselected preview point, and steps 1–3 are repeated
until the requirements are met, and the preselected preview point is output as the
preview point.
Machines 2021, 9, 304 5 of 19
Machines 2021, 9, x FOR PEER REVIEW 5 of 20

start

Transformation of longitude and latitude coordinates into

discrete point coordinates in geodetic coordinate system

Traverse to find the nearest point

Preliminarily determine the preview distance according

to the vehicle speed

Starting from the nearest point, calculate the distance l1

between the point on the path and the workshop in turn,
Reduce preview distance
and take the path point closest to the preview distance as
the primary preview point

Determine the maximum distance point and calculate the Is the curvature too
curvature large？
N

Calculate the connection between the self vehicle

position and the preview point

Traverse all points in the nearest point and preview

point, and calculate the distance l2 between these points
and the connecting line

Select the maximum distance point

MAX(Distance)<Threshold？
as the preview point N

Determine final preview point

Figure
Figure 3.
3. Judgment
Judgment and
and selection
selection of
of preview
preview points.
points.

The minimum value of x and its location is represented by fmin(x), ffit((x1,y1),(x2, y2))
3. used
is Lateral Motion
to fit Control
a straight line Based
accordingon Variable Universe
to two points, andFuzzy Sliding
fdis is the Mode
distance from the point
to theConsidering
line. The finaltheoutput
complex (x(maxdisIndex), y(maxdisIndex))
working environment of a military is the previewvehicle,
unmanned point coor-
and
dinates aftercoupling
the strong the above process cycle
relationship judgment
between is repeated
the steering manyand
system times.
chassis components,
sliding mode control with strong robustness adaptability, and no dependence on the
5. Lateralof
accuracy Motion Control
the model, Based
is very on Variable
suitable Universe
for the lateral Fuzzy
motion Sliding
control Mode At present,
algorithm.
the chattering
Considering phenomenon
the complexhas been solved,
working to a certain
environment extent, within
of a military which
unmanned the use
vehicle, of
and
FSMC is more common. Compared with other deep learning algorithms,
the strong coupling relationship between the steering system and chassis components, such as neural
networks,
sliding modefuzzy control
control withhas the advantage
strong robustness of adaptive approximation
adaptability, and no dependence and noonneed of
the ac-
accurate prior knowledge, which makes it easier to be widely used.
curacy of the model, is very suitable for the lateral motion control algorithm. At present,
the chattering phenomenon has been solved, to a certain extent, within which the use of
 ciple of plane geometry, the following applies:
2 ∗ 𝑅 ∗ sin 𝛼 = 𝑙 (6)
where 𝑙 represents the distance between the center point of the rear axle and the preview
Machines 2021, 9, 304 point, that is, the preview distance; 𝛼 is the angle between the UAV heading and preview
6 of 19
line, and its value can be solved by the following formula:
𝐺 −𝐶
𝛼 = arcsin −𝜑 (7)
3.1. Calculation of Expected Steering Angle Based 𝑙on Pure Tracking Algorithm
𝜑 is the
whereWhether it isheading
manually input
angle of or
thecollected
unmanned in advance,
vehicle. the final actual path information
is a According
series of GPS longitude andsteering
to Ackermann’s latitude principle,
coordinatethepoints. Therefore,
following these longitude and
is true:
latitude coordinates (lon, lat) need to be converted first, and then the discrete coordinate
tan 𝛿 = 𝐿⁄𝑅 (8)
points (X, Y) in the geodetic coordinate system are obtained.
whereSpecifically,
L representsthe thecoordinate
wheelbase ofof
the center
the point ofvehicle.
unmanned the rear axle of the unmanned vehicle
is set as ( C ,
By substituting
xr C yr ) in
Formulas (5) and
(6) into Equationand
the geodetic coordinate system, (7), the coordinates
we obtain of the preview
the following:
point (i.e., target point) are Gxr , Gyr , The front wheel angle of the unmanned vehicle is
tan 𝛿is= ⁄𝑙
δ f , and the current turning radius R.2As
∗ 𝐿shown
∗ sin 𝛼in Figure 4, according to the principle (9)of
plane Thegeometry,
control law the of
following
the pureapplies:
tracking controller can be obtained by transforming For-
mula (8), as follows:
2 ∗ R ∗ sin α = ld (6)
𝛿 = arctan(2 ∗ 𝐿 ∗ sin 𝛼⁄𝑙 ) (10)
where ld represents the distance between the center point of the rear axle and the preview
𝛿 is the input 𝛿 of the motion control system. In particular, for the pure tracking
point, that is, the preview distance; α is the angle between the UAV heading and preview
controller,
line, and its thevalue
selection
can beofsolved
preview distance
by the 𝑙 isformula:
following particularly important to the perfor-
mance of the controller. In practical applications, the preview distance needs to be ad-
justed by a large number of tests. The basicGcalculation yr − Cyr formula is shown in the formula
 
α = arcsin − ϕ pv (7)
of preview distance (Formula (5)) in Chapter 1 of ld this paper. After obtaining the expected
angle by the pure tracking algorithm, the actual vehicle rotation angle is obtained by using
where
the ϕ pv is
variable the heading
universe FSMC angle of the unmanned vehicle.
algorithm.

Target δf
Path (Gxr , G yr )

ld L
R
α
2α
R (Cxr , C yr )

Figure 4. Schematic diagram of pure tracking algorithm.

According to Ackermann’s steering principle, the following is true:

tan δ f = L/R (8)

where L represents the wheelbase of the unmanned vehicle.

By substituting Formulas (5) and (6) into Equation (7), we obtain the following:

tan δ f = 2 ∗ L ∗ sin α/ld (9)

The control law of the pure tracking controller can be obtained by transforming
Formula (8), as follows:
δ f = arctan(2 ∗ L ∗ sin α/ld ) (10)
δ f is the input δ of the motion control system. In particular, for the pure tracking
controller, the selection of preview distance ld is particularly important to the performance
of the controller. In practical applications, the preview distance needs to be adjusted by a
large number of tests. The basic calculation formula is shown in the formula of preview
distance (Formula (5)) in Section 1 of this paper. After obtaining the expected angle by the
pure tracking algorithm, the actual vehicle rotation angle is obtained by using the variable
universe FSMC algorithm.
Machines 2021, 9, 304 7 of 19

3.2. Design of Sliding Mode Controller

Assuming that the front wheel of the vehicle turns by a small angle, the vehicle
dynamic equation can be described as follows:
 .
mvy + mv x β = Fy f + Fyr
. (11)
Iz ω = aFy f − bFyr

In Equation (11), v x represents the longitudinal speed of the vehicle, vy is the lateral
speed, m is the mass of the whole vehicle, ω represents the yaw rate of the vehicle, lf and
lr are the distance from the front and rear wheels to the vehicle centroid, respectively, Fx f
and Fyr represent the lateral forces of the front and rear wheels, respectively, β is the side
slip angle of the mass center, and Iz is the moment of inertia to the center of mass of the
vehicle. When the tire is in a linear region, the relationship between the lateral force and
the tire can be approximately expressed as follows:

Fy f ≈ α f ∗ C f
(12)
Fyr ≈ αr ∗ Cr

Cf and Cr are the lateral stiffness of the front and rear wheels, respectively. The side
slip angle of the front and rear wheels of the vehicle can be expressed as follows:
v
(
α f ≈ vyx + aω
vx − δ
vy bω (13)
αr ≈ v x − v x

In the case of a small center of mass, the steering angle of the front wheel can be
approximately expressed as follows: β = v x /u. The 2-DOF vehicle model equation about
yaw rate ω and β can be obtained by Formulas (5)–(7) simultaneously, as follows:
2 2

 ω. = l f C f +lr Cr ω + l f C f −lr Cr β − l f C f δ
Iz v x Iz Iz
. (14)
 β = ( l f C f −lr Cr − 1)ω + C f +Cr β − C f δ
Mv2x Mv x Mv x

As shown in Figure 5, for an uncertain system, when the system is subject to both
internal parameter perturbation and external disturbances, the system state equation can
be expressed as follows:
.
x = Ax + B[u(t) + E(t)] (15)
Machines 2021, 9, x FOR PEER REVIEW 8 of 20
where A and B are parameter perturbations within the system, x is the state quantity, u is the
control quantity, and E(t) includes the uncertainty and external disturbance of the system.

Twodegree
Figure5.5.Two
Figure degreeof
offreedom
freedomdynamic
dynamicmodel
modelof
ofvehicle.
vehicle.

According to Formula (8), the following can be concluded:

ω a11 a12  b11 

x =   , A=  , B = b  (16)
β  a21 a22   21 
l C −l C C +C
lf Cf Cf
Machines 2021, 9, 304 8 of 19

According to Formula (8), the following can be concluded:

     
ω a11 a12 b11
x= , A= , B= (16)
β a21 a22 b21

l f 2 C f +lr 2 Cr l f C f −lr Cr l f C f −lr Cr C f +Cr lf Cf

where a11 = Iz v , a21 = Mv2x
− 1, a12 = Iz , a22 = Mv x , b11 = Iz ,
Cf
b21 = Mv x .
Therefore, the state equation of the controlled system studied in this paper can be
expressed as follows:
.
x = f (ωr , β) + b11 [u(t) + E(t)] (17)
where f (ωr ) can be expressed by the side slip angle of the center of mass and the yaw rate,
as follows:
f (ωr , β) = a11 ωr + a12 β (18)
The gain of reaching law of sliding mode control is selected as follows:

K (t) = max(| E(t)|) + η (19)

where η > 0.
Because the side slip angle of the center of mass cannot be directly controlled, and the
ideal value and the actual value of the side slip angle of the center of mass cannot be directly
controlled, the ideal value of the side slip angle of the center of mass will have a large error
when the vehicle is unstable. Moreover, the yaw rate can reflect the overall movement
trend of the vehicle better than the wheel angle. Therefore, the difference between the
actual yaw rate and the ideal yaw rate is selected as the tracking error of the system.

e = ωr − ω d (20)

The ideal yaw rate ωd in Equation (19) is calculated according to the expected angular
value and the 2-DOF model; ωr is the actual yaw rate. The switching function of the
controller is selected using the following equation:

Zt
s = e+λ e(t)dt (21)
0

By deriving the switching function, we can obtain the following results:

. . . .
s = e + λe = ω r − ω d + λ(ωr − ωd ) (22)

The sliding mode control law is designed as follows:

1 .
u= [− f (ωr ) + ω d + λ(ωr − ωd ) + K (t)sgn(s)] (23)
b11

The Lyapunov theorem is used to prove the stability of the designed lateral sliding
mode controller, as follows:
1
V = s2 (24)
2
The derivation of Equation (23) shows the following:
. . . .
V = ss = s[e + λe] = s[ f (ωr ) + b11 [u(t) + E(t)] + ω d + λe] (25)
Machines 2021, 9, 304 9 of 19

The sliding mode control law of Equation (22) is substituted into Equation (24) to
obtain the following:
.
V = s[−K (t)sgn(s) − E(t)] = −K (t)|s| − E(t)s ≤ −η |s| (26)

where η > 0, the designed controller meets the stability requirements.

By substituting f (ωr ) into Equation (21), the expression composed of parameters of
the vehicle dynamic model can be obtained as follows:

. l f 2 C f + lr 2 Cr lf Cf .
s= ωr − δ − ω d + λ ( ωr − ω d ) (27)
Iz v Iz
. .
Let s = 0. According to the hypothesis, we can know the yaw acceleration ω d = 0 at
this time. Therefore, the control input front wheel angle can be expressed as follows:

Iz l f 2 C f + lr 2 Cr .
δ= [ ωr − ω d + λ0 (ωr − ωd )] (28)
lf Cf Iz v

By substituting Formula (22), we obtain the following:

δd = δ + K (t)sgn(s) (29)

K (t) is the rate of the system moving point approaching the switching surface; when
K (t) is larger, the speed of the approaching moving point will be faster, and the response
speed of control will be faster, but it will also cause greater jitter. Therefore, the jitter of the
sliding mode control is mainly caused by the reaching term K (t)sgn(s); in order to ensure
the control effect of the sliding mode controller, it is necessary to further optimize it.

3.3. Design of Fuzzy Sliding Mode Controller

According to Equation (28), it can be observed that the existence of the symbolic func-
tion can effectively eliminate the unknown interference, but it inevitably leads to chattering.
Therefore, this paper uses adaptive fuzzy control, based on switching fuzzification, to
improve the trajectory of the system near the sliding surface, which not only enables the
system to adapt to various working conditions, but also enables good elimination of differ-
ent interference intensities. By using fuzzy approximation, the discrete symbolic function
can be continuous, which can effectively reduce the chattering phenomenon. Therefore,
a variable switching gain is designed in this paper. Fuzzy rules are used to adjust the
switching gain according to the relative position and movement trend of the system and
sliding surface.
.
Taking s and s as the fuzzy input, and k as the output, the corresponding fuzzy
.
linguistic variables are {NB NM NS ZO PS PM PB}. If ss > 0, the current state of the sliding
mode function is the same as the change trend, and there is a trend far away from the
.
sliding surface, the switching gain k(t) should be increased at this time. If ss < 0, then
the state of the sliding mode function is opposite to the change trend, and the system is
approaching the sliding surface, so the switching gain k should be reduced. At the same
.
time, the size of |ss| also needs to be considered, so as to further design fuzzy rules
.
reasonably. When ss is large, |k| should also have a larger change; otherwise, the change
should be reduced.
In the design of the fuzzy controller, the design of fuzzy rules is an important part to
determine its performance (Table 1).
Machines 2021, 9, 304 10 of 19

Table 1. Fuzzy rule table.

.
s
NB NM NS ZO PS PM PB
s k
NB NB NB NM ZO PM PB PB
NM NB NM NS ZO PS PM PB
NS NM NS NS ZO PS PS PM
ZO ZO ZO ZO ZO ZO ZO ZO
PS PM PS PS ZO NS NS NM
PM PB PM PS ZO NS NM NB
PB PB PB PM ZO NM NB NB

The output y(x) of the fuzzy system can be written as the product inference engine,
single-valued fuzzy controller, and center average defuzzer, as follows:
m n
∑ y j ( ∏ µ Aj (x ) )
j =1 i =1 i i
y( x ) = m n (30)
∑ ( ∏ µ Aj (x ) )
j =1 i =1 i i

In this section, the switching function s(t) is used as the input of the fuzzy system. In
j
this section, the switching function is used as the output of the fuzzy system, where Ai is
its fuzzy set {NB NM NS ZO PS PM PB}, µ j is the membership function of si , where
Ai ( si )
µ NB adopts the Z-type membership function, µ PB adopts the S-type membership function,
and the others adopt the triangle membership function; the membership function is shown
in the following Figures 6–8.

3.4. Design of Variable Universe Fuzzy Sliding Mode Controller

According to the principle of fuzzy control, if the fuzzy system is to approach a
changing function infinitely accurately, it needs infinite fuzzy rules to realize in theory,
which is almost impossible to complete. Therefore, under the premise that the number of
Machines 2021, 9, x FOR PEER REVIEW 11 of 20
rules and membership functions remain unchanged, further accurate approximation can
be achieved by adaptive adjustment of the universe.

NB NM NS ZO PS PM PB
1

0.8

0.6

0.4

0.2

-1 -0.5 0 0.5 1
s

Figure 6.
Figure 6. S membership
membership function
function diagram.
diagram.

NB NM NS ZO PS PM PB
1

0.8

0.6

0.4
 0.2

-1 -0.5 0 0.5 1
Machines 2021, 9, 304 s 11 of 19

Figure 6. S membership function diagram.

NB NM NS ZO PS PM PB
1

0.8

0.6

0.4

0.2

-1.5 -1 -0.5 0 0.5 1 1.5

ds

Figure
Figure 7.
7. dS
dS membership
membership function
function diagram.
diagram.

NB NM NS Z PS PM PB
1

0.8

0.6

0.4

0.2

-1.5 -1.125 -0.75 -0.375 0 0.375 0.75 1.125 1.5

K

Figure 8.
Figure K membership
8. K membership function
function diagram.
diagram.

In this
9. Design ofpaper, theUniverse
Variable exponentialFuzzyexpansion
Slidingfactor
Modein the function model is selected to
Controller
adjust the fuzzy universe of three variables adaptively.
According to the principle of fuzzy control, if the fuzzy system is to approach a
changing function infinitely accurately, it (− needs2 infinite fuzzy rules to realize in theory,
α( x )= 1−λe 0.5∗ x ) λ ∈ (0, 1) (31)
which is almost impossible to complete. Therefore, under the premise that the number of
rules
whereandλ ismembership functions
the scale factor, remain unchanged,
which determines the degreefurther accurate
of scaling. approximation
Parameter can
x is the input,
.
be achieved by
such as s, s, or k. adaptive adjustment of the universe.
In
In this paper,the
this paper, theexpansion
exponential expansion
factor λ of the factor in the function
input universe model
is 0.6, and that is
ofselected to
the output
adjust
universetheisfuzzy universe
0.3. Since of three variables
the expansion factors ofadaptively.
the input and output domains are all reduced,
it is necessary to adjust the initial universe, designed in the previous section, so that the
(-0.5*x 2 )
α ( x)=1- λ e
product of the adjusted universe and the expansion λ ∈factor
(0,1)is centered on the initial universe.
(31)
At the same time, because the fuzzy logic module in Matlab/Simulink cannot realize the
variable universe, this paper uses the S-function module to design the variable universe
fuzzy control, which replaces the original fuzzy logic module.

4. Simulation Analysis
To preliminarily verify the trajectory tracking control algorithm designed in this paper,
based on the Simulink and CarSim joint simulation platform, the tracking algorithm is
tested in a variety of working conditions. The road centerline is used to simulate the GPS
path as the input of Simulink, and the front wheel rotation angle, obtained by algorithm
processing, is the output to CarSim. The structure is shown in the Figure 9.
This paper mainly compares PID, traditional FSMC, and variable universe FSMC. The
weakening of the chattering phenomenon is obvious, and is not the focus of this paper.
Therefore, we will focus on the comparison of the effects before and after the variable
universe. For the comparison of the weakening of the buffeting phenomenon, please refer
to [26].
 10. Simulation Analysis
To preliminarily verify the trajectory tracking control algorithm designed in this pa-
per, based on the Simulink and CarSim joint simulation platform, the tracking algorithm
is tested in a variety of working conditions. The road centerline is used to simulate the
Machines 2021, 9, 304 12 of 19
GPS path as the input of Simulink, and the front wheel rotation angle, obtained by algo-
rithm processing, is the output to CarSim. The structure is shown in the Figure 9.

Calculation of preview point Expected angle

δf Adjustment of
γ
αf
variable fields
vy
δ
β
αr 0 vx
Fxr
Fyr Fxf
Fyf

lr lf Ꞷd Fuzzy controller
l

2-DOF model _ Sliding

K
mode
+ controller
Ꞷr δd
CarSim modular
Simulation condition

Figure9.9.Overall
Figure Overallarchitecture
architecturelayout
layoutofofsimulation
simulationverification
verificationtest.
test.

4.1. Oval
ThisLane
paperTestmainly compares PID, traditional FSMC, and variable universe FSMC.
The The
weakening
ellipse of thecan
lane chattering
be usedphenomenon is obvious,
to test the rapidity and is not
(transient) andthe focus of(steady
accuracy this pa-
per. Therefore,
state) we error
of the lateral will focus on the comparison
convergence of vehicles of in the effects tracking.
trajectory before andItsafter
lanethe variable
shape and
universe. For
parameters arethe comparison
shown of the weakening
in the Figures 10 and 11. of the buffeting phenomenon, please refer
In the simulation test under this condition, the average absolute lateral error of the
to [26].
vehicle under the control of the variable universe FSMC algorithm is 0.0089 m, which is far
less than the
11. Oval Lane0.2025
Testm of PID and 0.1499 m of FSMC. Compared with the FSMC algorithm,
the PID algorithm cancan
The ellipse lane stabilize rapidly
be used to testafter
the arapidity
long period of oscillation,
(transient) and, finally,
and accuracy (steadykeep a
state)
lateral deviation of about 0.5 m. Although the FSMC algorithm can be effectively
of the lateral error convergence of vehicles in trajectory tracking. Its lane shape and pa- adjusted,
the controlare
rameters accuracy
shown is inobviously
the Figures not
10as
and good
11. as that of the variable universe FSMC, and
Machines 2021, 9, x FOR PEER REVIEW
the final error is about 0.2 m. Therefore, both the fast convergence performance in transient 13 of 20
state and the adjustment accuracy in steady state are better than the other two algorithms.

Figure 10.
Figure 10. Oval
Oval lane.
lane.

1.5
FSMC
PID
1 Variable fields FSMC

0.5
ation/m

0
Machines 2021, 9, 304 13 of 19

Figure 11. Comparison of simulation results under elliptical lane.

4.2. FHWA Lane Test

Similarly to the double-lane change condition, the Federal Highway Administration
(FHWA) has developed a more stringent test path, as shown in the Figures 12 and 13.
In the simulation test under this condition, the average absolute lateral error under
the control of the variable universe FSMC algorithm is 0.0173 m, which is far less than
the 0.1212 m under the PID method and 0.0248 m under the FSMC algorithm. From the
perspective of the control process, the variable universe FSMC can converge to zero error
faster because of the continuous adjustment of the universe, and can be closer to zero error
than FSMC in most cases, and the adjustment is more accurate.

4.3. “8” Lane Test

Machines 2021, 9, x FOR PEER REVIEW 14 of 20
2 the following performance of vehicles under continuous
The “8” lane test can assess
and large-scale curves. The structure of the “8” lane test is shown in Figures 14 and 15.

Figure 12. FHWA lane.

Figure 12. FHWA lane.

0.6
FSMC
PID
0.4 Variable fields FSMC
teral deviation/m

0.2

0
 Machines 2021, 9, 304 14 of 19

Figure 12. FHWA lane.

0.6
FSMC
PID
0.4 Variable fields FSMC
Lateral deviation/m

0.2

-0.2

-0.4

-0.6
0 1 2 3 4 5 6 7 8 9 10
021, 9, x FOR PEER REVIEW Time/s 15 of 20104

Figure13.
Figure Comparisonof
13.Comparison ofsimulation
simulationresults
results under
under FHWA
FHWA lane.
lane.

In the simulation test under this condition, the average absolute lateral error under
the control of the variable universe FSMC algorithm is 0.0173 m, which is far less than the
0.1212 m under the PID method and 0.0248 m under the FSMC algorithm. From the per-
spective of the control process, the variable universe FSMC can converge to zero error
faster because of the continuous adjustment of the universe, and can be closer to zero error
than FSMC in most cases, and the adjustment is more accurate.

13. “8” Lane Test

The “8” lane test can assess the following performance of vehicles under continuous
and large-scale curves. The structure of the “8” lane test is shown in Figures 14 and 15.

Figure 14. “8” lane.

Figure 14. “8” lane.

1.5
FSMC
PID
1 Variable fields FSMC
Lateral deviation/m

0.5

-0.5

-1

-1.5
0 1 2 3 4 5 6 7 8 9 10
Time/s 104
Figure 15. Comparison of simulation results under “8” lane.
Figure 15. Comparison of simulation results under “8” lane.

In the simulation test under this condition, the average absolute lateral error under
the control of the variable universe FSMC algorithm is 0.0883 m, which is far less than the
0.4601 m under the PID method and 0.1846 m under the FSMC method. It can be observed
that when the FSMC is close to zero error, it will not be fine-tuned. This is due to the
setting of fixed fuzzy universe, which makes the input error map to the minimum fuzzy
Machines 2021, 9, 304 15 of 19

In the simulation test under this condition, the average absolute lateral error under
the control of the variable universe FSMC algorithm is 0.0883 m, which is far less than the
0.4601 m under the PID method and 0.1846 m under the FSMC method. It can be observed
that when the FSMC is close to zero error, it will not be fine-tuned. This is due to the setting
of fixed fuzzy universe, which makes the input error map to the minimum fuzzy language
variable. To avoid frequent chattering, the error is not adjusted, so that the system cannot
finally approach zero. The improved variable universe FSMC enables the system to make
subtle changes without chattering, which greatly improves the accuracy of system control.

4.4. Real Vehicle Test

The speed, acceleration, wheel speed and other basic state signals of the vehicle,
acceleration/deceleration, and steering control are transmitted through the power CAN
bus of the original vehicle, and the positioning information of the high-precision GPS
positioning equipment is transmitted through another CAN. Finally, the analysis of the
CAN message is completed in the built model, and the real-time vehicle status is 16
Machines 2021, 9, x FOR PEER REVIEW
fed back
of 20
to the model and controlled. The control algorithm model built in this paper is compiled
and generated, and runs in MicroAutobox. The data information interaction between
MicroAutobox generated,
and the upper
and runscomputer is carried
in MicroAutobox. The out
datathrough TCP/IP.
information Finally,
interaction betweentheMicro-
control
desk software on Autobox and PC
the test the upper
is usedcomputer is carried outcalibration,
for debugging, through TCP/IP.
andFinally, the control deskThe
user interaction.
software on the test PC is used for debugging, calibration, and user interaction. The whole
whole shown in Figure 16.
shown in Figure 16.

SIMULINK Model

TCP/IP

MicroAutoBox
Upper computer
CAN1 CAN2

Test vehicle

EMS EPS ESP positioning system

Figure 16. System architecture.Figure 16. System architecture.

The transformation part of the communication system is shown in Figure 16. The test
vehicle used in this project has the body CAN and the power CAN. In the test, the power
CAN is mainly used to obtain the vehicle running state parameter information, such as
the wheel speed, steering angle, and gear position. High-precision GPS integrated navi-
gation also uses the CAN interface for communication. The two CAN buses of vehicle
Machines 2021, 9, 304 16 of 19

The transformation part of the communication system is shown in Figure 16. The test
vehicle used in this project has the body CAN and the power CAN. In the test, the power
CAN is mainly used to obtain the vehicle running state parameter information, such as the
wheel speed, steering angle, and gear position. High-precision GPS integrated navigation
Machines 2021, 9, x FOR PEER REVIEW 17 of 20
also uses the CAN interface for communication. The two CAN buses of vehicle
Machines 2021, 9, x FOR PEER REVIEW power
17 of 20
CAN and the GPS positioning system are connected to the two-way CAN of MicroAutobox,
respectively. MicroAutobox
running the pre-burned model, analyzes
so as totheobtainCAN thedata of each
required channel
relevant byquantity
state runningtothe pro-pre-
burned
running model,
the so as
pre-burned to obtain
model, the
so
vide real-time vehicle feedback information. asrequired
to obtain relevant
the state
required quantity
relevant to
state provide
quantity real-time
to pro-
vehicle Tofeedback
vide real-time information.
furthervehicle
test thefeedback
performance information.
of the algorithm, the real vehicle test is carried out in
To
two cases further
To further test
test the performance
the performance
of rectangular field turningof of the
the
left algorithm,
algorithm,
and thereal
the
rectangular realvehicle
field vehicle
turning test
test is is
right, carried
carried
as outout
shown in in
in
two
two cases of
cases of
the figure rectangular
below. To more field
field turning
turning left
intuitively left and rectangular
and rectangular
compare the error size field
field turning
at turning right,
differentright,
times, as
asthisshown
shown
paperin in
the
the figure
figure
selects below.
thebelow. To more
absolute value of intuitively
intuitively compare
the errorcompare the error size
the error sizeindex,
as the measurement at different
at different times,
storestimes,
the GPS this paper
thiscoordi-
paper
selects
selects
nates ofthe absolute
thethe absolute value
valueofofthe
predetermined the error
track errorasasthe
and the
themeasurement
measurement
GPS coordinates index,
index, stores
of thestoresthetheGPS
following GPS coordinates
coordi-
track, and
of the predetermined
nates
draws of
thethetrajectory track and
predetermined
comparisontrack the GPS
and
chart and coordinates
the GPS
the of thecomparison
coordinates
lateral error following track, respectively,
of the following
chart, and draws
track, andthe
trajectory
draws
as shownthecomparison
trajectory
in chart and chart
Figurescomparison
17–20. the lateral
and the error comparison
lateral chart, respectively,
error comparison as shown
chart, respectively,
in
asFigures
shown in 17–20.
Figures 17–20.
39.068
Predetermined trajectory
39.068
Following trajectory
Predetermined trajectory
Following trajectory
39.0675
39.0675

39.067
Longitude

39.067
Longitude

39.0665
39.0665

39.066
39.066

39.0655
117.3535
39.0655 117.354 117.3545 117.355 117.3555 117.356
117.3535 117.354 117.3545 Latitude 117.355 117.3555 117.356
Latitude
Figure17.
Figure 17. Comparison
Comparison of rectangular
rectangularright
rightturns.
turns.
Figure 17. Comparison of rectangular right turns.
0.5
0.5
error/m

0.4
error/m

0.4

0.3
distance

0.3
distance

0.2
Centroid

0.2
Centroid

0.1
0.1

0.0
0.0 0 100 200 300 400 500 600 700
0 100 200 300 400 500 600 700
Data point
Data point
Figure 18. Comparison of rectangular right turns errors.
Figure 18. Comparison of rectangular right turns errors.
Figure 18. Comparison of rectangular right turns errors.
Machines 2021, 9, x FOR PEER REVIEW 18 of 20
Machines 2021,
Machines 9, x304
2021, 9, FOR PEER REVIEW 1817ofof 20
19

39.068
39.068 Predetermined trajectory
Predetermined trajectory
Following trajectory
Following trajectory
39.0675
39.0675

39.067
Longitude

39.067
Longitude

39.0665
39.0665

39.066
39.066

39.0655
39.0655
117.3535 117.354 117.3545 117.355 117.3555 117.356
117.3535 117.354 117.3545 Latitude 117.355 117.3555 117.356
Latitude
Figure
Figure 19. Comparison
Comparison of
of rectangular
rectangular left turns.
Figure 19. Comparison of rectangular left turns.
0.5
0.5
error/m

0.4
distanceerror/m

0.4

0.3
0.3
Centroiddistance

0.2
0.2
Centroid

0.1
0.1

0.0
0.0 0 100 200 300 400 500 600
0 100 200 300 400 500 600
Data point
Data point
Figure 20. Error chart of rectangular left turns errors.
Figure 20. Error chart of rectangular left turns errors.
Figure 20. Error chart of rectangular left turns errors.
It can be observed, from the error diagram in Figures 18 and 20, that the algorithm
keepsIt the
can be observed, from the errorof diagram in Figures 18 and 20, that the algorithm
It canlateral error (absolute
be observed, from thevalue) the vehicle
error diagram within 18
in Figures 0.5and
m, while four
20, that thelarge “wave
algorithm
keeps
peaks” thearelateral error
generated (absolute
in the value)
whole of
runningthe vehicle
process,within
which 0.5 m, while
correspond four
to large
the four“wave
right
keeps the lateral error (absolute value) of the vehicle within 0.5 m, while four large “wave
peaks”
angle are generated
bends of the in the whole
rectangular track.running
Although process,
the which
error is correspond
larger than to of
that the four right
straight line
peaks” are generated in the whole running process, which correspond to the four right
angle bends
driving, it of still
can the rectangular
ensure high track.
trackingAlthough
accuracy theon
error
the is larger than that of straight line
whole.
angle bends of the rectangular track. Although the error is larger than that of straight line
driving, it can still ensure high tracking accuracy on the whole.
driving, it can still ensure high tracking accuracy on the whole.
5. Conclusions
15. Conclusions
Firstly, the optimal preview point is obtained according to the data points of the
15. Conclusions
Firstly, GPS
differential the optimal preview
signal. The point is reflected
curvature obtained from
according to the
the side, bydata points of the
the calculation ofdif-
the
Firstly, the optimal preview point is obtained according to the data points of the dif-
ferential
maximum GPS signal.point,
distance The curvature is reflected
and the preview from
points arethe side, by
selected the calculation
iteratively. Secondly,of the
ferential GPS signal. The curvature is reflected from the side, by the calculation of the
maximum
pure tracking distance point,
algorithm and the
is used preview the
to calculate points are selected
expected steeringiteratively. Secondly,
angle. Finally, the
a variable
maximum distance point, and the preview points are selected iteratively. Secondly, the
pure tracking
universe fuzzyalgorithm is usedcontroller
sliding mode to calculate is the expected
designed steeringthe
to control angle. Finally,
lateral a variable
motion of the
pure tracking algorithm is used to calculate the expected steering angle. Finally, a variable
vehicle by
universe measuring
fuzzy slidingthe deviation
mode in yaw
controller rate. Theto
is designed variable
controluniverse
the lateral FSMC algorithm
motion of the
universe
designed fuzzy
thissliding ismode controller is designedFSMC,
to control the oflateral motion of and
the
vehicle byinmeasuring
paper the superior
deviationto the traditional
in yaw rate. The variablein terms
universe theFSMC
arrival time
algorithm
vehicle by measuring the deviation in yaw rate. The variable universe FSMC algorithm
Machines 2021, 9, 304 18 of 19

the steady-state accuracy, and further optimizes the tracking accuracy of the system on the
premise of weakening the chattering phenomenon. Due to the limited time and conditions,
this paper only carries out the real vehicle test on the more conventional road, and does not
test the effect of other control algorithms. In the future, the real vehicle test will be carried
out on more roads and compared with other algorithms.

Author Contributions: Conceptualization, X.Y. and J.L.; methodology, J.L. and X.S.; software, J.L.;
validation, X.Y., H.Z. and X.X.; formal analysis, X.Y. and J.L.; investigation, X.Y. and J.L.; resources,
X.Y. and J.L.; data curation, X.Y.; writing—original draft preparation, J.L.; writing—review and
editing, X.Y.; visualization, X.Y.; supervision, H.Z.; project administration, X.X.; funding acquisition,
X.S. All authors have read and agreed to the published version of the manuscript.
Funding: National Key Research and Development Program of China under Grant (Zhejiang Wan’an
Technology Co., Ltd. 2018YFB0104803).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Acknowledgments: Thanks for the test conditions provided by China Automotive Technology &
Research Center Co. Ltd. (Tianjin, China) for my experimental research.
Conflicts of Interest: There is no conflict of interest regarding the publication of this paper.

References
1. Lu, W.; Dong, J.; Zhang, P. Research on vision enhancement system of military vehicle driver based on uncooled infrared detection
technology. Appl. Opt. 2019, 40, 1033–1038.
2. Xia, X.; Wang, H.; Li, X. Research on high precision automatic leveling control system of a military radar vehicle. Mech. Des. Manuf.
2019, 9, 158–160.
3. Lu, H.; Xing, Y.; Zhuo, G. Review on Motion Control of Autonomous Vehicles. Available online: http://kns.cnki.net/kcms/
detail/11.2187.th.20200416.1728.010.html (accessed on 6 June 2021). [CrossRef]
4. Ding, N.; Ran, X.; Zhang, H. Driver model for single track vehicle based on single point preview optimal curvature model. J.
Mech. Eng. 2008, 44, 220–223, 228. [CrossRef]
5. Dong, T. Lane keeping system based on improved preview driver model. Pract. Automob. Technol. 2019, 24, 143–145.
6. Ren, D.; Cui, S.; Wu, H. Preview control of lane keeping and its steady state error analysis. Automot. Eng. 2016, 38, 192–199.
7. Xu, S.; Peng, H. Design, analysis, and experiments of preview path tracking control for autonomous vehicles. IEEE Trans. Intell.
Transp. Syst. 2020, 21, 48–58. [CrossRef]
8. Zhao, X.; Chen, H. A study on lateral control method for the path tracking of intelligent vehicles. Automot. Eng. 2011, 33, 382–387.
9. Pan, Z. Research on Motion Control Approaches of Autonomous Vehicle in Urban Environments; University of Science and Technology
of China: Hefei, China, 2012.
10. Marino, R.; Scalzi, S.; Netto, M. Nested PID steering control for lane keeping in autonomous vehicles. Control. Eng. Pract. 2011,
19, 1459–1467. [CrossRef]
11. Silva, M.; Garrote, L.; Moita, F. Autonomous electric vehicle: Steering and path-following control systems. In 2012 16th IEEE
Mediterranean Electrotechnical Conference; IEEE: Manhattan, NY, USA, 2012; pp. 442–445.
12. Guo, J.; Luo, Y.; Li, K. An Adaptive Hierarchical Trajectory Following Control Approach of Autonomous Four-Wheel Independent
Drive Electric Vehicles. IEEE Trans. Intell. Transp. Syst. 2017, 19, 2482–2492. [CrossRef]
13. Ackermann, J.; Guldner, J.; Sienel, W. Linear and nonlinear controller design for robust automatic steering. IEEE Trans. Control.
Syst. Technol. 1995, 3, 132–143. [CrossRef]
14. Yu, Z.; Zhang, X.; Xiong, L. Dynamic Control for Unmanned Skid-steering Vehicle with Conditional Integartors. J. Mech. Eng.
2017, 53, 29–38. [CrossRef]
15. Tagne, G.; Talj, R.; Charara, A. Higher-order sliding mode control for lateral dynamics of autonomous vehicles, with experimental
validation. In 2013 IEEE Intelligent Vehicles Symposium (IV); IEEE: Manhattan, NY, USA, 2013; pp. 678–683.
16. Ji, X.; He, X.; Lv, C. Adaptive-neural-network-based robust lateral motion control for autonomous vehicle at driving limits.
Control. Eng. Pract. 2018, 76, 41–53. [CrossRef]
17. Falcone, P.; Borrelli, F.; Asgari, J.; Tseng, H.E.; Hrovat, D. Predictive Active Steering Control for Autonomous Vehicle Systems.
IEEE Trans. Control. Syst. Technol. 2007, 15, 566–580. [CrossRef]
18. Falcone, P.; Borrelli, F.; Tseng, H.E.; Asgari, J.; Hrovat, D. Linear time-varying model predictive control and its application to active
steering systems: Stability analysis and experimental validation. Int. J. Robust Nonlinear Control. 2008, 18, 862–875. [CrossRef]
19. Falcone, P.; Tseng, H.E.; Borrelli, F.; Asgari, J.; Hrovat, D. MPC-based yaw and lateral stabilisation via active front steering and
braking. Veh. Syst. Dyn. 2008, 46, 611–628. [CrossRef]
20. Gong, J.; Jiang, Y.; Xu, W. Model Predictive Control of Unmanned Vehicles; Beijing University of Technology Press: Beijing, China, 2014.
Machines 2021, 9, 304 19 of 19

21. Guo, H.; Liu, J.; Cao, D.; Chen, H.; Yu, R.; Lv, C. Dual-envelop-oriented moving horizon path tracking control for fully automated
vehicles. Mechatronics 2018, 50, 422–433. [CrossRef]
22. Kim, E.; Kim, J.; Sunwoo, M. Model predictive control strategy for smooth path tracking of autonomous vehicles with steering
actuator dynamics. Int. J. Automot. Technol. 2014, 15, 1155–1164. [CrossRef]
23. Nam, H.; Choi, W.; Ahn, C. Model predictive control for evasive steering of an autonomous vehicle. Int. J. Automot. Technol. 2019,
20, 1033–1042. [CrossRef]
24. Ge, S.S.; Lee, T.H.; Hang, C.C.; Zhang, T. Stable Adaptive Neural Network Control; Kluwer Academic Publishers: New York, NY,
USA, 2001.
25. Wai, R.-J.; Lin, C.-M.; Hsu, C.-F. Adaptive fuzzy sliding-mode control for electrical servo drive. Fuzzy Sets Syst. 2004, 143, 295–310.
[CrossRef]
26. Zhang, H.; Liang, J.; Jiang, H.; Cai, Y.; Xu, X. Stability Research of Distributed Drive Electric Vehicle by Adaptive Direct Yaw
Moment Control. IEEE Access 2019, 7, 2169–3536. [CrossRef]