Automatic control strategy for catching a fixed-wing

drone using a ground vehicle

A. Alatorre, Pedro Castillo Garcia, Rogelio Lozano

To cite this version:

A. Alatorre, Pedro Castillo Garcia, Rogelio Lozano. Automatic control strategy for catching a fixed-
wing drone using a ground vehicle. 26th IEEE International Conference on Intelligent Transportation
Systems (ITSC’23), Sep 2023, Bilbao, Spain, Spain. �hal-04342801�

HAL Id: hal-04342801

https://hal.science/hal-04342801
Submitted on 13 Dec 2023

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 Automatic control strategy for catching a fixed-wing drone using a
ground vehicle
Armando Alatorre1,2 , Pedro Castillo1 & Rogelio Lozano1,2 .

Abstract— This paper presents a control strategy designed in For improving safe landing, some sensors, and embedded
3D for landing a fixed-wing drone on a moving ground vehicle. devices have been included in the aerial vehicle. These
The control strategy focuses on leading the fixed-wing vehicle sensors allow the measurement of variables to improve the
towards a desired trajectory while the control algorithms for the
ground vehicle regulate its speed and tracks the relative position control algorithms such as the Global Position System (GPS)
of the aerial vehicle in the x − y plane. The desired trajectory [9], and the Inertial Measurement Unit (IMU). In addition,
for the aerial vehicle is based on a hyperbolic tangent function some perception solutions are also included and they are
to perform a soft descending reaching the ground vehicle’s based on computational vision algorithms, identify markers
altitude. The strategy allows the rendezvous of both vehicles to align the drone towards a runway [10] or estimate its
obtaining a safe landing for the airplane. The methodology to
determine the control laws is based on the Lyapunov analysis, position and speed with respect to a target [11].
guaranteeing the stability on each control stage. The strategy is In the literature, it is possible to find the flight stages
evaluated in numerical simulations for validating the systems
performance in closed loop.
for landing a fixed-wing drone on a runway. For example
in [12], the authors divided the landing process into three
I. I NTRODUCTION stages: a descending flight, flare maneuver, and taxiing. The
The growing development of new technologies has encour- conditions required to carry out a landing on a runway or
aged the study of Unmanned Aerial Vehicles (UAVs) thanks flat zone are very strict to reduce the likelihood of accidents
their capabilities to perform missions such as monitoring in the landing process.
high-risk areas [1], target identification [2], tracking of a Considering the previously, some fixed-wing drones per-
moving ground vehicle [3], and agriculture activities. form missions in order to navigate during a lot of hours
The research community has been working to innovate or transport payload. Therefore, they need to reduce their
and improve autonomous navigation in different types of weight and delete the landing gear, this problem increases
environments and conditions. The field of automatic control the risk of damage to the vehicle’s structure.
has a high impact on the development of novel applications,
One solution could be the usage of recovery systems such
maneuvers, and theoretical validation of control algorithms.
as [13]. In this solution, the aircraft would have integrated a
Diverse research projects have focused on missions of path
hook in its structure to take a tense or elastic cable. However,
planning [4], guidance strategies [5], cooperative control [6],
sometimes these kinds of devices are not convenient for
payload transportation, and obstacle avoidance.
aerial vehicles. Another approach consists of support to the
UAVs are commonly classified into multirotor and fixed- aerial vehicle for landing on a ground vehicle. In this case,
wing vehicles. Multirotor configuration is composed of two the challenge will be to develop a rendezvous synchroniza-
or more rotors. It can fly at hover, allowing high-precision tion of both vehicles.
maneuvers to be performed such as the manipulation of ob-
jects, and the exploration of high-risk areas. Otherwise, fixed- Research projects working on cooperative control or ren-
wing vehicles are suitable for missions of high altitudes, dezvous guidance control are focused on keeping a distance
long distances, and high speeds [7]. The main advantage is or reach to the target position such as in a refueling mission.
the energy consumption [8], since its aerodynamic properties In [14], the authors propose a guidance control strategy to
allow the generation of lift to maintain the flight. align two aerial vehicles. A tanker aircraft and a receptor
aircraft are controlled to maintain constraints of speed to
Focusing our research on fixed-wing vehicles, one critical
avoid collisions. Similarly, a ground vehicle was used as
flight stage is the landing. Since, the aircraft must perform
mobile refueling unit, the aircraft is guided to a defined
the maneuver with precision, guiding the aircraft to the
altitude with the same speed as the ground vehicle [15].
desired target and avoiding wind disturbances, which are the
main causes of accidents. Thus, this increases the motivation In [16], the authors proposed a cooperative control to
of the researcher community to perform a safe landing by land a fixed-wing vehicle on a ground vehicle. The aircraft
applying the control theory. descends its altitude, and the landing is carried out once
both vehicles reach the same speed and position. The project
1 Université de technologie de Compiègne, CNRS, Heudiasyc (Heuristics
evolution was reflected in [17], where the authors proposed
and Diagnosis of Complex Systems), CS 60319 - 60203 Compiègne Cedex, a Model Predictive Control (MPC) for landing the aerial
France. (aalatorr,pcastillo, rlozano)@hds.utc.fr
2 Center of Research and Advanced Studies of the National Polytechnic vehicle in a finite time on the ground vehicle. However, it
Institute (CINVESTAV). [email protected] becomes a complex process for the control strategy.
 In this work, we propose a cooperative control strategy In addition, (Au , , Av , Aw ) represent the aerodynamic
for landing a fixed-wing vehicle on a moving ground vehicle. forces, which are described as
The control strategy consists of guiding the fixed-wing drone  ca q

Cx (α) + Cxq (α) 2V + Cxδe (α)δe
 
following a hyperbolic trajectory. The desired trajectory is Au 2
Av= ρVa Sa
a
Cy bp Cyr br
Cy0+Cyβ β+ 2Vpa + 2V +Cyδaδa +Cyδrδr(13)

designed to perform a soft descending flight until reaches 2 a
Aw ca q
the ground vehicle’s altitude. At the same time, the ground Cz (α) + Czq (α) 2V a
+ Czδe (α)δe
vehicle needs to control its speed and direction to reach the
fixed-wing drone’s position to receive it for landing. The where the surface area of the wing is denoted as Sa , the mean
closed-loop stability of both systems has been determined chord of the wing is represented as ca , and the wingspan is
using the Lyapunov theory. given by b. The control inputs for an airplane with classical
configuration are given by the engine input δt , the elevator
The manuscript is organized as follows: mathematical deflection δe , the ailerons δa , and the rudder δr .
preliminaries are presented in Section II. The problem state-
ment is given in Section III. The landing control strategy The aerodynamics forces coefficients for the longitudinal
composed by the controllers of the fixed-wing vehicle and subsystems, Cx and Cz , depend on the angle of attack α,
ground vehicle is described in Section IV. Main graphs from the pitch angular rate q, and elevator deflection δe .
simulation results when validating the proposed strategy are The lateral coefficients are in function of the side-slip
shown in Section V. Section VI presents the concluding angle β, roll and yaw angular rates (p and r), the ailerons
remarks and future research directions. δa and rudder δr control inputs.
The moments related to rotational rates are described as
II. M ATHEMATICAL PRELIMINARIES
   ρVa2 Sa b 
The nonlinear motion equations for a fixed wing drone τp 2 Cp
2
 τq  =   ρVa 2Sa ca Cm  (14)

described in [18] can be written as
2
τr ρVa Sa b
Cr
ẋ = (cos θ cos ψ) u+(sin ϕ sin θ cos ψ−cos ϕ sin ψ)v 2

+(cos ϕ sin θ cos ψ+sin ϕ sin ψ) w (1) where

  bp br

ẏ = (cos θ sin ψ) u+(sin ϕ sin θ sin ψ+cos ϕ cos ψ)v Cp Cp0 +Cpβ β+Cpp 2V a
+Cpr 2V a
+Cpδaδa +Cpδrδr
ca
+(cos ϕ sin θ sin ψ−sin ϕ cos ψ) w (2) Cm= Cm (α) + Cmq 2V a
q + Cmδe δe (15)

Cr bp br
Cr0 + Crβ β+Crp 2Va +Crr 2Va +Crδaδa +Crδrδr
ż = u sin θ−(sin ϕ cos θ) v+(cos ϕ cos θ) w, (3)
Au ρSh Ch  The aerodynamic coefficients (13) and (15), and inertia
(kr δt )2 − Va2 ,

u̇ = rv − qw − g sin θ+ + (4) terms (10)-(12) are described with more details in [18].
m 2m
Av Finally, the airspeed, the angle of attack, and the side-slip
v̇ = pw − ru + g sin ϕ cos θ + , (5)
m angle can be denoted with the form:
Aw
ẇ = qu − pu + g cos θ cos ϕ + , (6) p
m Va = u2 + v 2 + w2 (16)
w
ϕ̇ = p+q sin ϕ tan θ+r cos ϕ tan θ (7) α = atan (17)
θ̇ = q cos ϕ−r sin θ, (8) u 
v
β = asin (18)
ψ̇ = q sin ϕ sec θ+r cos ϕ sec θ (9) Va
ṗ = Γ1 pq−Γ2 qr+τp (10) In this work, we will use also a ground vehicle (GV),
1 hence, its mathematical motion equations can be described
q̇ = Γ5 pr − Γ6 p2 + r 2


+ τq (11) as in [19]:
Jy
ṙ = Γ7 pq−Γ1 qr+τr . (12) ẋg = Vg cos ψg (19)
where (x, , y , z) represent the position in the inertial frame ẏg = Vg sin ψg (20)
{I}, the linear velocities (u, , v , w) are described in the Vg
ψ̇g = tan us (21)
body frame {B}. The Euler angles are described as: ϕ for lf
the roll angle, θ for the pitch angle, and ψ for the yaw angle. V̇g = ut (22)
(p , q , r) describe the rotational rates in the body frame. The
vehicle’s mass is given by m, and terms related to the inertia where (xg ,yg ) represent the ground vehicle’s positions, ψg
moments are denoted by Γ(·) . denotes its respectively yaw angle. The speed of the ground
The air density is denoted as ρ, the coefficient of motor vehicle is defined by Vg , the distance from the wheels to the
efficiency is represented by kr , the airspeed is defined as Va , mass center of the vehicle defines as lf . The steering control
the area and the aerodynamic coefficient of the propeller are input is denoted as us and the throttle input is represented
represented by Sh and Ch , respectively. by ut . Finally, the altitude of the vehicle is defined as zg .
 III. P ROBLEM STATEMENT The desired trajectory is designed to reach the altitude of
Landing stage is a critical maneuver for fixed-wing drones the ground vehicle, see Figure 2. Therefore, the landing is
imply a high number of accidents caused by the pilots’ inex- carried out since the ground vehicle is autonomous controlled
perience or crosswind disturbances. For solving this problem, to maintain its position with respect to the aircraft position
some solutions have been proposed such as recovery systems, in the x-y plane.
which are popular for creative maneuvers to capture aerial
vehicles. However, most of these techniques are considered
as aggressive maneuvers since the capture form or braking
the aircraft increases the risk of damaging its structure.
Our solution is based on a cooperative control strategy for
landing a fixed-wing drone on a moving ground vehicle, see
Figure 1. Both vehicles are navigating in autonomous mode
exchanging information to carry out the landing challenge.
The control strategy for the fixed-wing vehicle is will be
to perform an alignment stage in order to navigate straight
thorough the mission. After that, the aerial vehicle will
execute a trajectory tracking to perform a descending flight Fig. 2: Study of the parameters of the desired trajectory.
until reaches the ground vehicle’s altitude.
Otherwise, the ground vehicle will have the task of fol-
lowing and aligning itself to the aircraft’s position. Thus,
a control algorithm will be developed to orient the ground IV. L ANDING CONTROL STRATEGY
vehicle to the direction of the aircraft using the steering
control. Besides, the ground vehicle’s speed will be also Our control methodology for landing a fixed-wing drone
controlled to reach the aircraft’s speed. on a moving ground vehicle contains the control designs of
both vehicles. For the aerial vehicle, its lateral subsystem is
controlled to be aligned towards a direction in the longitudi-
nal plane (x,z). These lateral motions will not be aggressive
maneuvers, i.e., the aircraft will execute slow displacement
reaching small angles. The last stage for the aerial vehicle,
will be a tracking control algorithm, which is proposed to
perform a descending flight reaching the ground vehicle’
altitude. The control strategy for the ground vehicle will be
designed to track the aerial vehicle’s position, so that, the
ground vehicle aligns and controls its speed relative to the
drone.

A. Fixed-wing vehicle controllers

1) Side-slip angle stabilization: The aircraft alignment

Fig. 1: Representation of the control strategy for landing a with respect to the airspeed vector is carried out stabilizing
fixed-wing drone on a moving ground vehicle. the side-slip angle. Using the Lyapunov stability analysis, we
propose a positive function as V1 = 12 β 2 . Then, differentiat-
ing the previous function and using (18), it yields
A. Desired trajectory
ρVa2 Sa
!
The desired trajectory is proposed to perform a descend- pw − ru + g sin ϕ cos θ + 2m CY
V̇1 = eβ < 0 (24)
ing flight. The trajectory is based on a hyperbolic tangent Va cos β
function, which depends on the time t.
Therefore for making β → 0, we propose the rudder control
The desired trayectory is defined as follows
  input δr as
hD hD t − tm
zd = − tanh + zg (23)  h
bp br
i
2 2 µ − ClatY + CY0 +CYp 2V a
+CYr 2Va +C δ
Yδa a
δr = (25)
where hD represents the distance between initial reference C Yδ r
altitude hT and the ground vehicle’s altitude zg . The time
2m
tm is related to the mean altitude of the trajectory, and µ where ClatY = ρVa2 Sa [pw − ru + g cos θ sin ϕ]+k1 eβ , with
modifies the inclination of the altitude descending. k1 > 0.
 2) Lateral alignment: Notice that firstly, the aircraft Introducing (4) into the above equation, the engine control
focuses on aligning with the y-axis, maintaining straight input can be determined as
throughout the descending flight. Thus, considering small
V2
 
1 Sa
angles for the aircraft lateral maneuvers, we can express the δt2 = a2 − Au (36)
lateral dynamic related to the y-axis as follows kr cos α Sh Ch
!
2m V̇ad − eVa ẇ tan α
ẏ = ψ (u cos θ + vϕ sin θ + w sin θ) − wϕ (26) + qw + g sin θ + −
ρSh Ch kr2 cos θ 2m
Defining yd as a constant desired position, we can express
that ey = y − yd . Then, proposing a positive function as Therefore, V̇4 = −e2Va < 0. Then, it follows that Va → Vad ,
V2 = 12 e2y , it follows that implying that ẇ → 0 and u̇ → 0.

V̇2 = ey ėy 4) Altitude stabilization: The altitude analysis is focused

on the Lyapunov theory with the goal to determine a de-
= ey [ψ (u cos θ + vϕ sin θ + w sin θ) − wϕ] (27) sired pitch angle, which will guarantee altitude stabilization.
Expressing the desired yaw angle as Therefore, an elevator control input is designed to modify
the pitch angle of the aircraft.
wϕ − k2 ey
ψd = (28) Defining a desired altitude zd , the altitude error can be
u cos θ + vϕ sin θ + w sin θ
expressed as ez = z − z d . Thus, a positive function is
where k2 > 0, implies that if ψ → ψd then ey → 0 and proposed as V5 = 21 e2z , which must satisfy the Lyapunov
y → yd . stability properties, V5 > 0 and V̇5 < 0.
Using the Lyapuov stability analysis for the yaw error, Differentiating V5 , it yields that V̇5 = ez ėz < 0. Once
which is given by eψ = ψ − ψd , a positive function is the lateral dynamics are stabilized, then the longitudinal
proposed as V3 = 21 e2ψ , such that its differentiation must equations can be simplified. Thus, we can rewrite V̇5 using
satisfy the Lyapunov condition V̇3 = ėψ eψ < 0. Analyzing ż from (3),
the previous equation it follows that
V̇5 = ez (u sin θ − w cos θ − żd ) < 0. (37)
(qϕ + r) eψ < 0, (29)
Observe that, (37) can be simplified dividing the expression
Considering the above equation, the desired roll angle ϕd by cos θ,
can be described as
−r − eψ V̇5 żd
ϕd = . (30) = ez (u tan θ − w − ) < 0. (38)
q cos θ cos θ
The previous is satisfied for − π2 < θ < π2 . Therefore, using
From (10), the aileron control input can be proposed as
(38) for defining θd , it follows
h i
bp br
−ClatP − Cp0 +Cpp 2V +Cpr 2V +Cpδrδr −ηϕ
!
żd
a a −1 w + cos θ − ez
δa = , (31) θd = tan . (39)
Cpδa u
where Substituting (39) into (38), we will obtain that −e2z < 0.
2 Then, the goal will be to propose a controller such that θ →
ClatP = (Γ1 pq − Γ2 qr) . (32)
ρVa2 Sa b θd , this will imply that ez → 0 and then z → zd .
Notice that equation (31) involves the dynamic behavior Computing the first and second derivative of the equation
(39), it yields
ηϕ = kpϕ (ϕ − ϕd ) + kdϕ ϕ̇, (33)  
u z̈d cos cos
θ+żd θ̇ sin θ
2θ − ėz
where, kpϕ and kdϕ are the proportional and derivative gains, θ̇d = (40)
żd

2
respectively. u2 + w + cos θ − ez

3) Airspeed control: From the airspeed expression in (16),  

it is possible to determine its derivative based on the rotation u 2θ̇2 cos θ sin2 θ
2


u −ëz + cos θϱ1
from wind frame to the body frame as follows θ̈d = + 2
u2 + (w + cosż θ − ez )2 u + (w + cosż θ − ez )2
2
V̇a = u̇ cos α + ẇ sin α (34)

2u(w + cosż θ − ez ) z̈ coscos
θ+ż θ̇ sin θ
2θ − ez
using again the Lyapunov analysis,and defining the airspeed −
2 (41)
u2 + (w + cosż θ − ez )2
error as eVa = Va − Vad . We can propose a positive function
as V4 = 21 e2Va . It implies that where
...
ϱ1 = z d (cos θ+ θ̇ sin θ)+ z̈d (θ̇2 cos θ+ θ̇ sin θ+ θ̈ sin θ) (42)
 
V̇4 = eVa u̇ cos α + ẇ sin α − V̇ad (35)
 5) Pitch control: A tracking controller is designed to We propose a feedback state controller as
modify the pitch angle using the elevator control input based 1

1

on the feedback state approach. ut = Vg ψ̇g sin ψg + ẍ + ėζx + k3 ϑx (55)
cos ψg 2
The pitch dynamics can be described as
introducing (55) into (54), it yields
θ̇ = q (43)
V̇7 = −k3 V7 (56)
ρVa2 Sc
q̇ = Cm . (44)
2Jy Therefore, eζx → 0 and ėζx → 0, that is, xg → x and
The pitch error can be expressed as eθ = θ − θd . Then, ẋg → ẋ.
proposing the following positive function 2) Steering controller: Define the error related with the
1 y-axis as eζy = y − yg . Then, a positive function is proposed
V6 = (θ − θd )2 (45) as V8 = 21 ϑ2 , where ϑy = eζy + ėζy .
2
Computing the derivative of the above function, it yields
It implies that !
h  i Vg2
V̇6 = (θ − θd ) ėθ + θ̈ − θ̈d V̇8 = ϑy ėζy + ÿ − V̇g sin ψg − cos ψg tan us (57)
lf
ρVa2 Sa ca
 
= (θ − θd ) ėθ − θ̈d + Cm (46) Proposing the steering control input as
2Jy   
Therefore, the pitch dynamics can be stabilized using the lf 1
us =atan 2 −V̇g sin ψg + ÿ + ėζy + k4 ϑ (58)
following controller Vg cos ψg 2
  
1 ca where k4 > 0, it implies that V̇8 ≤ −k4 V8 < 0. Thus,
δe = − Cm0 + Cmα α + Cmq q
Cmδe 2Va eζy → 0 and ėζy → 0, that is, yg → y and ẏg → ẏ.
 
2Jy 1
+ 2 −ėθ + θ̈d − (θ − θd ) ,(47) V. S IMULATION RESULTS
ρVa Sa ca 2
where kθp and kθd are positive gains. Substituting (47) into The landing control strategy is implemented in numerical
(46), yields simulations to validate the control algorithm performances
V̇6 = −V6 < 0 (48) in closed loop. The strategy is designed to align the aerial
drone in a certain direction for performing a descending
Notice that eθ → 0 and ėθ → 0, that is, θ → θd and θ̇ → θ̇d . flight. Moreover, the ground vehicle is controlled to reach
Therefore, it implies that q → 0, ez → 0, and ż → 0. the reflected position of the aircraft in the x-y plane.
B. Ground vehicle controllers For simulations, the initial position of the ground ve-
The goal of the control design for the ground vehicle is hicle is defined in (xg (0), yg (0)) = (0, −2) in meters
to track the position of the fixed-wing drone in the x − y with , ψg (0) = 0.57◦ , and the aircraft’s position is given
plane. First, we present the throttle control input ut in order in (x(0), y(0), z(0)) = (−10, 2, 10)m with ψ(0) = 0◦ .
to stabilize the speed of the ground vehicle with respect to The fixed-wing drone regulates its airspeed to a constant
the drone’s airspeed. Then, the steering control input us is minimum phase (6 m/s) with the goal that the ground vehicle
developed to align the ground vehicle to the aircraft. will be able to reach the aircraft position.
1) Throttle controller: Considering the error between the The aerodynamic parameters of the aircraft are described
aircraft and ground vehicle positions in the x-axis, then the in the Table I.
following expression can be written The desired trajectory’s parameters are described as fol-
lows: hD = 9 meters, tm = 16 seconds, µ = 6, zg = 1 m
eζ x = x − xg (49) and lf = 0.3 m. The steering control input is saturated in
ėζx = ẋ − ẋg (50) the region [−20◦ , 20◦ ] and the motor by [0 , 3].
ëζx = ẍ − ẍg (51) The simulation result of the landing strategy is illustrated
where in a 3D space in Figure 3. Notice that the aerial vehicle
ẍg = V̇g cos ψg + Vg ψ̇g sin ψg (52) aligns to the x-axis. Then, the fixed-wing vehicle performs
the descending flight until reach the ground vehicle position.
Now, the Lyapunov analysis will be carried out considering However, the ground vehicle follows the aircraft position,
the following positive function which is tracking a long the mission.
1 2 1 In Figure 4, the performance of the fixed-wing vehicle to
V7 = (eζx + ėζx ) = ϑ2x (53)
2 2 carry out the trajectory tracking is presented. The descending
Differentiating the previous equation, it yields trajectory imposes the aircraft to reduces its altitude from 10
  meters to 1 meter, allowing the landing on the GV’s top.
V̇7 = ϑx ėζx + ẍ − V̇g cos ψg + Vg ψ̇g sin ψg (54)
 Parameters Value Parameters Value
CL0 0.4029 CD0 0.0256
CLα 4.64 CDα 0.1749
CLq 7.4431 CDq 0
CLδe -0.42108 CDδe 0.00528
Cm0 -0.0408 Cmα -1.0454
Cmq -8.9585 Cmδe -1.09407
Sh 0.0314 Ch 1
Kr 8 CDp 0.027
m 0.824 Jy 0.02453
AR 6.54 ca 0.168
M 50 α0 0.4712
Sa 0.185 b 1.1
Cy0 =Cl0 =Clr 0 Cn0 =Cnδa =Cnδr 0
Cyb -0.16451 Cnβ 0.058246
Cyp -0.12525 Cnp -0.016292
Fig. 3: Landing control strategy performance in a 3D space. Cyr 0.13822 Cnr -0.046933
Cyδa 0.04111 Jx 0.02628
Cyδr 0.09977 Jz 0.04811
Clβ -0.054443 Jxz -0.0009316
10
Cl p -0.48364 Clδa 0.0287
9 Desired trajectory
Fixed-wing drone
Clr 0.069133 - -
8

7
TABLE I: Aerodynamic parameters of the reference aircraft.
Altitude [m]

5
7
4

3
6

2
5 Va
1 Vg
Speed [m/s]

0 5 10 15 20 25 30 35 4
Time [s]
3
Fig. 4: Tracking results of the descending trajectory.
2
2.5
Fixed-wing drone 1
2 Ground vehicle

1.5 0
0 5 10 15 20 25 30 35
1
Time [s]
0.5
y [m]

Fig. 6: Regulation performance of the speed of both vehicles.

0

-0.5

-1
Besides, the aircraft presents a minimum airspeed change
-1.5
in the descending period but the control law compensates
-2
-50 0 50 100 150 200
quickly the speed. The behavior of the control input signals
x [m] for the ground vehicle are presented in Figure 9.
Fig. 5: Simulation result of landing strategy in the x-y plane. Observe that the control inputs start the mission with a
value of saturation maximum (us = 20◦ and ut = 3[N ]),
In Figure 5, we present the alignment in the x − y plane.
and then the system is stabilized when us and ut tends to 0.
Observe in this figure, the aircraft stabilizes its position to
keep its reference equal to zero. Similarly, the ground vehicle In Figure 10, the performance of the longitudinal control
aligns to the relative position of the airplane with the goal inputs in order to evaluate the behavior during the descending
to guarantee that the ground vehicle catches the drone. flight is presented, the lateral controllers were not considered
As previously mentioned, the aircraft regulates its airspeed because their signals are almost zero. Thus, analyzing the
to a constant minimum value. Therefore, the ground vehicle elevator deflection in this figure, notice that it is almost at its
increases its speed to be able of reaching the aircraft position. limit to generate more lift because the airplane is maintaining
The speed performance of the ground vehicle to reach 6 a minimum airspeed. In addition, the behavior of the motor
m/s, i.e., the aircraft’s airspeed can be observed in Figure 6. input tends to reduce its value in the descending flight, taking
In Figure 7, we can analyze the convergence of the ground advantage of the gravity force.
vehicle to the drone’s position in the x and y axes. The speed
error is presented in Figure 8. Notice from this figure that
speed error required to align the ground vehicle is minimum.
 4 3
25
us
2 ut 2.5
20
0

Angle [deg]
2
15
-2
1.5

-4 10
1

-6 5
0.5
-8
0 0
-10
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
Time [s] Time [s]

Fig. 7: Positions errors between both vehicles. Fig. 9: Performance of us ans ut control inputs.

6 -17
1.4
e
5 t
-17.5 1.2
4

Angle [deg]
1
3 -18
0.8
2
-18.5
1 0.6

0 0.4
-19

-1 0.2

-2 -19.5
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
Time [s] Time [s]

Fig. 8: Speed tracking error. Fig. 10: Behavior of the elevator and engine control inputs.

ACKNOWLEDGMENT
VI. CONCLUSIONS
This paper was supported by the RPV project - UTC
In this paper, an autonomous control strategy for landing a foundation and the Mexican National Council of Science and
fixed-wing drone on a moving ground vehicle was presented. Technology - CONACyT.
The strategy focused on lead the problem to a simple solution
R EFERENCES
where simple tasks were assigned to each vehicle allows
the rendezvous of both vehicle to guarantee a safe landing. [1] Š. Čerba, J. Lüley, B. Vrban, F. Osuský, & V. Nečas. “Unmanned
radiation-monitoring system”. IEEE Transactions on Nuclear Science,
The control strategy for the fixed-wing drone was based on 67(4), 636-643, 2020.
the trajectory tracking for a descending flight. The desired [2] Barber, D. B., Redding, J. D., McLain, T. W., Beard, R. W., &
trajectory was designed to carry out a soft descending, Taylor, C. N. (2006). Vision-based target geo-location using a fixed-
wing miniature air vehicle. Journal of Intelligent and Robotic Systems,
steering the aircraft to reach the ground vehicle’s altitude. 47(4), 361-382.
In addition, the ground vehicle executes the task of tracking [3] T. Oliveira, & P. Encarnação. “Ground target tracking control system
the aircraft position, i.e., it aligned and regulated its speed for unmanned aerial vehicles”. Journal of Intelligent & Robotic
Systems, 69, 373-387, 2013.
to maintain the same position in earth plane. The controllers [4] T. M. Cabreira, C. D. Franco, P. R. Ferreira and G. C. Buttazzo.
were designed using the Lyapunov theory. “Energy-Aware Spiral Coverage Path Planning for UAV Photogram-
The simulation results allowed the validation of our pro- metric Applications”. IEEE Robotics and Automation Letters, vol. 3,
no. 4, pp. 3662-3668, Oct. 2018.
posal obtaining a satisfactory result for a safe landing. [5] P. Iscold, G. A. S. Pereira and L. A. B. Torres. “Development of
a Hand-Launched Small UAV for Ground Reconnaissance”. IEEE
Transactions on Aerospace and Electronic Systems, vol. 46, no. 1,
A. Future work pp. 335-348, Jan. 2010.
[6] Z. Qu. “Cooperative control of dynamical systems: applications to
The main goals for future work are to improve the control autonomous vehicles” London: Springer. (Vol. 3). 2009.
designs to compensate external disturbances and system [7] M. Varga, J. C. Zufferey, Heitz, G. H. M. Zufferey, & D. Floreano.
“Evaluation of control strategies for fixed-wing drones following slow-
uncertainties. Moreover, the experimental validation will be moving ground agents”. Robotics and Autonomous Systems, 72, 285-
carried out to lead with more realistic conditions. 294, (2015).
 [8] T. Elijah, R. S. Jamisola, Z. Tjiparuro, & M. Namoshe. “A review on
control and maneuvering of cooperative fixed-wing drones”. Interna-
tional Journal of Dynamics and Control, 9(3), 1332-1349, (2021).
[9] Jantawong, J., & Deelertpaiboon, C. (2018, July). “Automatic landing
control based on GPS for fixed-wing aircraft”. In 2018 15th Inter-
national Conference on Electrical Engineering/Electronics, Computer,
Telecommunications and Information Technology (ECTI-CON) (pp.
313-316). IEEE.
[10] M. Ruchanurucks, P. Rakprayoon, & S. Kongkaew. “Automatic Land-
ing Assist System Using IMU+ P n P for Robust Positioning of Fixed-
Wing UAVs”. Journal of Intelligent & Robotic Systems, 2018, 90(1),
189-199.
[11] O. A. Yakimenko, I. I. Kaminer, W. J. Lentz, & P. A. Ghyzel.
“Unmanned aircraft navigation for shipboard landing using infrared
vision”. IEEE Transactions on Aerospace and Electronic Systems,
2002, 38(4), 1181-1200.
[12] D. Zhang,& X. Wang. “Autonomous landing control of fixed-wing
uavs: from theory to field experiment”. Journal of Intelligent &
Robotic Systems, 2017, 88(2-4), 619.
[13] Landing A Drone Aircraft With No Runway: Unique ’SkyHook’
Recovery System. [Online]
https://www.youtube.com/watch?v=J4uJ4yShEDA&t=2s
[14] A. Tsukerman, M. Weiss, T. Shima, D. Löbl, & F. Holzapfel. “Optimal
rendezvous guidance laws with application to civil autonomous aerial
refueling”. Journal of Guidance, Control, and Dynamics. 2018; 41(5):
1167-1174.
[15] A. Rucco, P. B. Sujit, A. P. Aguiar, J. B. De Sousa, & F. L. Pereira.
“Optimal rendezvous trajectory for unmanned aerial-ground vehicles”.
IEEE Transactions on Aerospace and Electronic Systems. 2017; 54(2):
834-847.
[16] T. Muskardin, G. Balmer, L. Persson, S. Wlach, M. Laiacker, A. Ollero
& K. Kondak. “A novel landing system to increase payload capacity
and operational availability of high altitude long endurance UAVs”.
Journal of Intelligent & Robotic Systems, 2017, 88(2), 597-618.
[17] L. Persson, T. Muskardin, & B. Wahlberg. (2017, December). “Coop-
erative rendezvous of ground vehicle and aerial vehicle using model
predictive control”. 56th Annual Conference on Decision and Control
(CDC), 2017, (pp. 2819-2824). IEEE.
[18] R. W. Beard and T. W. McLain, “Small Unmanned Aircraft: Theory
and Practice”, 1st ed. Princeton, NJ: Princeton Univ. Press, 2012.
[19] R. Rajamani, “Vehicle dynamics and control”. Springer Science &
Business Media 2011.