Proceeding Paper

Particle Swarm Optimization for Target Encirclement by a

UAV Formation †
Tagir Muslimov

Ufa State Aviation Technical University, Ufa 450077, Russia; [email protected]

† Presented at the 15th International Conference “Intelligent Systems” (INTELS’22), Moscow, Russia, 14–16
December 2022.

Abstract: This paper presents an idea of using particle swarm optimization (PSO) to tune the control
system of a decentralized unmanned aerial vehicle (UAV) formation. Simulations were run on
a consensus-based decentralized UAV formation. Vector field guidance was used to control the
formation. A fitness function is proposed that is based not only on the error of distance to the
circular path, but also on the relative inter-UAV distance error. To demonstrate the effectiveness of
the proposed method, the obtained results of such tuning are compared to those obtainable by the
conventional trial and error method.

Keywords: UAV formation flight; collective circumnavigation; target tracking; vector field guidance;
drone flocking

1. Introduction
Use of autonomous robots in groups is a promising area of research in today’s mobile
robotics, and it receives much attention. Decentralized control of autonomous robots is
one of the more complex yet effective approaches. Thus, many papers cover decentralized
control applications in ground-based robots [1–3] and autonomous unmanned aerial
vehicles (UAVs) alike [4,5]. Most papers cover decentralized control of rotary-wing UAV
groups, mainly quadcopter formations [6,7].
Currently, control of decentralized swarms of autonomous robots [8–10] and un-
manned aerial vehicles (UAVs) [11,12] is a promising area of research. However, beside con-
Citation: Muslimov, T. Particle
trol and trajectory-planning algorithms, these formations require optimizing their transient
Swarm Optimization for Target
trajectories. Since autonomous robot formations, including UAV formations, are complex
Encirclement by a UAV Formation.
nonlinear interconnected systems, soft computing could be effective for such optimization.
Eng. Proc. 2023, 33, 15. https://
Control optimization research focuses on evolutionary algorithms [13] including
doi.org/10.3390/engproc2023033015
genetic algorithms [14,15] and particle swarm optimization [16,17]. Some papers show
Academic Editors: Askhat Diveev, implementation of particle swarm optimization for single-quadcopter controllers [18–20],
Ivan Zelinka, Arutun Avetisyan and UAV formations [21–23], UAV trajectory optimization [24], UAV movement planning [25]
Alexander Ilin and UAV formations [26] in an uncertain environment. However, swarm optimization
Published: 9 June 2023
for vector field-controlled UAV formations remains under-researched. Thus, the goal
hereof is to test the feasibility of applying particle swarm optimization to a decentralized
consensus-based UAV formation controlled via vector field guidance.

Copyright: © 2023 by the author. 2. Preliminary Remarks and Statement of Problem

Licensee MDPI, Basel, Switzerland. The assumptions here are the same as in [12,27]: no wind, inter-UAV communica-
This article is an open access article tion enabled and sufficiently accurate computation of the relative inter-UAV distance in
distributed under the terms and
the group.
conditions of the Creative Commons
The dynamics of a decentralized UAV formation can be tested via full models as well
Attribution (CC BY) license (https://
as via high-level models that approximate the movement of the formation provided that the
creativecommons.org/licenses/by/
UAVs are equipped with fine-tuned onboard autopilots. Full models are preferable, e.g., for
4.0/).

Eng. Proc. 2023, 33, 15. https://doi.org/10.3390/engproc2023033015 https://www.mdpi.com/journal/engproc

Eng. Proc. 2023, 33, 15 2 of 8

final testing of control algorithms or for testing formation-wide stability. High-level models,
also referred to in the literature as guidance models, are more suitable for simulating spatial
movement planning algorithms as well as for trajectory optimization. However, full models
are still useful when the trial and error method is used to find the initial values for the
formation controller coefficients that are further to be used for trajectory optimization.
UAV formation trajectory optimization is a subtask of cooperative target tracking.
This task is sometimes referred to as collective circumnavigation or target encirclement.
The idea is to maintain a certain preset distance not only between the UAVs (through
specified angular values) but also to the target encirclement orbit, which is a moving path.
Formal statement of the problem can be found in Section 4. We covered a similar problem
in [28], where we used a genetic algorithm to solve it.

3. Consensus-Based UAV Formation Control Algorithm

This paper uses a decentralized neighbor-to-neighbor interaction topology, where
UAVs only receive data from the neighboring aircraft. The topology can be defined as a
graph referred to as the “interaction graph” hereinafter. Paper [12] shows a mathematical
model of interaction in a consensus-based UAV formation. The same model is described
later in this section. Let N be the set of all UAV agents.
Let eθ ∈ R N ×1 be this vector, where R N ×1 is a space of N × 1-dimensional matrices
with components from R. To find this vector,
use some elements of the vector of all possible
relative phase shift angle errors ēθ = êi, j ∈ R N −1)×1 , where êi, j is the value of error for
N (

the directly interacting ith and jth agents. The choice is dictated by the interaction architec-
ture; in this research, the control action vector is set as such for open-chain interaction in
the same manner as described in [12,27]:

e1 ê12
   
 ..   .. 
 .   . 
 = M̂θ ^
   
eθ =  e
 k  
 =  − ê k −1, k + êk, k +1  e θ + D, (1)
 .   ..
 ..  

. 
eN −ê N −1, N
 T T
where D = −Mθ H−
θ
1
Pdθ , P̂θ is a system control vector in the space of relative
distances (an ( N − 1)-dimensional space generated by the interaction graph incidence
matrix columns), and Hθ is a matrix that specifies the agents for agent-to-agent distance
measurements, defined as follows:
 .. T
  .  T
q1   1
 1 
 1
 q2  
..
 
Hθ =  , qi =   , i < N, q N =  ..  ,
     
.. .
 .     . 
 −1 
qN 
..
 1
.

where Hθ ∈ R N × N , qi ∈ R1× N and the positions of “1” and “-1” in qi are determined
according to the structure of the interaction graph.
N
Pdθ ∈ R( N −1)×1 is the vector of the desired inter-UAV phase shift angles and P̂θ = ∑ ϕk
k =1
is the totalof the 
current UAV phase angles in an inertial coordinate system;
^ ^
eθ = e i,i+1 ∈ R( N −1)×1 is the vector of current phase shift angles for directly
i =1,N −1
co-engaged agents, calculated by the triple scalar product, e.g., when the final movement is
directed clockwise, the following applies:
Eng. Proc. 2023, 33, 15 3 of 8

^ (d , d ) ^
If n · (di × di+1 ) ≥ 0, then e i, i+1 = β = arccos kd ikkdi+1 k and e i, i+1 = 2π − β in other
i i +1
cases, where dk , k ∈ N is the vector of aircraft-to-moving-target distance at a given time,
n = (0, 0, 1)T ;
Mθ ∈ R N × N is an interaction matrix that in cases of decentralized neighbor–neighbor
interactions as herein is as follows:

−1 1 0 ··· 0
 
. .. .. 
 1 −2 . . .

 . 

Mθ =  0
 . .. . .. 1 
;
 0  
 . ..
 ..

. 1 −2 1 
0 ··· 0 1 −1

M̂θ ∈ R N ×( N −1) is a matrix derived from the matrix Mθ H− 1

θ by removing the Nth column.
For collective target encirclement, this paper uses the same control laws as in [12,27].
For control based on angular errors, it uses the speeds of the UAVs in the formation.
The following control command vector vc is set for UAV speeds:
T
vc =

v1 v2 ... vN = v1 N + L, (2)
T
R N ×1

where

1N = 1 1 ...

1 ∈ and the vector
R N ×1

L = v f (2/π ) arctan k θ ei + k θ̇ ėi ∈ is found given (1), k θ is the positive
i =1,N
tuning coefficient, k θ̇ is the positive tuning coefficient for the derivative signal, v f is the
maximum norm of the additional velocity vector that is to be adjusted for the constraints of
the real-UAV dynamics, and v is the ultimate linear cruise speed of the UAVs provided that
the target is stationary.
Path error-based control relies on the heading angles of the UAVs in the formation.
The following control law from [12,27] is applied to the heading-angle command vector Øc
with slightly modified coefficients:
 hπ  i
Ø c = ϕi + λ + arctan kio (di − ρ) + kiȯ d˙i ∈ R N ×1 , (3)
2 i =1,N

where di is the ith UAV-to-target distance, d˙i is the corresponding derivative signal, kio is the
tuning coefficient for the distance-to-circular-path signal for the ith UAV, kiȯ is the tuning
coefficient for the distance-to-circular-path derivative signal for the ith UAV, ρ is the radius
of the circular path that the UAV follows whilst encircling the target, ϕi is the phase angle
of circumvolution around the target for the ith UAV.

4. Implementation of Particle Swarm Optimization

The standard particleswarm solver from MATLAB2015b with default parameters
was used for particle swarm optimization. A four-UAV formation was tested for this paper,
hence eight tuning parameters. These parameters are coefficients in the control law (3):

K = [ k1o k1ȯ ... k4o k4ȯ ] ∈ R1×8 .

We also added upper and lower bounds as follows:

[ 0.1 0.1 . . . 0.1 0.1 ] ≤ K ≤ [ 30 30 . . . 30 30 ]

These constraints were chosen in order to preserve the UAV formation stability. Sta-
bility can be lost if the control law coefficients go beyond certain limits in the absence of
adaptive control.
Eng. Proc. 2023, 33, 15 4 of 8

The initial guess for the vector K was as follows:

Kinitial = [ 1 2 ... 1 2 ] ∈ R1 × 8 .

For the fitness function, we chose the following:

Z tn
∆
Ff itness =
0
t ∑ (|di − ρ| + |ei |) dt, (4)
i =1,...,4

where tn is the particle swarm optimization time; the remaining parameters are defined in
Equations (2) and (3). Thus, this solution optimizes not only for the error of each UAV’s
distance to the ultimate orbit of target encirclement but also for the relative neighbor-to-
neighbor distance errors.
The formal statement of the goal would be as follows:
Z tn
∆
minimize Ff itness =
0
t ∑ (|di − ρ| + |ei |) dt,
i =1,...,4

s.t. [ 0.1 0.1 . . . 0.1 0.1 ] ≤ K ≤ [ 30 30 . . . 30 30 ]

5. Simulation Results
5.1. Simulation Parameters
For simulation, we ran a high-level UAV model from [29]. The formation consisted
of four UAVs of the same type. To make an initial guess, we also ran full UAV models
of this formation. This allowed us to find, by trial and error, a controller coefficient that
would keep the entire formation system stable. Control laws (2) and (3) were used in the
simulation. The simulation parameters are shown in Table 1.

Table 1. Simulation parameters.

Parameter Symbol Values

 T
Initial UAV heading vector (rad) Ø (0) 0 0 0 0
 T
Initial UAV speed vector (m/s) v (0) 13 13 13 13
Circular path radius (m) ρ 200
 T  T
Initial target coordinates (m) c = c e (0) c n (0) h (0) 500 500 100
Ultimate cruise speed of the formation in case of a stationary v 13
target (m/s)
Target speed (m/s) vtarget 2
Target course angle (rad) χtarget π/4
 T
p1 (0), 180 910 −100 ,
 T
p2 (0), 245 247 −100 ,
Vector of initial UAV coordinates in the ICS (m) T
p3 (0),

710 135 −100 ,
p4 (0) 
915 760 −100
T
T
Desired inter-UAV phase shift angles (◦ ) Pdθ

270 260 290
Initial tuning coefficients k θ , k θ̇ , kio , kiȯ 35, 15, 1, 2

5.2. Simulation Results and Discussion

Figure 1 shows how the fitness function changed during optimization. Apparently,
the function’s value stopped declining drastically after 25 iterations. However, a signifi-
cantly increasing number of iterations would be required to further reduce the value.
Running the optimization algorithm returned the following K = Kopt values that were
used in the simulation:
 T
Kopt = 4.7426, 25.1807, 2.6214, 23.8265, 20.6034, 29.4922, 18.7241, 26.8984
Eng. Proc. 2023, 33, 15 5 of 8

×105 Best Function Value: 160129

1.8

1.78

1.76

1.74

1.72
Function value

1.7

1.68

1.66

1.64

1.62

1.6
0 5 10 15 20 25 30 35 40 45 50
Iteration

Figure 1. Fitness function value change during simulation.

Figure 2 shows UAV formation angular errors before and after optimization. As can
be seen in the graphs, optimization enabled the formation to reach the pre-specified relative
angular positions somewhat faster. Figure 3 shows how path errors changed in the UAV
formation. As can be seen from the graphs, UAV1 and UAV2 showed the most drastic
changes. Apparently, transient trajectories before and after optimization are different
(Figure 4). Even though the trajectories look similar in the figure, they are still different.
That is especially noticeable in the trajectories for UAV1 and UAV2.

(a) (b)
Between 1st and 2nd UAVs Between 1st and 2nd UAVs
40 30 Between 2nd and 3rd UAVs
Between 2nd and 3rd UAVs
Between 3rd and 4th UAVs Between 3rd and 4th UAVs
30 20
Angle errors [°]

20
Angle errors [°]

10
10
0
0
−10
−10

−20 −20

−30 −30
0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90
Time [s] Time [s]

Figure 2. UAV angular errors at time t = 90 s. (a) Angular errors before particle swarm optimization;
(b) angular errors after particle swarm optimization.
Eng. Proc. 2023, 33, 15 6 of 8

(a) 550
(b)
UAV No. 1 550 UAV No. 1
UAV No. 2 UAV No. 2
500 UAV No. 3 500 UAV No. 3
UAV No. 4 UAV No. 4
450 450
Path errors [m]

Path errors [m]

400 400

350 350

300 300

250 250

200 200

0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90
Time [s] Time [s]

Figure 3. UAV path errors at time t = 90 s. (a) Path errors before particle swarm optimization;
(b) path errors after particle swarm optimization.

(a) UAV No.1 trajectory UAV No.1 trajectory

UAV No.2 trajectory
(b) UAV No.2 trajectory
UAV No.3 trajectory UAV No.3 trajectory
UAV No.4 trajectory UAV No.4 trajectory
Vector field for UAV No.1 Vector field for UAV No.1
Target trajectory Target trajectory
Altitude position [m]

Altitude position [m]

200 200

1000 1000
100 100
800 800

0 0 600
600
200 200
400 400 400 400
600 600
800 200 800 200
1000 1000
North-South position [m] North-South position [m]
East-West position [m] East-West position [m]

Figure 4. UAV formation trajectories at time t = 90 s. (a) Trajectories before particle swarm
optimization; (b) trajectories after particle swarm optimization.

Notably, although the controller coefficients were tuned only for the control law (3),
the fitness function included the total angular error of the formation (4). The reason for
this was that control by path errors is tied to control by angular errors in a decentralized
UAV formation system. This connection can be seen, among other things, in the simulation
results in Figure 2.

6. Conclusions
The paper demonstrates a successful use of particle swarm optimization for target
encirclement and tracking by a UAV formation. The formation itself was a vector field-
controlled decentralized formation. Simulations showed a reduction in the proposed fitness
function as well as a change in the pattern of transient trajectories. A close connection
was found between optimizing the path error controller optimization and the quality of
transient trajectories for angular errors in the UAV formation.

Funding: This work was supported by the Ministry of Science and Higher Education of the Russian
Federation (Agreement No. 075-15-2021-1016).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Eng. Proc. 2023, 33, 15 7 of 8

Conflicts of Interest: The authors declare no conflict of interest.

References
1. Darintsev, O.V.; Yudintsev, B.S.; Alekseev, A.Y.; Bogdanov, D.R.; Migranov, A.B. Methods of a Heterogeneous Multi-agent Robotic
System Group Control. Procedia Comput. Sci. 2019, 150, 687–694. [CrossRef]
2. Ivanov, D.Y. Distribution of roles in groups of robots with limited communications based on the swarm interaction. Procedia
Comput. Sci. 2019, 150, 518–523. [CrossRef]
3. Veselov, G.; Sklyrov, A.; Mushenko, A.; Sklyrov, S. Synergetic Control of a Mobile Robot Group. In Proceedings of the 2014 2nd
International Conference on Artificial Intelligence, Modelling and Simulation, Madrid, Spain, 18–20 November 2014; pp. 155–160.
[CrossRef]
4. Bennet, D.J.; McInnes, C.R.; Suzuki, M.; Uchiyama, K. Autonomous Three-Dimensional Formation Flight for a Swarm of
Unmanned Aerial Vehicles. J. Guid. Control Dyn. 2011, 34, 1899–1908. [CrossRef]
5. Milyakov, D.A.; Merkulov, V.I. The Approach to Managing a Group of UAVs as a System with Distributed Parameters. Procedia
Comput. Sci. 2019, 150, 39–45. [CrossRef]
6. Diveev, A.I.; Shmalko, E.Y.; Hussein, O. Synthesized Optimal Control of Group Interaction of Quadrocopters Based on Multi-Point
Stabilization. Instrum. Eng. 2020, 133, 114–133. [CrossRef]
7. Titkov, I.P.; Karpunin, A. Collision-aware formation assignment of quadrotors. Procedia Comput. Sci. 2021, 186, 727–735. [CrossRef]
8. Pavlovskii, V.E.; Pavlovskii, V.V. A mathematical model of a 2D homogeneous swarm of robots. Sci. Tech. Inf. Process. 2016, 43,
306–314. [CrossRef]
9. Senotov, V.D.; Aliseychik, A.P.; Pavlovsky, E.V.; Podoprosvetov, A.V.; Orlov, I.A. Algorithms for swarm decentralized motion
control of group of robots with a differential drive. Keldysh Inst. Prepr. 2020, 123, 1–39. [CrossRef]
10. Zakiev, A.; Tsoy, T.; Magid, E. Swarm robotics: Remarks on terminology and classification. In Proceedings of the Interactive
Collaborative Robotics: Third International Conference, ICR 2018, Leipzig, Germany, 18–22 September 2018; pp. 291–300.
11. Ollervides-Vazquez, E.J.; Rojo-Rodriguez, E.G.; Garcia-Salazar, O.; Amezquita-Brooks, L.; Castillo, P.; Santibañez, V. A sectorial
fuzzy consensus algorithm for the formation flight of multiple quadrotor unmanned aerial vehicles. Int. J. Micro Air Veh. 2020,
12, 1756829320973579. [CrossRef]
12. Muslimov, T.Z.; Munasypov, R.A. Multi-UAV cooperative target tracking via consensus-based guidance vector fields and fuzzy
MRAC. Aircr. Eng. Aerosp. Technol. 2021, 93, 1204–1212. [CrossRef]
13. Diveev, A.; Shmalko, E. Hybrid evolutionary algorithm for synthesized optimal control problem for group of interacting robots.
In Proceedings of the 2019 6th International Conference on Control, Decision and Information Technologies (CoDIT), Paris,
France, 23–26 April 2019; pp. 876–881. [CrossRef]
14. Bożko, A.; Ambroziak, L.; Pawluszewicz, E. Genetic Algorithm for Parameters Tuning of Two Stage Switching Controller for
UAV Autonomous Formation Flight. Adv. Intell. Syst. Comput. 2021, 1390, 154–165.
15. Kim, M. Error Dynamics-Based Guidance Law for Target Observation using Multiple UAVs with Phase Angle Constraints via
Evolutionary Algorithms. J. Control Autom. Electr. Syst. 2021, 32, 1510–1520. [CrossRef]
16. Ali, Z.A.; Zhangang, H. Multi-unmanned aerial vehicle swarm formation control using hybrid strategy. Trans. Inst. Meas. Control
2021, 43, 2689–2701. [CrossRef]
17. Ali, Z.A.; Han, Z.; Masood, R.J. Collective Motion and Self-Organization of a Swarm of UAVs: A Cluster-Based Architecture.
Sensors 2021, 21, 3820. [CrossRef] [PubMed]
18. Derrouaoui, S.H.; Bouzid, Y.; Guiatni, M. PSO Based Optimal Gain Scheduling Backstepping Flight Controller Design for a
Transformable Quadrotor. J. Intell. Robot. Syst. 2021, 102, 67. [CrossRef]
19. Kiyak, E. Tuning of controller for an aircraft flight control system based on particle swarm optimization. Aircr. Eng. Aerosp.
Technol. 2016, 88, 799–809. [CrossRef]
20. Saribas, H.; Kahvecioglu, S. PSO and GA tuned conventional and fractional order PID controllers for quadrotor control. Aircr.
Eng. Aerosp. Technol. 2021, 93, 1243–1253. [CrossRef]
21. Biantoro, N.; Halim, M.; Nazaruddin, Y.Y.; Juliastuti, E. PSO-based Optimization of Formation Control and Obstacle Avoidance
for Multiple Quadrotors. In Proceedings of the 2021 International Conference on Instrumentation, Control and Automation (ICA),
Bandung, Indonesia, 25–27 August 2021; pp. 133–137. [CrossRef]
22. Hoang, V.T.; Phung, M.D.; Dinh, T.H.; Zhu, Q.; Ha, Q.P. Reconfigurable Multi-UAV Formation Using Angle-Encoded PSO. In
Proceedings of the 2019 IEEE 15th International Conference on Automation Science and Engineering (CASE), Vancouver, BC,
Canada, 22–26 August 2019; pp. 1670–1675. [CrossRef]
23. Sruthy, A.N.; Jacob, J.; Ramch, R. PSO Based Integral Backtepping Control for Leader-Follower Quadrotors. In Proceedings
of the 2020 Fourth International Conference on Inventive Systems and Control (ICISC), Coimbatore, India, 8–10 January 2020;
pp. 466–471. [CrossRef]
24. Vijayakumari, D.M.; Kim, S.; Suk, J.; Mo, H. Receding-Horizon Trajectory Planning for Multiple UAVs Using Particle Swarm
Optimization; AIAA Scitech 2019 Forum; American Institute of Aeronautics and Astronautics: Reston, VA, USA, 2019. [CrossRef]
25. Patley, A.; Bhatt, A.; Maity, A.; Das, K.; Ranjan Kumar, S. Modified Particle Swarm Optimization Based Path Planning for Multi-Uav
Formation; AIAA Scitech 2019 Forum; American Institute of Aeronautics and Astronautics: Reston, VA, USA, 2019. [CrossRef]
Eng. Proc. 2023, 33, 15 8 of 8

26. Skrzypecki, S.; Tarapata, Z.; Pierzchała, D. Combined PSO Methods for UAVs Swarm Modelling and Simulation. In Proceedings
of the Modelling and Simulation for Autonomous Systems: 6th International Conference, MESAS 2019, Palermo, Italy, 29–31
October 2019; pp. 11–25.
27. Muslimov, T.Z.; Munasypov, R.A. Coordinated UAV Standoff Tracking of Moving Target Based on Lyapunov Vector Fields. In
Proceedings of the 2020 International Conference Nonlinearity, Information and Robotics (NIR), Innopolis, Russia, 3–6 December
2020; pp. 1–5. [CrossRef]
28. Muslimov, T. Application of Genetic Algorithm for Vector Field Guidance Optimization in a UAV Collective Circumnavigation
Scenario. In Robotics in Natural Settings: CLAWAR 2022; Springer International Publishing: Cham, Switzerland, 2022.
29. Beard, R.W.; McLain, T.W. Small Unmanned Aircraft: Theory and Practice; Princeton University Press: Princeton, NJ, USA, 2012.

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