TESIS DOCTORAL
Guidance Navigation and Control
Algorithms for High Dynamics Vehicles
Autor:
Ra´
ul de Celis Fern´
andez
Director:
Luis Cadarso Morga
Programa de Doctorado en TECNOLOG´IAS DE LA INFORMACI ´ON Y LAS COMUNICACIONES
Escuela Internacional de Doctorado
Abstract
Accuracy and precision are the cornerstone for ballistic projectiles from the earliest days of
this discipline. In the beginnings, impact point precision in artillery devices deteriorated
when range was extended, particularly for ballistic artillery rockets and shells, which are
not propelled except during the launch. Later, inertial navigation and guidance systems
were introduced and precision was unlinked from range increases. In the last thirty
years, hybridization between inertial systems and GNSS devices has improved precision
enormously.
Unfortunately, during the last stages of flight, inertial and GNSS methods (hybridized
or not) feature big errors in attitude and position determination. Low cost devices, which
are precise on terminal guidance and do not feature accumulative error, such as quadrant
photo-detector, seem to be appropriate to be included in the guidance systems. Hybrid
algorithms, which combine GNSSs, IMUs and photo-detectors, are required to implement
these novel techniques.
The acceleration autopilot with a rate loop is the most commonly implemented
au-topilot, which has been extensively applied to high-performance missiles. Nevertheless,
for high speed spinning rockets, the design of the guidance and control modules is a
chal-lenging task because the rapid spinning of the body creates a heavy coupling between the
normal and lateral rocket dynamics.
Hybridized measurements are implemented in modified proportional navigation law
par-ticularized for a high spinning ballistic rocket, has been developed to perform simulations
to prove the accuracy of the presented algorithms.
The research process developed to obtain the final results implied the following steps:
1. The development of a flight model in order to simulate the dynamics of a highly
spinning rocket which features a decoupled fuse.
2. The development of a novel 3D guidance law, based on a modular rotatory force, for
gyroscopically stabilized artillery rockets (i.e., spin rates in the hundreds of rotations
per second during the launch), which is derived from proportional navigation.
3. A model for a quadrant photo-detector based on a real-time area intersection
algo-rithm was developed and the subsequent development of a novel algoalgo-rithm which
improves the precision of spot center determination for a Semi-active Laser quadrant
detector in the terminal guidance of artillery rockets.
4. The integration of this photo-detector spot center determination algorithm and the
hybridization with GNSS/IMU in order to improve the precision of the line of sight.
5. The development of an algorithm, based on an estimation method in order to obtain
the gravity and velocity vectors in a different pair of triads, which aims at avoiding
gyroscopes for attitude determination.
6. The integration of these attitude determination methods together with the aid of
filtering techniques, into the previously photo-detector, GNSS, IMU, and control
rotatory force, developed algorithms.
Finally, nonlinear simulations based on ballistic rocket launches were performed to
demonstrate the applicability of the proposed solution for flight navigation, guidance and
control, for ballistic rocket terminal guidance.
Resumen
La exactitud y la precisi´on han sido la piedra angular de los proyectiles bal´ısticos desde los
primeros d´ıas de esta disciplina. En los comienzos, la precisi´on del punto de impacto en los
dispositivos de artiller´ıa se deterioraba a mediada que el alcance del proyectil se extend´ıa,
especialmente para aquellos cohetes y proyectiles bal´ısticos que no est´an propulsados
excepto en el lanzamiento. Posteriormente, fueron introducidas la navegaci´on inercial y
los sistemas de guiado y consecuentemente la precisi´on se deslig´o del alcance. En los
´
ultimos treinta a˜nos, la hibridaci´on entre dispositivos GNSS y los sistemas inerciales ha
aumentado la precisi´on enormemente.
Desafortunadamente, durante las ´ultimas etapas del vuelo, los m´etodos inerciales y
GNSS (hibridizados o no) desencadenan grandes errores en la determinaci´on tanto de la
posici´on como de la actitud. Dispositivos de bajo coste, que pueden ser precisos durante
el guiado terminal y que no est´an sometidos a errores acumulativos, tales como los
foto-detectores de cuadrante, parecen ser adecuados para ser incluidos dentro de los sistemas de
guiado. Se requiere, por tanto, el desarrollo de algoritmos de hibridaci´on que combinen
sensores de GNSS, IMUs y foto-detectores, para la implementaci´on de estas novedosas
t´ecnicas.
Por otra parte, el autopiloto giro-acelerom´etrico, es el tipo de autopiloto m´as com´un en
estos dispositivos, y ha sido ampliamente implementado en misiles de altas prestaciones.
Sin embargo, para cohetes con altas velocidades de rotaci´on, el dise˜no de los m´odulos de
alto acoplamiento entre la din´amica normal y lateral del cohete.
Se presenta como soluci´on un m´etodo en el que una hibridaci´on de medidas obtenidas
de los sensores son introducidas en una ley de navegaci´on proporcional en la que act´ua
una fuerza rotatoria. Para validar la precisi´on de los algoritmos implementados se ha
desarrollado un modelo no lineal de mec´anica de vuelo, particularizado para un cohete
bal´ıstico de alta rotaci´on.
El proceso de investigaci´on se desarroll´o siguiendo los pasos mostrados a continuaci´on:
1. Desarrollo de un modelo de mec´anica de vuelo que simula el comportamiento f´ısico
de un cohete bal´ıstico de alta rotaci´on con una espoleta desacoplada.
2. Desarrollo de una novedosa ley de guiado tridimensional, basada en una fuerza
rotatoria modulable, para cohetes de artiller´ıa estabilizados girosc´opicamente (tasas
de rotaci´on de cientos de revoluciones por segundo en el lanzamiento), derivada de
la navegaci´on proporcional.
3. Realizaci´on de un modelo para un foto-detector de cuadrante basado en un m´etodo
de intersecci´on de ´areas, calculado en tiempo real, que desemboc´o en el consecuente
desarrollo de un algoritmo para la mejora de la precisi´on en la determinaci´on del
centro de la huella de un l´aser en un foto-detector de cuadrante para el guiado de
cohetes bal´ısticos de artiller´ıa
4. La integraci´on e hibridaci´on del foto-detector y el algoritmo de determinaci´on de la
posici´on del centro, junto con sensores GNSS e IMUs con el objetivo de mejorar la
precisi´on en la determinaci´on de la l´ınea de mira.
5. El desarrollo de un algoritmo, basado en la estimaci´on del vector gravedad y el
vector velocidad en una pareja de triedros diferentes, con el objetivo de eliminar los
gir´oscopos en la determinaci´on de la actitud del veh´ıculo.
6. La integraci´on de esos m´etodos de determinaci´on de actitud, con la ayuda de t´ecnicas
de filtrado, en el desarrollo previo de algoritmos h´ıbridos del foto-detector, GNSS e
IMU y la fuerza rotatoria de control.
Finalmente, se realizaron simulaciones no lineales basadas en lanzamientos de cohetes
bal´ısticos para demostrar la aplicabilidad de la soluci´on propuesta para la navegaci´on,
control y gu´ıado terminal del cohete bal´ıstico.
Acknowledgments
First, I would like to thank my research adviser, Professor Luis Cadarso Morga. This
research would have not been possible without his advice and contributions. He was
always available to discuss the research related topics.
I am fortunate to have worked with the researchers in ’Instituto Tecnologico La
Mara-nosa’ at INTA and for the help, data and solid modeling they provided in the very early
stages of this research.
Special thanks to Manuel Fuentes Gonzalez, who helped me in the first stages of the
development of the non-linear simulation model.
I have to thank the Department of Signal Theory, Communications, Telematic systems
and Computation from the Higher Technical School of Telecommunications Engineering
from Universidad Rey Juan Carlos, Higher Technical School of Telecommunications
En-gineering from Universidad Rey Juan Carlos, the Universidad Rey Juan Carlos and the
European Institute For Aviation Training And Accreditation from Universidad Rey Juan
Carlos for their support.
I also want to thank my friends who have suggested me minor and/or major changes
to this dissertation.
Finally, my deep and sincere gratitude to my family for their continuous and
unpar-alleled love, help and support. I am grateful to Johanna for always being there and for
putting up with me all these years. I am forever indebted to my parents for giving me the
me to explore new directions in life and seek my own destiny. This journey would not
have been possible if not for them, and I dedicate this milestone to them.
Contents
Abstract iv
Resumen ix
Acknowledgments xiii
I
INTRODUCTION
3
1 GNC Algorithms for High Rotating Speed Artillery Vehicles 5
1.1 Introduction . . . 5
1.2 Navigation, Guidance and Control . . . 8
2 Thesis Outline 11
2.1 Chapter 3: Thesis Contributions . . . 12
2.2 Chapter 4: Flight Dynamics, Navigation, Guidance and Control for High
Dynamic Rotating Artillery Rockets . . . 12
2.3 Chapter 5: Algorithms for Spot-Center Determination in Semi-Active
Laser-Detector for Terminal-Guidance . . . 13
2.4 Chapter 6: GNSS/IMU Laser Quadrant Detector Hybridization Techniques
2.5 Chapter 7: Aircraft Attitude Determination Algorithms through
Accelerom-eters, GNSS Sensors and Gravity Vector Estimator . . . 15
2.6 Chapter 8: Hybridized Attitude Determination Techniques to Improve
Bal-listic Projectile Navigation, Guidance and Control . . . 15
2.7 Chapter 9: Thesis Conclusions . . . 16
2.8 Chapter 10: Future Research . . . 16
3 Thesis Contributions 17
3.1 Chapter 4: Flight Dynamics, Navigation, Guidance and Control for High
Dynamic Rotating Artillery Rockets . . . 17
3.2 Chapter 5: Algorithms for Spot-Center Determination in Semi-Active
Laser-Detector for Terminal-Guidance . . . 18
3.3 Chapter 6: GNSS/IMU Laser Quadrant Detector Hybridization Techniques
for Artillery Rocket Guidance . . . 19
3.4 Chapter 7: Aircraft Attitude Determination Algorithms through
Accelerom-eters, GNSS Sensors and Gravity Vector Estimator . . . 20
3.5 Chapter 8: Hybridized Attitude Determination Techniques to Improve
Bal-listic Projectile Navigation, Guidance and Control . . . 21
II
GUIDANCE AND CONTROL FOR HIGH DYNAMIC
ROTATING ARTILLERY ROCKETS
23
4 Flight Dynamics, Navigation, Guidance and Control for High Dynamic
Rotating Artillery Rockets 25
4.1 Introduction . . . 25
4.2 Rocket Flight Dynamic Model . . . 28
4.2.1 Rocket . . . 28
4.2.2 Flight Dynamic Model . . . 30
4.3 Navigation and Guidance Law . . . 35
4.4 Control System . . . 37
4.5 Simulation Results . . . 38
4.5.1 Example Trajectory . . . 39
4.5.2 Monte Carlo Simulations . . . 40
III
LASER QUADRANT PHOTO-DETECTOR SPOT-CENTER
DETERMINATION AND HYBRIDIZATION TECHNIQUES
FOR ARTILLERY ROCKET TERMINAL-GUIDANCE
49
5 Algorithms for Spot-Center Determination in Semi-Active Laser-Detector for Terminal-Guidance 51 5.1 Introduction . . . 515.2 Problem Description . . . 54
5.3 Semi-Active Laser Sensor Model . . . 56
5.3.1 Spot and Photo-detector Area Intersection . . . 58
5.3.2 Spot Center Position Estimation . . . 60
5.3.3 Interpolation Improvement . . . 70
5.4 Simulation Results . . . 72
6 GNSS/IMU Laser Quadrant Detector Hybridization Techniques for Ar-tillery Rocket Guidance 87 6.1 Introduction . . . 87
6.2 System Modeling . . . 91
6.2.1 Rocket Definition . . . 91
6.2.3 Sensors . . . 97
6.3 Navigation, Guidance and Control . . . 100
6.3.1 Navigation . . . 100
6.3.2 Guidance Law . . . 103
6.3.3 Control System . . . 105
6.4 Numerical Simulations . . . 108
6.4.1 Nominal Trajectories . . . 108
6.4.2 Monte Carlo Simulations . . . 109
6.4.3 Discussion . . . 110
IV
NON-GYROSCOPES AIRCRAFT ATTITUDE
DETER-MINATION ALGORITHMS AND HYBRIDIZATION
TECH-NIQUES FOR ARTILLERY ROCKET TERMINAL-GUIDANCE
113
7 Aircraft Attitude Determination Algorithms through Accelerometers, GNSS Sensors and Gravity Vector Estimator 115 7.1 Introduction . . . 1157.2 Problem Description . . . 118
7.2.1 Triad Definition . . . 118
7.2.2 Involved Vectors Determination . . . 119
7.3 Gravity Vector Estimation . . . 120
7.4 Attitude Determination Algorithm . . . 125
7.5 Flight Dynamics and Sensor Model . . . 128
7.6 Simulation Results . . . 129
7.7 Application of algorithms on an Unmanned Air Vehicule . . . 139
7.7.1 UAV Definition . . . 139
7.7.2 Flight Dynamics and Sensor Model for UAV application . . . 142
7.7.3 Simulation Results for UAV application . . . 144
8 Hybridized Attitude Determination Techniques to Improve Ballistic Pro-jectile Navigation, Guidance and Control 155 8.1 Introduction . . . 155
8.2 System Modeling . . . 159
8.2.1 Rocket . . . 159
8.2.2 Flight Dynamics Model . . . 160
8.2.3 Sensors . . . 161
8.3 Navigation, Guidance and Control . . . 162
8.3.1 Navigation . . . 162
8.3.2 Guidance Law . . . 165
8.3.3 Control System . . . 166
8.4 Numerical Simulations . . . 166
8.4.1 Nominal Trajectories . . . 167
8.4.2 Monte Carlo Simulations . . . 167
8.4.3 Discussion . . . 168
V
CONCLUSIONS, FUTURE RESEARCH &
BIBLIOGRA-PHY
173
9 Thesis Conclusions 175 9.1 Chapter 4: Flight Dynamics, Navigation, Guidance and Control for High Dynamic Rotating Artillery Rockets . . . 1759.3 Chapter 6: GNSS/IMU Laser Quadrant Detector Hybridization Techniques
for Artillery Rocket Guidance . . . 177
9.4 Chapter 7: Aircraft Attitude Determination Algorithms through
Accelerom-eters, GNSS Sensors and Gravity Vector Estimator . . . 178
9.5 Chapter 8: Hybridized Attitude Determination Techniques to Improve
Bal-listic Projectile Navigation, Guidance and Control . . . 179
10 Future Research 181
BIBLIOGRAPHY 183
List of Figures
4.i Thrust curve depending of flight time. . . 29
4.ii Aerodynamic coefficients for the rocket under study depending of the Mach
number. . . 30
4.iii Reference systems for a 140 mm axisymmetric rocket with wrap around
fins and a roll-decoupled fuse. . . 31
4.iv A ballistic trajectory for an initial θ of 40 degrees. . . 35
4.v Initial elevation angle and ACorr depending on target distance to launch
point. . . 36
4.vi Controller scheme. . . 38
4.vii Example Trajectory. . . 39
4.viii Response of the rocket to different step commands in the control variable. . 40
4.ix Complete dispersion area of all the flights. . . 41
4.x Impact point dispersion patterns for ballistic and controlled flights for a
target distance onf 18500 m. . . 43
4.xi Impact point dispersion patterns for ballistic and controlled flights for a
target distance onf 19000 m. . . 44
4.xii Impact point dispersion patterns for ballistic and controlled flights for a
4.xiii Impact point dispersion patterns for ballistic and controlled flights for a
target distance onf 20000 m. . . 46
4.xiv Impact point dispersion patterns for ballistic and controlled flights for a
target distance onf 20500 m. . . 47
4.xv Histogram of number of impacts for each distance to the target. . . 48
5.i Standard (left) and cross (right) quadrant detector configuration. . . 56
5.ii Photo-detector Scheme . . . 57
5.iii Intensity registered under foggy laboratory conditions tests . . . 57
5.iv Error for Non-dimensional Intensity registered under foggy laboratory
con-ditions vs proposed area method . . . 58
5.v Illuminated area for spot-detector ratio of 0.25. . . 61
5.vi Illuminated area for spot-detector ratio of 0.50. . . 62
5.vii Illuminated area for spot-detector ratio of 0.75. . . 63
5.viii Illuminated area for spot-detector ratio of 1.00. . . 64
5.ix Photo-detector spot coordinates geometric composition . . . 65
5.x Transformation developed by method 1 . . . 66
5.xi Transformation developed by method 2 . . . 67
5.xii Transformation developed by method 3 . . . 68
5.xiii Transformation developed by method 4 . . . 69
5.xiv Example of trajectory and grid of 81 impact points . . . 73
5.xv Detector footprint for a shot angle of 30◦ with a spot-detector ratio of 0.25. 74 5.xvi Detector footprint for a shot angle of 30◦ with a spot-detector ratio of 0.50. 75
5.xviiDetector footprint for a shot angle of 30◦ with a spot-detector ratio of 0.75. 76
5.xviiiDetector footprint for a shot angle of 30◦ with a spot-detector ratio of 1.00. 77
5.xix Detector footprint for a shot angle of 45◦ with a spot-detector ratio of 0.25. 78
5.xx Detector footprint for a shot angle of 45◦ with a spot-detector ratio of 0.50. 79
5.xxi Detector footprint for a shot angle of 45◦ with a spot-detector ratio of 0.75. 80
5.xxiiDetector footprint for a shot angle of 45◦ with a spot-detector ratio of 1.00. 81
5.xxiiiDetector footprint for a shot angle of 60◦ with a spot-detector ratio of 0.25. 82
5.xxivDetector footprint for a shot angle of 60◦ with a spot-detector ratio of 0.50. 83
5.xxvDetector footprint for a shot angle of 60◦ with a spot-detector ratio of 0.75. 84
5.xxviDetector footprint for a shot angle of 60◦ with a spot-detector ratio of 1.00. 85
6.i 140 mm axisymmetric spinning rocket with wrap around fins, decoupled 2
by 2, a roll-decoupled fuse and its actuation force. . . 91
6.ii Reference systems. . . 93
6.iii Ballistic trajectories for initial pitch angles (θ0) of 20, 30 and 45 degrees. . 94
6.iv Incidence aerodynamic speed decomposition, local angle of attack (αi), and
fin deflection (δi). . . 95
6.v dax(M) and dlat scheme. . . 97
6.vi Quadrant photo-detector configuration used. . . 99
6.vii Results for hybridization algorithm. . . 103
6.viii Control system scheme. . . 107
6.ix Ballistic shots for 20◦, 30◦ and 45◦ initial pitch angles. . . 110
6.x Detailed shots for different algorithms. . . 111
7.i Trajectories for initial pitch shot angle of 15◦, 45◦ and 75◦ and 8 different azimuths. . . 128
7.ii Comparison between estimated and real magnitudes for Euler Angles for a
shot angle of 45◦ and azimuths of 0◦ (left), 45◦ (center) and 90◦ (right). . . 130
7.iii Comparison between estimated and real magnitudes for gravity vector for
7.iv Comparison between estimated and real magnitudes for aircraft angular
velocity for a shot angle of 45◦ and azimuths of 0◦ (left), 45◦ (center) and 90◦ (right). . . 132 7.v Representation of euler angles RMSE along the whole trajectory for each
shot angle and for the 8 azimuths tested. . . 135
7.vi Representation of gravity vector RMSE along the whole trajectory for each
shot angle and for the 8 azimuths tested. . . 136
7.vii Representation of zB angular velocity vector RMSE along the whole
tra-jectory for each shot angle and for the 8 azimuths tested. . . 137
7.viii Scheme of UAV modeled for Simulations. . . 140
7.ix Trajectories 1 (top-left), 2 (top-right), 3 (center-left), 4 (center-right), 5
(bottom-left) and 6 (bottom-right) . . . 143
7.x Comparison between estimated and real magnitudes for Euler Angles for
an inclination path angle of 5◦ and trajectories 1 (left), 3 (center) and 5 (right). . . 145
7.xi Comparison between estimated and real magnitudes for gravity vector for
an inclination path angle of 5◦ and trajectories 1 (left), 3 (center) and 5 (right). . . 146
7.xii Comparison between estimated and real magnitudes for aircraft angular
ve-locity for an inclination path angle of 5◦ and trajectories 1 (left), 3 (center) and 5 (right). . . 147
7.xiii Representation of Euler angles RMSE along the whole trajectory for each
inclination path angle and for the 6 trajectories tested. . . 150
7.xiv Representation of gravity vector RMSE along the whole trajectory for each
inclination path angle and for the 6 trajectories tested. . . 151
7.xv Representation of angular velocity vector RMSE along the whole trajectory
for each inclination path angle and for the 6 trajectories tested. . . 152
List of Tables
4.I Representative data of the studied rocket. . . 29
4.II Monte Carlo initial condition distribution parameters. . . 41
4.III The CEP for each of the targets and for ballistic and controlled flights. . . 42
5.I Relationship angles between ideal and calculated spot center position . . . 71
5.II Correlations between ideal and calculated radial components . . . 72
5.III Quadratic error for the combination of each shot angle, spot-detector ratio
and methods . . . 86
6.I 140 mm axisymmetric spinning rocket aerodynamic coefficients. . . 92
6.II Interpolation between measured radial distance, rquad, and real radial
dis-tance, rc. . . 99
6.III Nominal trajectories’ parameters. . . 109
6.IV Monte carlo simulation parameters. . . 109
6.V Values for the constants on each flight phase. . . 110
6.VI Circular Error Probable in different cases. . . 112
7.I Root Mean Squared Errors along the whole trajectory for a set of shot
angles and for the 8 azimuths tested and calculation of a representative
mean parameter. . . 138
7.III Root Mean Squared Errors along the whole trajectory for a set of
incli-nation path angles and for the 6 trajectories tested and calculation of a
representative mean parameter. . . 153
8.I 140 mm axisymmetric rocket main parameters versus time. . . 160
8.II Nominal trajectories’ parameters. . . 167
8.III Monte carlo simulation parameters. . . 168
8.IV Values for the constants on each flight phase. . . 169
8.V Circle error probable for the different cases. . . 171
Part I
Chapter 1
GNC Algorithms for High Rotating
Speed Artillery Vehicles
This chapter gives a brief overview about the role that GNC Algorithms research play in
High Rotating Speed Artillery Vehicles.
1.1
Introduction
Artillery is the set of weapons of war designed to fire large projectiles over long distances
using an explosive charge or a rocket as a driving element. By extension, the military unit
that manages them is also called artillery. Every artillery piece has a firearm, a metal
tube of a certain caliber and length and a frame where it rests.
The invention of gunpowder - together with that of another artifact closely linked to
the former - the cannon - would constitute the first milestone that would begin the history
of artillery, well differentiated from the history of mere siege devices. In Europe, there
are several references in the 14th century to the use of primitive artillery pieces by the
Arabs at the site of Baza, and it is known that the army of Alfonso XI used it in 1312 at
is evidence of the use of a cannon that used stone balls as ammunition. The ammunition
used until the seventeenth century consisted usually of balls of stone or metal, suitable for
breaking down walls or attacking ships at sea, but with very little effect on the infantry or
cavalry, other than scaring the horses. They were very dangerous weapons to use which,
often, the nobles preferred to take to the battlefield mainly to intimidate.
During the seventeenth century the artillery did not change too much, since it remained
a dangerous tool. These weapons were a hindrance for the generals, who often had to
use two trios of horses to carry them. This is why the guns would remain during the
eighteenth century as a weapon to disorganize enemy troops, rather than a weapon of
major destruction. Shortly after the Napoleonic wars, the howitzer appeared, a weapon
similar to the cannon but that allows for the first time what is called indirect shot in a
primitive form, that is, to attack positions that, being in the line of reach, are hidden by
elements of the land, walls, etc. thanks to that it allows inclinations of 45 deg or more.
In the second half of the nineteenth century, the artillery undergoes a revolution,
indi-rect shooting is generalized by topographic maps thanks to the improvement of shooting
control, using observers who have the position to beat in sight and that by telephone or
radio are providing to the command of the artillery the information to correct the shot.
These and not others, are the first attempts at improvement in the accuracy of artillery
weapons. In World War I, and thanks to the control of the recoil and the improvement of
the propulsion loads, artillery bombings are made at distances of more than 20 km and
even special cannons are mounted on railway rails that can bombard cities To 100 km of
distance, although the wear of the pieces is enormous and one must be changing the cane
continuously in this case.
From World War II until today, the main innovations have been the use of computers
to give a rapid calculation of the trajectory, including the embedded computing devices
as the subject that competes the development of this thesis, whereas before there was
to make several test shots and correct them, using observers if the target was at a great
distance.
In World War II rocket artillery appears, although it had previously been used in very
primitive forms, for example, in China since the thirteenth century, in India against the
British in the eighteenth century or Paraguay in the nineteenth century in its war against
the Triple Alliance. The British adopted the Congreve rocket as an incendiary weapon
and for its psychological effect on enemy rather than physical capabilities against the
infantry, at least at these times. In the nineteenth century rocket engineering continued
to study and improved especially so that after the launch maintained a regular trajectory
and increase its destructive capacity. Even in World War I, aviation rockets were used in
a limited way.
Rocket, unlike the missile, lacks a guidance system after its launch. It is used as
a weapon of saturation, to completely destroy an area, with heads of high explosive,
incendiary. For this purpose, several rockets are mounted on a rail or pipe guidance
system, and the whole assembly, on a moving vehicle or platform which is pointed at the
area to be destroyed and simultaneously fired by an electrical system. The classic Russian
Katiusha rockets of World War II, launched from platforms mounted on trucks are still
used today in modern versions, and they showed their potential by destroying a certain
range of firing. Even armies such as the United States, which for decades despised the
use of rockets as a crude weapon of antiquated armies, have in recent years incorporated
vehicles that allow them to launch a certain amount of rockets to saturate an area. It is
also that in recent years, the concept of asymmetric warfare has imposed cost reduction
and the consequent adaptation of these saturation rockets by intelligent rockets, equipped
1.2
Navigation, Guidance and Control
The purpose of a Guidance, Navigation and Control (GNC) system is to make the
ve-hicle move following a certain reference path, which meets certain requirements. These
requirements can be of different nature: impact on a target, arrival at a certain point,
orbit around a point, etc. To achieve this goal, the automatic GNC systems perform three
fundamental functions, which correspond to their acronyms:
- Guidance.
- Navigation.
- Control.
The guidance function, whose purpose is to calculate a reference state vector such that
the trajectory complies with what is required.
The navigation function is responsible for measuring or estimating the vehicle’s state
vector. Depending on the type of trajectory required, it will be necessary to know certain
kinematic variables of the vehicle in order to know the deviations that the actual trajectory
is undergoing with respect to the one of reference required and to be able to calculate the
necessary actions to approach them. Depending on the specific case, it will be necessary to
know the position, speed, acceleration or combinations of the components of the previous
ones. This function is developed by various devices, such as:
- Gyroscopes.
- Accelerometers.
- GPS Systems.
- Terrain Reference Systems (TRN or TFN).
- Doppler navigation systems.
- ...
The control function, which aims to make the state vector itself evolve so that it
approaches the reference. The control function commands these actions based on the
deviation that exists, at any moment, between the state vector and the reference vector.
The execution of these control actions will be carried out by the actuators available on
Chapter 2
Thesis Outline
This dissertation is divided in five main parts. The first part presents an introduction
to the the role that GNC Algorithms research play in High Rotating Speed Artillery
Vehicles; then, the thesis outline and contributions are presented. The second part of
this dissertation presents the guidance and control methods for high dynamic rotating
artillery rockets, including a dissertation on the mathematical model employed and the
presentation of the rotating force control method. The third part presents the algorithms
for spot-center determination in a GNSS/IMU laser quadrant photo-detector and its
hy-bridization techniques for artillery rocket terminal-guidance. The forth part, describes a
novel aircraft attitude determination algorithm which aims on avoiding gyroscopes and
the related hybridization techniques for artillery rocket terminal-guidance. Finally, the
fourth part enumerates the conclusions of the research presented in this dissertation and
the future research.
In the following sections, we briefly outline the contents of the chapters in this
2.1
Chapter 3: Thesis Contributions
This chapter briefly describes the main contributions to the literature of the research
presented in this dissertation.
2.2
Chapter 4: Flight Dynamics, Navigation,
Guid-ance and Control for High Dynamic Rotating
Ar-tillery Rockets
The acceleration autopilot with a rate loop is the most commonly implemented autopilot,
which has been extensively applied to high-performance missiles. However, for spinning
rockets, the design of the guidance and control modules is a challenging task because
the rapid spinning of the body creates a heavy coupling between the normal and lateral
rocket dynamics. Nonlinear modeling of the rocket dynamics, control design as well as
guidance algorithms are performed and discrete-time guidance and control algorithms for
the terminal phase, which is based in proportional navigation, are performed. Finally,
complete nonlinear simulations based on realistic scenarios are developed to demonstrate
the robustness of the proposed solution with respect to uncertain launch, environment
and rocket conditions. The performance of the proposed navigation, guidance and control
system for a high-spin rocket leads to significant reductions in impact point dispersion.
2.3
Chapter 5: Algorithms for Spot-Center
Determi-nation in Semi-Active Laser-Detector for
Terminal-Guidance
Precision of guided projectiles is dependent both on the precision of the measurement
system used for location determination and the precision in setting the coordinates of the
target. Development of algorithms for low-cost high-precision terminal guidance systems
is a cornerstone in research in this field. Semi Active Laser Kits (SAL), and particularly
quadrant detector devices, have been developed in order to improve precision in guided
weapons. Photo-detection system can be functionally divided into two main parts:
sens-ing and processsens-ing. The sensed signal is processed to estimate the spot coordinates, i.e.,
the laser footprint, and to obtain the needed information for the navigation and guidance
algorithms. Using an interpolation algorithm based on the four electrical intensities
ob-tained in a semi active laser quadrant photo-detector, laser footprint center estimation
is improved for artillery applications. The electrical intensities that real sensor provides
under laboratory conditions are compared to a mathematical model based on area
in-tersection calculations in order to simulate the intensities on real flights. From these
intensities, four different processing algorithms are tested for different spot sizes so as to
obtain correlation between real spot center position and estimated position calculated by
algorithms. Finally, an example illustrating a nonlinear real-life projectile is employed
to compare real and calculated laser footprints in order to select the best algorithm for
2.4
Chapter 6: GNSS/IMU Laser Quadrant
Detec-tor Hybridization Techniques for Artillery Rocket
Guidance
In the past, ballistic rockets impact point precision deteriorated at the same time as
range was extended, specially for those rockets which were non-propelled and non-guided
amid the greater part of their trajectories. Once that inertial and GNSS navigation and
guidance systems were introduced, precision was unlinked from range increments. The
fundamental issue from these inertial and GNSS strategies (hybridized or not) is the
enormous errors on attitude and position determination during last phases of flight as
the movement is governed by aerodynamic forces and moments. Consequently, the
aero-dynamic forces model has a deeply nonlinear character. Choosing another kind of low
cost sensors, independent of accumulative errors and precise on terminal guidance, for
example, quadrant photo-detector semi-active laser, is crucially essential. Hybridization
nonlinear algorithms, such as extended Kalman filter, joining measurements from sensors
such as Global Navigation Satellite System (GNSS), Inertial Measurement Units (IMUs)
and photo-detectors are described in this paper to be utilized on modified proportional
navigation techniques and novel control methods. The results are tested on rocket
non-linear flight simulations in order to prove accuracy of proposed algorithms.
2.5
Chapter 7: Aircraft Attitude Determination
Al-gorithms through Accelerometers, GNSS Sensors
and Gravity Vector Estimator
Aircraft and spacecraft navigation precision is dependent on the measurement system for
position and attitude determination. Rotation of an aircraft can be determined
measur-ing two vectors in two different reference systems. Velocity vector can be determined
in inertial reference frame from a GNSS-based sensor and by integrating the
accelera-tion measurements in body reference frame. Estimating gravity vector in both reference
frames, and combining with velocity vector, determines rotation of the body. A new
approach for gravity vector estimation is presented, and employed in an attitude
deter-mination algorithm. Nonlinear simulations demonstrate that, using GNSS sensors and
strap-down accelerometers, aircraft attitude determination is precise, especially in
ballis-tic projectiles, allowing for substitution of precise attitude determination devices, which
are usually expensive and forced to bear high solicitations as for instance G forces.
2.6
Chapter 8: Hybridized Attitude Determination
Techniques to Improve Ballistic Projectile
Navi-gation, Guidance and Control
Precise rotation determination is an expensive task in aircraft, as it is usually determined
by strap-down sensors such as fiber optic gyros or MEMS. Particularly in ballistic
projec-tiles, these gyro determination devices increase their price as they need to bear enormous
accelerations during the initial stages but not during the ballistic flight. A new approach to
At-titude determination methods and gravity vector estimation method, is presented in this
paper. Measurements of accelerometers, GNSS-sensors and Semi-Active photo-detectors
are hybridized to get such a result. The attitude determination method, avoiding the use
of gyroscopes, measures pairs of vectors, i.e., gravity, velocity and line of sight vectors, in
a pair of reference systems, i.e., body fixed and north-east-down reference frames.
Grav-ity vector estimation is based on flight mechanics of a ballistic projectile, but it may be
extrapolated to any aircraft, and later employed in an attitude determination algorithm.
Modified proportional navigation techniques and previously developed control methods
are employed during flight. The presented approach is tested on realistic non-linear flight
simulations to prove accuracy of proposed algorithms.
2.7
Chapter 9: Thesis Conclusions
This chapter presents the main conclusions of this dissertation for each chapter in a
separate way.
2.8
Chapter 10: Future Research
This chapter presents future research it is going to be embarked in the near future.
Chapter 3
Thesis Contributions
In the following, they are described the major contributions of this thesis to the existing
literature. they are also detailed the contributions of each chapter in a separate way.
3.1
Chapter 4: Flight Dynamics, Navigation,
Guid-ance and Control for High Dynamic Rotating
Ar-tillery Rockets
The main contributions of this chapter are the development of a flight model in order
to simulate the dynamics of a highly spinning rocket which features a decoupled fuse.
The The previous model is employed in the development of a novel 3D guidance law for
gyroscopically stabilized artillery rockets (i.e., spin rates in the hundreds of rotations per
second), which is derived from proportional navigation. This novel control law allows
the development of a simple but effective and robust single-input single-output controller,
which is able to handle the heavy coupling between the normal and lateral rocket
dy-namics. This new concept allows simple and effective control algorithms which enables
in this chapter on a 140 mm rocket equipped with a rotating force mechanism and
non-controlled thrust. The rotating force mechanism consists of a decoupled fuse from the aft
part of the rocket. The main advantage of the overall setup is that, on the one hand, it
maintains the inherent dynamic stability properties of a rapidly spinning body due to the
aft part, while at the same time, the front part, remains easy to be fit to any unguided
rocket, hence transforming it into a guided one. Nonlinear simulations based on
realis-tic scenarios are performed to demonstrate the robustness of the proposed solution with
respect to uncertain launch, environment and rocket conditions. The chapter proceeds
as follows. First, the nonlinear rocket dynamic model is defined. Second, an integrated
navigation and guidance approach is presented, followed by design of a controller.
Ex-ample of ballistic and controlled flight simulation results are presented, which analyze
performance.
3.2
Chapter 5: Algorithms for Spot-Center
Determi-nation in Semi-Active Laser-Detector for
Terminal-Guidance
The main contribution of this chapter is the development of an algorithm which improves
the precision of spot center determination for a SAL quadrant detector for terminal
guid-ance of artillery rockets. Also, a model for the quadrant detector based on real-time area
intersection algorithm is developed.
Relevant parameters which influence sensor precision and performance are determined
and studied. Also, sensitivity analysis is performed on some of them. It is shown that
the transformation from the real to the estimated spot footprint made by the methods is
conformal. Based on this, an interpolation algorithm is proposed to improve the
perfor-mances.
Nonlinear simulations based on ballistic rocket launches are performed to obtain this
ideal spot position and compare it to the position obtained from the interpolation
algo-rithms, and to demonstrate the applicability of the proposed solution for artillery final
flight stages guidance.
3.3
Chapter 6: GNSS/IMU Laser Quadrant
Detec-tor Hybridization Techniques for Artillery Rocket
Guidance
The main contribution of this chapter is the development and integration of a novel
algorithm, based on an Extended Kalman Filter (EKF) hybridization between GNSS/IMU
and semi-active laser quadrant photo-detectors, which improves the precision of line of
sight (the vector between rocket center of masses and target) determination during the
terminal guidance of ballistic rockets, and in consequence the precision on impact point.
Laser quadrant photo-detector footprint centroid calculation is developed based on a
cubic spline interpolation. Note that the advantage of such a combined system over the
individual GNSS/IMU, even when the error of GNSS/IMU is reduced to extremely small
quantity by virtue of some technique, is the ability to avoid jamming and also impact on
points in a vertical plane.
The algorithm is based on based on an EKF hybridization which determines the
ter-minal line of sight to be used on a modified proportional navigation law and on a rotatory
control technique. The proposed control approach is based on a robust double-input
double-output controller. This controller is able to handle the heavy coupling between
the normal and lateral rocket nonlinear dynamics. The use of a flight mechanics model,
which takes in account the non-linearity in aerodynamic forces and moments, with the
spinning rocket, demonstrates the precision and applicability of these algorithms under
uncertain launch, environment and rocket conditions.
3.4
Chapter 7: Aircraft Attitude Determination
Al-gorithms through Accelerometers, GNSS Sensors
and Gravity Vector Estimator
The main contributions of this chapter are twofold. First, the development of a novel
algorithm which aims at avoiding gyroscopes for attitude determination. The central
idea is to decrease attitude sensors costs and even to improve attitude determination
by applying filtering techniques, especially for artillery device and UAV applications,
where high acceleration conditions increase the cost of precise attitude determination
sensors such as gyroscopes or low cost requirements are imposed, respectively. Second, the
development of an estimation method in order to obtain the gravity vector in body axes.
This estimator is motivated by the need of having two vectors expressed in two different
triads in order to determine attitude changes. Flight nonlinear simulations are performed
to determine real attitude and compare it to the estimated one. The applicability of the
proposed solution for aircraft flight navigation, guidance and control, or for ballistic rocket
terminal guidance, where attack and side-slip angles or total angle of yaw are relatively
small, is also demonstrated.
3.5
Chapter 8: Hybridized Attitude Determination
Techniques to Improve Ballistic Projectile
Navi-gation, Guidance and Control
The main contributions of this chapter are the development of a new algorithm which
substitutes gyroscopes in favor of lower cost sensors for attitude determination and the
employment of an estimation method in order to obtain the gravity vector in body axes,
which is only based on the aerodynamic parameters of the cell and the measurements
provided by accelerometers. The objective is to get simplicity in attitude sensors and
even to increase precision by applying filtering techniques, especially for artillery device
purposes, where high solicitation acceleration conditions increase the price of precise
at-titude determination devices such as gyroscopes. In order to get high precision at impact
points, multiple sensors are employed and a hybridization algorithm is employed so as to
to handle information. Mixing inaccurate signals, e.g., from GNSS and accelerometers,
and precise signals, e.g., from semi-active laser quadrant detector, enables the
determina-tion of a high fidelity line of sight. Non-linear flight simuladetermina-tions are performed in order to
prove the applicability of the proposed approach for ballistic rocket navigation, guidance
Part II
GUIDANCE AND CONTROL FOR
HIGH DYNAMIC ROTATING
Chapter 4
Flight Dynamics, Navigation,
Guidance and Control for High
Dynamic Rotating Artillery Rockets
4.1
Introduction
Precision has always been recognized as an important attribute of weapon development.
One of the greatest advantages of the precision weapon is the confidence that it can
minimize ”‘collateral damage”’. In addition, using force in some circumstances, would
be either unacceptable or call into question the viability of continued military action in
absence of precision weapons (Hamilton, 1995). A precision-guided munition (PGM) is
a guided munition intended to precisely hit a specific target, and to minimize collateral
damage. Because the damage effects of explosive weapons decrease with distance, even
modest improvements in accuracy enable a target to be attacked with fewer or smaller
bombs. The precision of these weapons is dependent both on the precision of the
measure-ment system used for location determination and the precision in setting the coordinates
is accurate. If the targeting information is accurate, satellite-guided weapons (including
inertial navigation in the event of signal loss) are significantly more likely to achieve a
successful strike in any given weather conditions than any other type of precision-guided
munition. Development of low-cost navigation, guidance and control technologies for
unguided rockets is a unique engineering challenge. Over the past several decades,
nu-merous solutions have been proposed, primarily for large artillery projectiles or for slowly
rolling airframes. Guided projectiles may be divided into three categories in terms of
the control mechanisms employed: aerodynamic surfaces (Morrison, Amberntson, 1977),
(Rogers,Costello, 2010), jet thrusters (Burchett,Costello, 2002), (Lloyd,Brown, 1979), and
inertial loads (Murphy, 1981), (Ollerenshaw, Costello, 1981). Guided projectile concepts
involving aerodynamic control surfaces can be divided into two categories: fin-stabilized
and spin-stabilized. Spin-stabilized guided projectiles are generally equipped with a
roll-decoupled trajectory correction fuse, designed to provide trajectory correction, while at
the same time they rely on the high spin rate of the aft part for airframe stability. But
the high spin rate creates an important coupling between the normal and lateral axes of
the body, which makes the dynamic characteristic rather complex. For such projectiles,
previous work has proven that flight instabilities occur for spin-stabilized projectiles
ma-neuvering perpendicular to the gravity field when the control effectiveness is sufficiently
high (Lloyd,Brown, 1979). In (Lin et al., 2000) it is stated that in a guidance system,
it is important to choose a suitable guidance law and navigation constant.
Investiga-tion and comparison of the system behavior of guidance laws under different navigaInvestiga-tion
constants is developed. In (Tyan, 2015) it is analyzed the capture region of the general
ideal proportional navigation guidance law. In (He et al., 2015) it is presented a new
longitudinal autopilot to address the finite-time tracking problem. In (Yeh, 2010) it is
addressed a nonlinear terminal guidance/autopilot controller with pulse-type control
in-puts. In (Wang et al., 2016) a three dimensional integrated guidance and control law with
impact angle constraint, is developed, using the dynamic surface control and extended
state observer techniques. In (Mohammadi et al., 2016) a decrement of coupling effects
of a tactical missile by designing a robust autopilot for its roll channel is developed. In
(Chen, 2016) a description of an application of a new terminal guidance law for missiles in
three-dimensional (3D) environment during the terminal phase is performed. A guidance
and control strategy for a class of 2D trajectory correction fuse with fixed canards for
a spinning projectile is developed in (Yi et al., 2015). Correction control mechanism is
researched through studying the deviation motion and impact point deviation prediction
based on perturbation theory. In (Creagh, Mee, 2010) it is detailed the design and
simu-lation of an attitude guidance and control scheme for a spinning aerospace vehicle. Two
single-input/single-output controllers are used, which in turn issue flap deflection
com-mands. In (Li et al., 2012) it is noticed that the stable region of the design parameters
for the autopilot shrinks significantly under the spinning condition. It is also observed
that the stable region for design parameters is further narrowed when an integrator is
introduced into the acceleration loop while the steady-state accuracy is dramatically
im-proved. In (Zhou et al., 2013) it is also observed that an unexpected and unstable coning
motion occur for spinning missiles after burnout. To address this unstable motion, the
governing equation of the coning motion, and the dynamics of the fin actuators under
the associated hinge moment is derived. The necessary and sufficient conditions of the
coning motion stability are then analytically derived and further validated through
non-linear six degrees-of-freedom simulations for a representative scenario. A discrete-time,
proportional-derivative navigation guidance law, for the terminal phase of an engagement,
with emphasis on the effect of a digital implementation is proposed in (Lechevin,
Rab-bath, 2012). In (Theodoulis et al., 2013) it is presented a complete design, concerning
the guidance and autopilot modules for a class of spin-stabilized fin-controlled projectiles.
The proposed concept is composed of two sections: the rapidly spinning aft part contains
4.2
Rocket Flight Dynamic Model
This section describes the nonlinear flight dynamic model used in this study, including
dy-namics and aerodydy-namics, and control actuation. An example rocket concept is described,
followed by a description of the mathematical model for each relevant component.
4.2.1
Rocket
The guidance and control formulation proposed in this study applies to a 140mm
axisym-metric rocket with wrap around stabilizing fins. This system features a supersonic launch
and a rocket spin rate of approximately 150 Hz. The maneuver mechanism is placed at
the front part of the spin stabilized rocket, with a roll-decoupled fuse. It consists of a
section composed of fixed canard surfaces in order to generate a rotating control force
and its associated moment. This mechanism consists of a rotary motor linked to the front
section (see 4.iii). The most representative data of the rocket are shown in 4.I. The thrust
curve depending on flight time is shown in 4.i. The propellant mass decreases according
to (4.1).
m(t) =m0−
Z t
0
T(τ)
2348.2dτ (4.1)
Where m(t) is rocket mass function of time,m0is rocket initial mass and T(t) is thrust modulus function of time.
These curves have been fitted using data from actual shots. Note that once the rocket
has been launched there is no control in the thrust force. Finally, all the aerodynamic
coefficients for the rocket under study are acquired by numerical computing and tested
against actual shots. Their variation with the Mach number is shown in 4.ii.
Table 4.I Representative data of the studied rocket.
Parameter
Value
Maximum thrust
29160 N
Burn-out time
2.70 s
Initial mass
77.40 kg
Propellant mass
21.60 kg
I
x00.36 kg
m
2I
y035.63 kg
m
2X
CG01.30 m
Caliber
0.14 m
Figure 4.ii Aerodynamic coefficients for the rocket under study depending of the Mach number.
4.2.2
Flight Dynamic Model
Three axes systems are defined in order to express forces and moments: earth axes, body
axes, and working axes. Earth axes are defined by sub index e. xe pointing north, ze
perpendicular to xe and pointing nadir, and ye forming a clockwise trihedral. Working
axes are defined by sub index w. xw pointing to the target, yw perpendicular to xw and
pointing zenith, and zw forming a clockwise trihedral. Body axes are defined by sub
index b. xb pointing forward and contained in the plane of symmetry of the rocket, zb
perpendicular to xb pointing down and contained in the plane of symmetry of the rocket,
and yb forming a clockwise trihedral. The origin of body axes is located at the center of
mass of the rocket and they are severely coupled to the roll-decoupled fuse. Earth, body
and working axes systems are illustrated in 4.iii.
Total forces and moments of the rocket are given by (4.2) and (4.3), respectively.
−−→
Fext=
− →
D +−→L +−M→+−→P +−→T +−W→+−→C , (4.2)
Figure 4.iii Reference systems for a 140 mm axisymmetric rocket with wrap around fins and a roll-decoupled fuse.
where −→D is drag force, −→L is lift force, −M→ is magnus force, −→P is pitch damping force,
− →
T is thrust force, −W→ is weight force and −→C is Coriolis force.
−−→
Mext=
− →
O +−→PM +
−−→
MM +
− →
S , (4.3)
where −→O is overturn moment, −→PM is pitch damping moment,
−−→
MM is magnus moment
and −→S is spin damping moment.
Rocket forces in working axes include contributions from drag, lift, magnus, pitch
damping, thrust, weight and Coriolis forces, which can be described by the following
(4.4), (4.5), (4.6), (4.7), (4.8), (4.9) and (4.10):
− →
D =−π
8d 2ρ C
D0 +CDα2α
2
k−→vwk−v→w (4.4)
− →
L =−π
8d 2ρ C
Lα·α+CLα3α
2
k−v→wk2−x→w−(x−→w· −v→w)−v→w
(4.5)
− →
M =−π
8d 3ρCmf
Ix
−→
Lw· −x→w
− →
P = π
8d 3ρCN q
Iy
k−→vwk2
−→
Lw× −x→w
(4.7)
− →
T =T(t)−x→w (4.8)
−→
W =m−g→w (4.9)
− →
C =−2m−→Ω × −v→w, (4.10)
Where d is rocket caliber, ρis air density,CD0 is drag force linear coefficient,CDα2 is
drag force square coefficient,α is total angle of attack,CLα is lift force linear coefficient,
CLα3 is lift force cubic coefficient, Cmf is magnus force coefficient,
−→
Lw is rocket angular
momentum expressed in working axes, Ix and Iy are rocket inertia moments in body
axes, CN q is pitch damping force coefficient, −x→w is rocket nose pointing vector expressed
in working axes, −g→w is gravity vector in working axes,
− →
Ω is earth angular speed vector,
and −v→w is rocket velocity expressed in working axes.
Likewise, rocket moments in working axes include contributions from overturning,
pitch damping, Magnus, and spin damping moments and can be described by the following
(4.11), (4.12), (4.13) and (4.14):
− →
O = π
8d 3
ρ CMα +CMα3α
2
k−v→wk2 (v−→w× −x→w) (4.11)
−→
PM = π
8d 3ρ
Iy
CMqk− →
vwk
−→
Lw−
−→
Lw· −x→w
−→
xw
, (4.12)
−−→
MM =−
π 8d 4ρ Ix Cmm −→
Lw · −x→w
((−v→w· −x→w)−x→w)− −→vw
(4.13)
− → S = π 8d 4ρ Ix
Cspink−v→wk
−→
Lw · −x→w
−→
xw, (4.14)
Where CMα is overturning moment linear coefficient, CMα3 is overturning moment
cubic coefficient, CMq is pitch damping moment coefficient, Cmm is magnus moment co-efficient and Cspin is spin damping moment coefficient.
(4.15) and (4.16) model the control forces and moments in body axes, respectively.
−→
CF = ρSexpkvwk 2
4 CN(α)
0
cos(φc−φ)
−sin(φc−φ)
(4.15) −−→
CM =−x−→cgb × −→
CF , (4.16)
where−→CF is control force, −−→CM is control moment,Sexp is reference surface for fins,
CN(α) is fin normal force coefficient, φc is control force angle, φ is roll angle and −x−→cgb is the center of gravity position expressed in body axes.
In order to solve the motion of the rocket a body reference frame, which is coupled
to the fuse, is used. Because, the fuse is uncoupled from the rear part and the rear part
spins at high rates, it is required to model this spinning motion in order to account for
rear Magnus force and moment and gyroscopic effects. The spin rate of the rear part of
the rocket can be modeled as shown in (4.17), note that initial spin speed is modeled as
an impulse which correlates to experimental data.
pr =−
Z
200δ(t0)−
π
8d 4ρ
Ix
Cspink−→vbkk
−→
Lb· −→xb
−→
xbk
dt, (4.17)
whereδ(t0) is dirac’s delta,
− →
Lb is rocket angular momentum expressed in body axes,−→vb
is rocket velocity expressed in body axes and−→xb is rocket nose pointing vector expressed
It is assumed that fuse mass is negligible, which involves no appreciable reactions are
involved between fuse and aft part. Taking this in account, aft effect can be expressed
as an extra addition of external forces and moments that can be expressed by equations
(4.18), (4.19) and (4.20). This simplification is obtained from a development of Euler
equations, −−→Fext = dm
− →v b dt + − →
ωb ×m−→vb on where gyroscopic contributions of aft part are
computed separately and moved to the left part of expression.
−→
Mr =−
π
8d 3
ρCmf Ix
−→
Lb· −→xb
(−→xb × −→vb) (4.18)
−−→
MM r =− π 8d 4ρ Ix Cmm −→
Lb· −→xb
((−→vb · −→xb)−→xb)− −→vb
(4.19)
−→
Gr =−
Ix 0 0
0 Iy 0
0 0 Iy
d dt
pr+p
q r + − →
i −→j −→k
p q r
Ix(pr+p) Iyq Iyr
, (4.20)
Where −→Mr is magnus force of the rotating part of the rocket,
−−→
MM r is magnus moment
of the rotating part of the rocket, −G→r is gyroscopic moment of the rotating part of the
rocket, p, q and r are angular speed components of the rocket and pr is angular speed
of the rotating part of the rocket.
Given the force and moment models above, the equations of motion for the rocket are
formulated using a Newton-Euler approach. The inertial, flat-Earth coordinate system
(denoted by frame e) and the body-fixed coordinate system b are related by Euler roll φ,
pitchθ, and yaw ψ angles.
−−→
Fext+
−→
Mr =
dm−→vb
dt +
− →
ωb ×m−→vb (4.21)
−−→
Mext+
−−→
MM r +
−→
Gr =
d−→Lb
dt +
− →
ωb ×
− →
Lb (4.22)
The equations of motion given by (4.21) and sumMaft are integrated forward in time
using a fixed time step Runge-Kutta of fourth order to obtain a single flight trajectory.
Note that control action is input to this system through specification of the roll angle of the
uncoupled fuse, φc. The result for a non-controlled trajectory, i.e., a ballistic trajectory,
calculated by this model is shown in 4.iv for an initialθ of 40 degrees.
Figure 4.iv A ballistic trajectory for an initialθ of 40 degrees.
4.3
Navigation and Guidance Law
In order to determine the elevation angle of the shot and to correct lateral deviations
produced by Coriolis, Magnus and Gyroscopic effects depending on target location a set
of non-controlled simulations were done and an initial Azimuth is added toAZ0 in order to obtain the initial Azimuth: AZinitial=AZ0+ACorr . Initial elevation angle and ACorr
depending on target distance to launch point are showed in 4.v.
After this correction, navigation and guidance is provided by a modified proportional
law defined in (4.23) and (4.24). Guidance is activated if and only if the variableGN CAct
takes value 1; it takes value 1 if flight time is greater than 5 seconds andθ ≤-15 degrees.
Figure 4.v Initial elevation angle andACorr depending on target distance to launch point.
ballistic flight because only small lateral deviations are pretended to be corrected. Errors
to be introduced in the controller are defined by the following equations:
θerr =−N ·
d dt
−Xp1w −Xp2w −Xp3w
d dt atan
Xp2w
q
X2
p1w +X 2
p3w
(4.23)
ψerr =atan
Xp3w
Xp1w
−atan
xw3
xw1
, (4.24) where,
Xp1w
Xp2w
Xp3w
=−−−−−→XT argetw− −→
Xw, (4.25)
where Xp1w, Xp2w, Xp3w are the line of sight vector components expressed in working axes,−−−−−→XT argetw is target position expressed in working axes,
−→
Xwis rocket position expressed
in working axes, θerr is pitch error, ψerr is yaw error and N is Proportional Navigation
Law Constant.
4.4
Control System
Control is processed by a double loop feedback system, which uses accelerations and
angu-lar speeds in body axes. The inner loop is only used as a system of stability augmentation.
The control angle for the rotating force is defined in (4.26), taking Pitch and Yaw errors
as inputs. 4.vi shows the philosophy of the controller. It has three main inputs: the
accel-eration of the rocket, the pitch error and the yaw error. Roughly speaking, the controller
calculates the needed pointing angle of the aerodynamic force calculating the arc-tangent
of the quotient of the pitch and yaw error. This gives an angle at which the aerodynamic
force, in the yb-zb plane, must point to reach the target. However, the gyroscopic effect
due to the spinning part of the rocket makes the response difficult to govern, i.e.,
impos-ing a φc = 90 degrees will not make the rocket to respond upwards. Therefore, it is also
required to measure the acceleration of the rocket, without accounting for gravity, and
make zero the difference between the angle that forms the projection of the aerodynamic
force in the yb-zb plane with yb and φc .
φc =GN CAct
h
KP
h
atanacczb
accyb
−atan θerr−C1·θ
C2(ψerr−C1ψ)
i
+
+KI
R h
atanacczb
accyb
−atan θerr−C1·θ
C2(ψerr−C1ψ)
i
dt+
+KDdtd
h
atanacczb
accyb
−atan θerr−C1·θ
C2(ψerr−C1ψ)
i
−atan θerr−C1·θ
C2(ψerr−C1ψ)
i
(C1 = 0.01, C2 = 100)
(4.26)
where accxb, accyb, acczb are rocket accelerations expressed in body axes, θ is pitch
angle,ψ is yaw angle, GN CAct is activation parameter for GNC,KP is PID proportional
constant,KI is PID Integral Constant, KD is PID derivative constant.
In order to develop the discrete-time guidance and control algorithms for the terminal
phase the following process is utilized: Model-In-the-Loop (MIL), Software-In-the-loop
Figure 4.vi Controller scheme.
all the models in Sections 4.2, 4.3, and 4.4 are implemented and tested; here, continuous
time is used for all the components of the system. Once the MIL stage is successfully
accomplished, the autopilot, that is, the models in Sections III and IV are enabled for
discrete time and tested together with the model of the plant in Section II (i.e., SIL stage).
Then, in the PIL stage, the autopilot is loaded in the final hardware (e.g., processor or
FPGA) and tested again together with the model of the plant. Finally, a gyroscopic
table of three degrees of freedom is employed to perform tests with real hardware such as
processor and sensors (i.e., HIL tests).
4.5
Simulation Results
Simulation results are presented using the nonlinear flight dynamic model to demonstrate
closed-loop performance of the presented navigation, guidance and control novel approach
and contrast performance with ballistic flight. MATLAB/Simulink R2015a was employed