Volume 4, Number 3, September 2013
Ground Speed Calculation Using Wind Component
Information for Trajectory Prediction
Yong-Kyun Kim1, Deok Gyu Lee1†, Jong Wook Han1
1 2
Electronics & Telecommunications Research Institute
Software Research Laboratory Dept. of Cyber Security
Daejeon, Rep. of Korea
{Ykkim1, deokgyulee, hanjw}@etri.re.kr
Abstract—Ground speed calculation is basic work for trajectory
prediction, conflict detection and air traffic flow management.
This paper proposes a novel algorithm based on Vincenty’s
formulas for ground speed calculation. Our experiments used
simulations with wind components and our experimental results
show that our ground speed calculation exhibits much better
accuracy performance.
Keywords—Ground speed, True Airspeed, Trajectory Prediction,
ATFM
I. INTRODUCTION
The air traffic control (ATC) system improves the safety
and efficiency of air traffic by preventing collisions against
other aircraft and obstacles and managing an aircraft’s
navigation status [1]. Air traffic demand is expected to more
than double over the next 20 years [2]. The accuracy of
trajectory predictions in en-route airspace impacts ATM
conflict predictions and estimated times of arrival (ETA) to
control fixes. For the airspace user, inaccurate trajectory
predictions may result in less-than-optimal maneuver advice in
response to a given traffic management problem [3][4]. These
include missed advice and false advice. Missed advice refers to
the lost opportunity of resolving a traffic management problem
in a manner most efficient to the airspace user. False advice
refers to the suggestion of an unnecessary maneuver that may
cause an aircraft to depart from its most efficient or user-
preferred trajectory. In this paper, we propose a ground speed
calculation using wind component information. The remainder
of this paper is organized as follows. In the next section,
ground speed calculation techniques and the theoretical
background about ground speed calculations are presented. We
describe our ground speed calculation algorithm in Section 3.
In Section 4 we present some experimental results of our
proposed scheme, and finally give our conclusions in Section 5.
II. GROUND SPEED CALCULATION TECHNIQUES
A. GROUND SPEED CALCULATION THEORY
Let us first consider a fairly simplified model for our trajectory
design problem. The set of flows shall be arbitrarily chosen. A
flow is defined as a set of flights between a departure airport and
1
This research was supported by a grant (code# 07aviation-navigation-03)
from the Aviation Improvement Program funded by the Ministry of Construction
& Transportation of the Korean government.
†
Deok Gyu Lee, Corresponding Author, deokgyulee@etri.re.kr
Copyright ⓒ 2010 Future Technology Research Association International
Journal of Convergence
Hyodal Park2
Electronic and Electrical Engineering, Inha University,
#100, Inharo, Nam-gu, Incheon, Rep. of Korea
hdpark@inha.ac.kr
an arrival airport. The following simplifications are made. The
airspace is considered as an Euclidean space, where all airports
are at altitude 0. Latitudes and longitudes on the ellipsoidal
Earth’s surface are converted into (x, y) coordinates by a
stereographic projection, and the altitude in feet shall be our z
coordinate [5].
Fig. 1. Basic design of trajectory
All aircraft fly with identical performances and follow linear
slopes of climb and descent.
B. PARAMETERS FOR GROUND SPEED CALCULATION
For ground speed calculation, we must consider the concept
of speed, speed variation due to changes in altitude and wind
parameters.
First, airspeed is the ground speed calculation relative to the
air. Among the common conventions for qualifying airspeed
are: indicated airspeed (IAS), calibrated airspeed (CAS), true
airspeed (TAS), and ground speed (GS).
IAS is the airspeed indicator reading (ASIR) uncorrected
for instrument, position, and other errors. From current
EASA(European Aviation Safety Agency) definitions, IAS
means the speed of an aircraft as shown on its pitot static
airspeed indicator, calibrated to reflect standard atmosphere
adiabatic compressible flow at sea level uncorrected for
airspeed system errors.
Most airspeed indicators show the speed in knots (i.e.
nautical miles per hour). Some light aircraft have airspeed
indicators showing speed in miles per hour
CAS is IAS corrected for instrument errors, position error
and installation errors. CAS values of less than the speed of
sound at standard sea level (661.4788 knots) are calculated as
follows:
1
Journal of Convergence
2 difference between thrust and drag is the greatest (maximum
é ù
ê æ Q ö 7
Vc = A0 5ç ç c
+ 1÷ - 1ú
÷ (1)
êè P0 ø ú
êë úû
where
VC is the calibrated speed.
QC is the impact pressure sensed by the pitot tube.
P0 is 29.92126 inches Hg; static air pressure at
standard sea level
A0 is 661.4788 knots; speed of sound at standard sea
level
This expression is based on the form of Bernoulli’s
equation applicable to a perfect, compressible gas. The values
P0 and A0 are consistent with the International Standard
Atmosphere (ISA).
TAS is the physical speed of the aircraft relative to the air
surrounding the aircraft. The TAS is a vector quantity. The
relationship between the TAS (Vt) and the speed with respect to
the ground (Vg) is
Vt = v g - Vw (2)
where
Vw is the windspeed vector.
Aircraft flight instruments, however, do not compute TAS
as a function of groundspeed and windspeed. They use impact
and static pressures as well as a temperature input. Basically,
TAS is CAS that is corrected for pressure, altitude and
temperature. The result is the true physical speed of the aircraft
plus or minus the wind component. TAS is equal to CAS at
standard sea level conditions.
The simplest way to compute TAS is to use a function of
the Mach number
T axis.
Vt = A0 × M (3)
Tc
where M is the Mach number, T is temperature (kelvins) and T0
is standard sea level temperature (288.15 kelvins)
Second, speed variation due to changes in altitude means
when the aircraft climbs or descends.
The rate of climb (RoC) is the speed at which an aircraft
increases its altitude. This is most often expressed in feet per
minute and can be abbreviated as ft/min. Elsewhere, it is
commonly expressed in meters per second, abbreviated as m/s.
The RoC of an aircraft is measured with a vertical speed
indicator (VSI) or instantaneous vertical speed indicator (IVSI).
The rate of decrease in altitude is referred to as the rate of
descent or sink rate. A decrease in altitude corresponds with a
negative RoC.
There are two airspeeds relating to optimum rates of ascent,
referred to as Vx and Vy. Vx is the IAS for best angle of climb.
Vy is the IAS for best RoC. Vx is slower than Vy.
Climbing at Vx allows pilots to maximize the altitude gain
per unit of ground distance. That is, Vx allows pilots to
maximize their climb while sacrificing the least amount of
ground distance. This occurs at the speed for which the
2 Copyright ⓒ 2010 Future Technology Research Association International
Volume 4, Number 3, September 2013
excess thrust). In a jet airplane, this is approximately the
minimum drag speed, or the bottom of the curve of drag vs.
speed. The climb angle is proportional to the excess thrust.
Climbing at Vy allows pilots to maximize the altitude gain
per unit of time. That is, Vy allows pilots to maximize their
climb while sacrificing the least amount of time. This occurs at
the speed for which the difference between engine power and
the power required to overcome the aircraft’s drag is the
greatest (maximum excess power). Climb rate is proportional
to excess power.
Vx increases with altitude and Vy decreases with altitude.
Vx = Vy at the airplane’s absolute ceiling, the altitude above
which it cannot climb using just its own lift.
Last, we consider wind parameters. Wind parameters can
be divided into two components (weather fronts and thermal
wind) on a large scale.
Weather fronts are boundaries between two masses of air of
different densities, or different temperature and moisture
properties, which are normally convergence zones in the wind
field and are the principal cause of significant weather. Within
surface weather analyses, they are depicted using various
colored lines and symbols.
The air masses usually differ in temperature and may also
differ in humidity. Wind shear in the horizontal occurs near
these boundaries. Cold fronts feature narrow bands of
thunderstorms and severe weather, and may be preceded by
squall lines and dry lines.
Cold fronts are sharper surface boundaries with more
significant horizontal wind shear than warm fronts. When a
front becomes stationary, it can degenerate into a line which
separates regions of differing wind speed, known as a shear
line, though the wind direction across the feature normally
remains constant. Directional and speed shear can occur across
the axis of stronger tropical waves, as northerly winds precede
the wave axis and southeast winds are seen behind the wave
Horizontal wind shear can also occur along local land
breeze and sea breeze boundaries.
Thermal wind is a meteorological term not referring to an
actual wind, but a difference in the geostrophic wind between
two pressure levels p1 and p0, with p1 < p0; in essence, wind
shear. It is only present in an atmosphere with horizontal
changes in temperature.
In a barotropic atmosphere, where temperature is uniform,
the geostrophic wind is independent of height. The name stems
from the fact that this wind flows around areas of low (and
high) temperature in the same manner as the geostrophic wind
flows around areas of low (and high) pressure.
f tt = K ´ Ñ(f 2 - f 0 ) (4)
where φx are geopotential height fields with f2 > f0 , f is the
Coriolis parameter, and k is the upward-pointing unit vector in
the vertical direction. The thermal wind equation does not
determine the wind in the tropics. Since f is small or zero, such
as near the equator, the equation reduces to stating that
Ñ(f2 > f0 ) is small. This equation basically describes the
existence of the jet stream, a westerly current of air with
Volume 4, Number 3, September 2013 Journal of Convergence

maximum wind speeds close to the tropopause which is (even
though other factors are also important) the result of the
temperature contrast between the equator and the poles.
III. PROPOSED GROUND SPEED CALCULATION ALGORITHM
This section describes the method of computing the various
parameters used to compute the flight plan and our GS
calculation algorithm.
First of all, we need coordinates of the waypoint and to
create the aircraft’s path and then the wind components and
the aircraft’s airspeed used to calculate the GS of the actual
aircraft.
Given the coordinates of the two points (φ1, λ1) and (φ2, λ2)
Vincenty’s inverse method finds the azimuths α1, α2 and the
ellipsoidal distance s by calculating the reduced latitude U1
( arctan[(1 - f ) tan j1 ] , U2( arctan[(1 - f ) tan j 2 ] , and L, and setting the
initial value of λ= L, then iteratively evaluating the following
equations until λ converges.
sin s = (cos U 2 sin l )2 + (cosU 1 sin U 2 - sin U 1 cos U 2 cos l ) 2 (7)
cos s = sin U 1 sin U 2 + cosU 1cos U 2 cos l (8)
s = arctan(sin s / cos s ) (9)

cos(2s m ) = cos s - (sin U 1 sin U 2 / cos 2 a ) (10)

sin a = [(cos U 1 cosU 2sin l ) /(sin s )] (11)

cos 2 a = 1 - sin 2 a (12)

C = (1 / 16) cos 2 a [4 + f (4 - 3 cos 2 a )] (13)

2
l = L + (1 - C ) f sin a {s + C sin s [cos( 2sm ) + C cos s ( -1 + 2 cos ( 2sm )]} (14)

When λ has converged to the desired accuracy, the

following are evaluated:

Fig. 2. GS calculation methods U 2 = COS 2a [(a 2 - a1 ) / b2 ] (15)

Waypoints consist of latitude and longitude. Accordingly, to
measure the distance between waypoints over a flight A = 1 + (u 2 / 16384){4096 + u 2 [-768 + u 2(320 - 175u 2 )]} (16)
information region, the curvature of the Earth must be taken
into consideration. B = (u 2 / 1024){256 + u 2 [-128 + u 2 (74 - 47U 2 )]} (17)
As a rough estimate, we could assume the Earth is a sphere.
Ds = B sin s {cos(2sm) + 0.25B[cos s (-1 + 2 cos 2 (2sm) (18)
dN = Rqf (5)
(-3 + 4 sin 2s )(-3 + 4 cos 2(2sm))]}
dE = R cos qfl (6)
s = bA(s - Ds ) (19)
where R is the radius of the Earth (average of 6378.1 km) and
the differences in latitude and longitude are in radians, the a1 = arctan[(cos U 2 sin l ) /(cos U 2 sin U 2 - sin U 2 cos U 2 cos l ) (20)
distance is 447.47 km. This method is a valid assumption over
very small distances; however, over large distances we need to a = arctan[(cos U sin l ) /( - sin U 2 cos U 2 + sin U 2 cos U 2 cos l ) (21)
account for the non-uniformity of the Earth. The Earth is not 2 2

actually a sphere; it is an ellipsoid of revolution, 21 km shorter

in the north–south direction than the east–west direction. Then the azimuths (α1, α2) and distance s can be computed.

The flattening at the poles is caused by the centrifugal force

of the spinning Earth. Because of this flattening, the radius if IV. EXPERIMENTAL RESULTS OF PROPOSED SCHEME
the Earth is not a constant value as we assume for Spherical The performance of the GS calculation algorithm is
Earth coordinates. measured using wind components and the aircraft’s airspeed.
A more accurate method of measuring the distance between
two points on the surface of the Earth is Vincenty’s Formula.
TABLE 1. SIMULATION DATA
It is accurate to 0.5 mm over a distance of a few centimeters to
nearly 20,000 km. Aircraft’s Airspeed (TAS) 100 [Knot]
Wind Component Speed: 15 [Knot]
Direction: 320 [Degree]

Copyright ⓒ 2010 Future Technology Research Association International 3

Journal of Convergence Volume 4, Number 3, September 2013

Start Fix Location SEL(N372449, E1265542)

End Fix Location BELMI(N371249, E1265929)

First of all, the start position and airspeed are needed. The
start position consists of the latitude and longitude.
To use the GS algorithm, the predicted GS and estimates
time are as shown in Figure 3.

Fig. 5. Simulation results (climb rate [degree])

V. CONCLUSION

In this paper we proposed a GS calculation using wind

component information.

Experimental results show that our GS calculation exhibits

Fig. 3. Simulation result much better accuracy performance. The applicability of the
proposed algorithm is manifold during trajectory prediction
and modeling, such as in aeronautical traffic flow management
TABLE 2. SIMULATION DATA WITH CLIMB RATE
(ATFM) systems and vehicle trajectory prediction.
Aircraft’s Airspeed 100 [Knot]
(TAS) From now on, it is further suggested that the proposed
Wind Component Speed: 50 [Knot] algorithm may be extended to trajectory modeling, which may
Direction: 300 [Degree] further improve the 4D trajectory prediction.
Start Fix Location N372449, E1265542
End Fix Location N371249, E1265929 ACKNOWLEDGMENT
Climb Rate 7160.771 [ft/min], This research was supported by a grant (code# 07aviation-
45 [Degree] navigation-03) from the Aviation Improvement Program
funded by the Ministry of Construction & Transportation of
The following experiments were tested by adding the climb the Korean government.
rate.
To use the GS algorithm with the climb rate, the predicted
GS and estimated time are as shown in Figures 4 and 5 REFERENCES

[1] RTCA, Inc, “VHF Air-Ground Communications System Improvements

Alternative Study and Selection of Proposals for Future Action.”
RTCA/DO-255, 1994.
[2] David P. Thipphavong, Charles A. Schultz, Alan G. Lee, Steven H.
Chan, “Adaptive algorithm to improve trajectory prediction accuracy of
climbing aircraft”, Journal of Guidance, Control, and dynamics, Vol. 36,
No. 1, January-February 2013
[3] Nicolas Barnier, Cyril Allignol, “4D-Trajectory deconfliction through
departure time adjustment”, Eighth USA/Europe Air Traffic
Management Research and Development Seminar (ATM2009)
[4] Geraud Granger, Nicolas Durand, Jean-Marc Alliot, “Optimal resolution
of en route conflicts”, http://atm-seminar-97.eurocontrol.fr/durand.pdf.
[5] C.M. Thomas, W.E. Featherstone, “Validation of Vincenty’s formulas
for the geodesic using a new fourth-order extension of Kivioja’s
Fig. 4. Simulation results (climb rate[ft/min]) formula”, Journal of Surveying Engineering, pp.20–26,Feb., 2005

4 Copyright ⓒ 2010 Future Technology Research Association International

Volume 4, Number 3, September 2013 Journal of Convergence

BIOGRAPHIES

Yong-kyun Kim
2007: MS from the Dept. of Electronic Engineering, Inha
University
2010~: Researcher in the software research laboratory,
ETRI
Research interests: Air traffic management systems, 4D
trajectory modeling

Deok Gyu Lee

Dr. D. G. Lee received his PhD degree from the Graduate
School of Computer Science from Soonchunhyang
University, Korea. He is now a post-graduate doctor of
the R&D Institute, ETRI (Electronics and
Telecommunication Research Institute), Korea. Dr. Lee
has published many research papers in international
journals and conferences. Dr. Lee has served as the chairs
and on the program committees for many international conferences and
workshops. Dr. Lee’ s research interests include key management, signature
schemes, broadcast encryption, content security, wireless security, ubiquitous
computing, home networks, etc. He is a member of the KICS, KMMS, KIPS,
and IEEE, ACM.

Jong Wook Han

1991: MS from the Dept. of Electronic Engineering,
Kwangwoon University
2001: PhD from the Dept. of Electronic Engineering,
Kwangwoon University
1991~: Researcher in the Software Research Laboratory
Research interests: Home network security, convergence
service security, optical security

Hyodal Park
1987: PhD from the Dept. of Electronics Engineering,
ENSEA(École nationale supérieure de l'aéronautique et
de l'espace), France
1992~: professor at Inga University
Research interests: Avionics and microwave systems,
radar systems, antennae.
Member of Air Traffic Control Assiociation(ATCA)

Copyright ⓒ 2010 Future Technology Research Association International 5