Aerotecnica Missili & Spazio (2020) 99:263–273

https://doi.org/10.1007/s42496-020-00063-1

ORIGINAL ARTICLE

Performance Analysis of Space Surveillance Using Space Based Optical

Sensors
Federica Vitiello1

Received: 17 September 2020 / Revised: 17 September 2020 / Accepted: 8 October 2020 / Published online: 5 November 2020
© The Author(s) 2020

Abstract
This paper aims to describe the analysis of the performance of an electro-optical space-based sensor for space surveillance
purposes and space debris detection in the geostationary (GEO) ring. Such sensor is considered to be operating on a dawn–
dusk Sun-synchronous, circular low Earth orbit at an altitude of 630 Km, while its optical characteristics have been taken
from those of the Space-Based Visible (SBV) sensor. Two main simulations have been carried out through the use of the
MATLAB software. The first simulation deals with the detection capability of the sensor, which is discussed in terms of
detectable visual magnitude when the target of the observation is a diffuse sphere orbiting in the geostationary (GEO) orbit;
its minimum detectable size is then determined. In addition, the relative geometry between the Sun, the sensor and the target
has also been studied along with the configurations which can limit the visibility of the sensor over the target. The second
simulation has been used to evaluate the performance of the sensor in terms of number of detectable GEO targets and dura-
tion of the observation when a certain pointing strategy is adopted. In such strategy, two SBV-like sensors are placed on the
same orbit, thus creating a constellation in which each sensor points towards a fixed location in the inertial space. These
locations have been chosen to be the geosynchronous pinch points.

Keywords Space debris · Space-based optical Sensors · Visual magnitude · Solar phase angle

1 Introduction at a given epoch. An explanatory example of the breadth of

the space debris issue is given by the fragmentation debris
The space debris population is composed of objects of vari- population, which is composed of debris whose genesis
ous nature which have been brought in orbit by human space is related to fragmentation events. As a matter of fact, as
activities. Such objects are typically parts of bigger bodies referred to the state of affairs of January the 1st of 2019,
that have detached during collisions or explosions, and their it has been found that of the more than 19000 catalogued
position and velocity are, in most cases, unknown. Because RSOs orbiting the Earth, the fragmentation debris popula-
of the uncertainty in the pieces of information regarding this tion accounts for around the 55% [1]. The diameter of these
type of objects and because of their non-cooperative nature, objects can be as small as 1 μm ; nevertheless, the current
space debris poses a serious threat to the operative space detection capability of the SSN sensors is set on RSOs
assets which are currently orbiting the Earth and to the on- whose diameter is larger than 10 cm, at low Earth orbits
Earth population and activities. (LEO) altitudes and larger than 80 cm, at GEO altitudes
The current knowledge of the resident space objects [2]. This latter measurement is primarily obtained by optical
(RSO), among active satellites, inactive satellites and space ground-based telescopes.
debris, is mainly kept in the Satellite Catalog maintained It appears to be clear that in order to ensure the safety
by the US Space Surveillance Network (SSN), in which the of current and future space missions while, at the same
orbital elements of the so-called catalogued objects are listed time, protecting the on-Earth activities from the possible
re-entry of unknown, non-cooperative objects, an improved
* Federica Vitiello knowledge of the small debris population is needed. Even
federitiello@gmail.com though the majority of space assets operates at LEO alti-
tudes, because of the importance of the satellites which do
1
Department of Industrial Engineering, University of Napoli carry out their missions in the GEO ring, this space region is
“Federico II”, Napoli, Italy

13
Vol.:(0123456789)
264 F. Vitiello

looked at with great attention. For this very reason, it is the When the visual magnitude of a debris or, more gener-
purpose of this work to investigate the use of a space-based ally, of a satellite is considered, it is obviously not referred
optical sensor for the detection of sub-catalogue size objects to any intrinsic brightness of the object; indeed, it is referred
orbiting in the GEO belt in order to provide additional meas- to its capability to reflect light. The formula that has been
urements complementing the performance of the existing used for the evaluation of the visual magnitude during this
ground-based systems. work is showed below for a diffuse spherical object having
Due to its location above the atmosphere of the Earth, a diameter equal to d [4].
a space-based telescope is not affected by the same limita- [ 2( )]
tions of its ground-based counterpart. In fact, because of d 𝜌spec
mv,obj = mv,ref − 2.5 log10 2 + 𝜌diff pdiff (𝜙) (1)
the effects of the sky brightness and of the meteorologi- R 4
cal conditions, the operative time of a ground-based tele-
In such equation, the quantities 𝜌spec and 𝜌diff are, respec-
scope is reduced to nighttime and to those moments when
tively, the specular and diffuse reflectivity components of the
the observed sky window is clear. The measurements of a
surface of the observed sphere, R is its distance, measured
space-based sensor are immune from these constrains and
in Km, from the observer (sensor), mv,ref is a reference value
are, at the same time, independent from the geographical
in visual magnitude which has been chosen to be that of the
location of the sensor, thus implying that each nation could
Sun (-26.7), and pdiff is the phase function which depends
ideally create and maintain a catalogue of observed objects
on the solar phase angle 𝜙. An expression can be given for
with no need to rely on the measurements gained by other
the phase function of a diffuse sphere as written below [4].
nations for the computation of the collision probability of
their space assets. However, building and testing a space- 2
pdiff (𝜙) = [sin(𝜙) + (𝜋 − 𝜙)cos(𝜙)] (2)
based telescope are a great technological challenge and, 3𝜋
although the sensor is not affected by the presence of the
The solar phase angle, which is one crucial parameter for the
Earth atmosphere, it is still mainly limited by how close it
measurement of the visual magnitude, is an information on
can be pointed to the Sun and by the presence of the Earth
the illumination condition of an object in space. In fact, it is
in its field of view (FoV).
defined as the angle between the direction to the Sun and the
The characteristics of the sensor used for the development
direction to the observer as seen at the object which is being
of the simulations have been taken from those of the SBV
observed [5]. A more favourable illumination condition is
sensor which was launched on the Midcourse Space Experi-
obtained for a smaller phase angle. In fact, as this angle
ment (MSX) satellite in 1996 [3]. In particular, the FoV of
increases, the illuminated side of the observed object starts
such sensor has been considered to be conical with an angu-
facing away from the sensor. The worst illumination condi-
lar width of 1.4◦ and it is placed on a circular, dawn–dusk
tion is obtained when 𝜙 = 180◦, meaning that the observed
Sun-synchronous LEO orbit at an altitude of 630 Km and an
object lies exactly between the sensor and the Sun.
inclination of 98◦, thus optimizing both the power supply of
the hosting satellite and the relative geometry between the
Sun, the Earth, the sensor and the target. 2.2 Non‑Visibility Conditions

When a space-based telescope is taken into consideration,

2 Target Detectability the limits regarding its ability to continuously track a RSO
must also be considered. As a matter of fact, three differ-
2.1 Visual Magnitude and Solar Phase Angle ent configurations can limit the detection capability of the
sensor by partially or completely disabling the visibility of
The detectability of a target in space can be discussed in the target.
terms of its visual magnitude. In astronomy, the apparent
magnitude of a celestial body, such as a star or a galaxy, is 2.2.1 The Earth Shadow Region
used to define its brightness as seen by an observer which is
placed at a certain distance from the body itself. When such Depending on the position of the Sun, the shadow casted
observer is able to detect light within the visible range, as it by the Earth may intersect a portion of the orbit of a space-
is the case for an optical telescope, then the apparent mag- craft, promoting the so-called eclipse of the orbit. Since a
nitude is referred to as visual magnitude. The visual magni- RSO passing through the shadow of the Earth region is not
tude is measured on a reversed logarithmic scale, meaning illuminated by the Sun, such occurrence is considered to be
that the brightest the object is, the lowest its magnitude is; a non-visibility condition for the case of study of this paper.
the Sun, for example, which is one of the brightest celestial An approximated method has been used for the evalua-
bodies, is considered to have a visual magnitude of -26.7. tion of the eclipse portion of the GEO orbit where the target

13
PErformance Analysis of Space Surveillance Using Space‑Based Optical Sensors 265

debris, considered during this analysis, moves. The main are the unit vectors of the distance between the sensor and
hypothesis on which this method is based is: the debris moving along the GEO orbit at each given time.

– The Earth is considered as a sphere having a radius of ⎧ X(t) = Xsensor (t) + 𝜆̂𝐮𝐗 (t)
⎪ Y(t) = Ysensor (t) + 𝜆̂𝐮Y (t)
6371 Km. ⎨ Z(t) = Z (4)
– The penumbra (partial shadow region placed between ⎪ 2 sensor (t) + 𝜆̂
uZ (t)
the umbra and the full-light region) effects on the eclipse ⎩ X + Y 2 + Z 2 = RE
evaluation are ignored.
– The shadow casted by the Earth (umbra) is considered to
be a circular cylinder.
3 Numerical Simulation Results on Target
Detectability
Therefore, the intersection between the umbra region and
the GEO orbit can be seen as a semi-ellipse whose semi-
3.1 Simulation Set‑up
minor axis is equal to the Earth radius RE and semi-major
axis is equal to sin𝛿 [6], where 𝛿 is the angle between the
RE
The target considered in the first part of the simulation is
direction of the Sun rays and the orbital plane. However, in
a 200-cm aluminium diffuse sphere with a reflectance of
this case of study, since the orbital plane of the GEO orbit
18%, composed of a diffuse fraction of 95% ( 𝜌diff ) and a
lies on the equatorial one, such angle can be considered as
specular fraction of 5% (𝜌spec) [7]. The simulation has been
the declination of the sun. It is then possible to compute the
propagated for a time equal to the orbital period of the target
semi-angular width, 𝜃 , of the portion of the GEO orbit lying
(roughly 24 h) during which the sensor, whose orbital period
in the Earth shadow as shown below.
is around 1.5 h, will have completed around 15 orbits. The
� computation has been carried out for a fixed location of the
⎡ � �2 ⎤
⎢ 1 RE ⎥ Sun in the ECI reference frame equal to its vernal equinox
𝜃 = acos 1− (3)
⎢ cos𝛿 r ⎥ position, which means that both declination (𝛿Sun ) and right
⎣ ⎦
ascension (𝛼Sun ) of the Sun are null. The position of the Sun
Where r is the radius of the GEO orbit: r = 42157Km. is also used to determine the value of the right ascension of
the ascending node (RAAN) of the orbit of the sensor. In
2.2.2 The Solar Exclusion Region order to have a dawn–dusk Sun-synchronous orbit, when
the Sun is at the vernal equinox point, such value has to be
The solar exclusion region is defined as the portion of equal to 270◦.
the orbit of the debris, or RSO, where the object lies in a Finally, the location in the ECI reference frame of both
position between the Sun and the sensor. This condition is sensor and target at the beginning of the simulation (time t
related to a value of the solar phase angle which is close to = 0) has
( been chosen ) (to be, respectively,
) given by the ECI
180◦, meaning that the observed object appears to be in back- points 0, atarget , 0 and asensor , 0, 0 where asensor = 7001Km
light with respect to the sensor and is therefore impossible and atarget = 42157Km are, respectively, the semi-major axis
to detect. In this work, the solar exclusion region has been of the orbits of the sensor and the target.
considered to have an angular width of 45◦ [7].
3.2 Solar Phase Angle and Visual Magnitude Results
2.2.3 The Earth Blockage
The solar phase angle has been evaluated as follows:
( )
The Earth blockage event occurs when the line of sight of 𝜙 = acos uTarget̂- Sensor ⋅ uEartĥ - Sun (5)
the sensor is blocked by the presence of the Earth along its
path. where uTarget̂- Sensor is the target to sensor direction unit vec-
By assuming a spherical Earth, the computation of the tor and 𝐮𝐄𝐚𝐫𝐭𝐡−𝐒𝐮𝐧
̂ is the Earth to Sun direction unit vector.
Earth blockage time instants can be trivially reduced to This definition does not match the one given in 2.1 since
the computation of the time instants when the intersection the direction from the target to the Sun can be approximated
between a line (the line of sight) and a sphere of radius RE with the direction from the Earth to the Sun, based on the
(the Earth) occurs in the Earth-centred intertial reference fact that the distance between the target and the Earth is
frame (ECI). In mathematical terms, this is achieved by solv- negligible if compared to the distance between the Earth
ing in 𝜆 the system shown below, in which the ECI location and the Sun.
of the sensor at each given time t is identified by the compo- In Fig. 1, it can be noticed that during the description of
nents ( Xsensor (t), Ysensor (t), Zsensor (t)) and ûX (t), ûY (t) and ûZ (t) the first half of the target’s orbit, when the target itself gets

13
266 F. Vitiello

Fig. 3  Visual magnitude variation for 1 day with omitted Earth

Fig. 1  Phase angle variation for 1 day (orbital period of the target). shadow region. Sun at vernal equinox
Sun at vernal equinox

Fig. 4  Visual magnitude variation for 1 day with omitted solar exclu-
Fig. 2  Visual magnitude variation for 1 day (orbital period of the tar- sion region. Sun at vernal equinox
get). Sun at vernal equinox.
shown in Figs. 3 and 4, both the solar exclusion region and
farther from the Sun, the trend of the solar phase angle is a the Earth shadow region have been removed from the visual
decreasing one while, on the other hand, it gains an increas- magnitude diagram.
ing trend when the target completes the remaining half of In addition, the computation of the Earth blockage events
its orbit, getting closer to the Sun. The worst illumination has shown that, for the given orbit of the sensor, the overall
condition is obtained when the target passes through the duration of such occurrence is of roughly 11 h with a total
solar exclusion region with a correspondent “peak” in the number of distinct events equal to 22. In Fig. 5, each omit-
visual magnitude of around 15 (Fig. 2). Conversely, the best ted region in the visual magnitude diagram corresponds to
illumination condition is obtained when the target has com- a single Earth blockage event.
pleted half of its orbit, thus having its illuminated side fac-
ing the sensor. The correspondent visual magnitude shows 3.2.1 Variation in the Diameter of the Target
a constant trend with a mean value of around 11.5 (Fig. 2).
Still, this good visibility condition is partially interrupted by Published data on the SBV sensor performances have
the eclipse of the GEO orbit having a duration of 1.15 h. As reported that the sensitivity of such sensor is limited to a

13
PErformance Analysis of Space Surveillance Using Space‑Based Optical Sensors 267

3.2.2 Variation in the Location of the Sun

The last analysis carried out concerns the variation in the

location of the Sun in the ECI reference frame. The new cho-
sen position is the one corresponding to the summer solstice,
when 𝛼Sun = 90 and 𝛿Sun = 23.44◦.
◦

Because of the change of position of the Sun, a rotation

of the line of nodes of the sensor’s orbit has to be consid-
ered in order to simulate the dawn–dusk Sun-synchronous
condition. For this reason, the value of the RAAN has been
changed to 0◦. The aforementioned rotation promotes a vari-
ation in the initial ECI position of the sensor, which, at the
(beginning of) the simulation, is now given by the components
asensor , 0, 0 . In addition, a displacement of the solar exclu-
sion region in the ECI reference frame is also obtained, as
Fig. 5  Visual magnitude variation for 1 day with omitted Earth block-
Fig. 7 witnesses.
age events. Sun at vernal equinox By analysing the behaviour of the phase angle in Fig. 8,
it is possible to identify the worst illumination condition
when the target has roughly completed a quarter of its orbit
visual magnitude of about 15 [7]. In the following results, and is, in fact, passing through the solar exclusion region.
this value is used to remark the limit in the detection capabil- As expected, the corresponding visual magnitude reaches a
ity of the sensor when the diameter of the observed objects peak of about 14.5 (Fig. 9). The best illumination condition,
decreases. This is done with the purpose of identifying the instead, is obtained when the target lies in the opposite posi-
minimum detectable size of a debris orbiting in GEO. tion; the visual magnitude gains a more constant trend with
In Fig. 6, the green region is intended to be an illustra- a mean value of around 11.5 (Fig. 8).
tion of the visual magnitude values that the sensor is able to By comparing these results with those shown in sec-
detect as imposed by its limited sensitivity; such region ends tion 3.2 it has been found that the configuration involving
with the value mv = 15. It can easily be seen that a spherical the Sun at the summer solstice is more advantageous since,
diffuse object with a diameter of 200 cm is always visible for because of its high declination, the shadow of the Earth does
the chosen sensor. As the diameter decreases, the time inter- not intercept the equatorial plane where the GEO orbit lies.
val during which the object appears to be visible decreases As a consequence, the target is visible for a longer time.
too. When the diameter is equal to 50 cm, this time interval As shown in Fig. 10, the Earth blockage events disposi-
is roughly equal to 6 h while it reaches a null value when the tion in time is changed with respect to that of the previous
diameter falls off to 30 cm, meaning that the object is never case. Still, the number of total events and the overall dura-
visible. However, the results shown in Fig. 6 do not take in tion are unchanged.
account the non-visibility conditions.
In order to have a clearer idea of the results obtained,
an analysis on the fraction of time ( t∗ ) during which the 4 Numerical Simulation Results on Pointing
condition mv ≤ 15 is respected has been performed. A non- Strategy
dimensional parameter 𝜂 has therefore been defined as the
ratio between t∗ and the overall time used for the simula- 4.1 The Geosynchronous Pinch Points
tion (orbital period of the target, Ttarget ). The results of such
analysis are shown in Table 1 in terms of the behaviour of In the following section, the results of the performance anal-
𝜂 when the diameter of the target decreases. In the ideal ysis of a constellation of two SBV-like sensors operating on
case, when the non-visibility conditions are not considered the same orbit used in the previous simulations will be dis-
(Table 1 top), such parameter follows the results shown in cussed. For such sensors, a “fixed intertial” pointing strategy
Fig. 6, thus going from unity for a diameter of 200 cm to has been adopted; hence, both sensors are kept staring at two
the null value for a diameter of 30 cm. However, when the different fixed locations in the ECI reference frame with the
non-visibility conditions are taken into account, as shown aim of detecting debris passing through their fields of view.
in Table 1 bottom, the values of 𝜂 reduce drastically. This is In order to simulate the maximization of the number of
mainly due to the occurrence of the Earth blockage events detections, the chosen inertial locations have been consid-
which do remove more than the 45% out of the overall time ered as those of the pinch points on the GEO belt. These
t∗. points are characterized by a high density of passing objects,

13
268 F. Vitiello

Fig. 6  Visual magnitude variation for 1 day (orbital period of the tar- ▸
get) for a target having a decreasing diameter from 200 to 30 cm. The
green region is an illustration of the visual magnitude range that can
be sensed by the sensor. Sun at vernal equinox

both active and inactive, and are visualized in Fig. 11 which

shows the population of geosynchronous satellites over a
24-h period [8].
It can easily be seen that the pinch points are both located
on the equatorial plane (at a null declination 𝛿 ) and at two
different values of right ascension 𝛼 . In the following
description, the two points are referred to as “PP1,” located
at 𝛼PP1 = 65◦ and “PP2,” located at 𝛼PP2 = 245◦.

4.1.1 Genesis of the Pinch Points

The pinch points are the consequence of different factors

among which are the particular type of orbit and its evolu-
tion in time and the common strategy typically applied by
geosynchronous (GS) and geostationary satellites operators.
During its lifetime, a GS satellite faces an out-of-plane
force caused by the action of the perturbation promoted by
the Sun and the Moon (also referred to as lunisolar perturba-
tion), and the perturbation caused by the oblateness of the
Earth. This force causes a periodic variation in the inclina-
tion and RAAN of the orbit on which the satellite operates.
As shown in Fig. 12, it takes 27 years for the orbit’s incli-
nation of a GS satellite to increase up to 15◦ and nearly the
same time for it to get back to its initial value. During this
variation, the RAAN evolves from +90◦ to −90◦.
The evolution pattern shown in Fig. 12 is followed by the
majority of the inactive GS satellites which are no longer
manoeuvred as to keep their orbit’s inclination as close as
possible to 0◦, thus creating a geostationary satellite. GS
satellites operators choose to launch their satellites with an
initial value of inclination and RAAN that places the evolu-
tion of the satellite along the curve in Fig. 12.
It is the joint effect of the natural perturbations and the
choice of initial conditions decided by the operators that
promotes the creation of the pinch points congestion regions.

4.2 Simulation Set‑up

The illustration in Fig. 13 shows the initial geometry chosen

for the simulation, where the first sensor (S1) is located at
the ascending node of the orbit and the second sensor (S2)
is located in the opposite position. In particular, sensor S1 is
pointed towards the pinch point PP1 and sensor S2 is pointed
toward the pinch point PP2. In this case, the Sun is consid-
ered as fixed at the vernal equinox point.
The analysis of the performance deals with the capabil-
ity of each sensor to detect targets (debris) distributed on
the GEO orbit during a period of 24 h. For this reason, the

13
PErformance Analysis of Space Surveillance Using Space‑Based Optical Sensors 269

Table 1  Peformances of the Diameter [cm] t∗

𝜂=
sensor in terms of values of Ttarget
the adimensional parameter 𝜂 .
200 1.0
Non-visibility conditions not
considered (top), non-visibility 150 0.81
conditions considered (bottom) 100 0.66
50 0.31
40 0.028
30 0
200 0.22
150 0.39
100 0.10
50 0.050
40 0.020
30 0

Fig. 8  Phase angle variation for 1 day (orbital period of the target).
Sun at summer solstice

Fig. 7  Visual magnitude for 1 day (orbital period of the target) with
omitted solar exclusion region. Sun at summer solstice

simulation includes an ideal distribution of targets in GEO,

whose position is presented in Fig. 14. Such position is listed
in Table 2 in terms of the true anomaly 𝜈 of each target at Fig. 9  Visual magnitude for 1 day (orbital period of the target). Sun
at summer solstice
the beginning of the simulation. For the sake of clarity it
has to be noticed that for a circular orbit like the GEO one,
the definition of a true anomaly angle loses its meaning. to target direction is smaller than half of the angular width
Nevertheless, it has been used during this study as to give of the FoV (0.7◦).
an immediate idea of the position of an object along its orbit. It has been found that the pinch points do not fall within
The null value of 𝜈 along the GEO orbit has been set on the any of the non-visibility regions; hence, the only limit affect-
point of intersection between the ECI positive X axis and ing the detection of a specific target is the occurrence of
the orbital plane. an Earth blockage during the crossing of the target in the
For the detection to occur, the target must cross the FoV sensor’s FoV.
of one of the sensors, which has an angular width of 1.4◦
centred around the location of the observed pinch point. It is 4.3 Performance results
then considered that the target crosses the FoV of the sensor
when the angle between the direction of the line of sight of The performance of the constellation is mainly analysed in
the sensor (sensor to pinch point direction) and the sensor terms of number of detections achieved by each sensor and
duration of each observation.

13
270 F. Vitiello

Fig. 12  Evolution of the inclination and RAAN of geosynchronous

satellites (blue dots) caused by the lunisolar and J2 perturbations.
Reprinted with permission Courtesy of MIT Lincoln Laboratory,
Lexington, Massachusetts [8]
Fig. 10  Visual magnitude for 1 day (orbital period of the target) with
omitted Earth blockage events. Sun at summer solstice
out of 9 targets, five of which are observed for a time that
goes from three minutes to almost six minutes.
In Table 3, the observation duration is reported. This By using a constellation of sensors not only can we ensure
quantity has been evaluated as the difference between the a complementarity in the observation of the targets, but also
time instant when a certain target is last “seen” by the a greater number of observations (thus further measure-
sensor and the time instant when it is first seen by the ments) of the same target.
same. Because of the occurrence of the Earth blockage,
this duration is typically different for every target. In the 4.3.1 Variation in the Position of the Sun
worst scenario, the Earth blockage completely blocks the
detection of a target causing a null value of the observa- When the Sun is located at the summer solstice point, the
tion duration. rotation of the line of nodes of the orbit of the sensors in the
It is easy to understand how advantageous the use of ECI reference frame causes a different temporal disposition
a constellation of two sensors is in terms of number and of the Earth blockage events, as shown in Fig. 10. For this
duration of each observation. As a matter of fact, the per- reason, a change in the observation durations of the constel-
formance of sensor S1, which is only able to detect 3 out lation is expected.
of 9 targets, two of which are observed for less than one Still, the different ECI location of the solar exclusion
minute, is far to be as satisfactory as the one of sensor S2. region does not interfere with the detection of any target
This latter sensor is, on the other hand, able to observe 7 since none of the observed space windows (pinch points) fall

Fig. 11  Illustration of the

population of satellites on
the geosynchronous belt in a
24-h period as seen in the ECI
reference frame. The contours
show the number of objects
passing through a 1.4◦ x1.4◦
FoV. Reprinted with permis-
sion Courtesy of MIT Lincoln
Laboratory, Lexington, Mas-
sachusetts. [8]

13
PErformance Analysis of Space Surveillance Using Space‑Based Optical Sensors 271

Fig. 13  MATLAB illustration

of the geometry at the begin-
ning of the simulation. The line
of sights (red lines) of the two
sensors (black dots) are pointed
towards the pinch points

Table 3  Performance of sensors Target Observation

S1 (top) and S2 (bottom). Sun duration
at vernal equinox [min]

T1 0
T2 0
T3 0
T4 0
T5 0.15
T6 0.12
T7 2.8
T8 0
T9 0
T1 1.4
T2 0.15
T3 0.12
T4 5.7
T5 2.8
T6 0
Fig. 14  MATLAB illustration of the distribution of targets (red dots) T7 5.6
at time t = 0
T8 0
T9 3.7

Table 2  Location of targets Target 𝜈0

along their orbit at time t = 0 in
terms of true anomaly T1 30◦
T2 65◦
T3 70◦
T4 80◦
T5 245◦ fact, the sensor is now able to detect 6 out of 9 targets, four
T6 250◦ of which are also observed for the whole duration of their
T7 270◦ transit through the FoV of the sensor. The performance
T8 350◦ of this sensor is now comparable with that of sensor S2
T9 180◦ which does still show a greater productivity, being it able
to detect 7 out of 9 targets.
It appears to be clear that an improvement in the overall
within such region. The Earth blockage is, again, the only performance of the constellation is obtained when the Sun
limit for the detection of targets to be achieved. is located at the summer solstice since the Earth blockage
As Table 4 shows, the performance of sensor S1 is now events have been found to occur less frequently during the
improved with respect to the one of the previous cases. In passage of targets in the FoV of the sensors.

13
272 F. Vitiello

Table 4  Performance of sensors Target Observation However, a clear advantage in the implementation of a
S1 (top) and S2 (bottom). Sun duration space-based telescope in a space surveillance network is
at summer solstice [min] found when the goal of its mission is the detection of GEO
T1 0
debris. As a matter of fact, as the last simulation has shown,
T2 2.7 and 2.8
in a 24-h period, such sensor is ideally able to detect many
T3 0.3
objects orbiting in the GEO ring thus having a quick access
T4 5.5
to this particular space region. A ground-based telescope
T5 0
pointed towards GEO, because of the peculiarity of this type
T6 5.6
of orbit, would only be able to continuously follow those
T7 0
debris which fall in its FoV.
T8 5.3
The last simulation of this study has been developed as
T9 5.4
to identify the possible advantage of the use of a constella-
T1 5.4
tion of two SBV-like sensors operating on the same orbit.
T2 0
Such constellation was kept “staring” at a fixed region in
T3 5.6
the inertial space which was chosen to coincide with the
T4 5.5
geosynchronous pinch points region. Thanks to the inertial
T5 2.7 and 2.8
location of such points in addition to the two positions of
T6 0.3
the Sun, the two sensors do not experience the constrain
T7 5.5
given by the Earth shadow and the solar exclusion region.
T8 0
The performances, studied in terms of the duration of the
T9 3.7
observation of each target and number of detected targets,
showed that, as expected, the use of a constellation is a clear
advantage in terms of productivity.

5 Conclusions
Funding Open access funding provided by Università degli Studi di
Napoli Federico II within the CRUI-CARE Agreement..
The study developed during this project has shown that the
use of a space-based optical sensor placed on a dawn–dusk, Open Access This article is licensed under a Creative Commons Attri-
circular Sun-synchronous LEO orbit, having a conical FoV bution 4.0 International License, which permits use, sharing, adapta-
of 1.4◦, is advantageous based on its capability to comple- tion, distribution and reproduction in any medium or format, as long
ment the current ground-based optical sensors detection as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons licence, and indicate if changes
capability. were made. The images or other third party material in this article are
In fact, it has been found that in a 24-h period during included in the article’s Creative Commons licence, unless indicated
which the Sun is fixed at the vernal equinox position, such otherwise in a credit line to the material. If material is not included in
sensor, whose sensitivity limit has been set to a value of the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will
visual magnitude of 15, is ideally able to detect a 50-cm need to obtain permission directly from the copyright holder. To view a
diffuse spherical object in the GEO ring. copy of this licence, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/.
As far as the operative time is concerned, by summing
the duration in time of the targets transit through the non- References
visibility regions and the Earth blockage events, it has been
found that the time during which such sensor is not able to 1. ESA Space Debris Office, ESA’s Annual Space Environment
operate is about 15 h, when the Sun is located at the vernal Report (2019)
2. Bonnal, C., McKnight, D.S.: IAA Situational Report on Space
equinox point, and about 14 h when the Sun is located at the Debris-2016 (2016)
summer solstice point. It appears to be clear, that, even if this 3. Harrison, D.C., Chow, J.C.: The space-based visible sensor. John
type of sensor is not affected by the presence of the Earth’s Hopkins Apl. Techn. Dig. 17, 226–236 (1996)
atmosphere, its operations are still strongly limited by the 4. Shell, J.R.: Commercially-Hosted Payloads for Debris Monitoring
and Mission Assurance in GEO, (2011)
presence of the Earth and the Sun. Still, a ground-based 5. Kervin, P.W., Hall, D., Bolden, M., Toth, J.: Phase Angle: What
telescope is limited to operate during nighttime only (with is it good for?. Advanced Maui Optical and Space Surveillance
an ideal duration of 12 h) and to those moments when the Conference, Maui, Hawaii (2010)
observed sky portions is clear (with some sites being clear 6. de Iaco, V.A.: Shadow times of earth satellites. Rivista Italiana di
Compositi e Nanotecnologie. Springer, Berlin (2014)
no more than 25% of the time [7]). 7. Ackermann, R.M., Colonel Kiziah, R.R., Zimmer, P.C., McGraw,
J.T., Cox, D.D.: A Systematic Examination of Ground-Based and

13
PErformance Analysis of Space Surveillance Using Space‑Based Optical Sensors 273

Space-Based Aprroaches to Optical Detection and Tracking of Publisher’s Note Springer Nature remains neutral with regard to
Satellites, 31st Space Symposium. Colorado Springs, Colorado jurisdictional claims in published maps and institutional affiliations.
(2015)
8. Sharma, J., Stokes, G.H., von Braun, C., Zollinger, G., Wiseman,
A.J.: Toward operational space-based space surveillance. Linc.
Lab. J. 13, 309–334 (2002)

13