Page 1 of 42

FINAL REPORT: ASTEROID DEFLECTION

AAE 59000: AEROSPACE PROPULSION

PURDUE UNIVERSITY
AEROSPACE ENGINEERING DEPARTMENT

Shajed Ahmed

Professor: William E. Anderson

December 5, 2021
 Page 2 of 42

Abstract
In this report, we discuss the need for an asteroid deflection and some potential solutions that
exist, such as nuclear detonation, kinetic impactor and gravity tractors. We begin defining
mission requirements by determining the various asteroid sizes and which asteroids humanity
would need to focus on in the next 100 years. We identify some possible propulsion systems that
could satisfy these mission requirements and take a deeper look into a novel concept known as
the “Keplerian Gravity Tractor”. We detail a mission from Earth, launching from Cape
Canaveral to our asteroid body in question, 101955 Bennu. Through analysis, we check if
NASA’s DART, a similar asteroid designed as a kinetic impactor could theoretically deflect
Bennu via the Keplerian Gravity Tractor. Our required deflection is one Earth radius, and the
spacecraft must fit within current rocket configurations. Based on our results, we determine
NASA’s DART is not able to deflect Bennu, due to running out of fuel prior to mission
completion. We find that if we were able to theoretically increase the fuel mass by four times the
original, we could have substantial deflection within 10 years.
 Page 3 of 42

Table of Contents

Abstract .......................................................................................................................................... 2
Section 1: Introduction ................................................................................................................. 4
Section 2: Background.................................................................................................................. 8
Section 3: Design Requirements ................................................................................................ 13
Section 4: Analysis and Results ................................................................................................. 17
Section 4.1: Analysis Approach ............................................................................................. 17
Section 4.2: Ground to Geostationary Transfer Orbit ........................................................ 17
Section 4.3: Interplanetary Hohmann Transfer to Asteroid Body [20] ............................... 19
Section 4.4: Keplerian Gravity Tractor [8] ............................................................................ 24
Section 5: Summary & Conclusions .......................................................................................... 32
References .................................................................................................................................... 33
Appendix ...................................................................................................................................... 35
 Page 4 of 42

Section 1: Introduction
On February 15, 2013, a 20-meter asteroid entered Earth’s atmosphere and exploded in the air near
the Russian city of Chelyabinsk. Referred to as the Chelyabinsk meteor, the object exploded at a
height of 29.7 km (97,600 ft) from the Earth’s surface. The explosion produced additional
meteorite fragments that dispersed throughout Chelyabinsk and a massive shockwave when it
impacted the atmosphere. The kinetic energy of the shockwave was estimated from infrasound and
seismic measurements to be about the equivalent of 400-500 kilotons of TNT, which was more
than 30 times the estimated amount of energy released by the Hiroshima atomic bomb (~15
kilotons of TNT). There were no deaths reported from the meteor explosion, however
approximately 1,500 people suffered indirect injuries from the broken glass from blown in
windows from the shockwave. Figure 1 depicts the proposed damage if an asteroid of similar size
to that of Chelyabinsk, approximately 30 meters (100 feet) wide, were to hit New York City. [2] [18]

Figure 1: Proposed Damage if Chelyabinsk Asteroid had impacted NYC

(Image credit: National Science & Technology Council)

This is not the first time that an asteroid of this magnitude had struck Russia. In fact, it was
estimated the Tunguska event of 1908 (which occurred near the Podkamennaya Tunguska River
in Russia.) was approximately 12 megatons, almost 240 times of Chelyabinsk. Figure 2 depicts a
comparison of the Chelyabinsk and Tunguska asteroids with the Empire State building and the
Eiffel Tower. [4] [7]
 Page 5 of 42

Figure 2: Comparison of sizes of Empire State Building and Eiffel Tower to Chelyabinsk
(CM) and Tunguska (TM) meteoroids. Approximate Scale.
Image Credit: Wikipedia User Phoenix CZE

In 2014, NASA’s Near-Earth Object (NEO) program released a map, seen in Figure 3, of the
asteroids from 1994 -2013 that entered and disintegrated in the Earth’s atmosphere. The data shows
that approximately 500 small asteroids impacted Earth’s atmosphere, including the Chelyabinsk
asteroid (depicted with a red dot). Majority of these asteroids were harmless and disintegrated in
the atmosphere, but exceptions like Chelyabinsk cause concern. Our current methods for detecting
asteroids are not very good but hopefully with data such as this, scientists can better estimate the
sizes and distributions of NEOS, especially the ones that could pose a danger to Earth. [14]

Figure 3: Bolide Events from 1994 – 2013

Image Credit: Planetary Science.
 Page 6 of 42

So, what can we do?

Current strategies for asteroid deflection include nuclear interceptors, kinetic impactors, and
gravity tractors. [19]

One option is detonating a nuclear charge on/nearby the surface of an asteroid, known as a nuclear
interceptor. It is important to determine how the asteroid will respond to the nuclear detonation.
The asteroid could break up into smaller fragments and could make the situation worse. As a result,
it is very important to know the structural characteristics and material composition of the asteroid
to ensure appropriate detonation. [3]

Kinetic impactors aim to exert an impulse on the asteroid by colliding with it and changing its
direction. Like nuclear detonation, the composition and structure of the asteroid should be studied
to ensure that it doesn’t break up during the collision. Unlike nuclear detonation, the kinetic
impactor is much more controlled, and its results can be modeled with certainty. [9]

Gravity tractors is the strategy that this paper will follow specifically a novel concept theorized by
Yohannes Ketema, an Associate Professor at the University of Minnesota, he refers to as the called
the “The Keplerian Gravity Tractor”.

He proposes a method for asteroid deflection that uses a spacecraft moving back and forth on a
segment of a Keplerian orbit, as shown in the Figure 4 below. The spacecraft uses impulsive
motions within the orbit of the asteroid, tugging the asteroid and causing it to deflect from its
original path.

Figure 4: Schematic of Keplerian Gravity Tractor [8]

Image Credit: Yohannes Ketema

 Page 7 of 42

In his paper, he goes through the process of the derivation of a deflection formula for a Keplerian
gravity tractor, which can be seen below:
𝑁 Δ𝑉 𝑁 Δ𝑉
𝐾 −
𝐼𝑠𝑝 𝑔0
(𝑖−1) 𝐾 −
𝐼𝑠𝑝 𝑔0
(𝑖−1)
∆𝜁 = − (𝑡𝑒 − 𝑡𝑠 )𝜆𝑚𝑠𝑝𝑎𝑐𝑒𝑐𝑟𝑎𝑓𝑡 Δ𝑡 ∑ 𝑉𝑎𝑖 𝑒 + Δ𝑡 2 𝜆𝑚𝑠𝑝𝑎𝑐𝑒𝑐𝑟𝑎𝑓𝑡 ∑ 𝑉𝑎𝑖 𝑖𝑒
𝑚𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 𝑚𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑
𝑖=1 𝑖=1

Ketema assumes a spacecraft with a mass of 1500 kg, that is intended to the asteroid 2007
VK184, which was proposed to collide with Earth in 2048 at the time the paper was written. He
compares the performance of his Keplerian restricted motion gravity tractor to that of a
stationary gravity tractor. In the Figure 5 below, we can see the how the Keplerian gravity
tractor has phenomenal performance in deflecting the asteroid, as the lead time increases, it’s
effectiveness increases as well.

Figure 5: Snippet from Ketema’s Paper Highlighting the Performance of the Keplerian
Gravity Tractor Compared to a Stationary Gravity tractor. [8]

Image Credit: Yohannes Ketema

 Page 8 of 42

Section 2: Background
NASA has already started looking into kinetic impactors in the form of the Double Asteroid
Redirection Test (DART) mission, which is being planned to be launched later this year. DART
will be the first demonstration of the kinetic impactor technique. DART is intended to target 65803
Didymos, specifically Didymos’ moonlet which is about 160 meters in length. The moonlet
represents the typical asteroid size that could pose a realistic threat to Earth. Figure 6 depicts the
DART spacecraft along with its various subsystems. [12]

Figure 6: Two different views of the DART spacecraft.

Image Credit: NASA Planetary Defense

The DART spacecraft is powered by NASA’s Evolutionary Xenon Thruster (NEXT) engine
thrusters. The NEXT engine is a gridded-ion thruster and is a type of electric propulsion which
uses electricity to accelerate the propellant (in this case Xenon) to high speeds. The benefit of a
propulsion system like this is that the theoretical top speed is unlimited because it would only be
limited by the voltage applied to the system. Additionally, the engine can run for an extended
period in space because it can use solar arrays to generate the power needed. Table 1 shows the
NEXT Engine performance specifications. [13]

Table 1: NEXT Engine Performance Specifications

Top Speed 90,000 mph (145,000 km/h)

Thrust 236 mN
Minimum Thruster Power 0.5 kW
Thruster Power 6.9 kW
Specific Impulse 4,190 seconds
 Page 9 of 42

In gravity-tractor deflection, the spacecraft exerts a force on the asteroid through a gravitational
tugging effect. The spacecraft imparts a constant gravitational force that helps redirect the
trajectory of the asteroid. This method requires more fuel than the other proposed methods because
majority of the thrust from the fuel goes unused to help align the thrusters. However, with a
propulsion system like the NEXT ion thrusters the gravity-tractor method can be an extremely
feasible method for asteroid deflection. [6] [10]

The first step in our requirement definition would be to determine the sizes of asteroid that we
could realistically deflect with our current technology and knowledge. The European Space
Agency has a very helpful infographic that summarizes the various asteroid sizes and their
respective impact in Figure 7.

Figure 7: ESA Dangerous Asteroids

Image Credit: European Space Agency

One asteroid of interest on the infographic is Apophis. 99942 Apophis, a 340-meter-long asteroid,
depicted in Figure 8, caused significant concern in the scientific community since its discovery in
2004 and was identified as one of the most hazardous asteroids that could impact Earth. Fortunately
for us, as of March 2021, 99942 Apophis has been recently determined by NASA JPL that Earth
is safe from Apophis for at least another century. [16]
 Page 10 of 42

Figure 8: Comparison of possible size of Apophis asteroid to Eiffel Tower and Empire State
Building Approximate Scale.
Image Credit: Wikipedia User Phoenix CZE

However, there is no shortage of asteroids that would like to cause massive destruction or wipe
out the face of humanity. NASA JPL’s Center for Near Object Studies (CNEOS) has an online
tool called Sentry, which is an automated collision monitoring system that continually scans the
most current asteroid catalog for possibilities of future impact with Earth over the next 100 years.
[17]

The Sentry tool uses two scales for categorizing the impact hazard of NEOs, the Torino scale and
the Palermo scale. The Torino scale was developed to communicate the risk associated with an
asteroid to the general public. The Torino scale, shown below, has integer values from 0 – 10 and
considers the likelihood of the event occurring and the potential impact energy. [11]

Figure 9: Torino Scale

Image Credits: CNEOS NASA JPL
 Page 11 of 42

Meanwhile the Palermo Scale was intended to be used by scientists and engineers and is the base-
10 logarithm of the relative risk, which is defined as

𝑃𝑆 = 𝑙𝑜𝑔10 𝑅

The relative risk R is given by

𝑃𝐼
𝑅=
𝑓𝑔 ∗ 𝐷𝑇

Here 𝑃𝐼 is the impact probability of the event in question and DT is the time (in years) until the
potential event. The annual background impact frequency, 𝑓𝑔 = 0.03 ∗ 𝐸 −4/5 is the annual
probability of an impact event with an equivalent energy (E, in megatons of TNT) could occur
compared to the one in question.

The cumulative Palermo Scale value takes the base-10 logarithm of the sum of the individual
relative risk values to reflect how serious the impact event could be. [1]

𝑃𝑆𝑐𝑢𝑚 = 𝑙𝑜𝑔10 (10𝑃𝑆1 +10𝑃𝑆2 +10𝑃𝑆3 + ⋯ )

For our reference, 99942 Apophis held the record for the highest Palermo scale values, with a
value of 1.10 for impact in 2029. [5]

For this study we will use the Palermo scale to determine possible asteroids that could impact Earth
within the next 100 years. Using the Sentry tool and filtering for the largest Palermo we can
generate Table 2. Highlighted in red are NEOs that are more than 100 meter in size and could
cause potential planet extinction and highlighted in yellow are less than 100-meter diameter but
could cause immense damage, like the Chelyabinsk and Tunguska impact events. [17]

Table 2: Sentry: Earth Impact Monitoring – Impact Risk Data [17]

Impact Estimated
Palermo Scale
Object Designation Year Range Probability Diameter
(cum.)
(cumulative) (km)
101955 Bennu (1999
2178-2290 5.7e-4 0.49 -1.41
RQ36)
29075 (1950 DA) 2880-2880 1.2e-4 1.3 -1.42
(2009 JF1) 2022-2022 2.6e-4 0.013 -2.58
(2000 SG344) 2069-2121 2.8e-3 0.037 -2.79
(2021 QM1) 2050-2118 1.5e-4 0.047 -3
(2008 JL3) 2027-2119 1.6e-4 0.029 -3.16
(2021 EU) 2024-2096 4.8e-5 0.028 -3.25
(2021 RR23) 2040-2110 1.3e-6 0.303 -3.28
(2010 RF12) 2095-2120 4.8e-2 0.007 -3.31
(2005 QK76) 2030-2059 7.2e-5 0.031 -3.39
(2021 GX9) 2032-2052 9.1e-5 0.029 -3.4
(2005 ED224) 2023-2064 2.5e-6 0.054 -3.46
 Page 12 of 42

(2015 JJ) 2111-2111 2.7e-5 0.13 -3.48

(2008 UB7) 2044-2101 3.3e-5 0.058 -3.67
(2007 DX40) 2035-2119 7.6e-5 0.04 -3.7
(2000 SB45) 2067-2118 1.6e-4 0.046 -3.75
(2012 QD8) 2047-2120 6.1e-6 0.081 -3.78
(2008 EX5) 2056-2090 5.3e-5 0.059 -3.81
(2020 VW) 2074-2079 7.0e-3 0.007 -3.85
(2017 WT28) 2083-2120 1.2e-2 0.008 -3.86
(2020 VV) 2044-2120 2.3e-3 0.012 -3.87
(2013 VW13) 2063-2095 4.4e-4 0.019 -3.89
(2012 HG2) 2052-2121 1.9e-3 0.014 -3.91
 Page 13 of 42

Section 3: Design Requirements

For this analysis, we will focus on the asteroid 101955 Bennu (1999 RQ36) and attempt to
deflect it within the requirements using the novel Keplerian gravity tractor concept highlighted in
Ketema’s paper.

The requirements are simple and open-ended:

1. The spacecraft must deflect the asteroid 101955 Bennu at least 1 Earth radius (6,738 km)
given a 10-year lead time.
2. Deflection spacecraft wet mass must be within the limits of existing rocket payload mass
configuration e.g., Atlas V, SpaceX Falcon Heavy, Delta IV Rockets.

Table 3: 101955 Bennu (1999 RQ36) Object Details: Earth Impact Risk Summary [17]

Torino Scale (maximum) 0 Vimpact 12.68 km/s

Palermo Scale (maximum) -1.59 Vinfinity 5.99 km/s
Palermo Scale (cumulative) -1.41 H 20.6
Impact Probability
5.7e-4 Diameter 0.490 km
(cumulative)
Number of Potential Impacts 157 Mass 7.4e+10 kg
Impact Search Technique MC Energy 1.4e+3 Mt
Analysis based on 554 observations spanning 7693.3
(1999-09-11.40624 to 2020-10-3.80160)

We can see in Table 2 (Impact Risk Table), that Bennu has the highest cumulative Palermo scale
out of all the Near Earth Objects the JPL Sentry tool has detected. Based on impact simulations,
Bennu has a 0.057% chance of impacting the Earth. If it were to impact, it would release ~1400
Megatons of TNT. This Tsar Bomba in comparison had a blast yield of only 50-58 Megatons of
TNT and is considered the most powerful nuclear weapon ever created and tested.
Let’s familiarize ourselves with 101955 Bennu’s orbit and physical parameters, as shown in
Table 4 and Table 5 respectively. Additionally, we can see Bennu’s orbit relative to the Sun and
where it intersects with Earth’s orbit, in Figure 10.
 Page 14 of 42

Table 4: Orbit Parameters of 101955 Bennu [17]

Element Value Units
Eccentricity 0.20375
Semi-major axis 1.12639 au

Perihelion Distance 0.89689 au

Inclination 6.03494 deg

Longitude of Ascending Node 2.06087 deg

Argument of Perihelion 66.2231 deg

Mean Anomaly 101.704 deg
Epoch 2455562.5 (2011-Jan-01.0) TDB
Reference: JPL 118 (heliocentric IAU76/J2000 ecliptic)

Table 5: Physical Parameters of 101955 Bennu [17]

Element Value Units
Velocity at impact 12.68 km/s
Hyperbolic Excess Velocity 5.99 km/s
Standard Gravitational
Parameter 4.89E-09 km^3/s^2
Diameter 0.49 km
Mass 7.4E+10 kg
Energy 1400 Mt
Bulk Density 1.194 g/cm^3
 Page 15 of 42

Figure 10: Orbit of Bennu Relative to the Sun and the Earth. Snapshot taken on December
6, 2021.

Image Credits: CNEOS NASA JPL

For our spacecraft, we will assume that a spacecraft like NASA’s DART has been designed, but
instead of being a kinetic impactor it is used as a Kepler gravity tractor.

Table 6: DART Physical Properties

Element Value Units
DART Wet Mass 720 kg
DART Dry Mass 610 kg
Combined Propellant
Mass 110 kg
Hydrazine Mass 50 kg
Xenon Mass 60 kg

However, one thing to note is that DART is meant to impact Dimorphos which is a small satellite
asteroid (0.017 km in diameter) that orbits the larger asteroid, Didymos. For this reason, we will
assume our hypothetical spacecraft has double the mass of DART, both structural mass and
propellant mass. This will hopefully allow us to successfully deflect 101955 Bennu.
 Page 16 of 42

Table 7: Hypothetical Spacecraft Physical Properties

Element Value Units
Spacecraft Wet Mass 1440 kg
Spacecraft Dry Mass 1220 kg
Combined Propellant
Mass 220 kg
Hydrazine Mass 100 kg
Xenon Mass 120 kg

We will assume in our calculations that our spacecraft is powered by 2 NEXT-C Engines. These
are attached on either side of the spacecraft and provide the impulsive thrusts required for the
restricted Keplerian motion. Since, we are only using the impulsive thrust to change the direction
of the spacecraft, this does not necessarily increase the performance of the spacecraft. We could
have 1 NEXT-C engine on our spacecraft but would require some sort of mechanism that
changes its direction to have the same behavior described above.
Table 8: NEXT-C Engine Performance Specifications

Element Value Units

Top Speed 145,000 km/h
Thrust 236 mN
Minimum Thruster
Power 0.5 kW
Thruster Power 6.9 kW
Specific Impulse 4190 sec
 Page 17 of 42

Section 4: Analysis and Results

Section 4.1: Analysis Approach

For the analysis we will break down the mission into three phases and as a result complete three
sets of calculations. These phases can be defined as:
1. Ground to Geostationary Transfer Orbit (GTO): Here we determine the increment
velocities required to place our spacecraft into GTO using readily available rockets.
2. Hohmann Transfer to the Asteroid Body – Patched Conics Method: Here we utilize the
patched conics methods to determine an approximate Hohmann transfer our spacecraft
can commit to arrive at 101955 Bennu. We determine the increment velocities that we
would need to be placed in Bennu’s orbit.
3. Keplerian Motion Gravity Tractor: Here we use Ketema’s deflection equation for a
Keplerian Gravity Tractor and see if we can use our hypothetical spacecraft to deflect
101955 Bennu while moving in a restricted Keplerian motion within the asteroids orbit.

Section 4.2: Ground to Geostationary Transfer Orbit

For this first stage of the mission, we can treat it similarly as one of our homework problems. In
this case, we are attempting to launch the spacecraft into the standard GTO, defined as 185 km
perigee by 35,788 km apogee. We will launch from Cape Canaveral, which is located
approximately 28.5 degrees North latitude. Additionally, we will be using the SpaceX Falcon 9
Block 1 Merlin 1C Configuration. We will determine the ∆𝑉 required for each stage, assuming
the 1st Stage provides 40% of the required ∆𝑉 and the second stage provides the remainder.

Table 9: SpaceX Falcon 9 Block 1 Merlin 1C Configuration

Falcon 9 - Merlin 1C
(est.) Stage 1 Stage 2
Diameter (m) 3.66 m 3.66 m
Length (m) ~30.1 m ~10.0 m
Empty Mass (tonnes) ~19.24 t ~3.1 t
Propellant Mass (tonnes) ~239.3 t ~48.9 t
Total Mass (tonnes) ~258.5 t ~52 t
Merlin
Engine Merlin 1C Vac
Engine Mfgr SpaceX SpaceX
Fuel RP1 RP1
Oxidizer LOX LOX
Thrust (SL tons) 387.825 t
Thrust (Vac tons) 442.938 t 42.18 t
ISP (SL sec) 266 s -
ISP (Vac sec) 304 s 336 s
Burn Time (sec) 180 s 346 s
No. Engines 9 1
 Page 18 of 42

Radius of the Earth, 𝑟𝐸𝑎𝑟𝑡ℎ 6378.1 𝑘𝑚

Standard Gravitational Parameter of Earth, 𝑚3
𝜇𝐸𝑎𝑟𝑡ℎ 3.986 ∗ 1014 2
𝑠
Height of Perigee, ℎ𝑝𝑒𝑟𝑖𝑔𝑒𝑒 185 𝑘𝑚
Height of Apogee, ℎ𝑎𝑝𝑜𝑔𝑒𝑒 35788 𝑘𝑚

Distances from Center of the Earth:

𝑟𝑝𝑒𝑟𝑖𝑔𝑒𝑒 = ℎ𝑝𝑒𝑟𝑖𝑔𝑒𝑒 + 𝑟𝐸𝑎𝑟𝑡ℎ
𝑟𝑝𝑒𝑟𝑖𝑔𝑒𝑒 = 6563.1 𝑘𝑚
𝑟𝑎𝑝𝑜𝑔𝑒𝑒 = ℎ𝑎𝑝𝑜𝑔𝑒𝑒 + 𝑟𝐸𝑎𝑟𝑡ℎ
𝑟𝑎𝑝𝑜𝑔𝑒𝑒 = 42166 𝑘𝑚
Semi-Major Axis
𝑟𝑎𝑝𝑜𝑔𝑒𝑒 + 𝑟𝑝𝑒𝑟𝑖𝑔𝑒𝑒
𝑎=
2
𝑎 = 24,365 𝑘𝑚
Velocity at Perigee
2 1
𝑉𝑝𝑒𝑟𝑖𝑔𝑒𝑒 = √𝜇𝐸𝑎𝑟𝑡ℎ ( − )
𝑟𝑝𝑒𝑟𝑖𝑔𝑒𝑒 𝑎
𝑉𝑝𝑒𝑟𝑖𝑔𝑒𝑒 = 10252 𝑚/𝑠
Velocity Increments
2𝜋𝑟𝐸𝑎𝑟𝑡ℎ cos (28.5°)
𝑉𝑖𝑛𝑖𝑡𝑖𝑎𝑙 =
24 ℎ𝑟𝑠
𝑉𝑖𝑛𝑖𝑡𝑖𝑎𝑙 = 407.62 𝑚/𝑠

∆𝑉𝑖𝑑𝑒𝑎𝑙 = 𝑉𝑝𝑒𝑟𝑖𝑔𝑒𝑒 − 𝑉𝑖𝑛𝑖𝑡𝑖𝑎𝑙

∆𝑉𝑖𝑑𝑒𝑎𝑙 = 9844.6 𝑚/𝑠

∆𝑉𝑠𝑡𝑎𝑔𝑒1 = 0.4∆𝑉𝑖𝑑𝑒𝑎𝑙
∆𝑉𝑠𝑡𝑎𝑔𝑒1 = 4922.3 𝑚/𝑠

∆𝑉𝑠𝑡𝑎𝑔𝑒2 = 0.6∆𝑉𝑖𝑑𝑒𝑎𝑙
∆𝑉𝑠𝑡𝑎𝑔𝑒2 = 7383.4 𝑚/𝑠
 Page 19 of 42

Section 4.3: Interplanetary Hohmann Transfer to Asteroid Body [20]

An important assumption to understand regarding the Hohmann transfers is that it is used

traditionally for co-planar movements. For the simplicity of this analysis and to have a
conceptual understanding for the design, we will assume that Hohmann transfer will be valid.

We can modify the concept of Hohmann transfer, so it becomes an interplanetary transfer. In

order to do this, we decompose the transfers into sets of two-body problems. For each celestial
body, we define its sphere of influence. This is the region where the celestial body in question
has a high gravitational attraction than the sun. So, in order to escape the celestial bodies SOI, we
implement a hyperbolic escape trajectory and the method of patched conics.

This phase is broken into subphases that are connected by a series of two body problem
“patches”. This allows the analysis to be simpler and give a close approximation to the actual
method. For this example, we will assume that the appropriate launch time has been selected.

1. Hyperbolic Departure
In this step, we will determine what the necessary ∆𝑉 is to escape Earth’s gravitational sphere of
influence. This will simulate a Hohmann transfer orbit from Earth’s orbit to our asteroid. This
escape trajectory is parallel to Earth’s velocity vector. We can see the departure hyperbola
depicted in FIGURE.

Figure 11: Hyperbolic Escape Trajectory

To enter the Hohmann transfer orbit, we must determine the hyperbolic excess speed (𝑉∞ ) in
order to place the spacecraft on the desired elliptical trajectory. This hyperbolic excess speed is
defined as the difference between the Earth’s orbital velocity around the Sun and the velocity the
 Page 20 of 42

spacecraft would need to transfer out of Earth’s orbit, or the velocity at the periapsis of the
transfer orbit.
𝑉∞,𝐸𝑎𝑟𝑡ℎ = |𝑉𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟,𝐸𝑎𝑟𝑡ℎ −𝑉𝑜𝑟𝑏𝑖𝑡𝑎𝑙,𝐸𝑎𝑟𝑡ℎ |

Where:
𝜇𝑠𝑢𝑛
𝑉𝑜𝑟𝑏𝑖𝑡𝑎𝑙,𝐸𝑎𝑟𝑡ℎ = √
𝑟𝐸𝑎𝑟𝑡ℎ,𝑆𝑢𝑛
and
2 1
𝑉𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟,𝐸𝑎𝑟𝑡ℎ = √𝜇𝑠𝑢𝑛 ( − )
𝑟𝐸𝑎𝑟𝑡ℎ,𝑆𝑢𝑛 𝑎𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟

Radius of Earth’s Orbit About the Sun, 1 𝐴𝑈 = 1.496 ∗ 108 𝑘𝑚

𝑟𝐸𝑎𝑟𝑡ℎ,𝑆𝑢𝑛
Radius of Asteroid’s Orbit About the Sun, 1.12639 𝐴𝑈 = 1.6851 ∗ 1011 𝑘𝑚
𝑟𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝑆𝑢𝑛
Gravitational Parameter of the Sun, 𝜇𝑠𝑢𝑛 20
𝑚3
1.327124400189 ∗ 10
𝑠2
𝑟𝐸𝑎𝑟𝑡ℎ,𝑆𝑢𝑛 − 𝑟𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝑆𝑢𝑛
Semi-Major Axis of Transfer Orbit, 𝑎𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟 = 9.4538 ∗ 109 𝑘𝑚
2

Therefore, we can find the hyperbolic excess speed:

𝑉∞,𝐸𝑎𝑟𝑡ℎ = 9.5962 ∗ 104 𝑚/𝑠

We now need to solve for the velocity that will occur at the periapsis of the hyperbolic escape
trajectory (𝑉𝑝𝑒𝑟𝑖𝑎𝑝𝑠𝑖𝑠,ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑖𝑐 ). This can be defined as

2 1
𝑉𝑝𝑒𝑟𝑖𝑎𝑝𝑠𝑖𝑠,ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑎 = √𝜇𝐸𝑎𝑟𝑡ℎ ( − )
𝑟𝑝𝑒𝑟𝑖𝑎𝑝𝑠𝑖𝑠 𝑎ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑎

For the sake of simplicity of the calculation, we will assume that 𝑟𝑝𝑒𝑟𝑖𝑎𝑝𝑠𝑖𝑠 = 𝑟𝑝𝑒𝑟𝑖𝑔𝑒𝑒 =
6563.1 𝑘𝑚 . As we saw from the analysis above, that is where the velocity is the highest which
we then use to determine our ideal velocity increment.
In order to determine the semi-major axis of the hyperbola (𝑎ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑎 ), we recognize that it is
the same semi-major axis for hyperbolic excess speed. Except in this case, the radius is
expressed as the radius of Earth’s sphere of influence (SOI). The sphere of influence of an
astronomical body is the region which it dominates the attraction of natural satellites. Earth’s
sphere of influence can be defined as:
2
𝜇𝑆𝑢𝑛 5
𝑟𝐸𝑎𝑟𝑡ℎ,𝑆𝑂𝐼 = ( ) 𝑟𝐸𝑎𝑟𝑡ℎ
𝜇𝐸𝑎𝑟𝑡ℎ
𝑟𝐸𝑎𝑟𝑡ℎ,𝑆𝑂𝐼 = 1.0319 ∗ 109 𝑘𝑚
 Page 21 of 42

−1
2 𝑣∞ 2
𝑎ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑎 = ( − )
𝑟𝐸𝑎𝑟𝑡ℎ,𝑆𝑂𝐼 𝜇𝐸𝑎𝑟𝑡ℎ
𝑎ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑎 = 4.3289 ∗ 104 𝑘𝑚
𝑉𝑝𝑒𝑟𝑖𝑎𝑝𝑠𝑖𝑠,ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑎 = 1.9284 ∗ 1012 𝑚/𝑠

We can now solve for our first burn, which is the difference between the velocity at the periapsis
and the location of the spacecraft, which is 𝑟𝑝𝑒𝑟𝑖𝑔𝑒𝑒 = 6563.1 𝑘𝑚.

𝜇𝐸𝑎𝑟𝑡ℎ
∆𝑉ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑖𝑐 𝑑𝑒𝑝𝑎𝑟𝑡𝑢𝑟𝑒 = |𝑉𝑝𝑒𝑟𝑎𝑖𝑝𝑠𝑖𝑠,ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑎 − √ |
𝑟𝑝𝑒𝑟𝑖𝑔𝑒𝑒
∆𝑉ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑖𝑐 𝑑𝑒𝑝𝑎𝑟𝑡𝑢𝑟𝑒 = 1.19284 ∗ 1012 𝑚/𝑠

We can additionally determine where we should burn. In Figure 11, there is an angle 'β'. This is
the angle between the apse line of the escape hyperbola, and the line parallel to Earth’s velocity
vector.

1
𝛽 = cos −1
𝑟𝑝𝑒𝑟𝑖𝑔𝑒𝑒 ∗ 𝑉∞ 2
1+
( 𝜇𝐸𝑎𝑟𝑡ℎ )
𝛽 = 89.6246 𝑑𝑒𝑔𝑟𝑒𝑒𝑠
 Page 22 of 42

2. Hyperbolic Arrival

With our spacecraft out of Earth’s sphere of influence we can now solve for arrival trajectory.
Like how we treated Earth as its own central body with the space craft orbiting it, we can assume
our asteroid to be the central body. Figure 12 shows the hyperbolic trajectory the spacecraft will
experience approaching the asteroid.

We can see this coincides with the flyby trajectory, so we will have a second burn that will
behave as an orbital insertion burn. This will slow down the spacecraft and allow it to enter the
orbit of the asteroid.

Once we enter the target planet's SOI, we can then assume the central body to be the target
planet. Our arrival trajectory should be hyperbolic like the diagram below:

Figure 12: Hyperbolic Arrival Trajectory

For the sake of simplicity, we will assume that the spacecraft burns into a circular orbit with a
radius that is equivalent to the periapsis distance of the hyperbolic arrival trajectory. Therefore,
the velocity we want to achieve would be equal to the circular velocity for an orbit about our
asteroid at the periapsis radius.

𝜇𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑
∆𝑉ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑖𝑐 𝑎𝑟𝑟𝑖𝑣𝑎𝑙 = |𝑉𝑝𝑒𝑟𝑎𝑖𝑝𝑠𝑖𝑠,ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑎 − √ |
𝑟𝑝𝑒𝑟𝑖𝑎𝑝𝑠𝑖𝑠
∆𝑉ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑖𝑐 𝑎𝑟𝑟𝑖𝑣𝑎𝑙 = 1.9284 ∗ 1012 𝑚/𝑠
 Page 23 of 42

Note that for this example, we used the periapsis distance upon arrival for simplicity. However, it
is possible to target a specific radius for our final orbit. In the next section, we will determine the
minimum radius from the asteroid the spacecraft should be to minimize plume impingement. If
we want, we could have set this as the radius instead of the periapsis distance. For now, let’s
assume that we want to capture pictures across the asteroid’s orbit, and this was the most
efficient method to enter the asteroids orbit.
 Page 24 of 42

Section 4.4: Keplerian Gravity Tractor [8]

Ketema defines the asteroid deflection formula using restricted Keplerian motion as:
𝑁 Δ𝑉 𝑁 Δ𝑉
𝐾 −
𝐼𝑠𝑝 𝑔0
(𝑖−1) 𝐾 −
𝐼𝑠𝑝 𝑔0
(𝑖−1)
∆𝜁 = − (𝑡𝑒 − 𝑡𝑠 )𝜆𝑚𝑠𝑝𝑎𝑐𝑒𝑐𝑟𝑎𝑓𝑡 Δ𝑡 ∑ 𝑉𝑎𝑖 𝑒 + Δ𝑡 2 𝜆𝑚𝑠𝑝𝑎𝑐𝑒𝑐𝑟𝑎𝑓𝑡 ∑ 𝑉𝑎𝑖 𝑖𝑒
𝑚𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 𝑚𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑
𝑖=1 𝑖=1
In order to determine the deflection, we must identify the various parameters within the equation.
We start with identifying the true anomaly on the asteroid’s orbit at the time of encounter.

Figure 13: Orbit of 101955 Bennu and Earth

Figure 13 depicts the orbit of 101955 Bennu and Earth, here we can see the position of the
projected encounter (𝑟𝑝 ) and the true anomaly (𝑓𝑒 ) on the asteroid’s orbit at the time of
encounter. We can determine the true anomaly using the equation of path.
𝑎𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (1 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 2 )
𝑟𝐸𝑎𝑟𝑡ℎ =
1 + 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 𝑐𝑜𝑠(𝑓𝑒 )
Which can be rearranged to solve for the true anomaly, note that 𝑟𝑝 = 𝑟𝐸𝑎𝑟𝑡ℎ :
1 𝑎(1 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 2 )
𝑓𝑒 = cos−1 ( ( − 1))
𝑒 𝑟𝐸𝑎𝑟𝑡ℎ

𝑓𝑒 = 11.0452 𝑟𝑎𝑑
The angle 𝜎 can be determined by subtracting the sum of the longitude of ascending node (Ω)
and argument of the periapsis (𝜔) by 2𝜋.
𝜎 = 2𝜋 − (Ω + 𝜔)
𝜎 = 0.5721 𝑟𝑎𝑑
Using the energy equation (vis-viva equation), we can determine the velocity of the asteroid at
the time of encounter:
 Page 25 of 42

𝜇𝑠𝑢𝑛 𝜇𝑠𝑢𝑛
𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝑒𝑛𝑐𝑜𝑢𝑛𝑡𝑒𝑟 = √2 ( − )
𝑟𝐸𝑎𝑟𝑡ℎ 2𝑎𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑
𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝑒𝑛𝑐𝑜𝑢𝑛𝑡𝑒𝑟 = 64,509 𝑚/𝑠

We can use the expression for angular momentum (ℎ) to find the flight path angle (𝛾) at the time
of encounter:
ℎ = 𝑟𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝑒𝑛𝑐𝑜𝑢𝑛𝑡𝑒𝑟 𝑐𝑜𝑠𝛾 = √𝑎𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (1 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 2 )𝜇𝑠𝑢𝑛
−1
√𝑎𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (1 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 2 )𝜇𝑠𝑢𝑛
𝛾 = cos
𝑟𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝑒𝑛𝑐𝑜𝑢𝑛𝑡𝑒𝑟
𝛾 = 5.2413 𝑟𝑎𝑑

Figure 14: Earth and Asteroid at the time of the “Closest” Approach

Referring to Figure 14, we can calculate the magnitude of the relative velocity of the asteroid
with respect to the Earth:

𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝐸𝑎𝑟𝑡ℎ = √𝑉𝐸𝑎𝑟𝑡ℎ 2 + 𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 2 − 2𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 𝑉𝐸𝑎𝑟𝑡ℎ 𝑐𝑜𝑠𝛾

𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝐸𝑎𝑟𝑡ℎ = 64,358𝑚/𝑠

Looking at the geometry we can solve for the incidence angle (𝜓) using a ratio:

𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝐸𝑎𝑟𝑡ℎ 𝑉𝐸𝑎𝑟𝑡ℎ
=
𝑠𝑖𝑛𝛾 𝑠𝑖𝑛𝜓

𝑉𝐸𝑎𝑟𝑡ℎ
𝜓 = sin−1 ( 𝑠𝑖𝑛𝛾)
𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝐸𝑎𝑟𝑡ℎ

𝜓 = −0.004 𝑟𝑎𝑑
 Page 26 of 42

We can then solve for the constant 𝐾:

3𝑎𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑
𝐾= 𝑉𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑,𝑒𝑛𝑐𝑜𝑢𝑛𝑡𝑒𝑟 𝑠𝑖𝑛𝜓
𝜇𝑠𝑢𝑛

𝐾 = 0.0181 𝑠/𝑚

Figure 15: Plume Non-Impingement Condition

We now solve for the smallest radius (𝑟𝑝𝑚 ) that satisfies the plume non-impingement condition,
as seen in Figure 15. Using the equation of a two-body problem we get:

(1 + 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 cos 𝜃𝑏 )
𝑟𝑝𝑚 = 𝑟
(1 + 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 )cos (𝜑 − 𝛾) 𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑

For simplicity of the formula, let’s define 𝜑 = 20 𝑑𝑒𝑔𝑟𝑒𝑒𝑠 and 𝜃𝑏 = 1. Solving for the smallest
radius:

𝑟𝑝𝑚 = 1141.2 𝑚

We can now determine the time of flight that occurs on the orbit segment. This is the time it
takes for one impulsive maneuver to complete before the next maneuver can start:

𝑟𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 3
Δ𝑡 = 2𝜃𝑏 √
𝜇𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑
 Page 27 of 42

Δ𝑡 = 3451.2 𝑠𝑒𝑐𝑠

We can also determine the velocity increments that occur at the ends of the orbit segment:

𝜇𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑
Δ𝑉 = 2√ √1 + 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 2 + 2𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 𝜃𝑏
𝑟𝑝𝑚 (1 + 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 )

Δ𝑉 = 0.1316 𝑚/𝑠

We can now solve for the constant (𝑞):

Δ𝑉
𝑞=
𝐼𝑠𝑝 𝑔0

𝑞 = 3.2010 ∗ 10−6

Which can used to determine the number of times the impulsive thrusts (𝑁) can be implemented
before we run out of fuel. For our case, we are assuming we only have the mass of Xenon,
because we used the hydrazine for the Hohmann transfer maneuvers:

𝑚𝑠𝑝𝑎𝑐𝑒𝑐𝑟𝑎𝑓𝑡 − 𝑚𝑓𝑢𝑒𝑙,𝑥𝑒𝑛𝑜𝑛 = 𝑚𝑠𝑝𝑎𝑐𝑒𝑐𝑟𝑎𝑓𝑡 𝑒 −𝑞𝑁

1 𝑚𝑠𝑝𝑎𝑐𝑒𝑐𝑟𝑎𝑓𝑡 − 𝑚𝑓𝑢𝑒𝑙,𝑥𝑒𝑛𝑜𝑛
𝑁 = − ln ( )
𝑞 𝑚𝑠𝑝𝑎𝑐𝑒𝑐𝑟𝑎𝑓𝑡

𝑁 = 32347
This helps us determine the duration of the mission, because we know the duration it takes to
complete one impulsive maneuver and how many maneuvers, we have the capability for:

𝑡𝑚𝑖𝑠𝑠𝑖𝑜𝑛 = 𝑁Δ𝑡

𝑡𝑚𝑖𝑠𝑠𝑖𝑜𝑛 = 1.1163 ∗ 108 𝑠𝑒𝑐𝑠 = 3.5395 𝑦𝑒𝑎𝑟𝑠

The last constant we solve for is (𝜆), which is defined as:

2𝐺𝑚𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 𝑠𝑖𝑛𝜃𝑏
𝜆=
ℎΔ𝑡

𝜆 = 1.0084 ∗ 10−4 𝑚/𝑠 2

Where (ℎ) is the angular momentum defined as the following:

ℎ = √𝜇𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 𝑟𝑝𝑒𝑟𝑖𝑎𝑝𝑠𝑖𝑠 (1 + 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 )

 Page 28 of 42

Now that all the constants are determined, we can revisit the deflection formula:
𝑁 Δ𝑉 𝑁 Δ𝑉
𝐾 −
𝐼𝑠𝑝 𝑔0
(𝑖−1) 𝐾 2
−
𝐼𝑠𝑝 𝑔0
(𝑖−1)
∆𝜁 = − (𝑡𝑒 − 𝑡𝑠 )𝜆𝑚𝑠𝑝𝑎𝑐𝑒𝑐𝑟𝑎𝑓𝑡 Δ𝑡 ∑ 𝑉𝑎𝑖 𝑒 + Δ𝑡 𝜆𝑚𝑠𝑝𝑎𝑐𝑒𝑐𝑟𝑎𝑓𝑡 ∑ 𝑉𝑎𝑖 𝑖𝑒
𝑚𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 𝑚𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑
𝑖=1 𝑖=1

We see that we must find 𝑉𝑎𝑖 , which is the velocity at each index or orbit segment. We compute
this for 𝑁 iterations. For a given index, we start by solving for Kepler’s equation for the eccentric
anomaly 𝐸𝑖 at the index using the known eccentric anomaly at the time of encounter 𝐸𝑒 . The
known eccentric anomaly is solved using the true anomaly 𝑓𝑒 at the time of encounter:

𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 𝑐𝑜𝑠𝑓𝑒
𝐸𝑒 = cos −1 ( )
𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑

𝐸𝑒 = 0.9244

Kepler’s equation for the eccentric anomaly at each index can be described as:

𝑎𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 3
𝑡𝑖 − 𝑡𝑒 = √ (𝐸𝑖 − 𝐸𝑒 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (𝑠𝑖𝑛𝐸𝑖 − 𝑠𝑖𝑛𝐸𝑒 ))
𝜇𝑠𝑢𝑛

If we assume that 𝑡𝑒 is the time of encounter, we can let it equal to zero, then 𝑡𝑖 = 𝑖. Note: This
greatly simplifies the problem. Ketema solved for this iteratively but given the time constraints
this was simplified for the ease of calculation. Ultimately, this allows us to move all the known
values to the left-hand side and leaves us:

𝑡𝑖 − 𝑡𝑒
+ 𝐸𝑖 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (𝑠𝑖𝑛𝐸𝑒 ) = 𝐸𝑖 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (𝑠𝑖𝑛𝐸𝑖 )
𝑎 3
√ 𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑
𝜇𝑠𝑢𝑛

We observe that if we introduce the constant 𝑀:

𝑡𝑖 − 𝑡𝑒
𝑀= + 𝐸𝑖 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (𝑠𝑖𝑛𝐸𝑒 )
𝑎 3
√ 𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑
𝜇𝑠𝑢𝑛

We get the Kepler equation in a familiar form:

𝑀 = 𝐸𝑖 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (𝑠𝑖𝑛𝐸𝑖 )

We can now solve Kepler’s equation for 𝐸𝑖 using Newton’s method:

𝑓(𝐸𝑖 )
𝐸𝑖+1 = 𝐸𝑖 −
𝑓′(𝐸𝑖 )
 Page 29 of 42

Where:
𝑓(𝐸𝑖 ) = 𝐸𝑖 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (𝑠𝑖𝑛𝐸𝑖 ) − 𝑀

𝑓(𝐸𝑖 ) = 1 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (𝑐𝑜𝑠𝐸𝑖 )

Now that we know what the value of 𝐸𝑖 for any index, we can find the corresponding radius 𝑟𝑖
using the equation of path:

𝑟𝑖 = 𝑎𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 (1 − 𝑒𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑 𝑐𝑜𝑠𝐸𝑖 )

With this we can determine the velocity at each orbit segment 𝑉𝑎𝑖 , using the energy equation:

𝜇𝑠𝑢𝑛 𝜇𝑠𝑢𝑛
𝑉𝑎𝑖 = √2 ( − )
𝑟𝑖 2𝑎𝑎𝑠𝑡𝑒𝑟𝑜𝑖𝑑

Finally, we set a range of values for (𝑡𝑒 − 𝑡𝑠 ), which is the starting time before the time of
project encounter. This can be expressed as an array of values, for our case we look from 1 year
to 20 years, to see the corresponding deflection. We can now finally solve numerically for the
deflection, once we run the MATLAB program in Appendix C, we get this plot for the
deflection:

Figure 16: Original Deflection Using Keplerian Tractor

 Page 30 of 42

We can see that around the 10-year mark we can only deflect 101955 Bennu around 1000 km,
which is well short of our target of 1 Earth radius (6378.1 km). Let’s try doubling the fuel
amount only from 120 kg to 240 kg:

Figure 17: Doubling the Xenon Fuel Mass, Spacecraft Mass is Constant

That only brought us closer to the required deflection to around 3000 km. What does doubling
the fuel and spacecraft mass do to the deflection? So now the spacecraft mass is 2440 kg and the
fuel mass is 120 kg.
 Page 31 of 42

Figure 18: Doubling Spacecraft Mass and Xenon Fuel Mass

As we can see, doubling both the spacecraft mass and fuel mass hinders the performance of the
gravitational tractor and places the new deflection somewhere in the middle. Keeping the mass
of the spacecraft constant and adjusting the fuel mass value it was determined at approximately 4
times original fuel mass, we have substantial deflection of the asteroid given a 10-year lead time.

Figure 19: Keeping Spacecraft Mass Constant and Quadrupling Xenon Fuel Mass
 Page 32 of 42

Section 5: Summary & Conclusions

From our analysis, we determined that our hypothetical spacecraft would not be able to deflect
10955 Bennu. It wasn’t until we quadrupled the Xenon fuel mass that the spacecraft was able to
achieve the appropriate deflection. This is no surprising because our hypothetical spacecraft was
sized based on NASA’s DART which has a different function as a low-cost kinetic impactor. This
spacecraft expended all its fuel in 3.5 years! For future studies, it will be good practice to size back
solve based on the deflection through means of an iterative analysis that tests different parameters.

Another detail to note is the difference in the plots generated between this analysis and Ketema.
This is due to immense simplification and assumptions made throughout the study. This resulted
in a linear plot versus Ketema’s oscillating plots. This potentially hindered the ability of our
hypothetical spacecraft because the oscillating motion of the Keplerian gravity helps it conserve
fuel and induce a greater gravity tractor force on the asteroid body. The simplification in this
analysis does not account for the oscillation and change in momentum.

Additionally, spending more time optimizing the code will be helpful. The current computing time
for the Newton Method and Riemann Sum took ~2.5 hours. This was primarily due to the N
iterations the calculation had to perform. Because of this troubleshooting was extremely difficult
and tedious. Towards the end I was able to reduce the computing time to a few minutes but more
work can definitely be done to optimize the code.

In future studies, we can look at comparing the Keplerian motion to other gravity tractors such as
displaced orbit and stationary orbit. There is also, the potential to add more space crafts in the
Keplerian motion, potentially increasing the force applied on the asteroid, reducing the lead time
and having smaller but cheaper spacecrafts. Finally, we can study the potential of having an ISS
type spacecraft but strictly for asteroid deflections. It would automatically sense asteroids that
are of great danger to us and dispatch the spacecrafts to deflect the celestial bodies.
 Page 33 of 42

References
[1] Chesley, S. R., Chodas, P. W., Milani, A., Valsecchi, G. B., & Yeomans, D. K. (2002).
Quantifying the risk posed by potential Earth impacts. Icarus.
[2] David, L. (2013, November 01). Russian Fireball Explosion Shows Meteor Risk Greater
Than Thought. Retrieved from Space.com: https://www.space.com/23423-russian-
fireball-meteor-airburst-risk.html
[3] Dearborn, D., Patenaude, S., & Managhan, R. (2007). The Use of Nuclear Explosives to
Disrupt or Divert Asteroids. Planetary Defense Conference.
[4] Farinella, P., Foschini, L., Froeschlé, C., Gonzci, R., Jopek, T. J., Longo, G., & Michel,
P. (2001). Probable asteroidal origin of the Tunguska Cosmic Body. Astronomy &
Astrophysics.
[5] Fischer, D. (2004, December 27). 2004 MN4 Earth Impact Risk Summary (computed 27
December 2004). Retrieved from The Cosmic Mirror:
https://web.archive.org/web/20050314032111/https://astro.uni-
bonn.de/~dfischer/mirror/285neo041227.html
[6] Gehler, M., Ober-Blobaum, S., Dachwald, B., & Marsden, J. (n.d.). Optimal Control of
Gravity Tractor Spacecraft Near Arbitrarily Shaped Asteroids. 1st IAA Planetary Defense
Conference.
[7] Jenniskens, P., Popova, O. P., Glazachev, D. O., Podobnaya, E. D., & Kartashova, A. P.
(2018). Tunguska eyewitness accounts, injuries, and casualties. Icarus.
[8] Ketema, Y. (2016). Asteroid Deflection Using a Spacecraft in Restricted Keplerian.
Minneapolis: University of Minnesota.
[9] Li, M., Wang, Y., Wang, Y., Zhou, B., & Zheng, W. (2020). Enhanced Kinetic Impactor
for Deflecting Large Potentially Hazardous Asteroids via Maneuvering Space Rocks.
Scientific Reports.
[10] Lu, E., & Love, S. (2005). Gravitational Tractor for Towing Asteroids. Nature.
[11] Morrison, D., Chapman, C., Steel, D., & Binzel, R. (2004). Impacts and the Public:
Communicating the Nature of the Impact Hazard. In Mitigation of Hazardous Comets
and Asteroids.
[12] NASA. (2021, June 04). Double Asteroid Redirection Test (DART) Mission. Retrieved
from nasa.gov: https://www.nasa.gov/planetarydefense/dart
[13] NASA Glenn Research Center. (2020, October 26). Gridded Ion Thrusters (NEXT-C).
Retrieved from grc.nasa.gov: https://www1.grc.nasa.gov/space/sep/gridded-ion-thrusters-
next-c/
[14] NASA Jet Propulsion Laboratory. (2014, November 14). New Map Shows Frequency of
Small Asteroid Impacts, Provides Clues on Larger Asteroid Population. Retrieved from
jpl.nasa.gov: https://www.jpl.nasa.gov/news/new-map-shows-frequency-of-small-
asteroid-impacts-provides-clues-on-larger-asteroid-population
[15] NASA JPL. (2014, November 14). New Map Shows Frequency of Small Asteroid
Impacts, Provides Clues on Larger Asteroid Population. Retrieved from jpl.nasa.gov:
https://www.jpl.nasa.gov/news/new-map-shows-frequency-of-small-asteroid-impacts-
provides-clues-on-larger-asteroid-population
[16] NASA JPL. (2021, March 25). NASA Analysis: Earth Is Safe From Asteroid Apophis for
100-Plus Years. Retrieved from jpl.nasa.gov: https://www.jpl.nasa.gov/news/nasa-
analysis-earth-is-safe-from-asteroid-apophis-for-100-plus-years
 Page 34 of 42

[17] NASA JPL CNEOS. (n.d.). Sentry: Earth Impact Monitoring. Retrieved from
cneos.jpl.nasa.gov: https://cneos.jpl.nasa.gov/sentry/
[18] Popova, O. P., Jenniskens, P., Emel'yanenko, V., Kartashova, A., & Biryukov, E. (2013).
Chelyabinsk Airburst, Damage Assessment,Meterite Recovery, and Characterization.
Science.
[19] Sanchez , P., Colombo, C., Vasile, M., & Radice, G. (2009). Multicriteria Comparison
Among Several Mitigation Strategies for Dangerous Near- Earth Objects. Journal of
Guidance, Control, and Dynamics.
[20] https://ocw.mit.edu/courses/aeronautics-and-astronautics/16-07-dynamics-fall-
2009/lecture-notes/MIT16_07F09_Lec17.pdf
 Page 35 of 42

Appendix
Appendix A – Ground to GTO
clc; close all; clear all;

%% Givens
G = 6.674*10^-11;% m3*kg-1*s-2 - universal gravitational constant

r_Earth = 6378.1; % km - radius of Earth

r_Earth_m = r_Earth*1000; % m - radius of Earth
mu_Earth = 3.986*10^14; % m3*s-2 - standard gravitational parameter of Earth

h_perigee = 185; % km - height of perigee

h_apogee = 35788; %km - height of apogee

%% Distances from Center of Earth

r_perigee = h_perigee + r_Earth; % km

r_apogee = h_apogee + r_Earth; % km

r_perigee_m = r_perigee*1000; % m

%% Semi Major Axis

a = (r_perigee+r_apogee)/2; %km
a_m = a*1000; % m

%% Velocity at Perigee

v_perigee = sqrt(mu_Earth*((2/r_perigee_m) - (1/a_m))); % m/s

%% Veloctiy Increments - delta_V

% Assuming Launch from 27 deg Inclination - Cape Carnaval has min 28.5 deg Inclination

v_initial = (2*pi*r_Earth_m*cosd(28.5))/(24*3600); % m/s

delta_V_ideal = v_perigee - v_initial; % m/s

% Assuming Required Veloctity Increment is 1.25x Ideal Velocity Increment

delta_V_required = 1.25*delta_V_ideal; % m/s

% Assuming 1st Stage provides 40% of Required Velocity Increment

delta_V1 = 0.4*delta_V_required; % m/s

 Page 36 of 42

% Assuming 2nd Stage provides Rest of Required Velocity Increment

delta_V2 = delta_V_required - delta_V1; % m/s

 Page 37 of 42

Appendix B – Interplanetary Hohmann Transfer

clear;
%% Sun Properties

mu_Sun = 1.327124400189*10^20; % m3*s-2 - standard gravitational parameter of Sun

%% Orbit Properties

r_EarthSun =1.495978707e11;% meter - Earth's orbit about Sun

r_asteroidSun = 1.12639 *r_EarthSun; %meter - asteroids orbit about Sun

%% Semi-Major Transfer Orbit

a_transfer=(r_EarthSun-r_asteroidSun)/2;

%% Velocities

V_orbitalEarth = sqrt(mu_Sun/r_EarthSun);
V_transEarth = sqrt(mu_Sun*((2/r_EarthSun)-(1/a_transfer)));

V_infinity = abs(V_transEarth-V_orbitalEarth);

%% SOI Earth Radius

mu_Earth = 3.986*10^14; % m3*s-2 - standard gravitational parameter of Earth

r_p = 6563.1*1000; %m
r_Earth = 6378.1; % km - radius of Earth
r_Earth_m = r_Earth*1000; % m - radius of Earth

r_EarthSOI = r_Earth_m*(mu_Sun/mu_Earth)^(2/5);
%% Hyperbolic Velocity
a_hyper = ((2/r_EarthSOI)-(V_infinity^2)/(mu_Earth))^-1;

V_peri_hyper=mu_Earth* sqrt((2/r_p)-(1/a_hyper));

%% Hyperbolic Departure

V_hyperdepart = abs(V_peri_hyper-sqrt(mu_Earth/r_p));

%% Hyperbolic Arrival
m_asteroid = 7.4* 10^10; % mass of asteroid - kg
G = 6.674*10^-11; % gravitational constant - m3kg-1s-2

mu_asteroid = m_asteroid * G; %m3s-2 - standard graviational parameter of asteroid

 Page 38 of 42

V_hyperarrival = abs(V_peri_hyper-sqrt(mu_asteroid/r_p));

%% Beta

B=acos(1/(1+((r_p*V_infinity^2)/mu_Earth)));
B_deg=rad2deg(B);
 Page 39 of 42

Appendix C – Kepler Tractor

clear; close; clc

%% General Givens

G = 6.674*10^-11; % gravitational constant - m3kg-1s-2

g0 = 9.81; % acceleration due to gravity - m*s-2

%% Asteroid Properties

m_asteroid = 7.4* 10^10; % mass of asteroid - kg

r_asteroid = 245; % radius of asteroid - m

a_asteroid = 1.126391; % asteroid semimajor axis - AU

a_asteroid_km = a_asteroid*1.496e8; % asteroid semimajor axis - km
a_asteroid_m = a_asteroid_km*1000; % m

e_asteroid = 0.5697; % asteroid eccentricity

longitude_ascendingnode = 253.64; % degrees

argument_periapsis = 73.58; % degrees

mu_asteroid = m_asteroid * G; %m3s-2 - standard graviational parameter of asteroid

%% Calculated Angles

longitude_perihelion = longitude_ascendingnode + argument_periapsis; % degrees

longitude_perihelion_rad = deg2rad(longitude_perihelion); % radians

theta = 2*pi-longitude_perihelion_rad; % radians

%% Earth Properties

r_Earth = 6378.1; %km

r_Earth_m = r_Earth*1000; % m
V_Earth = 30000; % orbital velocity of Earth - m/s - Mean Orbital Velocity is 30 km/s

%% True Anomaly

f_e = abs(acos((1/e_asteroid)* ((a_asteroid_km*(1-e_asteroid^2))/(r_Earth)) - 1)); % radians

%% Sun Properties

mu_Sun = 1.327124400189*10^20; % m3*s-2 - standard gravitational parameter of Sun

%% Velocity of Asteroid at Encounter

 Page 40 of 42

V_aencounter = sqrt(2*((mu_Sun/r_Earth_m) - (mu_Sun/(2*a_asteroid_m)))); % m/s

%% Solve for Constant K

K = ((3*a_asteroid_m)/(mu_Sun)) * V_aencounter * sin (0.829); % s/m

%% Flight Path Angle

gamma = abs(acos((sqrt(a_asteroid_m*(1-
e_asteroid^2)*mu_Sun))/(r_Earth_m*V_aencounter))); % radians

%% Magnitude of Relative Velocity of Asteroid wrt Earth

V_a_E = sqrt((V_Earth)^2 + (V_aencounter)^2 - (2*V_aencounter*V_Earth*cos(gamma))); %

m/s

%% Incidence Angle

psi = asin((V_Earth/V_a_E) * sin(gamma)); %%%%% Review - radians

%% Gravity Tractor Spacecraft - REPLACE with NASA DART Specs

m_spacecraft = 1220; % kg 1220

m_spacecraft_fuel = 120; % kg 120

Isp_spacecraft = 4190; % secs

%% Smallest Radius that Satisfies Plume Impingement

phi = 20; %degrees

phi_rad = deg2rad(phi); %radians

r_pm = ((1+e_asteroid*cos(1))/((1+e_asteroid)*cos(phi_rad-gamma)))*r_asteroid; % m

%% Time of Flight on Orbit Segment

delta_t = 2*1*sqrt(r_asteroid^3/mu_asteroid); % secs

%% Required Impulsive Velocity Change @ End Segments

delta_V = 2*sqrt(mu_asteroid/r_pm)*sqrt(1); % m/s

%% Constant q

q = delta_V/(Isp_spacecraft*g0);
 Page 41 of 42

%% Value of N - Amount of Impulsive Thrusts before all Fuel Expended

N = - (1/q) * log((m_spacecraft-m_spacecraft_fuel)/m_spacecraft);

%% Total Mission Duration

t_mission = N*delta_t; % seconds

t_mission_years = t_mission/3.154e7; % years

%% Constant Lambda

h = sqrt(mu_asteroid*argument_periapsis*(1+e_asteroid)); % angular momentum

lambda = (2*G*m_asteroid*sin(1))/(h*delta_t); %m/s2

%% Calculation of Riemann Sums

summe1 = 0.0; % starting value of Riemann 1

summe2 = 0.0; % starting value of Riemann 2
Ee = acos((e_asteroid + cos(f_e))/(1 + e_asteroid*cos(f_e))); % eccentric anomaly at time of
encounter [based on true anomaly (fe)] - radians
Kep_Ee = Ee - e_asteroid*sin(Ee); % Keplers Eqn for Eccentricity at Encounter - will be a
constant - radians
const = sqrt((a_asteroid^3)/(mu_Sun)); % constant to express square root constant in front
[sqrt(a^3/mu_sun)]

for i=1:N

t_i = i * delta_t; % time of interval in question - seconds

%% Newton Method for Eccentric Anomaly at Each Interval

M = (t_i/const)+Kep_Ee; % Definition of the Constant M

%% Newton's Method 2

x = M; % inital point and first guess of Newton

while eps>=1e-5 && n<=nmax

y = x - (x - e_asteroid*sin(x)- M)/(1 - e_asteroid*cos(x));
eps = abs(y-x);
x = y;n = n+1;
end

%% Solving for Velocity at Each Interval

 Page 42 of 42

Ei = x;

r_i = a_asteroid*(1-e_asteroid*cos(Ei)); % corresponding radius for each index value (i)

V_ai = 2*(sqrt((mu_Sun/r_i)-(mu_Sun/(2*a_asteroid)))); % Velocity of Asteroid for each

index value (i)

%% Summation of 1st Riemann Sum

summe1 = summe1 + V_ai*e_asteroid^-((delta_V/(Isp_spacecraft*g0))*(i-1)); % Summation

of Riemann Sum 1

%% Summation of 2nd Riemann Sum

summe2 = summe2 + V_ai*i*e_asteroid^-((delta_V/(Isp_spacecraft*g0))*(i-1)); %

Summation of Riemann Sum 2

end

%% Deflection Equation

time_interval_years = [-1:-1:-10]; % t_e-t_s - years

time_interval = time_interval_years*3.154*10^7; % seconds

deflection = -((K/m_asteroid)*(time_interval) * lambda * m_spacecraft * delta_t *

summe1)+(K/m_asteroid) * (delta_t)^2 * lambda * m_spacecraft * summe2;

%% Plotting
time_years = time_interval/3.154*10^7;
plot(time_interval/10^7, -deflection/10^7/1000);
xlabel('Time (years)')
ylabel('Deflection (km)')
title('Deflection - Keplerian Gravity Tractor')