Content uploaded by Ja'Mar A. Watson
Author content
All content in this area was uploaded by Ja'Mar A. Watson on Feb 26, 2019
Content may be subject to copyright.
2 Vol 71 No.1 January 2018 JBIS
CREWED MARS MISSION
Concept Development and Experimentation
JA’MAR A. WATSON Watson Institute for Scientific Engineering Research, 4201 Wilson Blvd Ste 110-444, Arlington, VA 22203, USA
email jamar.watson@sciengresearch.org
This paper presents the concept development and experimentation of human spaceflight missions to Mars. Utilizing
Acolytion’s Progspexion, analysis of alternatives is performed inclusive of 18 prominent mission concepts with a vast
experimentally designed tradespace exploration of crewed Mars mission architecture constituents. Results indicate that,
as of date, no proposed mission concept is capable of achieving acceptable mission utility to risk ratios; however, the
research produces 18 statistically guided developments to the concept of crewed Mars missions.
Keywords: Mission Utility Analysis, Space Mission Simulation, Human Spaceflight, Mars Exploration, Progspexion
1 INTRODUCTION
No prospective space mission has been stagnated greater than
the horizon goal of human spaceight (HSF): the Crewed Mars
Mission (CMM). Such a mission today is viewed as having in-
sucient scope to initiate exploration because the technolog-
ical state, understanding of risks, and experience with CMM
concept of operations (CONOPS) is too immature [1]. CMM
schedules have regressed to integrate the mission within the
Evolvable Mars Campaign [2] that aligns with the Global Ex-
ploration Roadmap [3] – eectively classifying the CMM as
currently unpursuable [4]. On the contrary, it can also be ar-
gued that this scheduling cannot achieve optimality with scope
and cost because the budget for a long-term pathways approach
is not sustainable without sizeable increases throughout the
program[5], the proposed exploration roadmaps would not
complete preparation for a CMM [1], it may be presumptuous
to believe that systems can be used for other missions without
much change [6], and many technologies will need to be de-
veloped for the campaign that do not contribute to the hori-
zon goal [5]. erefore, it is uncertain if the proposed CMM
designs will ever lead to a utilizable mission [7]. Consequently,
the CMM is nothing more than a concept described by a crude
problem statement and is devoid of any clear solution paths or
viable achievement schemes. Although many mission architec-
tures have been proposed to rectify the procrastination of a HSF
mission to Mars, none have achieved unanimous reception.
e objective of this research is to perform concept devel-
opment and experimentation (CD&E) of the CMM to inves-
tigate the viability of its immediate and direct pursuit amidst
prominent proposed mission concepts and the current state-
of-the-art in HSF. To do so, an analysis of alternatives (AoA)
is performed incorporating the Human Exploration of Mars
Design Reference Architecture (DRA) [8], [9], [10], and its
Austere edition [11], Concept 2-4-2 [12], [13] and its revised
version [14], [15], ESA’s Human Mars Mission (HMM) [16],
Mars Piloted Orbital Station (MARPOST) and its hybrid de-
sign [17], Mars Direct [18], [19] and its Semi-Direct and Hy-
brid-Direct architectures [20], [21], MarsOne [22], [23], and
the Human Exploration using Real-time Robotic Operations
(HERRO) architecture [24] [25]. In addition, constituents of
the Mars Project [26], Interplanetary Transport System (ITS)
[27], Cycling Pathways to Occupy Mars (CPOM) [28] [29], Red
Dragon [30], Inspiration Mars [31] [32], and Mars Base Camp
(MBC) [33] architectures are utilized to construct a tradespace
for concept experimentation. Mission utility analysis of these
proposed mission concepts produces 18 statistically guided de-
velopments to the CMM.
2. METHODS
2.1 Mission Utility Analysis
CD&E is enabled by the Holistic Methodology for Stochastic
Mission Utility Analysis, which has been described at length in
a previous publication [34]. e publication outlines the meth-
odology in which mission utility simulation quanties design
analysis for decision-makers [35] and introduces the primary
measure of eectiveness (mission utility) and gure of merit
(risk) used extensively in this CMM CD&E. Mission utility is
the ability to satisfy mission objectives during the mission and
risk is the probability of loss of mission (LoM) or loss of crew
(LoC) during the mission. ese metrics are quantied during
mission simulation in Progspexion.
e CMM objectives (Table 1, above right) have already
been established in a previous publication [36] that established
a baseline performance of the DRA in Progspexion. e publi-
cation also outlines the mission objectives shown here, which
are derived from Mars exploration community consensus, as
well as the scientic investigations conducted by the crew that
accomplish the mission objectives. Mission utility is therefore
quantied by the number of these mission objectives satised
during mission simulation, whose individual utility contrib-
utes are outlined in Table 1.
2.2. Progspexion
Acolytion’s Progspexion is utilized for mission utility simulation
in this CD&E. It is a soware suite of stochastic Monte Carlo
mission simulators that concurrently assess, explore, and devel-
JBIS VOLUME 71 2018 PAGES 2-18
JBIS Vol 71 No.1 January 2018 3
TABLE 1 Mission Objectives and Sub-Objectives
Objective Description Value
1Life on Mars 50%
1.1
1.2
Determine Prior Habitability Search for Extinct Life
Determine Current Habitability Search for Extant Life
25%
25%
2Martian Planetology 25%
2.1
2.2
2.3
2.4
2.5
Determine Current Climate Detail Atmosphere and Environment
Determine Geologic Record Detail Geologic Processes
Determine Climate History Detail Climate Changes
Characterize Planet Interior Determine Interior Evolution
Determine Planet Evolution Determine Planetary System Evolution
10%
5%
5%
3%
2%
3Colonization Preparation 20%
3.1
3.2
3.3
3.4
Obtain Knowledge for Orbital Missions
Obtain Knowledge for Surface Missions
Obtain Knowledge for Missions to Moons
Obtain Knowledge for Sustained Presence
7%
7%
3%
3%
4Ancillary Science 5%
4.1
4.2
4.3
Study Heliophysics from Mars
Study Astrophysics from Mars
Conduct Public Engagement Activities
2%
2%
1%
op statistically guided concepts of space exploration missions
via holistic mission utility analysis. e underlying stochastic
modeling is achieved by the synergistic integration of statisti-
cal mechanics, black-box, ansatz, and analytics techniques and
comprehensively accounts for uncertainty and variability at the
technology, payload, systems, architectural, operational, mis-
sion, and enterprise levels. e Monte Carlo simulation enables
multiple runs of the stochastic models, as well as alteration of
any desired mission concept variable amongst a vast tradespace
for concept experimentation. Descriptive, Predictive, and Pre-
Fig. 1 Progspexion’s CMM Tradespace.
scriptive (DP2) analytics [34] of Progspexion’s simulation da-
taset facilities concept development for the analyzed concepts.
Its mission utility analysis methodology has been veried and
validated through backtesting of Apollo 11-17 missions to the
Moon in a previous publication [37].
Particularly, Progspexion’s Mars mission simulator, Areo,
inherently provides a tradespace inclusive of the aforemen-
tioned derived CMM architecture constituents (Fig. 1) and
simulates the mission as shown in (Fig. 2). It holistically por-
CREWED MARS MISSION Concept Development and Experimentation
4 Vol 71 No.1 January 2018 JBIS
trays the entire mission by encapsulating crew ergonomics,
human factors, mission intelligence, mission phenomenolo-
gy, operational weather (OpWX), and stochastic uncertainty
in a rapid activity-based simulation environment integrated
with 63 possibilistic mission events (Table 2 opposite) and a
hierarchy of mission contingencies and aborts [38]. erefore,
simulating the CMM concepts of this research in Areo with
its inherent tradespace facilitates CMM concept experimenta-
tion. Areo’s analytics of its simulation dataset provides the uti-
lized metrics of mission utility and risk, as well as statistically
guided concept development of the CMM concept designs in
an operational context.
3. GROUND RULES, ASSUMPTIONS, AND DELIMITATIONS
Inherently, underlying ground rules, assumptions, and delim-
itations drive the quality of data produced in simulations [39].
Since the CMM CD&E method is underlined by simulation of
the CMM mission in Progspexion, they are stated below before
presenting the CD&E results.
3.1. Major Ground Rules
The Buddy System: e buddy system, always assuming crew-
members travel minimally as pairs, is critical to success and
safety in performing extravehicular activity (EVAs) [8]. ere-
fore, grouping of crewmembers shall always assume the incor-
poration of, at a minimum, two crewmembers.
Habitation: e volume (and subsequently mass) of habitation
is of utmost importance to ergonomics in the mission. NASA
research has settled upon the value of 25 m3 per crewmember
as minimum habitable volume for long-duration exploration
missions [40]. However, research also shows that it is possible
to make habitats smaller with quantication of their eects on
the crew. For this reason, 11 m3 per crewmember is derived as
a performance limit and 5.5 m3 per crewmember as a tolerable
limit [41].
Mass Budget: e mass budget derived from each cocnept shall
Fig. 2 Progespexion CMM Simulation Flow Diagram (ow of
contingencies and aborts not shown).
always reach its maximum; meaning, a reduction in subsystem
masses is accompanied by the addition of crew or tradespace
mass (or vice versa). is facilitates rapid calculation of size,
weight, and power (SW&P) budgets in tradespace exploration
while upholding the principles of the analyzed mission concept.
3.2. Major Assumptions
Commuter Operations: It is assumed the CONOPS follows the
"commuter” exploration strategy, in which there is a central
habitat used by the crew between sorties with mobility systems
proving access to sites of scientic interest [8]. Since real-time
operations are not possible from Earth-based mission control
centers, one group of the crew is always assumed to remain in
the habitat to serve this function and for emergency situations
[8] while the other groups follow the buddy system in perform-
ing traverses.
Consumables: In accordance with the DRA, the research as-
sumes a fully closed-loop environmental control and life sup-
port system (ECLSS). is brings the consumables require-
ment to 2.5 kg per crewmember per day [8]. In addition, a 20%
contingency of the total required consumables for the crew size
is included.
Launch Segment: All architectures assume a launch mass to
low Earth orbit (LEO) capability of 130 t or 40 t direct injection
to Mars. is is consistent with the launch vehicles of the afore-
mentioned proposed architectures.
Lossless Crew: It is assumed a successful mission is one in
which no crewmembers are lost; therefore, a mission involving
a LoC outcome is considered a failure.
Number of Missions: As an objective of the research, only one
single mission cycle is analyzed. Furthermore, it is assumed the
mission is immediately pursued, meaning the values for pos-
sibilistic events and stochastic variables assume current read-
iness levels.
Parking Orbits: All non-direct vehicles depart from an opti-
mized 407 km circular geocentric orbit by a two-burn escape
maneuver to reduce gravitational losses. At Mars, vehicles are
parked in a 250 km by 33,793 km, 1-sol, areocentric orbit to
optimize the trade between Mars orbit insertion (MOI), ascent
autonomous rendezvous and docking (AR&D), and landing
site geometry [8].
Special Regions: It is assumed the Committee on Space Re-
search’s (COSPAR) protection of special regions [42], [43] is
followed for placement of the habitat but not for scientic ex-
ploration. is means the crew will participate in any and all
scientic investigations. is also assumes that sensitive and
planetary protection sorties, such as the search for extraterres-
trial life (SETL), will require long traverses.
3.3. Major Delimitations
Degradation: ESA research shows that large assembly times
in LEO could lead to premature degradation of modules [16].
However, this research will not assume such degradation be-
cause this application relies on assumed launch-rate and cam-
paign capabilities. is, along with exclusion of potential deg-
radation for pre-deployed assets, exonerates this eect from
loss in reliability. It is deemed inherent in the system design to
account for such increased operational periods.
Low Energy Transfers: is research does not include low en-
ergy interplanetary transfer considerations which removes
electric propulsion systems and ballistic capture trajectories.
While electric propulsion could be included in the tradespace,
none of the proposed concepts considered in this research use
it and low energy transfer is prohibited in crewed vehicles due
JA’MAR A. WATSON
JBIS Vol 71 No.1 January 2018 5
TABLE 2 Mission Events Included in CMM Simulation
Phase (Event)
Earth Launch/Integration
Cargo Launch | Crew Launch | AR&D | LEO Repair Mission Required
Tra ns - Ma rs I nj ec ti on ( TM I)
Miss Insertion Window | Incapable of TMI | LoC | Unable to Abort
Mars Transit
LoV | Contingency EVA Required | Adequate Crew Skill Development | Onboard Problem Resolution | Unexpected Crew Deconditioning
Mars Orbit Insertion (MOI)
Mission Abort | Aerocapture LoC | Orbit Insertion Error | AR&D Failure | Extended Mars Vicinity Phase
Mars Entry, Descent, and Landing (EDL)
Crew Failure | Cargo Failure | Abort Required | Strenuous Activities Required | Crew Injury
Mars Surface Operations
Loss of Loitering Orbital Vehicle | Ascent Repair Required | ISPP Failure | Mission Abort |
Mission Constraints and Schedule Met | Habitat Power Failure | Loss of Habitat | Meet Go/No-Go Criteria for EVA
Mars Ascent/Integration
Ascent Failure | Ascent Delayed | Ascent Abort | Ascent Orbit Failure | AR&D Failure | Vehicle Transfer Failure
Trans-Earth Inject (TEI)
Delay | Incapable of TEI | LoC
Earth Transit
LoV | Contingency EVA Required | Onboard Problem Resolution | Address Planetary Protection Issues
Earth Entry, Descent, and Landing (EDL)
LoC During EOI | LoC During EDL | Loss of Payload
Common Cause Failures (CCF)
β CCF Factor | γ CCF Factor
Global
SPE |GCR | REID | Orbital Radiation Shielding Successful | Human Error
Emergent
Dust Storm | Habitat out of Range Due to Inaccurate EDL | Ascent Vehicle out of Range Due to Inaccurate EDL
| Missed Return Insertion Window | Abort Required Before ISPP Completed | Surface Abort Required During Colonization
KEY AR&D= Autonomous Rendezvous and Docking | EVA=Extravehicular Activity | GCR=Galactic Cosmic Radiation | ISPP= In-situ Propellant Production
| LoC=Loss of Crew | LoV=Loss of Vehicle | REID=Radiation Exposure Induced Death | SPE= Solar Particle Event
to extended periods of microgravity and space weather expo-
sure during interplanetary travel. While its utilization can be
argued for cargo vehicles, the lengthy transfer periods are not
consistent with mission design principles of the concepts in
this research – particularly when converting to single mission
consideration with potentially urgent resupplies.
Opposition Class Missions: Modern CMM designs are no
longer in favor of the opposition class mission because of its
considerable negative aspects. erefore, all missions are simu-
lated as conjunction class missions.
Rescue Missions: In construction of contingency options of
the mission phasespace, the possibility of a rescue mission is
removed. In this case, if the crew required rescuing, it is instead
assumed the mission resulted in LoC.
Surface Power Trade: A trade on nuclear and solar surface
power is excluded because analysis shows the choice to be
preference of mission deployment (except for landing sites in
extreme environments such as polar regions, underground,
craters, etc). Architectures that pre-deploy cargo prefer nuclear
power sources because they are easier to autonomously deploy,
provide consistent power for in-situ propellant production
(ISPP), and do not suer dust sensitivities to power produc-
tion the way solar arrays are anticipated [8]. On the other hand,
architectures utilizing the all-up approach will have humans
present when the surface power source is set up, meaning hu-
man operation can mitigate the need for complex autonomous
deployment of solar arrays. Crewmembers are also available to
provide cleaning of dust from solar array surfaces [12]. It was
found that, although solar array systems could require more
weight (dependent on potential emergency and nighttime
power modes), they are easier to repair and thus don’t require
as much redundant system mass as their nuclear counterparts.
Additionally, radiation hazard of the nuclear surface sources
requires operational restrictions and additional shielding mass
for the crew. All of the aforementioned considerations eec-
tively make the trade preferential for this single mission design.
CREWED MARS MISSION Concept Development and Experimentation
6 Vol 71 No.1 January 2018 JBIS
4. RESULTS
4.1. Defaults
CD&E is initiated by Default Simulations: mission utility sim-
ulation of the concepts as proposed (Fig. 3). As seen in the
box plot, the median value of utility, as well as the interquar-
tile range (IQR), of all mission concepts is zero, indicating
an overwhelmingly high number of mission failures. e red
outlier marks indicate both the distribution of remaining out-
lier outcomes as well as potential maximum achievement of
mission utility.
Fig. 4 displays the median and IQR only when the crew
launches. ese post pre-deployment (PPD) results show that
even if pre-deployment success could be guaranteed, the me-
dian result of utility is still zero, albeit with a slightly great-
er dispersion of outcomes. Overall, the PPD results remove
the eect of pre-deployment on mission concepts of the split/
pre-deployment mission class and show the potential of cargo
deployment improvement.
Fig. 5 displays the same results as shown in Figure 4 but rel-
ative to the DRA baseline. is relative measure is quantied
by mean of the mission utility and its standard deviation. By
concept design, concepts that do not pre-deploy have nearly
equivalent mean mission utility for all missions and missions
when the crew launches, since the crew launches in every in-
itial launch campaign. e results reinforce the low mission
utility of all CMM concepts as proposed.
Lastly for mission utility, the median value of simulations in
which the crew successfully completes the mission, (returning
to Earth without LoC or LoM) is presented in Fig. 6. Here the
bars represent the median absolute deviation (MAD). e g-
ure represents a measure of relative potential maximum utility
achievement, which are signicantly higher than the mean of
all default simulation outcomes. e results therefore indicate
the potential of CD&E in this research.
In addition to mission utility, the metric of risk, which pre-
sents the percent probability of LoM and LoC in the mission,
is shown in this paper (Fig. 7). It is worth noting that, although
there is also a LoM in LoC outcomes, the LoM percentages
are tabulated for simulations where the mission is loss but the
crew is not. is helps distinguish mission outcome factors.
erefore, absolute LoM outcomes is the cumulative percent-
age from the sum of LoM and LoC values. ese risk values
are also presented for PPD, removing the skew of signicant
predeployment failure, LoM outcomes. Even with the removal
Fig. 3 Mission Utility of Default Concepts.
Fig. 4 Mission Utility of Default Concepts When Crew Launches.
Fig. 5 Relative Mission Utility of Default Concepts.
Fig. 6 Mission Utility of Default Concepts During Successful
Missions.
Fig. 7 Risk of Default Concepts.
JA’MAR A. WATSON
JBIS Vol 71 No.1 January 2018 7
of pre-deployment, Fig. 8 shows that almost all simulations
across all mission concepts result in either LoM or LoC, with
a substantial shi to LoC outcomes in PPD since the crew is
involved in every simulation. Since a successful mission in
which the crew returns to Earth is an outliner, this explains
the zero mission utility values. Overall, results of investigating
the outcome of default mission concepts highlights the neces-
sity of CD&E for increasing the mission utility to risk ratio for
CMM concepts.
4.2. Best-Case Scenarios
Here the results of the best-case scenario are shown (Fig. 9).
Best-case scenario simulates the execution of an operationally
perfect mission. Architecturally, the concept is still limited by
SW&P budgets. Operationally, only factors from the constitu-
tive variables of OpWX (SPE, GCR, REID, and dust storms)
remain.
Additionally, the stochastic landing error ellipse is re-
duced to the selection of either the precise or future 1km x
10km EDL accuracy. e latter is included due to unavoida-
ble uncertainty of the topographically forced winds of Mars’
atmospheric environment, which can produce EDL errors
in excess of one kilometer even for perfectly executed EDL
phases [8].
Although the best-case scenario is an unrealistic simula-
tion, it does establish a maximum achievable mission utility
for the default concepts. In doing so, the simulations display a
segregation of results (Figure 9). is can be seen more clearly
in terms of the median and MAD (Fig. 10). Despite this being
a best-case scenario, some concepts still display relatively low
values of mission utility. It is determined these results stem
from inadequate consideration of ergonomics in the mission
architecture, resulting in overwhelming physiological and psy-
chological impacts to the crew. As such, despite operationally
traveling to Mars and returning to Earth, the crew is unable to
successfully complete the scientic investigations necessary to
obtain mission utility.
e metric of risk is also shown for these best-case scenarios
(Fig. 11). With the performance of an operationally perfect mis-
sion comes the reduction of overall percent risk by an order of
magnitude. e remaining LoC outcomes are due to OpWX en-
counters and radiation exposure induced death (REID), while
the LoM outcomes arise from losses of cargo systems/vehicles
due to the space environment. ese therefore also represent a
statistical minimum of percent risk for the default mission con-
cepts. e results of best-case scenario simulations present one
major impact to CMM CD&E: if the architecture is SW&P lim-
ited to the point it does not suciently accommodate the crew
the mission utility will remain low, even if operationally perfect,
due to human factor and ergonomic impacts that deplete suc-
cessful scientic investigations.
4.3. Crew Accommodation
Following best-case scenario simulations, the tradespace is
used to accomplish crew accommodations. It is the rst phase
in which simulated mission concepts veer from the default
denitions. is initial tradespace exploration is completed
using only architecture design trades that are directly able to
optimize crew ergonomics within the SW&P budget. e crew
size and surface systems are held constant during this phase of
concept experimentation.
Fig. 8 Risk of Default Concepts When Crew Launches.
Fig. 9 Mission Utility of Default Concepts’ Best-Case Scenario
Fig. 10 Median Mission Utility of Default Concepts' Best-Case
Scenario.
Fig. 11 Risk of Default Concepts' Best-Case Scenario.
CREWED MARS MISSION Concept Development and Experimentation
8 Vol 71 No.1 January 2018 JBIS
In Fig. 12, it can be seen that crew optimization in the
tradespace does indeed improve the mission concepts’ mission
utility in the best-case scenario. Fig. 13, however, shows this
improvement does not transfer to the mission utility of the
concepts in a realistic mission simulation. Fig. 14 does show
an increase to the IQR for pre-deployment concepts that reach
PPD and Fig. 15 represents a similar increase to the relative
mission utility results. erefore, there is indeed a slight im-
provement over the default results.
Greater potential of crew accommodation optimization is
realized in the results of simulations in which the mission is
successful (Fig. 16); however, the results of the risk metric are
relatively unchanged. Results indicate that even with improved
ergonomics from crew accommodation optimization, if the
crew cannot suciently complete the allotted investigations
in the available duration for surface exploration, then mission
utility will still experience reductions due to mission concept
denition. Although held constant in this phase of concept
experimentation, the architectural trade on surface systems,
which allow for mobility and scientic investigations on Mars’
surface or teleoperation from areocentric orbit, have the poten-
tial to rectify this shortcoming.
4.4. Architectural Optimization
To address the aforementioned lapse in surface exploration
measures of performance, the previously constant tradespace
variables of crew size and surface systems are included in the
concept experimentation to achieve overall architectural opti-
mization within the allotted SW&P budgets. Fig. 17 summa-
rizes simulation of architecturally optimized concepts, which
still result in a median mission utility of zero for all mission
concepts. e PPD mission utility is relatively unchanged;
however, architectural optimization does increase dispersion
of mission outcomes in the relative results (Fig. 18). e eects
of this optimization, however, are displayed in the results of
successful missions (Fig. 19). While many of the architectural
optimized concepts are still unchanged, Mars Direct shows sig-
nicant improvement in this optimization process. Again, the
metric of risk is relatively unchanged.
It now becomes clear that no architectural reconguration
can produce signicant increases in mission utility or decreases
in risk when considering all mission simulations. erefore, a
complete redenition of CMM concepts must be created from
the available tradespace to progress mission utility towards
best-case scenario results.
4.5. Apollo Benchmark
In addition to the best-case scenario, the architecturally op-
timized concepts are simulated against the benchmark of the
Apollo missions to further assess the impact of the CONOPS.
As the only HSF mission beyond LEO and one in which the
utility-to-risk ratio was suciently high to pursue 7 missions,
the Apollo benchmark provides insight into the mission con-
cepts to determine what mission utility could be obtained if the
Mars exploration portion of the CMM could be performed as
easily as the Apollo 11-17 missions to the Moon.
Merging the possibilistic event values of Progspexion’s lu-
nar mission utility simulator, Seleno [37] [38], with that of
the current CMM possibilistic event values in Aero enables
these simulations (Fig. 20). Although the median result for all
concepts is still zero, the decrease in diculty signicantly in-
Fig. 12 Mission Utility of Crew Optimized Best-Case Scenario.
Fig. 13 Mission Utility of Crew Optimized Concepts.
Fig. 14 Mission Utility of Crew-Optimized Concepts When the
Crew Launches.
Fig. 15 Relative Mission Utility of Crew-Optimized Concepts.
JA’MAR A. WATSON
JBIS Vol 71 No.1 January 2018 9
creased the IQR for concepts of limited mission phases (all-up
mission class or colonization end of mission (EOM) design) or
with transhipment during interplanetary transit (Concept 2-4-
2 and its revised version). e increase was large enough for
the Revised 2-4-2 concept that its dataset no longer contains
outliers for high values of mission utility.
While there is improvement in mission utility for some, the
concepts do not achieve comparable utility to that of the Apollo
missions. ey do, however, noticeably improve over the base-
line results (Fig. 21). e greater change in results is shown in
the next two gures (Fig. 22 and Fig. 23 overleaf), which dis-
plays the risk and PPD risk respectively. While the reductions
in risk across all concepts is modest, there is a signicant shi
from LoC outcomes to LoM. is is especially true of all-up
concepts.
ese results, however, again demonstrate the substantial
challenge with the pre-deployment phase [36] as only the PPD
results show pronounced reductions in risk. It should be noted
that the obtained mission utility in the Apollo benchmark anal-
ysis is not fully relevant to each concept because if the CMM
could truly be performed as easily as a lunar mission, the mis-
sion concept approach would also drastically change. However,
the Apollo benchmark results do show a decrease in necessity
of the pre-deployment phase as the mission’s probability of suc-
cessful execution increases.
4.6. Experimentation
With the above results providing rich insights into the eects
of limited tradespace components, a full tradespace explora-
Fig. 16 Mission Utility of Crew-Optimized Concepts During
Successful Missions.
Fig. 17 Mission Utility of Architecturally Optimized Concepts.
Fig. 18 Relative Mission Utility of Architecturally Optimized
Concepts.
Fig. 19 Mission Utility of Architecturally Optimized Concepts
During Successful Missions.
Fig. 20 Mission Utility of Architecturally Optimized Concepts at
Apollo Diculty.
Fig. 21 Relative Mission Utility of Architecturally Optimized
Concepts at Apollo Diculty.
CREWED MARS MISSION Concept Development and Experimentation
10 Vol 71 No.1 January 2018 JBIS
TABLE 3 Unique Concept Alterations for Optimization
Concept Mission
DRA
Chemical Propulsion
Propelled MOI for Cargo
Siphon Launch
Austere
Medium Mobility
Propelled MOI for Cargo
Siphon Launch
Concept 2-4-2 Inflatable Habitat
Siphon Launch
ESA
EDL Exploration
Medium Mobility
Siphon Launch
HERRO
EDL Exploration
Inflatable Habitat
Propelled MOI for Cargo
Siphon Launch
H/S Direct Aerial Mobility
Propelled MOI for Cargo
MARPOST
Aerial Mobility
Chemical Propulsion
EDL Exploration
Siphon Launch
8 Crew
Mars Direct To o M as s Li m it e d
Mars One
Aerial Mobility
Propelled MOI
Siphon Launch
Revised 2-4-2 Propelled MOI
Siphon Launch
tion is completed for a statistically guided experimental search
of the Pareto frontier. e exception is the trade on mission
class, which was found to violate the mass budget ground rule
and would deviate from the principals of each respective CMM
concept.
e results in obtaining these optimized solutions from
tradespace exploration are plotted (Fig. 24) and show a very
distinct multimodal result. Additionally, Hybrid-Direct and
Semi-Direct reduce to identical solutions and are expressed
as H/S Direct. is is the same for the MARPOST and MAR-
POST Hybrid mission concepts, referred to simply as MAR-
POST. e alterations required for each concept are listed in
Table 3 (right). In addition, every concept selected the mid
solar cycle mission epoch, initial mission intelligence of cer-
tain, all pre-cursor missions, luggage consumables, duplication
redundancy, no solar I&W, and colonize EOM design. While
it is possible many of the concepts made the above alterations
due to SW&P budget constraints, it represents a design opti-
mization from the tradespace while still maintaining the core
competencies of the concepts.
ough not achieving high values of mission utility, it is the
rst optimization in which the median mission utility and IQR
for the concepts displayed nonzero values when considering all
mission simulations. e process of experimentation was also
able to optimize the concepts suciently to exceed the Apollo
benchmark of the architecturally optimized concepts, as well
as the baseline performance of the default DRA (Fig. 25). e
PPD results in Fig. 26 shows the impact of pre-deployment on
the nal ranking of the optimized concepts. If concepts are still
mission utility limited in this stage of analysis, it is due to the
crewed portion of the mission.
When only showing results of the experimentally optimized
concepts when the crew successfully completes the mission
(Fig. 27), all concepts achieve high values of mission utility
with extremely low distribution. Additionally, the risk of these
experimentally optimized concepts (Fig. 28) is signicantly re-
duced overall; however, the initial LoM risk is greater, meaning
the risk of LoC is substantially decreased. e same can be seen
in the PPD results (Fig.29). Here LoC outcomes are slightly el-
evated since the crew is involved in all mission concept sim-
ulations. Overall the experimentation process utilizing a full
tradespace exploration statistically improved all concepts over
the baseline and default simulation results.
5. ANALYTICS
5.1. Sensitivity
To better understand the statistical signicance and inuence
of Progspexion on analytics, a sensitivity analysis is rst con-
ducted. is includes sensitivity of prominent input parame-
ters, as well as sensitivity of the possibilistic dataset and other
stochastically determined variables within Progspexion. To
perform sensitivity analysis of the possibilistic dataset, each
component of the 63 possibilistic mission event dataset is in-
dividually altered by ± 5%. e baseline DRA mission is then
simulated to detect changes in mission utility. Doing so iden-
tied the following mission events as producing greater than
1% change in mean mission utility: crew launch, incapable of
TMI, MT contingency EVA, MOI mission abort, Mars crewed
EDL failure, surface operations mission abort, areocentric
AR&D failure, and areocentric vehicle transfer failure. ere-
fore, these components should be investigated further in fu-
ture research to determine a potential increase in modeling
delity.
5.2. Descriptive
e descriptive portion of analytics provides statistical descrip-
tion of the mission through the aggregated simulation dataset
(Table 4 overleaf). e table shows that perfect mission exe-
cution is rare, as a mission dynamic (MD: the occurrence of
a mission phase/event that alters the mission’s phasespace tra-
jectory) occurs in nearly all missions for all concepts. e sta-
tistical mode calculated for the mission phase/event in which
the mission simulation ends determines that the mission ends
on Earth in most simulations, despite the fact that the concept
experimentation optimization selected colonization on Mars
as the EOM design. is outcome, and the high percentage of
missions in which an abort occurs, indicates a dire need to ad-
dress contingencies in the mission. In fact, all concepts’ abort
phase/event and end phase/event that occur the most in simu-
JA’MAR A. WATSON
JBIS Vol 71 No.1 January 2018 11
Fig. 22 Risk of Architecturally Optimized Mission Concepts at
Apollo Diculty.
Fig. 23 Risk of Architecturally Optimized Mission Concepts at
Apollo Diculty When the Crew Launches.
Fig. 24 Mission Utility of Experimentally Optimized Concepts.
Fig. 25 Relative Mission Utility of Experimentally Optimized
Concepts.
Fig. 26 Interquartile Range of Experimentally Optimized Concepts
When the Crew Launches.
Fig. 27 Mission Utility of Experimentally Optimized Concepts
During Successful Missions.
Fig. 28 Risk of Experimentally Optimized Concepts.
Fig. 29 Risk of Optimized Concepts When the Crew Launches.
CREWED MARS MISSION Concept Development and Experimentation
12 Vol 71 No.1 January 2018 JBIS
TABLE 4 Aggregates of Optimized Concepts
DRA Austere Concept 242 ESA HERRO H/S Direct MARPOST Mars Direct MarsOne Revised 242
Mission Dynamic 98.53% 98.11% 99.46% 96.61% 94.98% 98.03% 96.33% 96.56% 94.35% 95.84%
End Phase Mode Earth Earth Earth Earth Earth PD MOI Earh PD MOI Earth Earth
End Event Mode MT EVA MT EVA Aero MOI Surface Abort MT EVA PD Aero MOI MT EVA PD Aero MOI MT EVA MT EVA
Abort 62.35% 59.94% 59.64% 60.12% 39.28% 43.32% 56.42% 32.52% 38.78% 41.92%
Abort Phase Mode MOI MOI Launch Launch MOI PD MOI Launch PD MOI MOI Launch
Abort Events MOI Abort MOI Abort LEO Repair AR&D MOI Abort PD Aero MOI LEO Repair PD Aero MOI MOI Abort LEO Repair
Pre-deployment Failure 16.25% 11.15% N/A N/A 4.49% 32.08% N/A 19.24% 4.84% N/A
Dust Storm Encounter 94.02% 94.02% 94.05% 94.05% 94.03% 94.05% 94.05% 94.01% 94.01% 94.05%
LoS 10.21% 10.8% 10.80% 0% 0% 0% 0% 9.87% 0% 0%
LoV (Cargo) 71.37% 61.14% 61.14% 71.63% 31.50% 84.75% 5.90% 62.99% 32.00% 16.75%
LoV (Crew) 27.34% 25.17% 25.17% 7.04% 26.19% 35.88% 26.21% 42.71% 29.99% 29.24%
HAB Failure 0.94% 0.99% 0.99% 2.06% 5.52% 1.39% 1.44% 1.66% 2.25% 1.99%
HAB Power Failure 0.71% 0.74% 0.74% 0.76% 1.86% 1.03% 0.55% 1.23% 1.70% 0.73%
lations happens before the crew attempts EDL to the Martian
surface. With the execution of scientic investigations on Mars
being the primer mechanism for obtaining mission utility, this
therefore statistically describes the overarching reason for the
very low values of mission utility. Additionally, the fact that the
mission-ending events mostly do not fall within the nal mis-
sion phase of the simulations shows that most abort scenarios
are successful in bringing the crew back to Earth, and that there
are more successful aborts than there are LoC outcomes from
the MD.
e optimization from concept experimentation signicant-
ly decreased the percentage of pre-deployment failures com-
pared to the baseline DRA mission. is is due to the redun-
dancy selection in the tradespace, which integrates redundant
pre-deployed vehicles, as well as the launch trade selection of
the siphon method (the repetitive launching of a fully reusable
propellant tanker for the interplanetary transfer vehicle, based
on the method proposed in the ITS concept) for concepts with
non-direct TMI. e latter improves pre-deployment because
it reduces the number of mission critical launches that result
in LoM in the case of failure. By separating mission critical
systems from propellant, the mission critical systems can be
brought to orbit in fewer launches, and the failure of a propel-
lant tanker only assumes a delay in the mission.
e aggregate analysis also shows that the behavior and
interaction characteristics of and with dust storms may be of
signicant importance. is is because nearly all missions are
perturbed by a dust storm. e aggregated data also reveals the
LoV outcomes are highly individual to each concept. Contrari-
ly, the LoS and loss of habitat power values are low for all con-
cepts. is is due to the fact that these systems can be repaired
by the crew, and since concept experimentation prioritized
ergonomic optimization, more oen than not are xed by the
crew in the mission. It is worth noting that although the loss of
habitat percentage in missions is low, it still needs to be
addressed since a loss of this portion of the mission architec-
ture leads to LoC without a surface abort ability.
Finally, the descriptive analytics of the simulation data-
set determined that the concepts actually converge onto two
distinct solutions. is was not found directly from CD&E in
Progspexion because the framework does not include num-
ber of vehicles as a trade (except in redundancy trades). Also,
due to the mass budget ground rule that allowed rapid mass
budget calculations, the total size and mass of the vehicles was
not traded, only the composition of the payloads within. While
this allowed tradespace exploration of the core competencies of
each individual mission concept, this, along with the omission
of mission class in the tradespace, is the reason for the multi-
modal results.
In fact, all pre-deployed mission concepts converge into a
single solution, which closely resembles the Mars One mission
concept. e all-up mission concepts also converge into a sin-
gle solution obtained by the ESA HMM optimization which
resembles an abridged ITS mission concept. It is abridged be-
cause its scale is reduced to that considered in this research and
the ITS vehicle does not return back to Earth. erefore, the
descriptive analytics produces the Optimized Mars One and
Abridged ITS concepts that shall proceed into the following
analytics processes.
5.3. Predictive
ree types of predictive analytics are performed on the simu-
lation dataset. While the research produced many predictions
for each variant, only a limited number of exemplars of each is
discussed here. e rst type computes predictions that are not
part of the tradespace but can be calculated through manual
alteration of Progspexion’s inputs. e example of this predic-
tive analytics is the mass of EDL systems. While Progspexion
works to keep EDL mass at the concepts’ designated value, a
JA’MAR A. WATSON
JBIS Vol 71 No.1 January 2018 13
TABLE 4 Aggregates of Optimized Concepts
DRA Austere Concept 242 ESA HERRO H/S Direct MARPOST Mars Direct MarsOne Revised 242
Mission Dynamic 98.53% 98.11% 99.46% 96.61% 94.98% 98.03% 96.33% 96.56% 94.35% 95.84%
End Phase Mode Earth Earth Earth Earth Earth PD MOI Earh PD MOI Earth Earth
End Event Mode MT EVA MT EVA Aero MOI Surface Abort MT EVA PD Aero MOI MT EVA PD Aero MOI MT EVA MT EVA
Abort 62.35% 59.94% 59.64% 60.12% 39.28% 43.32% 56.42% 32.52% 38.78% 41.92%
Abort Phase Mode MOI MOI Launch Launch MOI PD MOI Launch PD MOI MOI Launch
Abort Events MOI Abort MOI Abort LEO Repair AR&D MOI Abort PD Aero MOI LEO Repair PD Aero MOI MOI Abort LEO Repair
Pre-deployment Failure 16.25% 11.15% N/A N/A 4.49% 32.08% N/A 19.24% 4.84% N/A
Dust Storm Encounter 94.02% 94.02% 94.05% 94.05% 94.03% 94.05% 94.05% 94.01% 94.01% 94.05%
LoS 10.21% 10.8% 10.80% 0% 0% 0% 0% 9.87% 0% 0%
LoV (Cargo) 71.37% 61.14% 61.14% 71.63% 31.50% 84.75% 5.90% 62.99% 32.00% 16.75%
LoV (Crew) 27.34% 25.17% 25.17% 7.04% 26.19% 35.88% 26.21% 42.71% 29.99% 29.24%
HAB Failure 0.94% 0.99% 0.99% 2.06% 5.52% 1.39% 1.44% 1.66% 2.25% 1.99%
HAB Power Failure 0.71% 0.74% 0.74% 0.76% 1.86% 1.03% 0.55% 1.23% 1.70% 0.73%
predictive analysis for the reduction of EDL mass produces
Fig.30. Here it can be seen that as the mass of EDL systems in
the Optimized Mars One concept decreases, the mission utility
slightly increases.
e second type is for variables which are part of the tradespace,
but do not extend across the desired range of analysis. An ex-
ample of this type of analytics is crew size. e tradespace in-
cluded crew sizes of 4, 6, and 8; however, predictive analytics
can provide the mission utility of a mission of greater crew
sizes. Fig. 31 shows the trend of mission utility as a multiplica-
tive of 8 crewmembers in the optimized Mars One concept. A
decrease in utility is evident for this variable. Another example
variable is the initial mission intelligence value, which quan-
ties the knowledge of Mars known by the crew prior to the
CMM. In Progspexion, only the values of certain, probable,
and neutral are allowed as initial values since it is assumed a
mission would not proceed with lesser knowledge of Mars. To
predict the value of mission utility for the Optimized Mars One
concept at lower initial missionintelligence values, a statistical
extrapolation can be used to determine its possible value (Fig.
32). e derivation is therefore able to predict mission utility as
a function of initial mission intelligence (Eq. 1).
Fig. 30 Mission Utility as a Function of EDL Mass Reduction for
the Optimized Mars One Concept.
Fig. 31 Mission Utility as a Function of Crew Size for the
Optimized Mars One Concept.
Fig. 32 Initial Intelligence Utility Function for the Optimized Mars
One Concept.
e third type of predictive analytics predicts values that are
both not in the tradespace and cannot be replicated by manual
manipulation of Progspexion’s inputs (assuming its code can-
not be altered). is is most oen useful for mathematically
predicting ‘what if ’ scenarios outside the scope of the simula-
tion framework. For example, predictive analytics is performed
to determine the eect of removing the mass budget ground
rule. is enables the alteration of vehicle numbers and sizes in
a mission concept, veering from its core principals.
Fig. 33 overleaf displays the probability of successful car-
go EDL at various levels of redundancy. “Single” is plotted as
reference of just one vehicle without the redundancy. is is
useful since only N+1 redundancy is coded in Progspexion’s
tradespace. “Split” is dened as the total mass being split be-
tween N+X vehicles. For example, a 40 t lander at N+1 would
convert to two 20 t landers, N+2 to three 13.33 t landers, etc.
e “duplicate” is dened as an identical number of N+1 ve-
hicles; meaning, N+1 is two 40 t vehicles, N+2 is three 40 t
(1)
CREWED MARS MISSION Concept Development and Experimentation
14 Vol 71 No.1 January 2018 JBIS
Fig. 33 Probability of Cargo EDL Success with Redundancy.
vehicles, and so forth. e name “duplicate” indicates parallel
redundancy in which only one of the N+X vehicles is required
to be successful.
In this predictive analytics, “duplicate” redundancy obtains
the greatest EDL success rate; however, transfer of these plots
to mission utility is signicantly greater for the Abridged ITS
concept than the Optimized Mars One concept since it is an
all-up mission design. For concepts like the Optimized Mars
One that pre-deploy cargo, the eects of adding additional in-
jections and orbit insertions (as well as the launches) for the
increasing number of vehicles also alters the mission utility.
As an example, according to the graphs, the directly injected
cargo lander should perform almost identically to splitting the
vehicle amongst two half-sized EDL systems at the mass of the
Optimized Mars One concept; however, due to the previous-
ly mentioned pragmatic eects in an operational context, the
single vehicle outperforms the split since less vhicles must be
deployed (Fig. 34). To note, it may still be possible to place mul-
tiple cargo landers on a single vehicle during pre-deployment.
e same prediction is performed for crewed EDL. While
the predictions are the same as for cargo vehicles, the dupli-
cate redundancy adds an additional layer of prediction since
the lossless crew ground rule would need to be removed to
allow for crewed parallel redundancy. Fig. 35 shows that for
lower massed vehicles, the single and split EDL performance
are roughly equivalent; but, for higher mass crew vehicles, it
is much more benecial to split them up into multiple smaller
vehicles. As such, the mission utility of the Optimized Mars
One concept when splitting up the 40 t crew lander is comput-
ed (Fig. 36). e gure shows that despite needing to propa-
gate additional crew vehicles, this disaggregation increases the
average mission utility up to N+2 redundancy before seeing a
decrease. In addition to the EDL eect, a great portion of this
increase is due to the introduction of transshipment capability
added to the mission architecture by splitting up the vehicles.
Also, to note, all of these mission utility values obtained during
predictive analytics exceed the Optimized Mars One mission
utility obtained during tradespace exploration.
erefore, this predictive analytics was not only able to
identify the limitations of the chosen ground rules, assump-
tions, and delimitations of the research, it was also able to
overcome them and produce a greater level of optimization.
is result is conceptually similar to the principal of the 2-4-2
concepts; however, the 2-4-2 concept is of the all-up mission
class (which underperformed its pre-deployed, split mission
counterpart) and mission class was not traded during concept
experimentation optimization. erefore, a new concept is ar-
chitected from this predictive analytics. It is a merger of the
Mars One and 2-4-2 concepts, hereinaer referred to as the
Enhanced 2-4-2 Concept. A comparison of the mission utility
and the risk with the Optimized Mars One concept is shown in
Fig. 37 and Fig. 38 respectively. Although the risk and median
mission utility do not improve much, the IQR does show a dis-
placement of the dataset more towards its maximum possible
achievement, which explains the greater mean mission utility
values in Fig. 36.
5.4. Prescriptive
Prescriptive analytics is a quantitative instruction for improv-
ing the mission concepts. From this it is possible to produce
statistically guided amendments, additional investigations,
contingencies, and preliminary principles for conducting
Fig. 34 Mission Utility of N+1 Redundant Optimized Mars One
Concept.
Fig. 35 Probability of Crew EDL Success with Redundancy.
Fig. 36 Mission Utility as a Function of Split Redundancy for the
Optimized Mars One Concept.
JA’MAR A. WATSON
JBIS Vol 71 No.1 January 2018 15
Fig. 37 Mission Utility of the Optimized Mars One and Enhanced
2-4-2 Concepts.
Fig. 38 Risk of Optimized Mars One and Enhanced 2-4-2
Concepts.
Fig. 39 Projection of Mission Utility with Redundant Crew.
Fig. 40 Mission Utility as a Function of Parallel Redundancy for
the Enhanced 2-4-2 Concept.
CMMs. It seeks to expand and give application to the previ-
ous steps of analytics. e exemplar shown here is leveraging
of the analysis concerning crewed EDL redundancy. It is clear-
ly demonstrated in Fig. 35 that duplication with parallel crew
redundancy would decrease the risk of EDL failure, which, in
these concepts, increases mission utility. e prescription is
therefore to remove the lossless crew ground rule.
Although an original assumption is that a LoC event is a failed
mission, the data used to produce Fig. 35 and Fig. 36 can be
set up as a comparative datasets. Doing so facilitates projec-
tion of mission utility values for the case of allowing parallel
redundancy in crewed EDL (Fig. 39). It can be seen that the
probability of EDL success continues to increase in parallel re-
dundancy of the crew and the point of diminishing returns is
beyond the investigated N+3 limit. Using this prescription, it is
possible to recreate Figure 36 but assume duplication of crewed
vehicles in the Enhanced 2-4-2 concept (Fig. 40). erefore,
within the range of N+3 redundancies, prescribing the disag-
gregation of crewed vehicles and removing the lossless crew
ground rule would improve mission utility by more than 10%.
is prescriptive analytics is therefore mathematically able to
determine the viability of partial LoC outcome acceptance.
6 DISCUSSION
Trades on precursor missions and EOM design oer the great-
est improvements on mission utility. Specically, selecting for
colonization EOM design is a primary driver in the improved
mission utility of experimentally optimized concepts. e sim-
ulations therefore suggest it is easier to resupply the crew on
Mars than to attempt to return them back to Earth, albeit with
the limited long-term consideration of colonization that is in-
cluded in this research. Earth return is further hindered by the
shortcomings of available precursor missions – none of which
perform developmental or operational testing and evaluation
of Mars ascent, areocentric AR&D, or TEI with CMM systems.
is uncertainty must be addressed if colonization is to be an
unfavorable EOM choice. As an acknowledgement, the selec-
tion of colonization also eliminated the possibility for teleop-
eration or telecollaboration to remain in optimized solutions
since Progspexion assumes colonization takes place on the
Martian surface.
Of the precursor missions available for selection, an un-
crewed landing has the greatest impact on mission utility, fol-
lowed closely by the crewed yby mission. e cis-lunar mis-
sion had a much more negligible impact to the success of the
CMM; therefore, its necessity should be investigated.
e siphon method for launch and integration in LEO is
highly selected during tradespace exploration for systems that
require multiple launches with LEO AR&D. is is especially
the case for larger vehicles such as crewed mission systems and
all-up concepts. Siphon method selection is due to the fact that
later mission critical launches, which result in LoM if failed, are
swapped with iterative launching of a propellant tanker, which
only results in a mission delay in the case of failure. erefore,
it is a trade that both increases payload capacity and decreases
LoM risk during launch and integration. It is worth mention-
ing that the cycler method also faired very well in tradespace
exploration and oered an alternative to the siphon method
for smaller vehicles that could still be directly injected. If the
readiness level of directly injected hyperbolic rendezvous can
be increased, the use of cyclers should be revisited for concepts
utilizing smaller crew vehicles.
CREWED MARS MISSION Concept Development and Experimentation
16 Vol 71 No.1 January 2018 JBIS
Alternative to crew launches, cargo launches favored the di-
rect injection method, both for pre-deployment missions and
for all-up missions simultaneously transiting multiple vehicles.
is is preferred for the smaller cargo vehicles in order to avoid
the added complexities of integration in LEO. While pre de-
ployment is currently needed in a CMM, the Apollo benchmark
results indicate an inverse proportionality between the necessity
to pre-deploy cargo and the ease in performing the mission.
Aerocapture should be avoided at all costs until its reliability
can be proven. is led to the selection of propulsive MOI in
all mission concepts that could aord the additional propellant
in the architecture’s SW&P budget. Consequently, this elimi-
nated nuclear propulsion as an optimal choice for in-space sys-
tems – the reason being attributed to the lower readiness of the
technology. Coupled with its increase to risk across numerous
mission phases in performing propulsive MOI and EOI ulti-
mately leads to its omission. However, with a strong capability
to alleviate payload budget constraints and reduce the number
of required launches, it is highly recommended that nuclear
propulsion be revisited for increasing its technology readiness
level (TRL).
As discussed in the analytics section, redundancy has a tre-
mendous eect on the design of the mission. e optimized
concepts indicated that some type of redundancy must be im-
plemented during interplanetary transit periods of the mission.
It is therefore pertinent that concepts reinstate this capability in
their mission denitions. Transshipment, as dened by Jean-
Marc Salotti in derivation from Wernher Von Braun, is the
current frontrunner in this solution since CD&E shows disag-
gregating the crew vehicles for EDL increases mission success
probability. Reinvestigating potential EOM designs will distin-
guish the trades between transshipment as a standalone feature
and conducting complete duplication of the architecture.
e TRL of the additive manufacturing / 3-D printing re-
dundancy option must be improved to make its selection vi-
able. is was also the case in selecting ISRU of consumables.
Even in the case of colonization, the mission utility to risk ratio
was optimized when resupply was chosen in place of in-situ
production reliance of consumables and redundant mass. e
usefulness of ISPP of return propellant, as well as the Mars
ascent siphon method, proved to be greater, but its purpose
was diminished when colonization became the primary EOM
choice. It may still be revisited if development of ascent and/or
return capabilities is desired.
e quantication of mission intelligence must occur to
make the estimative probability more accurate. Regardless of
such an eort, it is clear that initial intelligence of Mars alters
the mission utility even during execution of a perfect mission.
erefore, during the period of CMM incubation, the gain of
knowledge regarding Mars should not be stalled. Doing so will
require and place a heavy burden on precursor mission(s) to
replace this acquisition.
During surface exploration optimization, it was found that
crew size was inversely related to mobility system capability (or
rover system capability for teleoperation missions). For sur-
face exploration, crew sizes of 4 required advanced mobility
systems, such as those of the proposed rotocra, to overcome
only having one exploration group for investigations. Crews of
6 did not need these advanced systems, but still required ad-
vancements to current mobility systems, such as those present-
ed in the large to medium mobility systems. At a rew size of
8, advancements were deemed unwarranted and current mo-
bility systems of the medium size accomplished desired utility
values. During this portion of the research, it was also found
that if advanced unpressurized aerodynamic mobility systems
could be used for longer transits across the Martian surface,
pressurized mobility could be reduced to the small size re-
quirements and used only for backup or commuter habitation.
is combination saved on the mass budget in comparison to
large pressurized mobility systems for long regional transits ac-
companied by smaller unpressurized mobility systems for local
and vicinity investigations. e suborbital mobility option was
unnecessary in this research due to the range limit of 100 km
for regional investigations. e technology does, however, pos-
sess potential advantages if it can be shown to open up a new
range of global investigations while still allowing the central
habitat design. is would ultimately drive the reemergence of
ISPP necessity as well.
For investigations via teleoperations, aerodynamic rovers
enabled full mission investigations for smaller crew sizes. At
a crew size of 8, large rover systems were sucient for surface
exploration. ese large rovers, however, are still an advance-
ment to current technologies. ese results are based on the
current assumptions concerning the central habitat design of
investigations, the method of inclusion regarding the locale of
the investigations themselves, and the scheduling algorithms
developed for the pairs of crewmembers performing surface
exploration. A true mathematical comparison between the ef-
cacies of crew exploration to that of teleoperated rovers from
areocentric orbit must be continually developed to further so-
lidify these trades.
e addition of solar I&W capabilities had no eect on the
mission. Even when Progspexion guarantees the occurrence of
an SPE during the mission, Solar I&W did not produce changes
in mission utility or risk that were greater than the median ab-
solute deviation or standard deviation. is is also true for mis-
sions at solar maximum without any change to the solar I&W
capabilities. erefore, under the assumptions of Progspexion’s
OpWX model concerning SPEs, these systems are not of impor-
tance. is verdict quickly changes, however, when discussing
GCR. Simulations where a GCR occurs are substantially wors-
ened both in terms of mission utility and risk. For example, the
Enhanced 2-4-2 concept experienced roughly 50% reduction of
mission utility and increase in LoC risk for the mission. e
inclusion of specialized crew modules and in-situ surface hab-
itation barely eased this impact since GCRs can occur at many
portions of the mission in which the crew is not inside the hab-
itats. Since GCR is an area that is not very well understood, it is
best to further research GCRs as OpWX and the development
of any potential I&W schemes or technologies. erefore, GCR
could be a limiting factor of future CMMs.
7 CONCLUSIONS
CMM CD&E garnered insights into the mission, including: (1)
pre-deploy mission concepts perform statistically better than
all-up concepts, however the Apollo benchmark results indi-
cate the necessity to pre-deploy assets diminishes as the di-
culty of performing the mission decreases; (2) benets of the
uncrewed landing and crewed yby precursor mission transfer
well to the mission, while the crewed cis-lunar precursor mis-
sion has no impact on mission utility; (3) EOM design favors
colonization as it is easier to resupply the crew on Mars than
attempt to return them to Earth; (4) the TRL of ISRU is too
low to augment the Earth resupply need, however ISPP is ben-
JA’MAR A. WATSON
JBIS Vol 71 No.1 January 2018 17
ecial for missions requiring Mars ascent; (5) utilization of
the siphon method for Mars ascent is worthy of research and
development but is negligible with the suggested colonization
EOM design; (6) the EOM design impact also demonstrates
a signicant improvement to both mission utility and risk by
converting the mission contingency hierarchy to an abort-to-
Mars strategy; (7) redundancy is critical to mission success
and the research suggests instituting transshipment capabili-
ties, particularly during interplanetary transit and areocentric
exploration phases; (8) maturation of the iterative launching
of a fully reusable propellant tanker in the siphon method ef-
fectively increases the mission gear ratio while reducing mis-
sion-critical launches; (9) optimal launch congurations are
dierent for cargo and crew – cargo prefers direct injection
to Mars, rather crew uses AR&D in LEO utilizing the siphon
method; (10) maturation of the hyperbolic rendezvous could
advocate the cycler method for small crew vehicles; (11) there
is a strong coupling of MOI/EOI methods and propulsion sys-
tem selection – propulsive orbit insertion vastly outperforms
its aerodynamic counterpart which contributes to the prefer-
ence of chemical propulsion over nuclear systems due to the
repeated uses of the propulsion in missions relying on pro-
pulsive insertion methods; (12) the remaining mission gaps
and shortcomings are related to the omission of human-rated
Mars ascent and areocentric operation precursor evaluations,
as well as the lack of CMM CONOPS development; (13) both
dust storm perturbations/encounters and MD occur in almost
all missions and therefore are signicant in the generation of
pragmatic CMM concept solutions; (14) application of paral-
lel redundancy principles to crewmembers in acceptance of
partial LoC outcomes signicantly increases mission utility;
(15) solar I&W systems are not useful to the mission as alter-
ation of risk and mission utility, even during missions where
an SPE is guaranteed, is not statistically signicant; (16) Mars
mission intelligence clearly impacts mission outcome but its
improvement is beyond the scope of this research (except pre-
cursor missions) as it usually pertains to observations, scien-
tic missions, and rover exploration that are performed prior
to the CMM; (17) crew size is inversely proportional to surface
mobility size – crews of 4 require advanced mobility systems
and aerodynamic rovers during teleoperations, crews of 6 re-
quire upgraded mobility systems and substantial rovers dur-
ing teleoperations, and crews of 8 can use currently projected
mobility systems and teleoperation rovers; and, (18) GCR is
detrimental to the mission as its occurrence overwhelming in-
creases LoC outcomes, even in the case of employing in-situ
habitation shielding methods.
Acknowledgements
e author would like to thank Acolytion for enabling the use
of Progspexion in this research and all reviewers for their rec-
ommended edits to this article for publication.
1. U.S. Human Spaceight Plans Committee, "Seeking a Human
Spaceight Program Worthy of a Great Nation," NASA, Washington,
DC, 2009.
2. K. Goodli, P. Troutman, D. Craig, J. Caram and N. Herrmann,
"Evolvable Mars Campaign 2016 – A Campaign Perspective," in AIAA
SPACE Forum, Long Beach, CA, 2016.
3. International Space Exploration Coordination Group, "e Global
Exploration Roadmap," in Symposium of Human Space Endeavors, South
Africa, 2011.
4. Aerospace Safety Advisor Panel, "Annual Report for 2015," NASA,
Washington, DC, 2016.
5. National Research Council, "Pathways to Exploration: Rationales and
Approaches for a U.S. Program of Human Space Exploration," e
National Academies Press, Washington DC, 2014.
6. E. Conway, Exploration and Engineering: e Jet Propulsion Laboratory
and the Quest for Mars, Baltimore, MD: John Hopkins University Press,
2015.
7. D. Rapp, Enabling Technologies for Exploring the Red Planet, Berlin,
Germany: Springer, 2007.
8. National Aeronautics and Space Administration, "Human Exploration
of Mars Design Reference Architecture 5.0," NASA, NASA/SP 2009-566,
Houston, TX, July 2009.
9. National Aeronautics and Space Administration, "Human Exploration
of Mars Design Reference Architecture 5.0 - Addendum," NASA,
NASA/SP-2009-566-ADD , Houston, TX, July 2009.
10. National Aeronautics and Space Administration, "Human Exploration
of Mars Design Reference Architecture 5.0 - Addendum 2," NASA,
NASA/SP-2009-566-ADD2, Houston, TX, March 2014.
11. H. Price, A. Hawkins and T. Radclie, "Austere Human Missions to
Mars," American Institute of Aeronautics and Astronautics, 2009.
12. J.-M. Salotti, "2-4-2 Concept for Manned Missions to Mars," Cape
Town, S.A., 2010.
13. J.-M. Salotti, "Human Mission to Mars: e 2-4-2 Concept," Bordeaux
Cedex, F.R., 2011.
14. J.-M. Salotti, "Revised Scenario for Human Missions to Mars," Acta
Astronautica, vol. 81, pp. 273-287, 2012.
15. J.-M. Salotti, "Simplied Scenario for Manned Mars Missions," Acta
Astronautica, vol. 69, pp. 266-279, 2011.
16. European Space Agency. "Human Missions to Mars", CDF-20(A), 2004
17. M. Wade, "Marpost," 2001. [Online].
18. R. M. Zubrin, D. A. Baker and O. Gwynne, "Mars Direct: A Simple,
Robust, and Cost Eective Architecture for the Space Exploration
Initiative," American Institute of Aeronautics and Astronautics, vol. 91,
no. 0328, 1991.
19. R. Zubrin and R. Wagner, e Case for Mars: e Plan to Settle the Red
Planet and Why We Must, New York, N.Y.: Free, 2011.
20. R. M. Zubrin and D. B. Weaver, "Practical Methods for Near-Term
Piloted Mars Missions," in 29th AIAA/ASME Joint Propulsion
Conference, Monterey, C.A., 1993.
21. J.-M. Salotti and R. Heidmann, "Revisiting Mars Semi-Direct," Torino,
Italy, 2015.
22. Mars One, "Roadmap - Mission - Mars One," 2015. [Online].
23. S. Do, K. Ho, S. S. Schreiner, A. C. Owens and O. L. deWeck, "An
Independent Assessment of the Technical Feasibility of the Mars One
Mission Plan," in 65th International Astronautical Congress, Toronto,
2014.
24. S. R. Oleson, G. A. Landis, M. L. McGuire and G. R. Schmidt, "HERRO
Mission to Mars Using Telerobotic Surface Exploration from Orbit,"
Ameri can Institute of Aeronautics and Ast ronautics , Cleveland, OH, 2013.
25. S. R. Oleson, G. A. Landis, M. L. McGuire and G. R. Schmidt, "HERRO
Mission to Mars using Telerobotic Surface Exploration from Orbit,"
JBIS, vol. 64, pp. 304-313, 2011.
26. W. Von Braun, e Mars Project, Urbana, I.L.: University of Illinois,
1962.
27. Space Exploration Technologies Corporation, "Making Humans a
Multiplanetary Species," in 67th Annual Meeting of the International
Astronautical Congress, Guadalajara, Mexico, 2016.
28. B. Aldrin, S. Saika and J. Longuski, "Cycling Pathways to Occupy
Mars," in Inaugural Buzz Aldrin Space Institute and Florida Institute of
Technology Workshop, Melbourne, FL, 2016.
29. B. Aldrin and L. David, Mission to Mars: My Vision for Space
Exploration, Washington, D.C.: National Geographic, 2013.
30. Space Exploration Technologies Corporation, "Elon Musk's Mission to
Mars," 2013. [Online].
31. D. Tito and et al., "Feasibility Analysis for a Manned Mars Free-Return
Mission in 2018," IEEE, 2013.
32. Inspiration Mars Foundation, "Architecture Study Report Summary,"
2013.
REFERENCES
CREWED MARS MISSION Concept Development and Experimentation
18 Vol 71 No.1 January 2018 JBIS
33. T. Cichan and et al, "Science Possibilities Enabled By e Mars Base
Camp Human Exploration Architecture," Denver, C.O., 2017.
34. J. A. Watson, "Holistic Methodology for Stochastic Mission Utility
Analysis," Int. J. System of Systems Engineering, vol. 8, no. 2, pp. 174-188,
2017.
35. J. R. Wertz, "ORS Mission Utility and Measures of Eectiveness," in 6th
Responsive Space Conference, Los Angeles, C.A., 2008.
36. J. A. Watson, "Benchmarking the Human Exploration of Mars Design
Reference Architecture," In Press: Int. J. Space Science and Engineering,
2018.
37. J. A. Watson, "Stochastic Modeling for Concise Space Mission
Simulation," In Review: J. of Astronaut Sci, 2017.
38. J. A. Watson, "Progspexion: Simulators for the Prognostication of Space
Exploration Missions," Acolytion LLC, Cape Canaveral, FL, 2017.
39. R. d. Jonckheere and B. Preiss, "Using Simulations," in Space Modeling
and Simulation, El Segundo, C.A., 2004, pp. 95-128.
40. A. Whitmire and et al., "Minimum Acceptable Net Habitable Volume for
Long-Duration Exploration Missions," Houston, TX, 2015.
41. B. Woolford and R. Bond, "Human Factors of Crewed Spaceight," in
Human Spaceight: Mission Analysis and Design, W. J. Larson and L. K.
Pranke, Eds., McGraw-Hill, 1999, pp. 133-153.
42. J. D. Rummel and et al., "A New Analysis of Mars ‘‘Special Regions’’:
Findings of the Second MEPAG Special Regions Science Analysis
Group," Astrobiology, vol. 14, no. 11, pp. 887-968, 2014.
43. D. W. Beaty, J. D. Rummel and MEPAG, "Introduction to an Udated
Analysis of Planetary Protection "Special Regions" on Mars," in 45th
Lunar and Planetary Science Conference, 2014.
JA’MAR A. WATSON
Received 11 March 2018 Approved 27 April 2018
