Robotic Precursor Mission

Which would provide a greater return on investment: a 19-day mission or a costlier mission lasting nearly a year?

This study features the following innovations:

• It is the first study, to the best of our knowledge, to systematically analyze and compare return-on-investment for alternative durations of a NASA mission.
• It is the first of our studies to look at the entire life-cycle costs of capabilities, rather than just their development costs to TRL 6.

Before sending astronauts on a costly and hazardous mission to explore for lunar water, NASA deems it advisable (as of this writing) to send one or more robotic rovers on a precursor mission to confirm on the ground what remote-sensing instruments have suggested from orbit -- that frozen water is present in Shackleton crater at the lunar South Pole.

The START team has completed the first phase of an evaluation of two mission concepts which differ greatly in their durations; one would last for 19 days and the other would run for nearly a year. In addition to providing useful information for this specific mission, our analysis serves as a model for other, similar comparisons of the relative productivity (returned benefit divided by investment dollar) of short and long missions.

The question of how long the mission should last is complex. On the one hand, the costs of developing the needed technological capabilities and of transportation to the lunar surface are essentially the same whether the mission runs for a few days or a year. In that sense, longer missions -- which presumably would accomplish more -- would drive up the return on investment. On the other hand, there are operational costs, such as the ground crew that would monitor and control the rover and analyze the data it transmits, which would increase with mission duration.

The START team was asked to demonstrate how its methodology could be used to determine whether the added value of a lengthy mission would justify the added cost.

For this study, we looked at two options for providing power to the rovers that would search for lunar stores of frozen water. One option was to use a radioisotope thermal generator (RTG) to conduct a mission lasting 354 Earth days. The other option considered here was to use a methanol-LOX (liquid oxygen) fuel cell that could be expected to last about 19 days.

We began by defining 15 activities (Table 1) that would constitute this mission regardless of the power source, and assigning each a relative value. The most important activities included drilling holes along a spiral route on the lunar surface, extracting core samples, and analyzing the contents. Each value level was given a numerical weight: 0.50 for high, 0.33 for medium, and 0.16 for low.

Activity	Instruments	Weight
1. Spiral survey for ground truth; surface H variation	Neutron spectrometer	0.50
2. Regolith conductivity assessment	Ground-penetrating radar	0.33
3. Surface roughness (driving)	Laser ranging scanner	0.33
4. Topography (driving)	Pancam & flash	0.33
5. Radiation monitoring	Radiation monitor	0.16
6. Core acquisition	1-m drill, microbalance	0.50
7. Volatiles assessment, contamination, hydration, adsorption	TEGA & mass spectrometer	0.50
8. Regolith properties assessment	Microscope	0.33
9. Geophone survey for regolith density, cobble structure	Acoustic & thumper	0.33
10. Molecular, mineralogic assay	Raman spectrometer	0.16
11. Surface-ice assessment	Tunable diode laser	0.50
12. Free iron, ilmenite, contamination assessment	UV spectrometer	0.16
13. Color, mineralogy, ice, contamination assessment	VNIR spectrometer	0.50
14. Atomic composition, pierce varnish assessment	LIBS chemcam	0.16
15. Thermometer measurements (wheels)	Temperature gradient & probe	0.16

Next, we identified the technological capabilities required by each of the 15 activities, and mapped the capabilities to the activities (Table 2). The relationships were rather intricate, as each activity required more than one capability, and most capabilities benefited more than one activity.

Derived Capabilities		Required by Activity Number
Drive/gear motors (robotic operation)	Operational lifetime of dry lubricated gearmotors; environmental temperature maximum for gearmotors	1, 2, 3, 4, 6, 9
Surface mobility	Precision topographic map/ entry, descent, & landing	1 - 15
	Range of lateral traverse, autonomous traverse, terrain accommodation	1, 4, 6, 9
Power-delivery system	MMRTG continuous power, fuel-cell power source energy density	1, 2, 4, 6, 7, 9, 10, 11, 12, 13, 14, 15
Drive motors (thermal control)	Operational temperature	1, 2, 3, 4, 6, 9
	Temperature change to surface (V-grove thermal isolator)	2, 6, 7, 15
Communicate to and from the lunar surface in crate thermal environment	S-band radio transmit power (rover-lunar comm), S-band medium-gain antenna that can survive thermal environment, S-band cabling that can survive thermal environment	4, 8, 9
Drill system	Rate, depth, mass, power	6
	Sample preservation	5, 6, 7, 13
	Extreme environment	2, 3, 4, 6, 7, 10, 11, 12, 13, 14
	Sample volume	6, 7, 8, 12, 13, 14
	Sample number	1, 2, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
	Core-capture bits, regolith-capture bits	6
	Bit/drill stem exchange/release	6, 7
Sample-handling system	Mass, energy per sample placement, sample volume	6

Based on the values assigned to the activities (Table 1), we computed the value of each capability, taking into account the multiple relationships.

We then used the START software tool to compute optimal funding (based on expected return on investment) of capability areas by year in order to determine which of the potential missions would be enabled, given constraints on budget, mission launch dates, and workforce. The enabled missions were the following:

• Lunar campaign (traveling to and from the Moon)
• Robotic precursor (354-day or 19-day mission)
• Human-Robot prospecting (7-day mission)
• Human small-scale ISRU (7-day mission)
• Full-scale ISRU

The budget levels for each of the capability areas as a function of time are given in Fig. 1.

Figure 1. Optimized capability portfolio containing the longer (RTG) robotic precursor mission. Black curve is the anticipated total budget.

Looking at the capabilities included in the optimal portfolios, we computed the values of the 354-day option and the 19-day option, based upon the capabilities that could be actualized.

The next step was to estimate the life-cycle costs for each of the two mission concepts. The costs of the RTG option are presented in Table 3, normalized to 100 for the total cost:

Element	RTG cost element normalized to total cost of 100	Fuel-cell element cost divided by RTG element cost
Lander	20	1.0
Launch system	17	0.9
Development, demonstration, test, and evaluation	51	0.8
Operations	4	0.4
Technology (TRL<6)	8	1.1
Total	100	0.9

For each of the two options for the robotic-precursor mission, the total expected value of all the capabilities in the portfolio was divided by the total life-cycle cost to yield the option's productivity value.

The results: Productivity is 0.07 for the fuel-cell option, and 0.81 for RTG option. Thus, increasing the cost by about 11% (from fuel-cell to RTG) yields an RTG-option value that is 1200% that of the fuel-cell option, making the RTG option by far the more cost-effective option, given the constraints used in this particular analysis.

It is important to view these results in light of certain assumptions that were employed in this study. Key among these was the assumption, for the sake of simplicity in this first phase of the study, that each option would carry an identical payload. Thus, the fuel-cell option and the RTG option would have the same capabilities (mainly drilling holes and analyzing their contents), and the only real difference would be that the lengthier RTG option would perform these activities more times. It is easy to imagine, however, that a mission lasting nearly a year might be given different kinds of activities to accomplish rather than just more of the same activities -- perhaps beginning some ISRU experiments, for example, if frozen water is located. Such a change in the parameters might substantially increase the value calculated for the longer mission.

On the other hand, in most cases we assumed that the value of each activity would increase linearly with the number of times that activity was executed (this assumption was not made for drilling, where we knew that the value of an additional hole after many were already drilled may not equal that of the first few holes). If we were to replace this linear increase with an assumption of diminishing returns, the value of the longer mission would decrease.

Other important potential variations are under further study, including different kinds of fuel-cell technology (hydrogen/LOX, lithium/LOX, and Sterling engine); different, more-reliable autonomy (autonomous distance traversed per ground command); and different missions/ payloads for the candidate designs.

No matter how the parameters might be modified, however, it is important to realize that the methodology we employed would remain valid. In fact, the START methodology is designed to accommodate unlimited "what if" inquiries, in which the effect of changing various parameters can easily be seen.

Probable performance

Each of the capabilities included in this study carried a certain amount of uncertainty, with regard to both its value and its cost. Examples are displayed in Table 4 where, for instance, the first capability has a value of 3.2 plus or minus the square root of 0.2, and a cost of $3 million plus or minus the square root of $300,000. At the bottom of the table is the sum of all capability values (1151.9) followed by the uncertainty in that figure (33.0), and the sum of all capability costs (about $2.7 billion) followed by the uncertainty in that amount (plus or minus the square root of $30.5 million).

Table 4. Uncertainty in the value and cost of three example capabilities

Using these figures, we calculated the probability that the optimized portfolio we recommended for the 354-day RTG option would ultimately deliver the needed value within the available budget, given the uncertainty in the parameters on which the recommendations were based.

Figure 2 represents a 3-dimensional probability-density function over which we integrated to arrive at our answers, which are displayed in Table 5. They indicate a 46% probability that our portfolio would lie within the green area, where value equals or exceeds that which is required for the mission, and where the total cost is equal to or less than the available budget. There is a similar 46% probability that our portfolio would lie within the budget, but would not meet the value requirements. So, given the uncertainties in this particular study of an RTG option, there is a 92% probability of staying within budget and a 46% probability of also meeting or exceeding value requirements.

Figure 2. Probability density function. In the green region, value and cost both meet requirements; in the yellow region, cost is within budget but value falls short of requirements; in the orange region, value meets requirements but cost exceeds the budget; and in the red region, neither value nor cost satisfies the requirements.

Portfolio Performance	Probability
Value > Goal, Cost < Budget Cap	46%
Value < Goal, Cost < Budget Cap	46%
Value > Goal, Cost > Budget Cap	4%
Value < Goal, Cost > Budget Cap	4%
Cost < Budget Cap	92%
Cost > Budget Cap	8%

Table 5. Value and cost probabilities of the optimized capability portfolio

With this information, a decision-maker can determine how much of a budgetary margin to reserve in order to increase the probability of staying within the cost requirements, and whether further actions should be taken to increase the probability of meeting value requirements.