Comparing Architectures - Pressurized vs. Unpressurized Rovers

Variations and Non-Intuitive Solutions

Page:  1   2   3

Variation 1: Adding time margins

A margin of 25% is added to each science activity to represent time needed for documentation, and then that total is increased by 25% to allow for delays due to any engineering-related cause (e.g., equipment breakdowns that require the astronauts to stop and make repairs). The 25% "engineering" margin is also applied to all drive times and to the times the astronauts spend exiting and entering the lander-habitat and SPRs. No margin is added to the times allocated to sleep, unplanned activities or rover idleness. Results are shown below.

Table IV. Summary Results of Reference Problem + Time Margins

2 UPRs (Ref) 2 SPRs (Ref) 2 UPRs (Ref + Margins) 2 SPRs (Ref + Margins) Number of science activity sites visited 14 35 14 35 Tasks completed 75 186 72 186 Mass samples collected (kg) 86.6 216.5** 84.7 216.5** Vehicle distance traveled (km) 367 569 706 624 Total time required* 5 days, 8 hrs 5 days, 11 hrs 11 days, 8 hrs 7 days, 15 hrs EVA time (hrs) 118.6 53.4 238.2 79.0 Value 686 1700 662 1700 Mission productivity 4.1 12.9 1.9 9.3

  * Not including one day each for arrival and departure.
** This mass is likely to exceed the capacity of the return flight to Earth and so would require the astronauts to perform triage.

Adding these margins decreases productivity for both scenarios, since they increase cost without increasing value. Margins have an especially dramatic impact on the 2UPR scenario, where they more than double the total amount of time that would be required to complete the same tasks performed in the original analysis. (With time margins, the UPR scenario reaches the maximum time allowed for the mission without being able to complete all the tasks at locality #2.) The UPR is much more sensitive than the SPR to increases in the time required for driving (note that the UPRs' total driving distance now exceeds that of the SPRs) and for performing science tasks because everything done with a UPR counts as EVA time, which is subject to an 8-hour-per-day constraint. In this case, the time increases push the UPR into the need for a second 2-day rest/recharge period, severely consuming available mission time.

The SPR scenario fares much better, without the need for a second 2-day rest/recharge period, though with some backtracking which raises the vehicle distance traveled. With margins, the SPR scenario is found to be about 5 times as productive as the UPR scenario.

Variation 2: Adding science activities

Keeping the time margins in place, we next add 3 science tasks, to be conducted by the robotic rovers once at each of the 5 localities, to the list of activities: microscopic imaging of rocks, mapping of the lunar surface via lidar (a radar-like device that employs laser light), and ground-penetrating radar (GPR).

The results are shown in the table below. As expected, in the UPR scenario the astronauts and rovers are unable to travel beyond the first two localities or to complete the assigned science activities. The SPR scenario is able to complete all tasks. Nevertheless, relative productivity decreases for the SPR scenario because of the relatively low value/cost ratios of lidar mapping and GPR as shown in Table I.

Table V. Summary Results of Reference + Time Margins + New Tasks

2 UPRs (Ref + Margins) 2 SPRs (Ref + Margins) 2 UPRs (Ref + Margins + new tasks) 2 SPRs (Ref + Margins + new tasks) Number of science activity sites visited 14 35 14 35 Tasks completed 75 186 72 201 Mass samples collected (kg) 84.7 216.5** 75.5 216.5** Vehicle distance traveled (km) 706 624 633 584 Total time required* 11 days, 8 hrs 7 days, 15 hrs 11 days, 8 hrs 10 days, 7 hrs EVA time (hrs) 238.2 79.0 235.7 84.6 Value 660 1700 662 1840 Mission productivity 1.9 9.3 1.9 7.7

  * Not including one day each for arrival and departure.
** This mass is likely to exceed the capacity of the return flight to Earth and so would require the astronauts to perform triage.

As we saw, the addition of time margins brought total mission time to the upper limit in the UPR scenario, and so adding science activities bumps other activities rather than increasing mission time. Thus, cost remains the same. The UPR scenario gains value by performing all 3 new activities at locality 1 and one new activity at locality 2, but it now does not have enough time to go to the final science activity site at locality 2 and it therefore loses the value of the tasks it was previously able to do there. On balance, mission productivity remains about the same for the UPR scenario, and the SPR scenario is found to be about 4 times as productive.

Non-intuitive solutions

The optimal routes that HURON calculated for the reference problem and for each of the variations provide examples of the non-intuitive solutions that an automated planner can contribute. Given the constraint that the astronauts and rovers cannot work in the field for more than 3 days before they need to return to the lander-habitat for a 2-day rest/recharge period, and calculating that all of the tasks at localities 1, 2, and 5 could be performed within 3 days (in the SPR scenario), a human looking at the site geometry might intuitively say that the optimal solution would be to spend 3 days working at localities 1, 2, and 5 and then return to the lander-habitat for a rest/recharge period before traveling the longer distances required to reach localities 4 and 3.

"Intuitive" route	HURON route

However, HURON calculated that about 55 km of driving could be saved by taking a rest/recharge period before the constraint says it must. HURON's route takes the teams to locality 1, then 2, then back to the lander-habitat. After a rest/recharge period, the teams go to locality 5, then 4, then 3. The difference is that in the "intuitive" route, the teams drive from locality 2 to 5 (42.6 km) and from the lander-habitat to locality 4 (36 km), while in HURON's route, they drive from locality 2 to the lander-habitat (25 km) and from locality 5 to 4 (26 km) for a net savings of 27.6 km. Multiplied by the 2 rovers, that's a saving of about 55 vehicle km. Though the two work periods on each side of the rest/recharge period are unbalanced, the reduced driving time means lower cost and therefore a higher level of productivity.

When time margins are added, localities 5, 4, and 3 can no longer fit into one 3-day work period, so adhering to HURON's original driving plan would force a second rest/recharge period. To avoid this, HURON sacrifices the most efficient driving plan-another solution that might have escaped human intuition-and utilizes the route described as "intuitive" above (1, 2, 5, lander-habitat, 4, 3). When 3 new tasks are added, a second rest/recharge period becomes unavoidable and HURON returns to the itinerary that produces the shortest total driving distance. Thus, total driving distance in the reference+margins+new activities variation is 40 km shorter than in the reference+margins variation, saving the cost associated with the extra driving.