The past is prologue to the present
As background the Altair lander formed part of NASA’s now cancelled Constellation project and would have been used to land and takeoff from the lunar surface in a similar fashion to the Apollo Lunar Module (LM).
What’s interesting about MIT’s conceptual design is the way in which the original Apollo program decisions as to the split of responsibility between man and automation has subtly affected the resultant conceptual design of the human machine interface for the Altair spacecraft program. The best way to illustrate this impact of the original Apollo design is to look at how the landing trajectory ‘high gate’ (1) pitch up manouever affected the synthetic landing display concept.
If landing on the moon were simply a fuel and delta v management problem any landing trajectory would simply consist of a tail first braking phase with the final pitch up to vertical manouevre left as late in the landing as possible to conserve fuel. But due to the imprecision of mapping data, landing sites also need to be evaluated to ensure that a lander was not descending onto unsafe terrain such as an excessive slope, a crater rim or rock field. For Apollo only humans could do this job which meant in turn that the lander pilot needed to view the landing site during the approach and decide where and whether to land.
But to reduce mass the final configuration of the LM windows had ended up as two small angled windows providing only a 65 degree declination from the horizontal field of view and providing a greater degree of visibility was not be possible because the landers lower stage formed a blindspot. As an added complication the location of the radar on the opposite side to the LM’s crew windows meant that in order to obtain a radar ground track the crew would be forced to face upwards at the end of the braking phase.
In summary, given the final LM configuration both the need for a fuel efficient trajectory and for early radar acquisition conflicted with the crew’s need to see the lunar surface.
Deconflicting mission functions in Apollo
To deconflict these competing mission requirement NASA broke the landing trajectory into three functional segments, braking, approach and landing. The first or braking phase was designed primarily for the efficient propellant usage while reducing orbit velocity and guiding to ‘high gate’ conditions for initiation of the second phase via a pitch up to the vertical manoeuvre. The second,or approach phase was specifically designed for pilot visual monitoring of the approach to the lunar surface (2). The third or landing phase began at defined ‘low gate’ conditions and was designed to provide continual visual assessment of the landing site and to provide compatibility for the pilot takeover from automatic control for the final touchdown on the surface (NASA 1970).
The LM pitch attitude was varied from approximately 80 degrees from the vertical during braking to 40 degrees during the approach and finally to 10 degrees during landing. During landing the 25 degree blindspot underneath the lander also required that the lander maintain a greater than 25 degree look angle between the vehicle vertical axis and the line of sight to the landing sight.
Needless to say the astronaut corp were firmly in favor of such an approach as it maintained the ‘pilot in the loop’ approach (Mindell 2008) and this support is almost certainly one of the reasons that a fully automated approach to a pre-surveyed landing site was not seriously considered as an alternative mission architecture.
… I like computers and I believe in computers, but it ain’t going to land me on the moon. I’m going to do that. If something gets screwed up then it is going to be me, it isn’t going to be the computer…
David Scott, Apollo 15 LM Commander
Trading off fuel for knowledge
Now figuring out how high the high gate maneuver occured in the flight trajectory involved considering, how close could meaningful detail be discerned, how much time (and detail) was needed to make a decision, flight safety with regards to terrain altitude uncertainties, landing radar reliability and the ascent engines abort boundary.
The earlier the high gate occurred the more information and the greater the certainty, on the other hand the later the pitch over the less the certainty but more fuel reserve for a landing manoeuvre. In simple terms you trade fuel and flight time for knowledge. In the end for Apollo human perception and the lessening of uncertainty won the day and a high gate of about 9,000 ft was selected to give the crew time to view the landing site.
But even with a high gate set at 9,000 ft surprises and pilot disorientation could still occur. On Apollo 12 the crew found the initial view of the lunar surface difficult to encompass and critical seconds were taken up in orienting themselves, while on Apollo 15 the crew found themselves without any discernable land marks and again lost valuable time orienting themselves.
The problem was, when we pitched over and began to look out the window, there was nothing there . . . there were very few craters that had any shadows at all, and very little definition.
David Scott, Apollo 15 LM Commander
These recurring surprises are, in my opinion, a fundamental result of partitioning the flight timeline into a braking, approach and landing. When you can’t look out the window until late in the flight, surprises will occur.
Below the low gate
Another interaction between pilot FOV and trajectory during the approach is in the event of a landing re-designation. In order to re-designate short, or up range, the LM would need to pitch down to reduce the horizontal velocity more quickly and moving the landing site towards the bottom edge of the window. Diverting long, or down range, improves the pilots FOV by increasing the look angle. Apollo handled the constraint posed by the window edge on the look angle by only planning for down range diverts, which were also cheaper in fuel burn, but at a cost of decreased mission flexibility and safety by restricting the landing zone.
An alternate landing approach
As an interesting comparison the design of the Soviet LK lander design demonstrates a much greater focus on providing early visibility of the landing sight via the use of a scalloped cutaway and angled lower window for the pilot cosmonaut.
Due to launcher lift limits and the schedule pressure of racing against the americans the soviet designers had ended up with a simplified small single crew single stage lander with an extremely small fuel budget (100 seconds) for hovering and re-designation. To offset this ‘hat full’ of fuel reserve the soviet landing concept relied on a rugged landing system capable of handing a thirty degree slope and the use of ‘nesting’ rockets to hold the lander onto the surface and prevent a roll over.
On the plus side the single stage design inherently did not suffer from the 25 degree blindpsot formed by a large descent stage as did Apollo. In the LK design a pilot viewing window could provide a 83 degrees from the vehicles horizontal axis (7 degrees from vertical) view of the landing sight as opposed to the maximum declination of 65 degrees from horizontal (25 degrees from vertical) for Apollo.
As originally planned, an expendable stage (Block D) would provide the initial braking from orbit then separate at 13,000 ft. After separation the LK would use it’s own engine to continue braking then at a high gate pitchover at approximately 9800 feet for radar acquistion and descend to 1000 ft. At a 1000 ft low gate the lander would transition to a landing hover with 100 seconds of fuel reserve while displaying an automated landing point to the pilot. An earlier unmanned probe (or possibly an un-maned backup lander) of the Luna programme would act as a beacon for the manned LK.
With an ability to view to 83 degrees from the horizontal the LK lander only required a 7 degree look angle between vehicle vertical and the line of sight to the landing sight, thereby allowing a much steeper and more fuel efficient descent than Apollo.
The conclusion we can draw from this very different approach to landing is that the Apollo mission architecture was not the only viable approach to landing on the moon. The differences between the two approaches reflects the very different backgrounds of the design teams, the mission architectures selected and the constraints that these placed upon the subsequent detailed design and the allocation of mission functions.
And the sins of the fathers…
The original Apollo flight trajectory was in essence a compromise intended to deconflict competing requirements for crew visibility and fuel management. This conflict of course was introduced by the design constraints of the crew windows and the location of the lander radar.
Taking the opportunity presented by new technologies to decouple these functions is (for me) the actual justification for the use of a synthetic display in the Altair lander vehicle. For example the use of an external articulated boom mounted camera to provide the ‘view ahead’ before any pitch up to vertical could provide the crew with a dramatically improved capability to make decisions without the tradeoff in expenditure of fuel. Bottom mounted vehicle cameras would eliminate the blind spot and a requirement to maintain a minimum look angle during the landing sequence, Instead having selected a landing point a steeper and more fuel efficient landing phase could be used. The ability to perform up-range re-designation would be improved by a boom or tail mounted camera system as the pilots look angle would not be affected by the lower edge of the window during a pitch down manoeuvre to reduce delta V.
Using this sort of approach we could dispense with the Apollo segmentation of the landing into the three phases, potentially eliminating the approach phase. Unfortunately MIT’s study predicates that that the landing point forward-looking view is only available to the crew at the start of pitch-over thereby premising the current design upon the original functional partitioning of Apollo. As a result their conceptual design assumes a fixed set of cameras that recapitulate the original constraint of LEM forward facing windows and forward facing view.
To be scrupulously fair to the authors of what is otherwise an excellent report, they do note that what information should be provided during the preceding braking manoeuvre was to be a subject of future study, but I do think an opportunity to make significant improvements in mission safety has been missed. This is to me a beautiful example of the complexity of the design process in the real world and also illustrates why sometimes you need to consider the why of requirements and not just the what.
Cummings, M. L., Wang, E., Smith, C. A., Marquez, J. J., Duppen, M., & Essama, S. (2005).Conceptual Human-System Interface Design for a Lunar Access Vehicle (HAL2005-04), MIT Humans and Automation Laboratory, Cambridge, MA, 2005.
Mindell, D.A., Digital Apollo: Human and Machine in Spaceflight, MIT Press, May 2008.
NASA, Apollo Lunar Descent and Ascent Trajectories, NASA technical memorandum, NASA TM X-58040, March 1970, presented to the AIAA 8th Aerospace Sciences Meeting, NY, NY, 19-21 January, 1970.
1. The term ‘high gate’ and ‘low gate are both derived from aircraft pilot terminology for landing at airports. The use of such terms is of course a strong indicator of the ‘pilot’ centric approach adopted by NASA.
2. This statement was true for early Apollo missions, however on later missions the LEM commander would also lean forward to try to gain an early view ‘over the nose’ prior to the pitch over manouvre.
Apollo 15 landing sequence at Hadley Rille, from 104:40:06 at 3,000 feet to 104:42:40, just after touchdown. Hill 305 is visible in the background until 2,000 feet and Hadley Rille is visible until 400 feet altitude.