I just wanted to share the way I've been visualising the way this workgroup proceeds with its various tasks. I'm not really sure if this is the most sensible way to go about it or if I've overlooked anything, so I welcome input from those with more experience in the field. Here is how I see things going down:
1) Decide upon a standard coordinate system for our navigation work (or perhaps 2 or 3 and transforms between them). These systems need to be capable of expressing position and attitude. A natural choice for position seems to be spherical coordinates centred on the Earth, aligned with the Earth-moon plane. I'm not so sure about orientation.
2) Decide upon a "flight plan", which is basically just a series of (time, position, attitude) points, spaced sufficiently close together in time (we should be able to think of it as a continuous function in the coordinate space for most intents and purposes), with annotations indicating when burns are happening. Needless to say this series of points follows all relevant physical laws, taking into account our burns and other planned maneuvers. We can begin with "approximately physically correct" flight plans, though, which make various simplifying assumptions (patched conics trajectories, etc.), and incrementally work our way up to a full model with all orbital perturbations, etc.
3) Design a system ("the navigation system") by which our vehicle can at any time estimate its position and orientation (and change rates of these) in our standard coordinate system. This is where all our discussion on inertial navigation, star tracking, etc. comes in.
4) Design a system ("the guidance system") for determining which control signals to send to the RCS given the current time and estimated position and the flight plan, so as to keep the vehicle as on plan as possible. Since the RCS stuff is actually part of the Spacecraft Engineering Workgroup, these control signals will have to be specified by the NGW in fairly idealised terms, i.e. only specifying changes in velocity or orientation.
Is this a sensible big picture? The thing that worries me most is whether or not it actually makes sense to have a single ideal plan and attempt to follow it as closely as possible. If we discover half-way to the moon that we are slightly off course because our trans lunar injection burn was off in some respect, it may actually be more fuel efficient to stay on that "incorrect" course and make our lunar orbit burn a little different to compensate than it is to try to force ourselves back onto the One True Course. Of course, with this approach our guidance system needs to be able to quickly calculate alternative trajectories. That's by no means impossible but it may be more in keeping with the principle of doing "the simplest thing that could possibly work" to go with the One True Course approach.
Assuming this is a sensible big picture, perhaps the biggest open question is "how do we choose the One True Course"? It feels like if computing trajectories from a few parameters (regarding starting position, burn times and lengths, etc) can be done quickly enough we can specify some constraints on flight plans (maximum duration, etc.) and then use fairly standard optimisation techniques (simulated annealing, for example) to explore the parameter space and find the course which minimises total delta-v required subject to our constraints.
Of course, all of the above basically has to happen twice, once for the lunar lander's advance low energy transfer (assuming this idea sticks – so far it is based mainly on "wow, cool!" rather than hard numbers) and once for the CSM's Hohmann transfer.
Thoughts?
Login 
Register 
Search 
Home
Topic RSS 
