Space exploration from a shirtsleeve environment

Repeated navigation along a path without GPS is soon to be a COTS
capability for autonomous robots. See http://seegrid.com/
Many other capabilities that now require teleoperation will be
autonomous in 2018. ( H.P. Moravec, personal communication)

In article . com, Lee
Valentine wrote:

Repeated navigation along a path without GPS is soon to be a COTS
capability for autonomous robots. See http://seegrid.com/
Many other capabilities that now require teleoperation will be
autonomous in 2018. ( H.P. Moravec, personal communication)

In a sufficiently static environment. Of course, using GPS, or laying
down artificial landmarks (e.g. glyphing the floor), can often make
things a lot easier. If there is an easy way to do easy things, that
is often better than the hard way of hanging 8 cameras on a pole and
getting fragile and balky algorithms working properly sometimes. If
one method needs a Moravec and the other can use a Garmin, go for the
one that's available from the factory in million-lots.

What's really exciting is the Grand Challenge race in a few days, where
vehicles have to navigate previously-unseen terrain at high speeds.
Even using GPS it's not easy (which is what makes it a Challenge).
Much more impressive than rolling over the same path down the same
aisles over and over and over.

--
David M. Palmer (formerly @clark.net, @ematic.com)

Lee Valentine wrote:
Repeated navigation along a path without GPS is soon to be a COTS
capability for autonomous robots. See http://seegrid.com/
Many other capabilities that now require teleoperation will be
autonomous in 2018. ( H.P. Moravec, personal communication)

There might be four levels of operation
1. An astronaut has to EVA, and does the fixing / removing
2. A robot, teleoperated from the Hab Module, does the fixing /
removing
3. a robot, teleoperated from Earth, does the fixing / removing
4. A robot, controlled by computer does the fixing / removing

One would aim to do most operations according to the cheapest method.
(Can hitting a golf ball be done by autonomous robot? -

GPS (or LPS) would be an expensive investment for the moon, and
probably not necessary. It should work better than it does on the
Earth.

Most applications currently requiring repeated path navigation need to
do so indoors, where GPS is unavailable or insufficiently accurate. For
an outdoor operation, it would be foolish to rely on something that can
be jammed as easily as GPS.

Alex Terrell wrote:

Lee Valentine wrote:
Repeated navigation along a path without GPS is soon to be a COTS
capability for autonomous robots. See http://seegrid.com/
Many other capabilities that now require teleoperation will be
autonomous in 2018. ( H.P. Moravec, personal communication)

There might be four levels of operation
1. An astronaut has to EVA, and does the fixing / removing
2. A robot, teleoperated from the Hab Module, does the fixing /
removing
3. a robot, teleoperated from Earth, does the fixing / removing
4. A robot, controlled by computer does the fixing / removing

One would aim to do most operations according to the cheapest method.
(Can hitting a golf ball be done by autonomous robot? -

GPS (or LPS) would be an expensive investment for the moon, and
probably not necessary. It should work better than it does on the
Earth.


The existing GPS satellites are visible from the moon when the earth is
visible. Of course the geometry is atrocious, but a high gain antenna will
pick up signals as well as the omni-directional ones used on earth, and
there are no atmospheric effects, so it should have some use.
There is also a big black sky full of stars, so a set of automatic star
trackers could provide a good navigation system.


Most applications currently requiring repeated path navigation need to
do so indoors, where GPS is unavailable or insufficiently accurate. For
an outdoor operation, it would be foolish to rely on something that can
be jammed as easily as GPS.




--

John Halpenny

If you are what you eat...
I'd rather be a pig than a vegatable.

John Halpenny wrote:

The existing GPS satellites are visible from the moon when the earth is
visible. Of course the geometry is atrocious, but a high gain antenna will
pick up signals as well as the omni-directional ones used on earth, and
there are no atmospheric effects, so it should have some use.
There is also a big black sky full of stars, so a set of automatic star
trackers could provide a good navigation system.


Unless you're moving fast or require higher accuracy than seems
likely with all the reference satellitesin one dircetion, the star-tracker
application is nearly a solved problem. A few years ago, some of the
US Naval Observatory folks mentioned working on a black-box solution
for units in the field, which sounded as if it were to be anchored to
the ground and left alone to find enough stars for a quality location
solution (this may be in case of GPS jamming or some less obvious problem).
Change the rotational pole's celestial coordinates and it
should work really really well on the Moon.

Bill Keel

"John Halpenny" wrote in message
...

The existing GPS satellites are visible from the moon when the earth is
visible.

When (as opposed to where) is Earth not visible from the Moon?


John Halpenny

If you are what you eat...
I'd rather be a pig than a vegatable.

In article ,
John Halpenny wrote:
The existing GPS satellites are visible from the moon when the earth is
visible...

However, their *signals* are audible only when they are on the other side
of the Earth and not very far away from it in the sky. The GPS birds beam
their signals at Earth; they don't broadcast over a full sphere. The
beams are not overly tight, but even so, the Moon is outside most of them
at any given time.

GPS is of some limited use in GSO, especially if what you want is
after-the-fact orbit reconstruction rather than real-time orbit
determination, but even there the coverage is poor. And the Moon is
ten times farther away. If you want good GPS coverage on the Moon, it's
going to require some lunar-orbit satellites.

There is also a big black sky full of stars, so a set of automatic star
trackers could provide a good navigation system.

That gives you attitude fixes, but it doesn't determine location on the
Moon unless you either have some closer reference (like Earth) as well,
or can measure the direction of the local vertical accurately.
--
spsystems.net is temporarily off the air; | Henry Spencer
mail to henry at zoo.utoronto.ca instead. |

(Henry Spencer) wrote:

John Halpenny wrote:
There is also a big black sky full of stars, so a set of automatic star
trackers could provide a good navigation system.

That gives you attitude fixes, but it doesn't determine location on the
Moon unless you either have some closer reference (like Earth) as well,
or can measure the direction of the local vertical accurately.

Local vertical is easy: it's *down*.

Well-calibrated accelerometers are one option, but even a reasonably
simple pendulum will tell you which way gravity is pulling. A pendulum
with a lit point on it can even use the star tracker to read its
orientation.

"John Halpenny" wrote in message
...
Alex Terrell wrote:

Lee Valentine wrote:
Repeated navigation along a path without GPS is soon to be a COTS
capability for autonomous robots. See http://seegrid.com/
Many other capabilities that now require teleoperation will be
autonomous in 2018. ( H.P. Moravec, personal communication)

There might be four levels of operation
1. An astronaut has to EVA, and does the fixing / removing
2. A robot, teleoperated from the Hab Module, does the fixing /
removing
3. a robot, teleoperated from Earth, does the fixing / removing
4. A robot, controlled by computer does the fixing / removing

One would aim to do most operations according to the cheapest method.
(Can hitting a golf ball be done by autonomous robot? -

GPS (or LPS) would be an expensive investment for the moon, and
probably not necessary. It should work better than it does on the
Earth.


The existing GPS satellites are visible from the moon when the earth is
visible. Of course the geometry is atrocious, but a high gain antenna will
pick up signals as well as the omni-directional ones used on earth, and
there are no atmospheric effects, so it should have some use.
There is also a big black sky full of stars, so a set of automatic star
trackers could provide a good navigation system.


Most applications currently requiring repeated path navigation need to
do so indoors, where GPS is unavailable or insufficiently accurate. For
an outdoor operation, it would be foolish to rely on something that can
be jammed as easily as GPS.

For some applications, the distances traveled will be fairly short, say less
than 10 kilometers. There is the option of putting an antennae on top of a
mountain. This becomes not just your GPS system, but your high bandwidth
communication system.

In the early days of mining and manufacturing, there won't be very many
products, so they may try to consolidate their work into a fairly small
area.

In article ,
Alan Anderson wrote:
That gives you attitude fixes, but it doesn't determine location on the
Moon unless you either have some closer reference (like Earth) as well,
or can measure the direction of the local vertical accurately.

Local vertical is easy: it's *down*.

It's easy only if your sensor is absolutely motionless during the
measurement period. This is not that easy to achieve aboard vehicles,
which are one of the major applications of such navigation devices.

Also, when you start wanting to get seriously accurate, you need a gravity
model of the Moon to determine the relationship between (so to speak)
local gravity vertical and local geometric vertical.
--
spsystems.net is temporarily off the air; | Henry Spencer
mail to henry at zoo.utoronto.ca instead. |