Ramifications of a positioning and communications constellation in LMO or MMO

[nadreck]

At the request of baldusi I am creating a thread to discuss the ramifications of a positioning and communications constellation in low or medium Martian orbit.

I feel that the cost to do this would provide a service that can't exist by more ad hoc growth of capability as we have seen to date around Mars with various orbiters/orbital components of landers providing relay. While the cost of this service would be significant, it still would fit within the amount it saves both future robotic and eventual human exploration and exploitation activities.

My thought would be to install a constellation similar to what Teledesic was envisaged as, and given the smaller size of Mars a back of the envelop visuallization suggests 120 satellites in 12 planes at approximately 700km altitude in near polar orbits would ensure 9 satellites being line of sight from any flat ground at the equator at any time (more satellites where their orbits converge the further north or south). So at least six from the side of Olympus Mons at any point in time. If one only had the technology available today that they had when the Teledesic test satellite was built and launched in 1998 one would be looking at a total weight of 72t mass for the whole constellation, however I would argue that technology that has been demonstrated in current operational craft could have at least the same station keeping, communications, and positioning capabilities that was planned for Teledesic at 20% of the weight. By the time we actually build and deploy less by 50%, so lets just say we need to put 12 packages with 10 60kg satellites in TMI and that each package has the equipment to deploy the 10 satellites, the fuel and guidance for the trip and deployment, and Earth Link communications and in fact is a redundant full satellite itself as a spare for its plane plus one of the 12 Earth and Deep space links for the constellation. So these packages deliver 10 60kg satelittes, are one themselves plus the edge connector to Earth, and they do minor course corrections and atmospheric breaking to get to their separate planes, then use their resources to move into each spot in the plane and let a satellite go in it before stationing themselves slightly offset in the plane ready to move to where there has been an outage. Even if I presume that the "package" has a mass of 600kg over and above its payload of 10 satellites we are still looking at under 15t that has to be sent to mars to set up a full mobile data network with positioning.

So over to you baldusi for your input.

[baldusi]

It would be more logical to use something similar to Earth's based GNSS, which would be a 10,000km circular orbits (slightly more than 12hs period). Probably with 60 deg of inclination wrt the Equator on four planes. 20 birds should do it. That's not where there's problem.
The problem is spacecraft maintenance and orbit determination. There's a lot of surveying work don on Earth and from there there's the orbit determination. You have no infrastructure in Mars to set your datum. Much less all the ground support stations to do orbit determination and timing control. So you'd have not much better precision than if you simply let the star trackers do an approximate. And you still have the clocks stability issues. So in the end you'd not get that much better performance, without some massive infrastructure in Mars.

[nadreck]

Teledesic was to be the first commercial positioning system, it didn't happen, but the plan was for something more accurate than the public had access to with the original military network and it was originally planned for 700km Earth orbit, but ended up being reduced in satellite number and planned for 1200km. The design group felt that they could do positioning at that altitude. Given the power advantages for the communications aspects of this system I think 700km would probably be better than the orbits you have suggested, though you could have smaller more powerful satellites at that oribt.

However I agree about your issues on the datum, but there are solutions to that with time and future Martian developments.  Initially however it would be relying on its own data and references possibly to the Phobos and Deimos (or rather relays parked on them).

By the time humans were visiting Mars the data could very well have been refined to the level needed for navigation at surface, in the atmosphere but not for landing, and LMO. Before that there would be some positioning and a raft of valuable communications capability for the unmanned operations.

I don't get the issue of the clock stability, I was under the impression that current technology (Galileo) had accurate enough clocks flying, unlike the original USAF system.

[MP99]

It would be more logical to use something similar to Earth's based GNSS, which would be a 10,000km circular orbits (slightly more than 12hs period). Probably with 60 deg of inclination wrt the Equator on four planes. 20 birds should do it. That's not where there's problem.
The problem is spacecraft maintenance and orbit determination. There's a lot of surveying work don on Earth and from there there's the orbit determination. You have no infrastructure in Mars to set your datum. Much less all the ground support stations to do orbit determination and timing control. So you'd have not much better precision than if you simply let the star trackers do an approximate. And you still have the clocks stability issues. So in the end you'd not get that much better performance, without some massive infrastructure in Mars.

Agree re the datum.

GPS works by triangulating to multiple sats.

I wonder if there could be several ground stations setup, each broadcasting a simple GPS-type signal, and positioned so that at some point in their orbits each sat could see four stations at once?

The ground stations, of course, would just need to broadcast their fixed location, and the sats integrate that over the duration of their pass. Thinking about it, if the sats have GPS receivers for the ground stations, they could also receive fixes from other sats, especially those which have just refined their location from the ground stations.

Also, there is already a huge amount of mapping which has been done for the Martian surface. By the time of any missions it should be quite feasible to have these loaded into a rover navigation unit. The rover could have LIDAR and camera systems (matured by the huge investments in autonomous automobiles), which will sometimes be able to determine a really accurate location, starting from the inertial location. Just think of those boulder fields. If the inertial system can provide location even within a few hundred metres, that's a very small amount of mapping / picture information which would need to be interpreted to make an accurate fix.

The rover unit could then actually broadcast its fix *to* any GPS sats that happen to be in view. It could also provide a base station for one element of the fix once the Marsonaut leaves the rover (as long as it remains in range).

I'm assuming, of course, that signals in both directions will include statistics re the uncertainty of the position being provided.

Atmospheric effects will also be much less, I presume?



I also question whether so many sats are really required. Early Polaris subs used to take a GPS fix from a single satellite. They would take continuous fixes from the sat over its entire traverse of the sky, while integrating this with very accurate inertial navigation (and probably an internal clock).

One sat provides similar accuracy to a whole constellation of sats that just happen to follow each other in the same orbit. Of course, this will not give the same accuracy over all axes - that needs sats in different planes.

It's only because of the low sophistication of Earth receivers that they need to receive fixes from multiple sats at the same time to provide a fix. I really don't see an issue with including this level of sophistication in a Mars GPS receiver.

Perhaps the "consumer units" may even use primarily inertial navigation, with the system refining the accuracy of its position whenever a sat happens to be in the sky or it achieves a LIDAR/visual fix. I would imagine the accuracy ellipse from could be some way from circular until another sat comes across from a different direction and improves accuracy in that plane.

This ignores communication requirements, but that only needs a single sat to be visible.

Cheers, Martin

[MickQ]

And you always have Phobos above the horizon for 4 or so hours twice a day.

Mick.

[baldusi]

I don't get the issue of the clock stability, I was under the impression that current technology (Galileo) had accurate enough clocks flying, unlike the original USAF system.

The clock are controlled and tested constantly by the ground stations and checked against ground clocks. Here you are 20 light minutes away from the satellites, with barely enough power to reach 1500km.

[sdsds]

One sat provides similar accuracy to a whole constellation of sats that just happen to follow each other in the same orbit. Of course, this will not give the same accuracy over all axes - that needs sats in different planes.

It's only because of the low sophistication of Earth receivers that they need to receive fixes from multiple sats at the same time to provide a fix. I really don't see an issue with including this level of sophistication in a Mars GPS receiver.

This is a key observation. It draws into question the initial premise that location determination is, "A service that can't exist by more ad hoc growth of capability."

[nadreck]

One sat provides similar accuracy to a whole constellation of sats that just happen to follow each other in the same orbit. Of course, this will not give the same accuracy over all axes - that needs sats in different planes.

It's only because of the low sophistication of Earth receivers that they need to receive fixes from multiple sats at the same time to provide a fix. I really don't see an issue with including this level of sophistication in a Mars GPS receiver.

This is a key observation. It draws into question the initial premise that location determination is, "A service that can't exist by more ad hoc growth of capability."

Yes but I was referring to a full time standardized satellite mobile communications platform as well which I believe to be as important as the positioning.

I don't get the issue of the clock stability, I was under the impression that current technology (Galileo) had accurate enough clocks flying, unlike the original USAF system.

The clock are controlled and tested constantly by the ground stations and checked against ground clocks. Here you are 20 light minutes away from the satellites, with barely enough power to reach 1500km.

I reiterate that the Galileo system has more accurate clocks, drift of 0.45ns per 12 hours:

http://books.google.ca/books?id=peYFZ69HqEsC&pg=PA4&dq=galileo+positioning+clock+accuracy&hl=en&sa=X&ei=kCQgVMvOKeTCigL2wYHYAg&ved=0CDQQ6AEwAg#v=onepage&q=galileo%20positioning%20clock%20accuracy&f=false

Since these are side talking communications satellites they could either vote time corrections, or correct far less often than the GPS network by communications with a single mars surface or Phobos/Deimos station that has the radio power to interact with the earh based atomic clock stations.

[baldusi]

All problems are solvable. The issue is that's not as easy as sending a bunch of 3U sats and getting a GPS like service. The truth is that you can use radio signals, atomic clocks and star trackers to get a pretty good position... With a few days.
The only actual need for real time GNSS service, is when doing EDL, and for that you'd need the full infrastructure like on Earth. In the end, using a cheap set of ground based beacons would be much more efficient until a full blown GNSS is needed.

[nadreck]

Respectfully, this is the first mention of using it for EDL that I have seen. My discussion has been predicated on initial lower accuracy due to lack of survey data, that will increase as mobile surface units provide data to the overall martian topology model. The purpose of the constellation is to provide navigation positioning on the surface, positioning in space, navigation for powered (or maybe lighter than air) atmospheric flight, along with continuous coverage voice/data communications.

Also respectfully I was suggesting a 60KG satellite model replicated over 12 planes with 10 satellites per plane and 1 600kg satellite per plane as back up as well as relay to earth or other deep space location.  A 3 U satellite has a volume of 3 litres, I am talking more on the order of 10 - 12 litres plus deployed solar panels.

Finally the clock system on Galileo is 10 year old technology. We have not started designing the Mars system yet. Going from what I can find, standard consumer gps products need corrections to withing 20 - 50 nanoseconds for accuracy from 3 - 10M the highest accuracy known comes from devices that get a feed from one of the observatory clocks and get correction down to 0.1 nanoseconds.

The data provided by using GNSS services and feeding back first unmanned rover data, then later manned surface operations would increase the accuracy of the system, provide eventual better mapping data, and feed back to future users.  At the moment most craft sent to Mars need to be able to communicate directly with earth, though some optimize routine communications from the ad hoc set of orbiters I referred to in my first post. As a communications network it could offset significant complexity and mass of communications equipment (and more complicated and weightier positioning systems) on future craft.

[wes_wilson]

I think placing three radio towers on the surface would be much more economical and useful early on.   Do we really need precise positioning over the entire planet?  Or merely in the area of activity?  Three 500 foot towers would have a line of sight triangle of 40 miles per side (800 square miles) and there's no reason you couldn't use these for landing as well. 

A tower based system has the further advantages of being repairable and expandable by early colonists on the ground with fairly basic technologies.  For that matter, you might not even need towers initially.   Its possible you could find three high spots and drive rovers to them to act as towers.  Less range, but also less infrastructure. 

Nifty site to play with lines of sight:
http://www.math.tamu.edu/~dallen/valgebra/demos/lineofsight.html
Radius of mars: 2106 miles

[IslandPlaya]

I think placing three radio towers on the surface would be much more economical and useful early on.   Do we really need precise positioning over the entire planet?  Or merely in the area of activity?  Three 500 foot towers would have a line of sight triangle of 40 miles per side (800 square miles) and there's no reason you couldn't use these for landing as well. 

A tower based system has the further advantages of being repairable and expandable by early colonists on the ground with fairly basic technologies.  For that matter, you might not even need towers initially.   Its possible you could find three high spots and drive rovers to them to act as towers.  Less range, but also less infrastructure. 

Nifty site to play with lines of sight:
http://www.math.tamu.edu/~dallen/valgebra/demos/lineofsight.html
Radius of mars: 2106 miles

Good luck with constructing the towers. On Earth it usually involve helicopters...

[wes_wilson]

I think placing three radio towers on the surface would be much more economical and useful early on.   Do we really need precise positioning over the entire planet?  Or merely in the area of activity?  Three 500 foot towers would have a line of sight triangle of 40 miles per side (800 square miles) and there's no reason you couldn't use these for landing as well. 

A tower based system has the further advantages of being repairable and expandable by early colonists on the ground with fairly basic technologies.  For that matter, you might not even need towers initially.   Its possible you could find three high spots and drive rovers to them to act as towers.  Less range, but also less infrastructure. 

Nifty site to play with lines of sight:
http://www.math.tamu.edu/~dallen/valgebra/demos/lineofsight.html
Radius of mars: 2106 miles

Good luck with constructing the towers. On Earth it usually involve helicopters...

http://www.rohnnet.com/filedownload/downloadfile/fileid/78/filenum/0/src/@random4a7c7dd151123

In the low martian gravity a block and tackle is likely sufficient.  80 foot towers in the link above are about a thousand pounds total. 

[IslandPlaya]

I think placing three radio towers on the surface would be much more economical and useful early on.   Do we really need precise positioning over the entire planet?  Or merely in the area of activity?  Three 500 foot towers would have a line of sight triangle of 40 miles per side (800 square miles) and there's no reason you couldn't use these for landing as well. 

A tower based system has the further advantages of being repairable and expandable by early colonists on the ground with fairly basic technologies.  For that matter, you might not even need towers initially.   Its possible you could find three high spots and drive rovers to them to act as towers.  Less range, but also less infrastructure. 

Nifty site to play with lines of sight:
http://www.math.tamu.edu/~dallen/valgebra/demos/lineofsight.html
Radius of mars: 2106 miles

Good luck with constructing the towers. On Earth it usually involve helicopters...

http://www.rohnnet.com/filedownload/downloadfile/fileid/78/filenum/0/src/@random4a7c7dd151123

In the low martian gravity a block and tackle is likely sufficient.  80 foot towers in the link above are about a thousand pounds total.

Good luck with that too!

[nadreck]

I think placing three radio towers on the surface would be much more economical and useful early on.   Do we really need precise positioning over the entire planet?  Or merely in the area of activity?  Three 500 foot towers would have a line of sight triangle of 40 miles per side (800 square miles) and there's no reason you couldn't use these for landing as well. 

A tower based system has the further advantages of being repairable and expandable by early colonists on the ground with fairly basic technologies.  For that matter, you might not even need towers initially.   Its possible you could find three high spots and drive rovers to them to act as towers.  Less range, but also less infrastructure. 

Nifty site to play with lines of sight:
http://www.math.tamu.edu/~dallen/valgebra/demos/lineofsight.html
Radius of mars: 2106 miles

While I believe there is an advantage well before manned operations happen on Mars, I would point out that the costliest part of the system I propose is getting it to Mars, however I am talking about a system that needs roughly 18t injected into Mars Transfer Orbit.   I do believe that landing the mass required for the antennas and ground stations you are proposing would be at least as much as the constellation but the cost to land mass on Mars would be a lot higher than to put it in orbit around Mars.

One of the disadvantages of a tower based system is potential environmental damage and that seems far more likely than damage to satellites (and the system I propose would have coverage against outages).

[MP99]

One sat provides similar accuracy to a whole constellation of sats that just happen to follow each other in the same orbit. Of course, this will not give the same accuracy over all axes - that needs sats in different planes.

It's only because of the low sophistication of Earth receivers that they need to receive fixes from multiple sats at the same time to provide a fix. I really don't see an issue with including this level of sophistication in a Mars GPS receiver.

This is a key observation. It draws into question the initial premise that location determination is, "A service that can't exist by more ad hoc growth of capability."

Yes but I was referring to a full time standardized satellite mobile communications platform as well which I believe to be as important as the positioning.

I still don't see why this needs more than one sat over the horizon at any one time. (I'm assuming that a small increase in latency is not an issue.)

I don't get the issue of the clock stability, I was under the impression that current technology (Galileo) had accurate enough clocks flying, unlike the original USAF system.

The clock are controlled and tested constantly by the ground stations and checked against ground clocks. Here you are 20 light minutes away from the satellites, with barely enough power to reach 1500km.

I reiterate that the Galileo system has more accurate clocks, drift of 0.45ns per 12 hours:

http://books.google.ca/books?id=peYFZ69HqEsC&pg=PA4&dq=galileo+positioning+clock+accuracy&hl=en&sa=X&ei=kCQgVMvOKeTCigL2wYHYAg&ved=0CDQQ6AEwAg#v=onepage&q=galileo%20positioning%20clock%20accuracy&f=false

Since these are side talking communications satellites they could either vote time corrections, or correct far less often than the GPS network by communications with a single mars surface or Phobos/Deimos station that has the radio power to interact with the earh based atomic clock stations.

Yeah, I was thinking the same thing. If the sats talk to each other and the ground stations, then they can vote a concensus time.

Two problems, though.

1) ISTM that the sats would need to derive their own orbital elements in order to be able to project their orbit. ISTM that projecting the orbit may well need a good absolute time?

2) The user on the ground needs to know absolute position of the sats and time-of-flight.

Time-of-flight doesn't need an absolute time if you're receiving from multiple sats, but it does for a very sparse constellation.

At least, it does on Earth where the receiver is only a receiver. However, if the receiver can ping the sat and time the round trip, then it can also synchronise its clock with "constellation time".

Cheers, Martin

[nadreck]

One sat provides similar accuracy to a whole constellation of sats that just happen to follow each other in the same orbit. Of course, this will not give the same accuracy over all axes - that needs sats in different planes.

It's only because of the low sophistication of Earth receivers that they need to receive fixes from multiple sats at the same time to provide a fix. I really don't see an issue with including this level of sophistication in a Mars GPS receiver.

This is a key observation. It draws into question the initial premise that location determination is, "A service that can't exist by more ad hoc growth of capability."

Yes but I was referring to a full time standardized satellite mobile communications platform as well which I believe to be as important as the positioning.

I still don't see why this needs more than one sat over the horizon at any one time. (I'm assuming that a small increase in latency is not an issue.)

I envisage the system for positioning and two way communications. So yes I envisage a sat-phone positioning system on a chipset like any modern smartphone (but in many applications without the display). To make this reliable, efficient and a true advantage to direct to earth communications I want the satellites low so low power devices and minimal antennas work. Then there is the matter of redundancy. I want a redundant system that is fault tolerant, hoping for useful system life going from the era of unmanned probes to exploitive(resource exploitive, not the natives ;-) ) colonization.  So I see low orbit being key, no issues of hand off as cells change from one satellite to the next suggests that 4 or 5 above the  horizon at any one time seems reasonable. That communication requirement dovetails well with the positioning application to my mind.

I don't get the issue of the clock stability, I was under the impression that current technology (Galileo) had accurate enough clocks flying, unlike the original USAF system.

The clock are controlled and tested constantly by the ground stations and checked against ground clocks. Here you are 20 light minutes away from the satellites, with barely enough power to reach 1500km.

I reiterate that the Galileo system has more accurate clocks, drift of 0.45ns per 12 hours:

http://books.google.ca/books?id=peYFZ69HqEsC&pg=PA4&dq=galileo+positioning+clock+accuracy&hl=en&sa=X&ei=kCQgVMvOKeTCigL2wYHYAg&ved=0CDQQ6AEwAg#v=onepage&q=galileo%20positioning%20clock%20accuracy&f=false

Since these are side talking communications satellites they could either vote time corrections, or correct far less often than the GPS network by communications with a single mars surface or Phobos/Deimos station that has the radio power to interact with the earh based atomic clock stations.

Yeah, I was thinking the same thing. If the sats talk to each other and the ground stations, then they can vote a concensus time.

Two problems, though.

1) ISTM that the sats would need to derive their own orbital elements in order to be able to project their orbit. ISTM that projecting the orbit may well need a good absolute time?

2) The user on the ground needs to know absolute position of the sats and time-of-flight.

Time-of-flight doesn't need an absolute time if you're receiving from multiple sats, but it does for a very sparse constellation.

At least, it does on Earth where the receiver is only a receiver. However, if the receiver can ping the sat and time the round trip, then it can also synchronise its clock with "constellation time".

Cheers, Martin

As I stated above I think the communications aspect of this system outweighs the positioning only. I can't think of an application where using a non transmitting receiver makes significant sense over a two way one (marginal weight savings, slightly more substantial power savings) and still in any case I can't think of wanting positioning from something that you didn't want to communicate.

The edge connector (and spare) from each plane may well have to have the equipment to, eventually with the help of a station (unmanned for this function of course) on Phobos or Deimos, get far more accurate fixes on its position and from there the rest of the constellation will be able to calculate accurate position data.