budden@trout.nosc.mil (Rex A. Buddenberg) (01/08/90)
From: budden@trout.nosc.mil (Rex A. Buddenberg)
Several developments in the business of radionavigation should interest
readers in this newsgroup, so what follows is a review including current
events as I understand them. Because of my position, the disclaimer:
I am the planning officer in the Electronic Systems Division in USCG
Headquarters -- that is engineering, not policy. What follows is hopefully
informed, but spiced with some opinions and explanations that are definitely
personal and decidedly not official.
Loran:
The Mid-Continent Expansion project continues. Our stated goal is
to have the FAA-mandated service up by the end of 1990. Unfortunately
we lost a contractor recently for the New Mexico station and on-air service
will probably slip a few months. This project entails building 4 new
transmitting stations, 5 new monitor stations and dual-rating of 5
existing stations. The result will be expansion of the existing
coverage (which provides coastal coverage around North America and
incidentally terrestrial coverage over about half the continental
United States) to complete coverage of the continental US and significan
increases in coverage in Canada, especially over the population
concentrations in the southern tier.
GPS:
Recent policy change by the Joint Program Office will result in a
reconfiguration of the constellation. Currently the satellites are bunched
so you get complete 3-dimensional coverage part of the day and nothing
during other parts. Great for R&D, poor for operational use. The change,
plus a few more satellites launched (need twelve total, if I recall correctly)
will provide 2-d coverage continuously by the end of 1990 (although the
system will not be declared operational for some months after).
Coupled with the advent of reasonably priced receivers, this will make
GPS a viable system for real users fairly soon.
Zap-Sat. One of the defects in GPS that the FAA has been vocal about
for some time is the inability to block a sick satellite's signal quickly.
System integrity problem. JPO has agreed to field a monitor system that
will detect satellite bad data and scramble the 'gold code' which has
the effect of making a user's receiver drop the satellite from his nav
equation entirely. Not sure what the schedule for this is, but the
policy change was announced last month.
Loran/GPS integration. For aircraft use, system integrity is the critical
unfilled requirement -- both Loran and GPS are inherently accurate enough
for airway navigation and non-precision approaches. The problem is that
neither system is robust enough to shield the user from intermittent failures
at least some of the time. This is not a major problem for the mariner as
ships don't go that far in a minute of bad data, but an aircraft on approach
is a different story.
Part of the solution is user notification -- unannounced bad data is
worst of all. The Zap-Sat noted above brings GPS to the place where
Loran has been for years in this area (although we must tighten the
grace time up over the next few years).
The other part is to make Loran and GPS interoperable so that failures
in one system degrade, but do not destroy, the positional integrity.
In order to do this, Loran master stations are being synchronized more closely
to UTC than previously. By changing operating procedures, we feel we have
reliably reduced the variance from 2.5 useconds to 0.5 useconds. Engineering
changes in the time transfer mechansim will be required to reduce this
variance further to the goal of 100 nseconds. This should happen over
the next few years -- active engineering project now.
Availability numbers. Both Loran and GPS provide availabilities in the
neighborhood of 99.99, or 10e2 (the exponent equals the number of 9s after
the decimal point). Aviation folks are saying they need 10e8 [!!] kinds of
figures for the system to work right, safely, and to retire other obsolete
and costly radionav systems (VOR/DME). If the user (aircraft) has both
a Loran and a GPS receiver (no integration), then he boost his availability
figures a couple orders of magnitude -- to around 10e4. With the time
synchronization, then the user can take all of the data from both the Loran
stations (several chains worth -- anything he can track) and all visible
satellites. Throw them all into a Kalman filter and throw away any bad
data (sick satellites or Loran transmitters). This gets one or two more
orders of magnitude -- the analysts that I've talked to indicate 10e6 is
realizable without major changes to either of the systems (expensive).
Differential GPS:
dGPS operates by placing a (slightly specialized) GPS receiver over
a surveyed benchmark. This reference receiver tracks the GPS satellites
and compares where the satellites tell him he is and where he knows he is.
This <differential> is then communicated to a user 'in the neighborhood'
who applies it to what the satellites are telling him.
This technique (which also works for Loran) will correct GPS fixes
to within 10 meters (2dRMS). Current CG research is focussing on how
far away 'in the neighborhood' is with estimates running from 300NM to
3000NM (at which point you effectively fall off the edge of the earth).
While there is no requirement to provide public differential service,
the Coast Guard has in-house requirements (buoy tender positioning) and
we are talking with the Navy's mine warfare folks. (We do, informally,
have a weather eye out expecting that public requirements will eventually
get stated and validated -- so we are considering this eventuality in our
engineering.)
Differential techniques correct for several errors in the GPS signals
including orbit wobbles, clock drift ... and selective availability.
GPS selective availability:
In order to frustrate enemy use of GPS positioning, DoD invented
a concept of selective availability -- the satellite 'lies' about its
time base. This degrades the position to 100 meters unless you are
an 'authorized user' and have the crypto code to correct for the
satellite's lying. if you have this p-code, then you get positioning
accuracy to 18 meters (all these figures are 2 sigma).
Selective availability has aroused a lot of controversy for several
years and it seams to be drawing to a head for several reasons:
- we've demonstrated that we can defeat selective availability
by differential techniques. If a bunch of dumb Coasties can do it,
Ivan can do it.
- who follows the minesweeper (authorized user) through the swept
minefield? The convoy of merchant ships is who -- all unauthorized users.
These guys need the same reference and the same accuracy as the minesweepers
lest they start sweeping mines themselves (any ship can be a minesweeper...
once).
- would any Ivan in his right mind use the yanks' positioning system
to target ICBMs? The soviets we've talked to laugh uproriouly over that one.
There is some interesting history in the Loran world that is instructive.
Initial Loran deployments were for DoD use and the Coast Guard maintained
a reconfiguration schemes for wartime use. Our DoD customers (like FBM
submarines) has Loran tables so they could use the reconfigured chains.
After a few years, it became apparant that our sealift and especially
our airlift folks would also be dependent on Loran, so the base of users
needing the reconfiguration plans expanded by an order of magnitude. Then
the burgeoning commercial, allied and neutral user base started to get
considered. The last Security of Aids to Navigation (SCATANA) plan was
unclassified and said: don't change anything... My opinion is that
GPS selective availability is heading down this same road.
Shipboard developments.
I'll save this discussion, which is an active topic, for another
posting. Since the Coast Guard is both a service provider and a navigation
service consumer, we have a unique position in the business where we can
influence both. Look for a little feedback before I compose this one.
Rex Buddenbergroy@phri.nyu.edu (Roy Smith) (01/10/90)
From: roy@phri.nyu.edu (Roy Smith) In <12886@cbnews.ATT.COM> budden@trout.nosc.mil (Rex A. Buddenberg) writes: > Initial Loran deployments were for DoD use and the Coast Guard maintained > a reconfiguration schemes for wartime use. Do you mean Loran-A or Loran-C? In any case, it's interesting to note that the Russians operate a system very similar to Loran, called Chayka. So similar, in fact, that (according to the June 1989 Ocean Navigator) they are looking into the possibility of setting up a new North Pacific/Bering Sea chain, using an existing US transmitter on the island of Attu near the end of the Aleutian archapeligo and existing Soviet transmistters at Petropavlovsk on the Kamchatka peninsula and at Kurilsk, northeast of Japan. Apparantly the systems are so similar the existing commercial Loran recievers can lock in to and track the Chayka transmissions. -- Roy Smith, Public Health Research Institute 455 First Avenue, New York, NY 10016 roy@alanine.phri.nyu.edu -OR- {att,philabs,cmcl2,rutgers,hombre}!phri!roy "My karma ran over my dogma"