                        APRS/MIR Test 11 March 1998
                           Bob Bruninga, WB4APR
                         

On 11 March 1998 a special APRS/MIR test was conducted via the packet
system on the Space Station MIR.  The test was to show possible methods 
for improving the visibility of MIREX communications to students and 
schools.  The objectives were to:

   1) Demonstrate the use of a one-to-all protocol using UI frames 
   for more effective use of the MIR downlink to multiple 
   participating ground stations.

   2) Demonstrate the usefullness of spacecraft beaconing their
   GPS position a few times per footprint to identify their position.

   3) Use APRS map displays to display the moving position of the
   spacecraft and to display the locations of all participating
   ground stations.

   4) Demostrate the use of linked internet groundstations to 
   distribute the live downlink collected from ground stations all 
   over the US to users worldwide in real time.

BACKGROUND:  The Automatic Packet/Position Reporting System (APRS) is a 
connectionless protocol using UI frames to communicate information 
effeciently among a large group of stations in real time.  Each station
with information transmits, and all stations capture, sort and accumulate
the information on a variety of display pages or windows as follows:

   BULLETINS  Lists bulletins in sequence regardless of receipt order
   MESSAGES   Displays messages to and from your station
   TRAFFIC    Displays messages between other stations
   STATUS     Displays the current status of all stations
   POSITIONS  Displays the current position of all stations
   TELEMETRY  Displays any telelmetry packets
   HEARD LOG  Shows number of packets per station per hour, etc

Although most people associate APRS with GPS position reporting, a GPS
is not required unless a station is moving.  Moving stations
use a GPS to automatically update their position, instead of
requiring manual entry.   Otherwise, the primary advantage of APRS
is sharing information between the largest possible number of users
effeciently over a single shared channel.

MIR BULLETINS:  The MIR packet radio link is ideal for a one-to-all
protocol for distributing BULLETINS, ANNOUNCEMENTS and other information
to all users in the footprint.  Although MIR currently transmits a
few single line bulletins, these are sometimes lost among the 
hundreds of other ACKS, REJECTS, and BUSY packets in the downlink
that sometime overshadow any transfer of information.  By using a 
protocol like APRS to receive the MIR downlink, monitoring stations
can capture such properly formatted BULLETINS and ANNOUNCEMENTS of 
interest to everyone and display them in a user friendly manner.

MIR POSITION TRACKING:  Since APRS has a full maping capability, it can
not only display the location of all ground stations, it can also 
plot the moving position of a spacecraft with an onboard GPS.   But 
since a hardware upgrade to add a GPS to Mir is not likely, the
same effect can be accomplished by a few strategicly placed MIR
tracking stations that transmit up a pseudo-GPS position to be
digipeated by MIR to all ground stations.  Three such stations could
provide this service whenever MIR is over the USA.  These stations
transmit only a single 1/2 sec packet every minute (1% of channel
capacity) and the result is an accurately moving MIR ICON on all
ground station maps.

PARTICIPATING STATION DISPLAY:  Since the MIR PBBS is a single user
system, it can usually only be accessed by one or two users per 
footprint under good conditions.  Meanwhile, within the same footprint
are often dozens if not hundreds of users sending useless and 
competing CONNECT-REQUEST packets which only reduce channel effeciency.
The constant stream of DISCONNECT-BUSY packets in the downlink
conveys the number of users participating in the pass, but carries
no useful information.  If instead of useless connect requests, these
same stations simply transmitted a single compressed position report
it would not only show who is participating in the pass, but also
show the distribution of these stations on the APRS map displays at
no additional loading on the channel.  The flavor conveyed to the
users in this case is a feeling of camaraderie as a participant with
others in a pass instead of individuals fiercely competing with each
other for the single user access.  This in itself is a better 
attitude to convey to students and the map display is certainly more
visually appealing than a scrolling display of disjoint packets.


INTERNET INTERLINKED GROUNDSTATIONS:  With the worldwide connectivity
of the internet, the downlink packets from MIR can be received by
groundstations anywhere and made available to all users everywhere.
For over a year, the APRServe system has been providing that kind
of connectivity to APRS packets genearated nationwide.  During
the APRS/MIR test it allowed stations even out of the footprint to
observe the event.  Similarly for schools, it gives easy access 
even to individuals and classes with no amateur radio equipment.

THE APRS/MIR TEST:  Since the Mir packet system has been operating well
the last few weeks, Dr. Larson of the MIREX group suggested the APRS
Mir test be conducted as soon as possilbe since precession was taking 
Mir passes earlier every day and it would soon be out of view during
school hours.  He authorized the APRS/MIR test on only two orbits on 
the 10th of March.  Unfortunately the MIR packet system went off the 
air on these two orbits, so the test was extended to the next few 
orbits over the USA.  The test was limited to the USA only because 
it had the largest numbers of existing APRS ground stations ready to 
test in sufficient numbers to fully load the system.  But the delay 
further complicated matters since the next orbits over the USA were 
between 0230 AM and 0400 AM local time.  As a result, the test was 
extended for a full 5 orbits to allow testers to choose a pass and 
still get some sleep.

With the short notice of the test, and since the more effecient SPACE 
mode of APRS had not been used or tested since the STS-74 and 78 
SAREX and SPRE Missions over two years ago, APRS users were told to 
just operate normally on the MIR frequency but to shorten their 
position comments where possible.  The SPACE mode compresses the
station's position report into his gridsquare to save 16 bytes per
packet.  Due to the typical 10 minute duration of a pass, they were 
told to transmit their position reports once every 3 to 4 minutes 
to get a reasonable probability of success per orbit.

To make Mir appear to move on all groundstation maps, three special 
tracking-uplink stations beaconed the moving position of MIR via the 
MIR digipeater.  One from California using the callsign MIR-6, one 
from Michigan using MIR-8, and one in Maryland using MIR-3 to match 
their callsign areas.  West coast stations would see the moving 
MIR-6, midwest stations would see the incoming MIR-6 change to 
a MIR-8 and then east coast stations would see the moving ICON on their 
maps change to a MIR-3.

Although MIR was not programmed to transmit any special Bulletins,
several ground stations transmitted such BULLETIN, STATUS and 
MESSAGE packets.  Over 65 separate such packets were successfuly captured
during the test.  On average these packets were transmitted  
successfully twice (110 times).  The advantage of using the few 
specially authorized uplink sites to digipeat bulletins instead of 
MIR means that the bulletins may be updated instantly on the ground 
and then digipeated to ALL users at once without requiring an upload 
to the BBS nor hundreds of individual downloads.  Only bulletins 
originated by the MIR crew would need to be entered by the crew.

Each of the participating stations transmitted their brief position
packet once every 3 to 4 minutes.  These packets were far less frequent
than the usual once-every-10-seconds of all of the usual CONNECT-
REQUEST packets normally which congest the uplink.  Each time one of
these packets was successfully digipeated by MIR, all user map displays 
in the footprint would be updated with colorful Icons showing all
stations locations.  During the test 202 station position reports were
displayed averaging about 40 per pass.

To inject the downlink from Mir into the Internet, a few of the normal
APRS I-Gates tuned their radios from the normal APRS frequency to the 
Mir downlink frequency.  These Mir packets were intermingled with the 
normal stream of APRS packets into the APRServe Internet
system.  Although they would be seen on the main www.aprs.net maps
they would be hard to distinguish from the usual 1000 to 1200 or more
APRS stations on the air.  To provide a unique display of the APRS/Mir
packets alone, a special WEB page was designated to filter out only 
the APRS/MIR packets and display them spearately to users.  During 
the day of the event there were over 11,000 hits on the server system 
representing a peak load of 150 simultaneous users and as many as 
1000 users on the special MIREX page.

STATISTICS:  In raw numbers, the following table compares the APRS
packets during the test to the other packets observed on the downlink.
These statistics were mostly derived from observations on the 
east coast (Maryland) plus the APRS packets logged elsewhere.  

PASS  TOTAL   PBBS  PBBS  BUSY  R0MIR  R0MIR  APRS  APRS
      PKTS    PKTS  USERS REJ   PKTS   USERS  PKTS  USERS
----- ------- ----- ----- ----  -----  -----  ----  -----
0740z  229      88    3   18    30     1       89   44
0910z  153      46    3    7    14     0      102   41
1050z  122      47    3    5    11     0       75   51
1220z  170      96    3   15    35     4       72   39
1400z  188      94    5   12    41     3       53   27

Of the 202 APRS stations, 6 were associated with schools and 5 were 
mobiles inclusing one Naval Academy boat, and one railroad car.
Individual APRS packets were also reported by one station
in each of Tiawan, South Africa, Australia, Japan and Hawaii.  
Although APRS  users were encouraged to only send their position 
report until the total load on the channel could be assessed, there 
were still 65 Bulletins and Messages that were successfully relayed.
Since the test was conducted over multiple passes, the 202 successful
stations actually represent 104 different stations.

SUCCESS RATE:  Of these 104 different stations, all were running 10 
watts or more, but two were successful at 5 and 7 watts.  Of the
ten stations reporting a lack of success, 3 were running only 
1 watt, and two were running 4 and 9 watts.  One was only transmitting
once an hour, and one was transmitting once every 5 minutes.  Another 
was using an inside antenna.  Two were running 10 watts and 50 watts
but beacon rate was unreported.  If you consider a 10 watt baseline
and proper setup with a 3 to 4 miunte rate, then the success rate 
appears to be near 98%.  Although there may be many more stations that 
did not report their lack of success, these numbers clearly show that 
the channel was not saturated nor congestion limited.  In fact, during 
this test a nominal 3 Mir PBBS users logged onto the BBS per pass, but 
typically only one was successful at uploading and downloading any 
traffic.  Actually this is rather typical on most passes during user 
hours, but in the middle of the night, as in this test, these BBS 
users would have expected a less congested channel.


LESSONS LEARNED:

1)  Due to the short lead time to the test, there was insufficent
time to educate all users to use the short Gridsquare or compressed
SPACE mode, and to remove unnecessary verbage from their position
reports.  For this reason there were many quite verbose packets.
A 30% improvement could be expected here.

2)  Similarly, the SPACE mode was not used.  This further reduces 
bandwidth by automaticlly canceling all further transmission as 
soon as one's own packet is successfuly digipeated.  With the 3 to 4
minute cycle time used and without the automatic cancelation on
success, there were probably double the number of packets
transmitted than were actually required.  

3)  Non APRS stations were recommended to use the very efficient
grid square method of reporting position by placing their gridsquare
in the UNPROTO TOCALL or their packets.  Unfortunately, this does
not work on Kantronics TNC's which always send their BText UI frames 
to the callsign of BEACON.  THus their packets conveyed no position 
information.

4)  For the pseudo moving MIR-6, MIR-8, and MIR-3 uplinked posits, these
should have all used the same ICON name of just MIR.  Although the 
numbered MIR Icons made it possible to easily see which uplink station
was responsible for the ICON, the changing calls meant that as the
MIR-6 and Mir-8 uplink stations passed out of range, these posits
were stuck on the map at their last uplinked position and stopped
moving.  By using the same Icon name of Mir, then the single Icon
would have continued to move as long as there was at least one
uplink station in range.

6)  The uplinked MIR positions were more static than dynamic since 
only about 20 were recorded on all 5 passes.  At one per minute,
this shows only a 20 % success rate for the possible 100 pass minutes.
Higher power or tracking antennas may be required.

CONCLUSIONS:   The test was successful in meeting all of the original 
objectives.  The short notice and early morning hours helped to reduce 
the number of participants to about 104 stations.  We think this number 
is representative of the nominal number of schools that could be 
authorized to simultaneously participate in future such Mir experiments.  
A total of 12 stations associated with schools and students were 
reported.  One station even displayed 5 APRS stations while operating
with a whip antenna inside a motel room on travel.

     The test demonstrated the value of using a UI frame one-to-all 
packet protocol to improve the delivery of information to all ground 
stations.  Further, the test demonstrated the value of a few special 
MIREX ground stations to uplink the moving Mir position reports and 
to relay real-time MIREX bulletins and announcements that can be 
received by all stations in the footprint including receive-only 
school stations.  Finally, the test showed the value of multiply 
internet connected ground stations for not only providing a continuum 
of data from the downlink across the whole country, but also for 
providing WEB access to students and schools outside of the footprint 
or without amateur radio equipment.  You may see a replay of the event 
by using APRS to replay the file MIRTEST.HST or by visiting 
www.aprs.net/mirex.html.

All of the APRS stations want to thank the MIREX team and also those 
normal Mir BBS users who were inconvenienced by this test, for this 
opportunity to conduct this important experiment. 
