TRAKNET

An AMSAT Mobile Satellite System

Bob Bruninga, WB4APR
115 Old Farm Court
Glen Burnie, MD 21060

August 11, 1998

There are two distinct advantages of the global nature of satellites which cannot be easily met with terrestrial systems: wide-bandwidth point-to-point and mobile applications. With the availability of telephone, cable, and the internet to link hams at fixed sites to each other routinely, we are wasting a lot of potential of our very valuable AMSAT resources by ignoring mobile applications.

Ham radio is on the move. Many hams only have time to play radio while mobile, and similarly, whenever a ham contemplates a long trip, his ham radio is high on the packing list. Although many dream of taking along an HF mobile to play with and to report their progress back home, the $1000 to $2000 investment is just too much of a risk. Two meters is fun, and can bring emergency aid, but it just doesn't provide the nationwide coverage that is needed for the mobile ham traveler far from home, the offshore boater, or first-response teams arriving in a disaster area. In many cases, just a brief position/status report or message is all that is needed to assure the health and welfare of the traveler or to summon assistance or alert other communications channels.

What we need is a new perspective which takes advantage of some very unique capabilities to exploit a small portion of our AMSAT on-orbit capacity to the mobile requirement. This capability exists and the hardware is already in orbit! There are several AMSATs that have very easy uplinks which can be hit from the mobile using only a 2 meter 25 watt FM mobile radio. The ubiquitous radio that everyone has. These satellites are the under-utilized 1200 baud PACSATs.

Uplinks

Let's look at the 1200 baud PACSAT uplinks. These uplinks are unique for several reasons that make them ideal for the mobile environment as follows:

  1. A 2m omni antenna has a 9 dB advantage over a similar 435 MHz omni due to the 3 times larger antenna aperture
  2. There is no tuning or tracking required on the uplink since the Doppler shift on 2m is less than 3 kHz.
  3. Any 25 watt mobile 2m FM rig can be used as the transmitter
  4. Many TAPR TNC-2 compatible TNCs can be modified for the uplink for $3
  5. World Wide Coverage
  6. No software or hardware on-orbit modifications to the satellite!

Reportedly stations running as low as 7 watts into an indoor omni antenna have reported success with the 1200 baud PACSATs. This means that even back-packers with an HT and handheld gain antenna could get emergency or priority traffic into a PACSAT ... from anywhere on Earth! Figure 1 below shows one pass of the APRS/Mir experiment on 11 March 1998 when the Mir digipeater was activated for APRS UI packets. It shows how easily a space-based digipeater can be used for station position/status and brief messaging.

Mirtest.gif (14442 bytes)

Downlink

OK, so the 2m uplink is easy and anyone can do it, so what about the downlink? This is not so easy. The path loss omni-to-omni is 9 dB worse, the satellite is only transmitting a watt or so for another 13 dB worse performance, plus it requires Doppler tuning, a $250 PACSAT modem and a $1000 all-mode UHF receiver! In most cases, all successful PACSAT stations use high gain antennas and automatic tracking to make up for the more than 22 dB performance difference on the downlink. This is not something that most operators will want to add to their mobile. But what if the mobile application did not need to receive data via satellite, but only send it?

TRAKNET

The combination of easy uplinks, minimum downlinks, and an application that often only needs an uplink via satellite, such as the mobile position/status report is the whole idea behind TRAKNET. Only a few automated downlinks are needed every 1000 miles or so to receive the mobile data and to provide it into a nationwide system of linked ground stations. These ground stations relay the mobile position/status reports onto local mobile vehicle tracking channels and onto the internet. Anyone may access the data live on VHF, HF or via the internet.

Since this idea was first proposed in the May/June 1993 and 1995 issues of the AMSAT-NA Journal, the ground based tracking and worldwide distribution mechanism via the Internet has been fully implemented and is in daily use. This system, called APRServe, was pioneered by Steve Dimse, K4HG, to support the growing worldwide Automatic Position Reporting System, APRS. Thus, TRAKNET is not just a future idea, it is online now, awaiting only some more automatic PACSAT downlink stations.

There are two internet sites for APRServe. When you telnet to these sites, you will be connecting to a live feed of all packets captured everywhere!

www.aprs.net port 23 serves all APRS packets worldwide
www.aprs.net port 10001 serves all AMSAT/Mir packets.

Worldwide Message Infrastructure

It is important to note that not only can the internet sites feed data directly to on-line users live, but because of the APRS software used by these users while they are logged on, their stations automatically become gateways to all other VHF or HF users that their station sees. Thus we have worldwide mobile communications that is transparent to the users! Here is how it works:

  1. Every station logged on to the APRServe sites above, feeds everything they hear on their local VHF or HF nets into the APRServe system.
  2. APRServe then removes all dupes and sends it out to all logged on users. Thus everyone sees everyone and everything. The bandwidth of the internet is easily able to keep up with this throughput. Even with over a thousand users live, the total bandwidth is still on the order of about 120 baud average due to the low duty cycle of each station's transmissions in APRS.
  3. Now, if in the APRServe stream, the local station sees a message addressed to another station that is within RF range, then this station will act as an IGATE and will forward that single message out onto the local RF channel.

Thus any station that is near an active IGATE has transparent two-way message communications with any other similar station anywhere in the world where there are other IGates. It is important to note that this process is completely transparent to the callsigns of the IGates, and thus the mobile user does not need any a priori knowledge. In the case of TRAKNET, the PACSAT downlink stations just feed any UI packets from the birds into this existing infrastructure just like any other IGate station, and thus the packets are distributed worldwide.

Currently there are automatic feeds of Mir packets, and maybe some DOVE packets, but we still need some PACSAT stations. N3NRU in Pennsylvania is now on line with an automatic PACSAT downlink feeding APRS mobile position reports to the nearest Internet Gateway where they are being distributed nationwide.  The Naval Academy provides a display of the last 8 passes of Mir downlink packets.

But the problem is that we need more sites listening to the PACSATs, mostly because setting up an automatic PACSAT ground station is not trivial and the guys who play with the Web and APRS all day are not the same guys that are necessarily fully invested in PACSAT hardware. All we need are probably 5 stations scattered over the continental USA to implement a reasonable TRAKNET system.

Mobile Station

A mobile station consists of nothing more than a typical 2 meter FM radio and a modified TAPR TNC-2 compatible TNC and a laptop or GPS. Optional accessories are a GPS for moving position reports, and a laptop for entering messages. Most modern TNCs will accept the GPS data directly and will transmit the data in a timed packet burst. There is even a tiny handheld TNC called the APRS MicEncoder that includes front panel switches for selecting 1 of 7 pre-canned status messages without needing a laptop to change the status report. The modifications to the TAPR TNC-2 are to simply XOR (exclusive OR) the transmit data with its 1200 Hz clock and to filter the result to the microphone input of the radio.

The following circuit will do this with nothing but an 89 cent standard 400 quad 2-input NAND gate connected as an XOR gate or using a single XOR gate out of a 7486 chip. Just connect to the two points shown.

circuit.gif (2597 bytes)

TRAKNET Protocol

The problem with any amateur satellite is the very low bandwidth available compared to the very large worldwide amateur population. At first glance, the prospect of increasing the number of users on a PACSAT channel by a hundred fold raises lots of red flags in the minds of those stations who already find the 10 minutes of a satellite pass to be too short for any meaningful dialogue. But what if each of these hundreds of new users was limited to only a few seconds per orbit? Then as many as 200 stations per footprint could be tracked. That is the only objective of the TRAKNET protocol, to allow everyone to transmit a few single 1 second position/status reports during the closest point of approach over their location. If only one channel is designated for TRAKNET, then the other three channels are free to normal PACSAT use and no amount of congestion on the TRAKNET channel can interfere with existing users on the other three.

APRS/Mir Test

This capability to share a 1200 baud satellite channel was demonstrated very successfully with a 98% success rate for over 100 stations via Mir during the 11 March 1998 test. To keep the display uncluttered, only one pass out of several is shown in Figure 1.

TRAKNET Satellites

There are currently four 1200 baud PACSATs on orbit. One, WO-18 has actively invited UI frame digipeating and leaves digipeat on most of the time. The problem is that the WO-18 downlink is difficult to receive unattended due to a spur tone in the middle of the data which makes receiver lock a difficult and manual process. AO-16 has had its digipeater on for the last 6 months and welcomes APRS testing when there are no other users logged on. Blind uplink testing, however, is restricted to 145.94 and only on Tuesdays. Other PACSATs occasionally have digipeat turned on, but there is no formal policy. The purpose of this article is to encourage the designation of one good channel as a gathering point for TRAKNET experiments, and then progress can be made and the potential of TRAKNET can be evaluated. Here are the frequency plans of the existing traknet capable PACSATs:

Digipeater

FM Manchester Uplink Channels

Downlink
AO-16 .900 .920 .940 .960 437.051
LO-19 .840 .860 .880 .900 437.153
WO-18 .900 437.104
IO-26 .875 .900 .925 .950 435.822

Advanced TRAKNET Concepts

Even though TRAKNET is a fully functional system now, using only existing hardware and software, if it became very popular and lives up to its full potential, there are many things that can be done to improve the throughput of the system. Let's define these by phased improvements.

Phase 1

This is the existing system using existing PACSATs (WO-18 and AO-16) and existing flight software. The only requirement is that digipeating be turned on. This phase can be extended by gaining formal permission to use more of the PACSATs with digipeating on, such as LO-19 and IO-26. Also permission to use the single frequency of 145.900 on all satellites would make mobile operations easier.

Modification 1: Digipeat UI Only

This change is a simple modification to the PACSAT digipeater software routine to assure that the digipeater only digipeats UI frames. This prevents it from being abused by connect attempts between ground stations or new users. So far, this has not been a problem, and is not expected to be a problem due to the low efficiency of such use.

Modification 2: TRKNET Alias

This step greatly simplifies mobile scheduling by changing the digipeater alias on all of the TRAKNET PACSATs to the generic call of TRKNET. Thus, any TRAKNET PACSAT will digipeat any mobile position/status report on 145.900. This provides up to 24 different generic TRAKNET passes per day for the mobiles.

Modification 3: Digipeat Always

Since all packets to be digipeated by the PACSAT could all use the identical digipeater callsign of TRKNET, it is actually not needed. The PACSAT digipeater could simply be configured to digipeat any and all UI frames that it hears on the uplink. This saves 7 bytes per packet for a 10 to 20% channel improvement on the congested uplink.

Phase 2: Store and Forward

The straight digipeater mode for TRAKNET had the advantage of already existing with no on-orbit software changes required. But it also requires a ground station in every footprint worldwide and single downlink errors result in lost data. If the flight software were modified, however, to keep a continuous running buffer for all uplinked UI frames, then several advantages would accrue:

  1. Far fewer downlink stations would be needed. Two in North America, and one in Europe, the Middle East, Australia, and Japan.
  2. The repetitious downlink would assure 100% success on the downlink.
  3. Distant, DX and 3rd-world areas would be covered. Uplinks from the most distant regions of the world including the poles would all appear in the downlink.

Since a complete message or position/status report can easily fit in an 80 byte packet, only a 10K buffer would support over 120 mobiles. As usage grew, a worst case buffer size would need to be no bigger than about 32K over a full 20 minute USA pass assuming a 33% channel efficiency on the 1200 baud uplink. Plus, a new compressed APRS position format has been incorporated into APRS that reduces the size of a position report packet to only 22 bytes, thus gaining almost 400% in channel capacity!

Advanced Mobiles

While the preceding was written to emphasize the ease of using the PACSATs by anyone for emergency or priority status/position reporting, there is certainly no reason why a full two way PACSAT communications system cannot be added to most mobiles. Omni PACSAT downlinks are possible and the addition of only a modest gain antenna will certainly help. Advantages are the small size of a 6 dB two element UHF antenna and the short cable run found in a mobile. Rather than a $1000 SSB rig, a $90 QRP HF rig and a UHF downconverter could do just as well.

Conclusion

The advent of the handheld GPS unit for under $150 and availability of cheap, second-hand laptops has brought thousands of mobile amateur radio operators into the world of mobile data. Similarly, the state-of-the-art in automatic PACSAT ground station capability has been improving with many recent software packages to make unattended automatic ground station possible. And lastly, we have had a system of linked ground stations via the internet capability for over two years! The problem is that these two communities of expertise have so far had little cross-interests. It seems that the time is now to merge these technologies into a new amateur application that takes advantage of the unique capabilities of each and fuels the development of an Amateur Radio Mobile Satellite System.


Article by Bob Bruninga, WB4APR, wb4apr@amsat.org.   HTML conversion by KB5MU.