Main

AAUSAT 3 Home

(:includeurl http://mcc.space.aau.dk height=110 border=0:)

AAUSAT3 is our third CubeSat in orbit

News is further down on this page

  • AAUSAT3 was launched 25. February 2013 13:31CET and is in full operation.
  • It is in orbit 781 km above surface in a polar orbit. We are here just now
  • Our main payload is an SDR AIS receiver and we see in average 7000 AIS packages/hour
    • Up to +70 packages from one ship in one pass
  • Come and join us with our next cubesat (AAUSAT4) if you are looking for a fun and relevant education


Page TOC

Scanning...

Log for the mission

20140131 Still going strong

The other day we had AIS2 turned on for about 5 hours. This resulted in +40000 new ais msg. to download. We have now downloaded the first 13500. See the preview here:

Attach:uploads/20140131-13k5Preview.png | AIS2 13500 downloaded

20140118 Beacon turned off

Due to low power the beacon was turned off for now. We hope to turn it on again in one or two weeks. Also, after about 25 days uptime on the EPS, a reboot was needed - the EPS (PCU) did in error reboot the COM subsystem more than 500 times. However the system recovered correctly after reboot of EPS (RAM SEU? - single event upset).

AAUSAT3 in eclipse

AAUSAT3 was launched into a sun-synchronous dusk/dawn orbit which provides constant sun light. However, at the end of Nov. AAUSAT3 began to experience eclipse in its orbit. A forecast for the next 60 days can be seen in the figure below.

Attach:uploads/illum_dec_2013.png | 60-day eclipse forecast from the Illum software tool created by Mike Rupprecht (DK3WN)


The periods with eclipse is largest around winter solstice at Dec. 23rd. The eclipse period seems to be over by Feb. These extended periods of eclipse do mean that we need to be very careful with the power consumption on the space craft. We hope to be able to run more experiments once the eclipse period is over.

Morse Beacon Memory Corruption

Our friend from Japan, Toshio Kasei (JA1GDE), reported yesterday that the battery voltage in our morse beacon was suddenty prefixed with the '(KA)' prosign instead of 'B'. Instead of "OZ3CUB Bx.x Txx" the satellite transmitted "OZ3CUB (KA)x.x Txx". The radio stores the morse string as a regular ASCII string in RAM, with the '(KA)' and '(AR)' prosigns stored as the STX and ETX character codes. The difference between 'B' (0x42) and STX (0x02) is only a single bit, so it is plausible that bit 7 in the 'B' character was flipped due to radiation.

We rebooted the radio, which reinitialized the morse string from flash and this fixed the problem. However, this really makes you wonder how many bit flips like this go unnoticed :)

9 9 9 9 9

Today we passed 9 month 9 days 9 minutes and 9 seconds in space. We are still in full operation with the constraint that we have to be careful with power usage on board AAUSAT3.

Lately we have had a transient memory fail on our Linux GND computer in Greenland. The team has identified location of the specific memory addresses that fails (and done it from Aalborg on a living system) and have mapped the location out of usage so we can continue operation.

20131122 GOMX-1 up in space

Yesterday Gomspace located nearby had their first cubesat in orbit.

We were happy to help with initial tracking:

It provided some challenges in terms of finding the satellite. Even last night, about 12 hours after launch, NORAD had designated only about 6 objects from the launch , with a distance of several thousand kilometers apart . DK3WN also came with an offer for a number of TLEs ( http://www.dk3wn.info/p/?p=38644 ), so we overall had TLEs were so scattered that the first was out of range, before the last came .

We chose to place our AAUSAT3 earth station pointing toward the horizon in one of the good passes last night, in the direction of the objects showed up ( objects spreads only in the direction of flight , they are still in a row, so we knew ( azimuth ) where the would show up ) . So when we saw the first beacon from GOMX -1 , we did chose one of the TLEs had just emerged from the horizon and began to track it. Then we sat down to record the frequency and time spent on each of the received beacons using a spectrum analyzer ( they send a periodic beacon every 10 seconds , and the small range ( typically of 20-60 sec at cubesats ) was a great help here . From this information , we could then calculate from there that we were about 80 seconds after the last detected object. The information could be used for a better bearing on the next pass, where they really started getting telemetry down.

20131114 BEESAT3 and SOMP on way home

The BEESAT and SOMP people has been working hard the whole week but no real contact was ever obtained.

We say thanks for the visit - always nice to have nice visitors.

20131112 BEESAT3 and SOMP

Two guys from the German teams for BEESAT3 and SOMP arrived in Aalborg on Monday. They are going to borrow our antenna systems and 1 kW UHF Beko for a few days to see if they can obtain contact. Monday was used for setting up and testing. Tuesday morning at 750 the first passes will be conducted. Always nice to be able to help people.

20131025 Another Linux program in space

The AAU student space team is currently on a tight schedule integrating and testing AAUSAT4. However a new 100 KiB ADCS logger program has been uploaded to our Linux SDR platform. The file was uploaded during four passes, due to the fact that one ground station is off (oops) and the other has reduced performance. The program has stored 1.9 MiB of sensor data from ADCS2 (fine sunsensors, gyros, magnetometers) and from our EPS (power generation, MPPT sweeps etc.). The data was compressed to < 1 MiB, and once downloaded it will hopefully help us produce more more power.

This experiment shows the true value of a linux based system on a distributed network-based cubesat - our software defined radio is now also a easy programmable OBC with in-space debugging functionality.

Stay tuned for a data preview.

20131005 A good day in space

During 2013-09-22 we tested the latest version of our AIS receiver algorithm for 5 hours. A plot can be seen below. More than 52k ships were received correctly and subsequently downloaded via our two ground stations to the control center in Aalborg. Attach:upload/20130922-53k_in_5h.png | More than 52 000 ship positions received in 5 hours.

20130921 Single Event Latchup ?

We have had an autonomous reboot (watchdog initated ?) on our COM system. It did force it to 437.450MHz and 2400 baud. This was kindly reported by Toshio Kasei. We are now back on 437.425 MHz and at 4800 baud. The reason for not going back to 9600 is that we the last days has observed increased noise at UHF in Aalborg. So 4800 for having a ok link when we are trying to find this noise source in our neighboring area.

20130911 It's a long way to Aalborg -fixed

Trip time on our internet connection to Thule is close to one second. Our GND server is pushing AIS packages to Aalborg into our SQL database. A synchronous SQL call has been used locally at Aalborg but the trip time had a negative impact. So now asynchronous SQL write calls is used and we are now only limited by the bandwith which is not a problem.

So the night passes has downloaded and forwared up to more than 2000 AIS packages/pass. So we are on the right way to optimize Thule ground station. Just another lesson learned.

20130908 Some figures for 3.70B AIS receiver

From 01:22 to 01:55 CEST(33 minutes) AAUSAT3 covered the area from Greenland to the Southern part of Brazil. A total of 5989 AIS packages (from number 1427063 to 1421074) were received which we think is ok :-)

20130907 First realtime download from AIS2

Attach:uploads/130907ais2rt.png | less than 2 min measurement and download when passing Greenland

So here is the very first results with the new uploaded SDR AIS receiver. 200 packets recognized and downloaded to Greenland(Thule) all in less than 2 minutes.

This is real time AIS from space: only a few seconds delay from receiving the AIS packets until delivery in Aalborg and Copenhagen.

The previous SDR AIS receiver uploaded could find up to 10 AIS packets/second. We hope this will be even better.


20130907 Updated algorithm for AIS2

Last news - Upload has completed and testing has started. First positive results has been downloaded.

We are currently uploading a enhanced algorithm for the SDR. Changes includes, but are not limited to: 1) Less read-overflow in the sampling thread. Will improve online-time by 15%. 2) Better scaling of algorithm - significantly enhances the dynamic range. 3) Handling of Doppler with 0% degradation on sensitivity within +/- 5 kHz from nominal channels.

This new algorithm is due to students working hard, even in their summer vacation!

Stay tuned, as we reveal new key performance numbers for this low-power space based AIS receiver soon.

20130906 Latest info about charging

Attach:uploads/eps-production-2013-09-05-1.png | Energy production 3. Sep 00:00 to 6. Sep 00:00(CEST in all 72 hours)

Attach:uploads/eps-batvoltage-2013-09-05.png | Battery voltage during the charging

Since launch we have seen some slow degradation of our energy production. We think it is correlated to solar cells without cover glass - which means that atomic oxygen slowly erode the surface - if this is the cause. We do - of course - not know but is the best explanation we have.

On the positive side when comparing these measurements with the July measurements further down on this page it seems that energy production has stabilized so we are approx at same power production level as in July. AAUSAT3 is still tumbling a few rotation each orbit - as intended.

Charging is moderate, so we have to be careful about what we are doing. Just running AIS2 requires 1.1W or so and down link of data draws 2W so our decision about flying a battery powered system - despite we are in sun 24/7 - has been the right decision.

So operations will continue :-)


20130827 Ground station at Qaanaaq now running well

Today we did have a very nice pass and did receive AIS data. Some minor misconfiguration problems (sw) was solved so now everything seems to work correctly. Thanks to Omar and Sune and especially Svend Erik from DMI. Attach:uploads/a3-qaanaaq-gnd.jpg | Groundstation at Qaanaq Greenland

See a small movie for a 360 degrees tour on our antenna


20130825 First ships downloaded from Greenland

First ships plotted on map from first download of data from Greenland. It is live data because the ships positions are received and downloaded at once.

Attach:uploads/20130825-0909-liveAISatGND.png | First map plots


20130825 Reception at Greenland

Yesterday evening at 21:22:08 CEST, we did receive the first transmission from the satellite at our new groundstation at Greenland. Powered by DMA! (thanks to Omar and Sune) 4 bit-errors and no byte-errors.

The picture shows live received and downlinked data, during a short pass

Attach:uploads/08242013-093813firstGLtransmission.png | First package from Greenland Ground Station

130807 Northwestern

Today we again did receive a series of positions from Northwestern

like ... The Northwestern, 836.11 miles NE of Dutch Harbor

130730 EPS power production data

We found some link time to download a little data from the EPS. The following plot shows the power input from the last 24+ hrs. In the current configuration the satellite uses around 500 mW when the AIS2 receiver is off.

With a moving average of 100 mins (corresponding to one orbit), the power budget is thus positive, eventhough the margin might be a bit smaller than one could have wished for.

130725 Data preview from "the 1 mio." day

At 22nd of july 2013, we did receive about 30.000 position reports. These was gathered in the almost 6 hour time span, from 09:38 to 15:19 CEST. Here we present to you a small picture of the ship positions.

Note also the one red dot in Denmark.

Attach:uploads/130725-map-1mio-day.png | AIS from 09:38 to 15:19 CEST

130722 1 mio. crc ok ais msg received

In total we have received more than 1.024.000 ais position reports in space, using our <1kg cubesat. Just imagine what could be received (and downloaded) with a 2kg cubesat and deployable solar array...

130707 Raw sample with simultaneous transmitting ships over Europe

This new sample was downloaded with more than 50% of EU inside line-of-sight.

Attach:uploads/130707-eu-11.i.png | A raw sample from EU

130704 Waterfall plot of raw data

We have just downloaded a raw data sequence recorded over Greenland.

We think it looks nice - you can nearly hand decode ;-)

Attach:uploads/130704-greenlandwater.png | A raw signal from Greenland

130629 Brief analysis of DNOW 3.60b data

A brief analysis has been conducted on the 28000 frames received by the DNOW 3.60b algorithm on 12/06-13, 08:17-12:23. There's approximately 6300 unique MMSIs and 500 packets from class B transponders (although they transmit with only ~2W compared to ~12W for class A). Below images show the 28k frames and the route.


28k frames received by DNOW 3.60b the 12th of June, 08:17 to 12:23

Satellite position between 08:17 and 12:23 on the 12th of June


Inspecting the timestamps in the frames reveals, that the operation of this new algorithm is very much real-time. Actually, in ~87% of the time, one second with frames received is followed by a second with more data. More than one or two seconds of silent time is very rare (and is probably due to AAUSAT3 flying over silent areas) as illustrated on the figure below.


Distribution of "silent" seconds


Another result of interest is the number of packets received in the same second, refer to the figure below.


Number of packets received in the same second


With 28k frames and only 6.3k unique MMSIs, many are obviously from the same vessel. ~35% of the time, an MMSI has only been spotted once during the 4 hour flight, but as seen on the figure below, the algorithm picks up many packets from some vessels. Up to a stunning 76 packets from one vessel!


Histogram of messages received per ship


All in all, some quite good results from the new algorithm - newer results even show, that several packets have been received just north of Skagen as well as quite a few in the Baltic sea. Considering the enormous number of collisions when Europe is in range, this is very impressive!

130627 Raw data Greenland - II

First complete sample downloaded. The below picture shows the waterfall fft. The two AIS channels are at IF 175kHz and 225kHz.


Waterfall plot of the IF raw sample.


Taking a closer look, e.g. on the 4th packet on channel two starting about 90 ms in the sample and about 24 ms long, shows the following demodulator output and eye diagram. This raw sample shows that the new AGC algorithm of DNOW is working much better than the old. Now, we even have high quality recorded colliding AIS messages that we can play around with on ground.


Demodulator output from 4th packet on channel 2.

Eye diagram of 4th packet on channel 2.


130626 EPS reboot, COM reconfig

This morning the beacons indicated a situation similar to 24. May 2013. And in addition to that COM was gone to launch default config 437.500 MHz.

So ...

EPS does not respond on certain CSP packages - the situation did look very similar to the situation 24. May 2013 so after an analysis we did decide to use the same procedure as on 24. May.

At the 1904CET pass was ...

  • COM set to 437.425 and 9k6.
  • EPS was reset by sending the same hand crafted package as we did 24. May 2013.

And we were back in business and operation and download did continue.

We take this as a normal event you must expect when you are up in a space - of course it would be nicer not to have it but ... space is a tough playground.

Some observations:

  • Battery voltage 848CET and at 1905CET 7.96V.
  • At morning passes COM boot count was 292 and at 1900CET pass it was 318 - which indicates 26 COM reboots during the day.
  • EPS had at 0848CET boot count 3244 and at 1904CET 3272 - which indicates 28 reboots during the day.
  • A reboot of EPS should also reboot COM so two more EPS reboots could indicate some of the EPS reboot has been very close in time so COM has not been powered up.
  • It also indicates that COM has not been rebooting on by itself but only power cycled

but please take this as only a qualified guess.

Some selected beacons from the day


08:48 - all is normal



08:50 - COM cant get beaconinfo from EPS


130626 Raw data Greenland - I

We have sampled raw AIS spectrum on our SDR when passing Greenland this night. We are now starting to download them.

130624 UHF/COM reconfig

We have carried out a reconfiguration of the COM subsystem in space. As usual full FEC and running 9600 baud.

130613 10 min AIS2 data

12. June - Quick snapshot

A quick snapshot of downloaded AIS messages received from 20:32 CEST and 10 minutes ahead(12. June 2013), together with the position of the satellite at 20:32(around 2000 AIS messages). Not so bad ...


Data received with DNOW, a new uploaded SDR AIS algorithms

Line of sight at 20:32 CEST - and white dot position 10 min later.


130612 New AIS2 receiver

YES - First data from new uploaded AIS2 receiver

AAUSAT3 came by again after some hours on it own. We were waiting to see if the new AIS2 SW uploaded last night was running ok. so here it goes ...

From 8:17 to 12:23 CEST it did receive 28293 AIS packages with correct checksum

It seems the new AIS2 receiver receives three times more than we did receive in 12 hours of "receiver ON-time" on the previous AIS2 SW.

It is just so nice we can upload new SW :-) The true value of SDR in space!

So once again we can only say - we are more than happy with our little spacecraft :-)

So stay tuned - we will come with more information when we have a better evaluation of the results.


The track for the four hours data collection


The picture shows the track for AAUSAT3 during the 4 hours of operation from 8:17 CET. The distances between the track varies due to the polar orbit. At Equator the distance is approximately 3000 km and decreases when flying north or south. AAUSAT3 receives ships up to 3000 km on each side with majorities within 1500 km - which compared to the 3000 km distance between the tracks at Equator is ok. The yellow color of the path indicates AAUSAT3 is in the sun - even when there is dark below.(called a dusk dawn orbit). So AAUSAT3 is in the sun all time.

130612 CW Beacon wanted

06:15UTC Feedback much appreciated!

Today the new algorithm is running on our AIS SDR. Until 16:30 UTC we will only have one very low elevation pass, so if you decode a beacon (morse or FSK) please mail it to us. Monitoring the battery voltage will help us a bit. Thanks!

130611 Version b of DNOW

(21:00CEST) An optimized version uploaded this evening.

Its so nice to see we can just upload new SW to AIS2 without problems. However, remember to turn on the ADC power channel and make the program executable :)

130611 Yes! AIS2/DNOW ok!

AIS2 updated algorithm, a.k.a. DNOW, now tested 1 orbit in space(00:36CEST) Received 5071 crc ok AIS msg. Also detected 5434 crc error msg, but with correct HDLC flags.

More results to come after about 1 day of testing.

The figure shows a bit of the data downloaded in a 6 deg elevation pass this evening. We will have to wait for next pass for more data.


Data received with DNOW, the new software defined radio algorithm just uploaded.


130610 Oops power missing

-23:00CEST - forgot to turn ON the ADC

No status update for now, as we forgot to turn on the power channel for the ADC. Results will come tomorrow.

130610 New AIS2 upload succes

At 21:30 the last 18kB was uploaded during 129 sek. The checksum was correct so we have enabled the new software. We'll have the first results in one orbit (90 minutes). Stay tuned.

130610 AIS2 SW on way

Preparing AAUSAT4 + uploading new AIS program

Now we are on the way with the upload of new algorithm. 42 kB uploaded first pass. 60 kB in total, so only a bit left.


New DNOW uploaded during 1 pass. Still missing 12kB.


The bellow picture is live view from LAB. People are working on AAUSAT4: Including many improvements.


New activity in Satlab, working on AAUSAT4. This night @ 20:00 LT

130609 New AIS2 algorithm

...Preparing new algorithm for SDR

We are testing version 3.60a of our algorithm for AIS2, the Software Defined AIS receiver. Tests in satlab shows that the algorithm known as DNOW is an significant enhancement - is in running more than 95% of the time. Based on live AIS data from an antenna on the roof, combined with interference and noise from a signal generator, we are pretty sure that this is a good algorithm. At the same time, the AGC (Automatic Gain Control) is enhanced - we hope this will enable better reception when we EU inside our Field Of View (FOV).

DNOW was developed this semester by three 6th semester students from the Signal Processing line at Aalborg University.

We have turned off AIS2 to fully charge the batteries, and will upload the new algorithm on Monday. Stay tuned.

130605 100 days well spent

The following plot shows the latest ship positions downloaded during two evening passes, at the 100th day in space. During these two passes, more than 2400 ships were downloaded.


Plot of the ships downloaded during the 100 days anniversary pass - the plot contains 2400+ ships.


Based on most of the 700.000 ship position reports, received and downloaded during the first 100 days of the mission, the following density plot was generated.

Neat, right?


Density plot of the ships downloaded during the first 100 days gridded into 0.5 x 0.5 degree cells and with a limit of 20 ships pr. cell.


Soon, we will replace the original AIS SDR Algorithm with a new one, based on real data sampled by AAUSAT3 in space. Stay tuned for when we publish these results - we hope to improve performance significantly!

130605 100 DAYS IN ORBIT

Today - 5. June ( which is also the Danish Constitution Day) - we have been in orbit for 100 days. See Status for the first 100 days of AAUSAT3 mission where we walk through all our subsystems and evaluate current status. On the page is also a description of our two critical situations during this first period.

We'll try to keep the doors open into SatLab for the 19:00 pass today if anybody wants to have a look - or just knock on the windows.


130603 Phew traffic jam

That was close, but we're ok :-)

Today we had a close encounter with Iridium-45. In this case, a close encounter in space terms means passing each other at a distance of 834 meters with a relative velocity of 14.8 km/s. That would have made for one large explosion.

Luckily they passed each other without any problems at 16.20 UTC, and at the first evening pass of the AAU ground station at 16.30 UTC we received nominal beacons from the satellite, indicating that everything is going just fine.

130601 Older entries moved

to Log130601


(C) AAUSAT3 team - 2012 2013