Leaving Summit

The team is now in Kangerlussuaq after leaving this morning which was a good time to leave as 16 new people arrived.  It was nice having the place to mostly to ourselves as the only science group there.  Thanks to the crew for making our stay such a success and to my teammate as well, it was great trip! And thanks again to everyone that read my blog about Cool Robot's trip to Greenland! :)

The Cool Robot and all our luggage getting loaded last onto the plane at Summit.

Last Day at Summit

Well with the robot all packed up there was not a whole lot to do today.  We all just mostly just relaxed and took it easy.  Some looking at data, some working on reports, and some just doing nothing.  It was a nice day!  A plane was suppose to land today and dropping off a couple people and supplies but it was to windy to land and after seeing it circle three time it went back to Kangerlussuaq.  Tomorrow were are suppose to leave for Kangerlussuaq and hopefully there will be no problems with that plane landing and taking off.  As our last day at Summit here is a nice picture of the Cool Robot in front of the Summit sign.

Cool Robot at Summit Station in Greenland 2013.

Packing Up

We ran the robot almost all day doing the air sampler survey starting this morning after the robot sat outside all night.  Then it was time to pack the Cool Robot up!

Ben trying to figure out how to fit everything to the box.

Everything packed and ready to be put on a pallet.

Cool Robot chassis going into its box.

Cool Robot all crated up.

Day 21

What did the Cool Robot do today?  This morning and afternoon Jim and Alden ran a few grid surveys with the GPR.  Ben went for a ski collecting data point for a new fan survey for the air sampler. After Jim and Alden were done with their grids and before dinner, I manually drive the robot with GPR over where they thought the old freezer of the station was buried in the snow to see if we could find it.  It looks like there was defiantly something down there but that is all I could determine with my very limit knowledge of the GPR scans.

This evening we sent the robot on an night mission with the air sampler but the robot had some problems with the turns, getting itself really stuck.  We ran it through another couple of waypoints to see it would do on those turns; the second one went fine and third was a bit wonky but it made it through.  The robot also ran in slow mode (half speed) the whole time due to lack of power from the sun which it seemed to suit just fine except for the turns.  We decided to just leave the robot out there stopped and we will run it in the morning.  With more sun it will drive at its normal speed which should help the turning and hopefully it will not get stuck again.

Cool Robot taking off toward it next waypoint.

Ummm...what is this thing? Where did it come from in the middle nowhere?
Jim and Ben helping the CR out after getting itself stuck on a turn.

Skiway

Here are some pictures of the skiway that I keep referring too.  It is a little over 3 miles long.  It get groomed often to keep the snow hard, especially in summer time.  Sometime after a grooming on a Sunday you might see some of the crew taking off for a run on it. 

View of the Skiway

Big tracker grooming the skiway.

Day 20

Jim and I learned a lot from last night's mission even though it go stuck twice in the snow, stuck in stop mode, and lost all of its waypoints.  Yeah it sounds like a horrible run and it was in terms of not doing exactly what we thought but because of what we learned I would say it was one of our better ones.  Here is why:  When the CR got stuck I could first figure it out that it was stuck by looking at printouts the robot was giving me on the computer which I confirmed by seeing that the robot was not moving.  So setting the robot back to manual mode I first backed it up for about half a second and then ran it forward and turning for about 1 second before set it back to autonomous mode, which is all that it needed to be on its way again.  Jim also figured out that it was probably getting stuck because the robot was trying to make such a hard turn while going slow that the inside wheels would be probably stopped while the outside one where spinning, digging itself in.  Another thing discovered with the turns is if the robot decide to stop (the turn taking more power) that the robot could not restart itself because there was some kind of flag in the code that could not be reset for that condition therefore the robot would just sit.  The lost of the robot's waypoints we was a bit perplexing but we are guessing it was due to a lost of power to the rabbit board (the microcontroller) because the battery voltage dropped to low.  Not sure if that is what was as it is the first time we have seen it happen and hopefully it will happen again.

Today we brought the robot in for a few code changes: removed the possibility for the robot to get stuck in the program's dead spot, updated the panel voltage and current calibration curves (they where not much different) and increased the duty cycle for better power tracking of the solar panels due temperature changes.  Also we have noticed that panel 1 is hardly producing any power for some time now so we checked its boost converter and discovered that it took a hard hit at some point and melted.  Ben spent this evening finishing building up a replacement and installing it.  After which Jim and I ran a short test tonight to see how it worked and how the max power point tracking was working with the changed duty cycle.  Both looked great or better than they have looked in the past few days!  Tomorrow we will do a longer run with more sun.

Melted chip insert on the booster converter board.

Ben working the booster converter replacement.

Day 19

Part of the fanned survey we did yesterday, for the air sampler, ran through the night.  After collecting the data we discovered that the robot was stopped between midnight and 2am but did start back up on its own!  Just not at a point we would expected it too, but still happy it could restart after charging the batteries for awhile.  After the Cool Robot finish that survey this morning Jim and Alden did some more grid surveys with both GPR units.  The CR pulling the smaller of the two and Alden pulling the big one behind a snowmobile following the same path as the CR.  I went out for a little while to help out and take pictures.

Cool Robot and Alden pulling GPR units and a tracker grooming the skiway in the back ground.

Closeup of Cool Robot towing the small GPR unit.

As for the coding we decide it would be best to only fix a couple of things at a time after yesterday's episode.  So only fixing the logic error I found two night ago in the drive code we have load that up and the robot was running fine.  Now we are again sending it on night mission doing the same fan survey we did last night but this time the robot should able to slow down instead of just stopping when the power is low.

Day 18

Well after staring at the code for hours yesterday and still not getting anywhere I decide to take a nap and ended up waking up with an answer!  I had found a spot where the code might have been conflicting with itself.  We made the fixes by changing a variable here and there, then load it for another night run.  This time I launched the Cool Robot into autonomous mode from the Big House so that we could monitor it the whole time.  The robot went off with out a hitch!  As the robot went on its marry way I could see what waypoint it was on, how far it need to go to reach the next one, and what speed it was traveling.  I also could see what my code was doing by the print statements I added to it with regards to slowing down or stopping based on the robot's power conditions.  From this information I was able to tell that the robot wanted to drive at half speed but it wasn't, why?  Well I looked at the code I had just written and realize some of my logic was off, but the robot was still able to continue on, it would just have to stop instead.  Which it eventually did because it didn't have enough power.  Would it start back up again?  It did not, we had to restart it from the Big House, which wasn't hard to do but perplexing why the robot would not start back up on it own.

View of running CR from the Big House, the little dark box in the middle.

I made some changes to the drive code this morning, fixing the logic problem and moving a couple of things around.  Jim made some changes to the power code trying clear up some its logic.  But neither of us could figure out why the robot would not go again after stopping.  Anyways we load up our changes in hope to test on a fan survey for the air sampler...3 hours later and being very confused as to why autonomous mode would no longer work with the updates we went back to the code we used last night without any changes just so we could get the survey done.  Beside the code not doing exactly what we think it is suppose to be doing the Cool Robot is running great as long as it has enough power from the sun!

Happy Solstice!

SOB (Science and Operation Barn)

The building where we work on the Cool Robot is the SOB (Science and Operation Barn), basically a big garage.  The front part of the building contains the snow melter which make all of our drinking water by dumping snow from the snow mine into it and it melts and then it gets piped to the Big House.  On the side of the building are a bunch of containers that have different materials one might need while up here.  If they don't have it either figure something else out or you will have to wait until the next flight to get what need.

SOB snow melter side.

All the box stores one needs including Homedepot and Napa (inside building).

Day 17 - More coding

Well sent the Cool Robot out on a night mission last night!  It was so hazy last night that as soon as the robot got to the other side of the skiway we could no longer see it, all we could do was wish it good luck!  We ended up sending it on its way around 11pm last night.  When Jim woke up this morning he noticed that the robot had stopped and was not moving, but he was able to get it moving again from the Big House, which is nice that we don't have to fetch it.  From the data collected by the CR it appears that it only ran for about half an hour before it stopped because the batteries were too low, which it is suppose to do.  We are missing something in the code to make started back up again when the batteries have enough charge to go again where as the batteries did charge enough so that it could resume its course again.

Sending the Cool Robot off on a night mission.

Ben and I have been looking have been looking at the code all morning and have yet to find why the robot did not resume after stopping to charge its batteries.  Hopefully by this evening we will have found out why.  Meanwhile the CR is still hard at work with Jim and Alden doing a grid surveys with GPR.

Day 16 - Bright and Sunny!

Ben is currently eating dinner after chasing the robot on its last lap, of two, next to the skiway and is complaining about the titles of my posts saying they sound like a prison sentence, so to humor him I have added that it is Bright and Sunny!  Which it is pretty much ALL of the time here!  The sun had never set sense we have arrived.  Ok back to what or how the robot is doing today.  Well we have set Cool Robot to work doing the same two laps it did earlier this week while we continue to work on the coding of the battery limits and slowing down when there is not enough power.  Hopefully when the robot comes back we will give it the new code and can send it on a night (still sunny) mission while we all try to sleep in our tents.  I know these past couple of post have been kind of bland so for the next one I promise to post at least one picture with it.

Day 15

Spend the whole day making changes to code.  The morning was spent trying to get the datalogger to get the string we were sending to it from the drive code.  We eventually got the GPS coordinates, but still no motor currents and voltages, they will just have to wait unit another time.  The rest of the day was spent in implementing a slow down and stop mode that is signaled from the power board to the control board (drive board) and putting a few more limits on the battery voltage and currents as the robot gets more power or not enough power.  I think we almost have it we are just having a problem with the power board trying to control the driving before we, the user, want it too!  Hopefully we will have it fixed by tomorrow morning and we can run a longer endurance test with the air sampler in tow.

Day 14

Today we did an 8 hour long endurance test with the Cool Robot towing the air sampler behind it the whole time.  It's path is to navigate from camp to the far side of the skiway and travel north to the other end and back making a 11km box, which was done twice making for about a 22km run today.  See the Summit station maps to get a better idea of what I am talking about: http://www.summitcamp.org/site/.  The first go around Jim and Ben skied with the robot (just under 4 hours) to make sure it did not hit anything and the second time around we just watched it from the big house through binoculars and a telescope.  The robot did great on the run whereas the user computer kept loosing communication and every time we restarted the program it would stop the robot.  The first time it happened we didn't realize it for about half an hour.  We eventually figured it out and were able to get the robot to go again, which was great to discover that we could communicate with the robot that far out and not have to go chase it down on a snowmobile.

Ben and Jim skiing after the Cool Robot on its first circuit.

Outhouse

One might think that sleeping in a tent is bad but what about going to outhouse in the middle of the night?  It is cold here so that's gotta suck.  Yeah it does but not as bad as you think!  The outhouse here are actually pretty posh in the sense of the seat cover, when you sit on it it is not bone chilling cold!  You still get the typical outhouse smell which I actually thought you would not get with it being so cold.  Maybe they should make a Greenland outhouse calendar or maybe they have?

      

Day 13

Jim and I edited the drive code to try and get data to the datalogger, but still could not get it to work.  We believe our code is correct but there is some timing issue somewhere that we cannot seem to figure out as the as it is done exactly the same as the power code and the datalogger gets that information.  We are stumped for the moment, so we decided to take the air sampler we have for a test run around the taxiway with all 5 panels hooked up.  We also test towed a sled the contains about 250lbs of radar to see if the Cool Robot could even pull it, which it managed to do although a bit scary to test as the robot only weighs about 150lbs. Lets just say we will not be towing it autonomously!  This evening we gave a talk to the crew members about what we are doing and why.  Afterwards we gave a short autonomous demo with CR towing the air sampler.  It was also a test to see how the robot does sitting in the sun for awhile while we prepared and gave the talk.

CR towing the air sampler.

CR towing the radar monstrosity.

Day 12

Today has been busy with the added duties of dish washing and man did it feel like the dishes were never ending, but with the help of the crew they got done!  The Cool Robot was driven with all 5 solar panel hooked up and it didn't catch fire, but had some really bad steering happening which was not fun to drive.  While I was finishing up the dishes from lunch Jim and Ben drove the robot back to the barn.  When I joined up with the crew Ben told me the bad news that he snap the front support bracket off the support rib and just has he was telling that news Jim discovered that one of the wheels had come uncoupled from the motor, that is why it was steering like crap!  At first everything sounded really bad but taking a moment to think about the problem they turned out not that hard to fix.  The robot will be ready to drive again by tomorrow morning.  The motor coupling just need to be tightened, not easy to get to, but way better than removing the boost converters to get the motor mounts!  The bracket was just taken off and reworked and epoxy back on, good as new!  Later I worked on trying to get the control board to send the data of the motors and GPS over to the datalogger only to determine that there is some timing issue within the code that Jim and I will work through tomorrow morning when we are less frustrated with it.

Left: Jim work to tighten the motor coupling.
Top: Jim's view for getting to the coupling.

Left: Ben scraping the old epoxy off the support bracket.
Top: Allison scraping the old epoxy off the support.

Day 11

All 5 boost converter have been installed and ready to run all 5 solar panels at once!  Also the some parts of the drive code where changed, a longer front bumper was added, and a more ridged towing line was made for the GPR.  The test will have wait until tomorrow as we are going to take it easy tonight, I believe a well deserved break from working very long days to get to this point!

Big House

The Big House is where we eat and hangout.  There is a cook that make everyone lunch and dinner everyday except Sunday, which is leftover day.  The food is really good here so there is no worry about not getting enough to eat.  The big house also has a flush toilet, a shower, and laundry! But there is a water restriction here so we are only allowed one shower a week!  Also sense we are staying here for awhile we are put into the dish washer rotation and tomorrow is my day, but for the good food I am getting I guess I can wash a few dishes.

Big House!

Lunch time!

Day 10

A 3rd boost converter was installed and the code was updated to accommodate using the top panel.  Once that was done we button the robot back up and ran the same loop we ran yesterday with the top panel add into the mix and it was noticeably happier, aka was moving faster with the added power.  All good news!

The "Cool" gang walking behind the Cool Robot during the 3 panel test.

Day 9

We spent all day yesterday getting the robot ready to do autonomous, solar power test running on two panels.  The morning and afternoon were spend editing the code to fine tune the power tracking for the two panels.  Then in the even, with sun very low in the sky, we ran an autonomous run around the taxiway of the skiway that the planes land on.  The two powered solar panels were placed on the front and back of the robot.  It barely had enough power to do to the course and get back to the garage with the sun being so low and only have two panels working, but with that said it ran with just two panels!

Ben and I working preparing for the next solar test.

Ben walking the CR during solar power test.

CR in for the night waiting for the 3rd BC to be installed.

Day 8

We went back out this morning to run another autonomous grid in the same area but instead of east/west lines it did north/south lines.  Also this time the robot did not stop at each waypoint which smoothed out its line of travel.

CR doing waypoint following, can see the tracks from the previous day's run.

After lunch we took the solar panels back off and tested Ben's newly built inverter circuit which worked great once everything was connected properly!  I also started repair on the damaged solar panel which will hopefully look almost as good as new.  When we testing the boost converters (BC) we realized that something was wrong with our power tracking.  Today we did some simple test and discovered that some program channels were wrong which was an easy fix once we knew what the problem was.  With that last problem solved surrounding the BC we are confidant that we will be able to run on solar power real soon!  This evening we went through the power program to scrub out any other little problems that would improve its overall operation.  After doing that all of our brains hurt and we are ready for bed to start tomorrow fresh to solve the next set of problems, but hopefully very soon we will be watching CR work while we have our tea in the big house.

Day 7

Assembled the robot again and added the GPR instrument on the inside on top of the robot chassis while it drag it behind it.  We successfully ran a 50 meter grid with 5 meter between length autonomously today!  The only problem was when we hit a half buried bamboo pole that cause the Cool Robot (CR) to get stuck.  Some damage was caused to the side panel substrate a little, but NOT the cells.  The small bummer I built prior to going out did seem to help, I guess I need to make it the full length of the side panel.

Alden checking GPR instruments inside CR.

CR stuck on bamboo pole, circled in red.

Walking the CR home after the running the grid.

Sleeping in a Tent

You might be wondering how we live way up in the middle of nowhere with nothing surrounding us except for a LOT of snow?  Well I will try and cover different parts of day and what we do throughout this blog.  This time I will talk about where we sleep; in a tent!  What?!  I think they are called oven tents because they actually get hot during the day although at night they are not the warmest things to sleep in they are still warmer than sleeping outside.  Inside each tent is a cot with a some sleeping pads and on top of that you put your sleeping bags.  In my case I have one inside the other for when it get really cold at night!  The lowest temperature I have seen so far is -17F, but when it is windy it is colder and the tents are noisy.  Also with the tents it is bright orange inside them all of the time, as the sun never really sets during the summer, so you will need something to cover your eyes if you want any kind of darkness.

Tent city or the pumpkin patch.

The inside of my tent.

Day 6

We continue to test the boost converters (BC), with each test we do we grow more confidant that they are working!  As we are testing the BC in the robot with a solar panel we are also taking down many measurements to better calibrate the readings that the rabbit board (computer) uses to track power from the solar panels.  The problems we have left to fix appear to be just software fixes now which shouldn't be to hard to fix.

Ben and Jim taking various readings from the robot.

Side solar panel placed outside while robot stays indoors.

Day 5

Jim went out with Alden this morning to collect some GPS points and help Alden with finding a drainage channel in the snow by digging a pit and then running over it with the GPR to see if they can see using that device.  Mean while Ben continued to test the boost converters and I siliconed in the honeycomb gaps in the robot chassis.  Smelly and messy stuff!

Alden pointing to drainage channel in the snow.

In the afternoon Jim and I tested a side and top solar panel to find the power curve of each.  It was really a nice day for doing such a test with the panels getting about 1200-1300 W/m^2 of solar irradiance.

Side solar panel with me behind it getting ready to test.

Top solar panel.

This evening I finished tapping the edges of the robot chassis lid so the robot chassis make over is now done and it is ready to go as soon as we are ready with the solar power.

Day 4

We got one booster converter working!  Which means that we can run at least one solar panel right now, a very good thing.  I realized that I may have gotten a little to technical last time for most of the reader so I will try to keep it more simple.  First Ben fixed the inverter problem which was pretty easy and then a bench test of the boost converter was conducted to make sure that one was working the way we expected it too. While that was being done I finished taping the chassis box edges and created a stand for a side panel using Alden's camera tripod.

Bench testing boost converter.

The first time we tested a boost converter in the robot we took the whole setup outside which was very cumbersome.  For the next time we left the robot where it was and ran some very long wires to solar panel so that only the panel needs to go outside.  But in the case of a late evening test all we needed to do was just open the garage door cause the sun will shine directly in at that point.

Testing boost converter in robot
with solar panel connected.

Last Plane

Well the last plane that we will see until we leave, in 20 day, left today leaving 14 people behind; 10 crew members and us.  We are now officially stuck here!

Last plane until June 27.