Sensor Grid Trials

AMOS has been programmed to accept a pair of diagonally opposing GPS coordinates that define a rectangular area for collecting sensor data. It also requires the user to specify how many stops to make in the N-S direction over the grid, and how many stops to make in the E-W direction over the grid, and how many samples to collect at each stop. For example, the following script command:

grid_sample: 45.878336, -66.909070, 45.876592, -66.906151, 5, 5, 10

tells AMOS to make 25 stops (5 in the N-S direction by 5 in the E-W direction) down to the opposite corner at 45.876592, -66.906151. Here is what it looked like with the path to and from and all the sampling stops drawn on a Google Earth view:

The whole thing took a little over an hour to complete, and since the day was overcast and a bit rainy, the battery was nearly fully depleted. I had hoped that the water closer to the inlet of the small cove would be warmer and would make for a nice color map, but alas no. The water temperature was quite uniform throughout, either 14.06, 14.12, or 14.19 deg C everywhere. I don't have the turbidity or pH probes hooked up yet; I'm waiting until I finish building a mechanism for deploying them into the water so that they don't create excess drag when moving around. Here is a graphical representation of the temperatures collected, expressed as a bubble chart in Excel:

Today I tried to repeat the sensor grid test at a location more likely to show some sort of temperature gradient. The goal was to launch AMOS from a parking lot about 1 km north of where the Nashwaaksis River empties into the St. John River. I set up the GPS waypoints in a script file the night before with the intention that AMOS would first travel west about 100 m into the St. John River, then 1 km south where it would do two separate grid patterns around the end of the Nashwaaksis River. Both the wind and the current seemed really strong this morning, with the current running from north to south and the wind probably in excess of 30 km/hour (Andrew, I just bought an anemometer from Amazon so should have better data on that soon!) coming from the northwest. Poor AMOS didn't stand a chance. It struggled gamely for a little over an hour, before its battery quit and the wind quickly pushed it back to shore. I probably should have edited the script to just skip the first waypoint and proceed south instead but part of me kind of enjoyed standing there watching it struggle. At one point when the wind died down a bit it looked like it would almost make it, but in reality it was still pretty far off:

As this is a more populous spot than my usual Woolastook testing ground, there were a few curious onlookers that asked questions about it. Next time I'll have to bring my wallet with its home-made business cards 😃.

I really like the name AMOS, but I'm thinking of changing what the acronym stands for. Instead of "Autonomous Mini Ocean Surveyor", I'm thinking that "Acquatic Mini Observation System" might be more accurate, as it is not really an oceangoing vessel, unless perhaps you're talking about using it in the doldrums around the equator.

It turns out that I was not properly using the real-time clock that I had hooked up to AMOS a couple of months ago. It needed this line at the start of the /etc/rc.local file:

sudo hwclock -s

I had noticed that the ship log times were always wrong after a cold boot in the field without any available Internet, and that missing line was the reason why. AMOS might take a while to reach its destination against a stiff breeze, but at least it now knows exactly how late it is.