The Code Bender

It's interesting how the little things can bite you.  A new problem has arisen; SeaWolf likes going left, he's sticking to his southern heritage and apparently wants to drive a racecar.  This has the side effect of him missing our gate, at keeping us from qualifying. Hrm.  Well after discussing a few things about hydrodynamics and other possible reasons this could be, we've decided first we'll tackled a potential thruster imbalance.

So what are we doing to fix it? We're currently swapping out the thrusters to gives us the two "new" thrusters we just received from SeaBotix a few weeks ago on the port and starboard directional motors.  These motors have been recorded and should have basically the same characteristics when it comes to thrust.  The craft is naturally fairly stable when it comes to pitch and roll and we also have PIDs that will correct any errors that occur there so unbalanced thrusters aren't as big of an issue.  For our forward direction we currently rely on vision or acoustics pretty much solely, no magnetic compass or any other form of sensors rely.  This wouldn't be such a big issue except that once we're within about 10 ft of the gate we lose site of it and go into dead reckoning mode.

Assuming that the unbalanced thrusters aren't the issue we're currently dealing with, we are also doing what we can to get our IMU's magnetic compass working.  It's got quite some drift when we pole constantly but we're hoping that maybe by using a technique in which the engines are turned off, the IMU is polled and then the motor is kicked back on.  This gives us a brief period where the lightening storm of magnetic interference of the motors doesn't affect the compass values.

Update on Wednesday, July 29, 2009 at 11:05PM by Matthias

More info: After talking to a few more teams and some people from SeaBotix we are planning to try to implement a gyro for heading detection and some form of impeller for velocity tracking.  I'd like to tackle this as an Arduino project as between the two we should have functionality that resemlbes the holy grail of this kind of competiton, the Dopler Velocity Log.  A DVL basically allows for a craft to determine it's position in relation to a start point.  This kind of mapping would be extremely useful but they usually run in the area of 10+ grand.  I think we can manage to keep the cost a couple of orders of magnitude below what a real DVL would cost so it's our best bet at bang for the buck and leaves us without the drift problems our compass has.  So wish us luck, looks like we getting to hack something together in just a few short days!

Update on Thursday, July 30, 2009 at 2:49AM by Matthias

Our current IMU is responding at least semi correctly now, between the IMU yaw heading and the compas we should be able to come up with something to keep us going straight.  Simpler and more timely, but hopefully we can tackle the other project coming into the next year!