[Ian Mackenzie]

Thanks for the suggestions - the drive should be a lot smoother in a couple of days. There are a few code improvements that we were planning to make anyways, and we will be gearing up the potentiometers to get higher accuracy at the expense of reducing the number of revolutions the gearboxes can make (currently, we are using 10-turn potentiometers 1:1, which means an accuracy of about 3 degrees or so - we are going to change that to 2:1).

For the operator controls, the plan is to have the joystick control the translational velocity and a separate knob control the angular velocity of the robot - not only does this seem to have the simplest and most direct coding implementation, it makes sense because our driver is used to driving RC cars with the steering on a separate knob. So moving the joystick will simply move the robot without changing its orientation; turning the knob will spin the robot in place; straight ahead on the joystick and a turn of the knob will give tank-steering-like motion; sideways on the joystick and a turn of the knob will result in a "circle strafe", circling around an object while always pointing towards it.

We toyed with the idea of attempting a shifter, but that seemed a bit too ambitious - although next year, who knows?