I found a few bugs in the ConstantTurn transition model, which this PR aims to fix.
As detailed in #841, the documentation for Q_t did not match the implementation.
The bottom right term in the Q_t array was erroneously divided by the change in time, which would result in the noise of the turn rate being smaller for a larger change in time. I'm not really sure personally why this value is not instead multiplied by dt, but this seems to be a misunderstanding on my part and not a mistake.
There was an issue where if you put in an integer turn rate of zero, you would get an error for division by zero. This is a little bit academic as it's not really the intended use for this transition model, but I thought it was worth fixing for consistency.
I have also raised up a misbehaving comma in the first tutorial as it looked slightly messy.
I found a few bugs in the
ConstantTurn
transition model, which this PR aims to fix.Q_t
did not match the implementation.Q_t
array was erroneously divided by the change in time, which would result in the noise of the turn rate being smaller for a larger change in time. I'm not really sure personally why this value is not instead multiplied by dt, but this seems to be a misunderstanding on my part and not a mistake.I have also raised up a misbehaving comma in the first tutorial as it looked slightly messy.