|
QUESTION:PWS - powered wheel steering. Basically, this is where the left and
right wheels are independently powered and you steer by using the
difference in wheel speeds between the two. In other words, you steer
the robot like you would steer a caterpillar track tank rather than
how you would steer a normal motor car.
ANSWER: As far as the transport industry in Australia goes, it's also called "Skid
steer", since the turning action can only happen if the power transfer
components (tyres or tracks) "skid" in some way one the ground surface.
Note that the term "skid steer" applies equally to "caterpillar" style
vehicles (tanks, etc.) as it does to the four-wheeled variety of loaders
(also called Bobcats here), since they steer using the same idea (right hand
wheels/tracks forwards, left hand wheels/tracks reverse or stationary, means
left-hand-turn, etc).
In fact there are often six different types of turns that can be made, two
choices for direction (left or right) and three choices for turning point
(left-side, centre or right-side). If you want centre turning point, you
need to drive both sides at the same _speed_, but in opposite _directions_.
If you want edge turning point, one set of wheels/tracks should be
stationary, while the other set drives. The turning point (in theory) is
the centre of the stationary track/wheel-set.
Of course if you're a skilled operator in a proportionally controlled
vehicle (where the speed of each side can be independently and continuously
controlled from full-forward to full-reverse), you can put the turning point
anywhere on the line that cuts across the vehicle (call it an East-West line
when the vehicle is facing North) and passes through the centre turning
point (essentially the "centre" of the vehicle). This is done by controlling
the difference in left-right track/wheel _speed_ appropriately. The turning
point is always on the side that is driving slower _speed_ (magnitude). The
direction is always toward the side that is going least forwards (where
least includes going at a negative speed forwards, in otherwords, going in
reverse!). Whether the turning point is inside or outside of the vehicle
footprint is determined by the same-ness or opposite-ness of the
_direction_, I think. You'd have to verify this last part with an actual
vehicle and some testing. I think it's right though (at least it sounds
good!).
Which brings up another point. All this stuff about turning point (and
hence absolute position and heading) relies on the surface under both
tracks/wheel-sets being the same. Any difference in surface properties will
throw the turning point right off (and might also throw the direction in
really pathological cases!)
|
|
|
|
