Posted 6/12/2019 10:31 (#7554845 - in reply to #7554767) Subject: RE: Autonomous autosteer...
Cedar Rapids, Iowa
The answer to that question is very complex.
I'd split the solution into two parts: Determining the route to be traveled and actually driving said route.
The route calculation part needs to know how tight the machine can turn, implement width, offset, etc. Basically this builds a list of GPS waypoints to travel along.
The driving part needs to follow those waypoints. I'd personally use a dual-antenna GPS solution so you always know the absolute orientation of the vehicle. Single GPS solutions need to move to determine orientation, then need inertial sensors to maintain an estimate of the orientation when turning. With dual GPS, you can get orientation right at boot up without moving. Then you need the steering math and control circuits to steer the vehicle. Safety stuff to stop the vehicle when anything unexpected happens, and probably laser scanners to detect obstacles. When you're dealing with a machine that can run someone over, safety is very important.