Powersliding Rover using ROS
jesse brockmann
http://diydrones.com/profiles/blogs/teaching-rovers-to-powerslide-with-ros
They are using a high end RTK gps,but I think I can do the same thing on a budget using decawave setup. So, maybe I'll have to look into it more at some point. This is also the reason for the ROS question in my other email. It's an open source ROS solution, so I've already been looking at the code. But I'm hoping for a simpler source base to start from. My rover from last year is very similar to this one, so I think I could totally replicate it!
Michael Shimniok
Very cool. Their approach sounds like it is modeling and predicting in real time and is computationally expensive.
I have been wondering for some time how best to emulate the human approach of learning from experience where the limits are and driving to that. As well as reacting to loss of control.
From driving SCCA Solo II autocross for a couple years, all too often I couldn't react fast enough when I pushed too far. Sometimes I could. Surely a computer sampling various sensors very frequently would be better able to detect imminent loss of traction and recover before it became an issue.
So the robot basically follows the course but experiments with speed and turn radius to find limits at least for one particular surface. If one could get that working then leaning limits based on environmental factors (location, temp, surface type) could be added in.
Easy to type but I am guessing it'd be really hard to implement for me :)
--
You received this message because you are subscribed to the Google Groups "diyrovers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diyrovers+...@googlegroups.com.
To post to this group, send email to diyr...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/diyrovers/CANqFZg2M9T9gDD34ic9YP5GBKjMgDPta30T1pGMb8F69iRRqVg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.