Hi everybody,
For a little project we are planning to use the CMUcam5 but we are wondering if it could fit our needs.
We aim at localising a robot thanks to a known target.
The target : a 1m * 1m panel with 2 colors (say red and yellow) in order to use the Color Codes functionnality.
Assuming the robot starts its mission back to the target, drive for 6m then turn 90° right then drive 3m and finally turn 90° right again (see the image attached with a top view for better explaination).
Obviously the robot movement will never be perfectly accurate so we can not rely on the “expected distance travelled” or the “perfect 90° turn”. That is why we are planning to use CMUcam 5 to add some correction to this path.
Thanks to the CMUcam5 technical specifications we should be able to retrieve the distance between the robot and the target (range_in_meter = ((1280/2)*1_meter)/tan(75/2) see [[http://cmucam.org/boards/8/topics/2915]] for explaination.). In the image attached the distance should be around 9.5m. Are we correct at this point ?
Will the robot be able to known the yaw angle between itself and the target (expectedAngle/obtainedAngle) ?
Here [[http://cmucam.org/projects/cmucam5/wiki/Using_Color_Codes]] we can clearly see that the roll angle is returned.
To sum up our true question is : will the robot be able to position itself perpendicularly to the target but on a parallel axis ?
We hope this is clear enough…
Thank you in advance for your help !