Don't let the robot drive outside the edges of the field. Kinda like the aboutToFallOff but better and more thorough

Screw the copied/pasted class, and create a PI controller (don't really need D, never really used it anyway) based on that one screenshot on your phone. This should allow us to not miss waypoints

Do this by being able to turn around and head back to the bonus waypoint. We need to be consistent to be competitive, and as it is, bonus waypoint achievement is not consistent. The other part of this is #13

Should be easy now that most errors are figured out

exactly what it says. nothing fancy. just don't run into stuff

Subscribe in robot_control_node and publish a boolean which Robot subscribes to and uses to figure out if obstructed or not (because we use bumper for reversing)

By also looking at the longitude and comparing the difference between the goal and current coordinates to an error threshold or something

Far-future, yes, but time is still ticking

just to look cool. not necessary at all, and worthless, but I don't want to do actual stuff right now so here's an issue

Need to:

• subscribe to camera
• pass data to function in Robot
• take that information, and convert it to something OpenCV can use
• search for the waypoint using template matching
• if self.getDist() < self.threshold (temp value):
  # then use x pos of matched image minus center of the camera to PID to

what I need to do is written on a sheet of paper somewhere. Do it

That way we can be even more speed
Haha PC go whirr

By switching to Haversine formula instead of basic distance formula

Implement something resembling path planning. Needs to:

• detect obstacle
• turn angle away should be proportional to how far off we are from curr_goal (the heading, maybe?)
• but should never be lower than the minimum required to avoid an obstacle
• if there are obstacles only on one side (say, split self.stripped_data in half) then turn slightly away form that
• if we are obstructed then don't drive as fast

Path planning should use inequalities with angle + width of robot and stuff to see if we will hit something. Obstacles are stored as a list of points where they’ve been detected (x, y)

So it is easier to be used by other people. Ex. Functions should take filenames as params. Release on slack once done

Need to:

• subscribe to camera
• pass data to function in Robot
• take that information, and convert it to something OpenCV can use
• search for the object using template matching
• create an array of locations of the obstacles

Add in a part where you see if most of them are low which means reverse and then if reverse return -8. And then also some better turning values if obstacle detected. And make it more consistent

Should work if you just rotate the thing by 90 degrees