Lost in identification

15-16th Dec 2012 - The 9th and 10th meeting dealed with entering data about the game, distances and object recognition.

Identifying objects

This weekend we discussed about how to enter distances to objects and how can we identify the objects on the board. This lead to a wrong way and we lost some time because of this.

By some reason, there was this idea. Using the sonar we'd detect the existence of an object. If something was detected then we would tell the robot who was that object. That was of course wrong. We cannot tell to the robot explicitly who the objects are. He must acknowledge them by himself. And the code done for it was lost time, so it seems. Maybe not, maybe it can be reused in the future developments the parts for detection.

So the idea was to focus in techniques that helped identify the positions of the objects by their distances, while moving around on the board. We did case studies of several scenarios and saw a pattern on a way to determine possible locations and how to exclude hypothesis.
We learned that if we manage to pass by the line or the column of the Butter, then we will know exactly where it is.

The case of the Toaster is to be looked into after. The case for the Mold, because it moves around is still trickier. We will look into it later. The Phase 1 will deal with determining a simple route and following it. Setting a destination and generating the correct sequence of moves to reach it, considering barriers. 
After having this working correctly, Phase 2 will take in account what changes are necessary to account for the opponents.

As for the sonar,  it will be used when in case of doubt of the exact location of the opponents or the butter. It will allow better chances for a correct decision. Example, Toaster at 1 move distance, 50% chances between going forward or going right. The sonar can help determine where the object is. Of course if there's more than one object around this won't help.