troiwill / tekkodu Goto Github PK

The LeftOf, RightOf, InFront, and Behind modifiers do not work properly for perceptions such as see and bump. There is a bug in the code it references. So, I need to implement a temporary fix.

The parsing code (in the Parsing directory) was written in a hurry so that a user can write a Kodu program without having to recompile each time he or she altered it. Most of the code was not cleanly designed, which makes it hard to understand what the code is doing at times. There is also much copy-and-pasted code that makes the code unnecessarily longer and inefficient. Thus, the parsing code must be rewritten.

Currently, the robot's head is moved to a fixed position where it executes one (1) Map Builder Request, which takes a photo of what it's looking at and analyzed for specified objects such as red canisters. Then, it begins interpreting the Kodu program. But this method focuses the robot's attention to what's in front of it, not what's around it. Also, the user must reposition the camera each time he or she quits Tekkotsu, especially when Tekkotsu runs in conjunction with Mirage. To evaluate conditions such as WHEN see red apple behind, the robot must autonomously look around the environment first, then begin interpreting the Kodu program.

TODO

1. Learn how to set camera "gaze points" (Tekkotsu)
2. Write code that will make the robot looks at predefined "gaze points," then interpret the Kodu program