Vision and Game Integration

This module connects camera input, board-state recognition, game strategy, and robot actions.

Camera Setup

Use robotics/vision.py to identify which camera ID corresponds to the USB camera.

The course notes currently suggest changing the integer used to create BoardVision until the USB camera output appears. After identifying the correct camera, use that value consistently in the vision and game-vision scripts.

Game Vision

If gamevision.py freezes or uses the wrong camera, repeat the camera setup process with vision.py.

If Pygame font errors appear, reinstall Pygame inside the active environment:

pip uninstall pygame
pip install pygame --no-cache-dir

Board Calibration

The repository includes board calibration and mapping files:

• https://github.com/ripl/hs-robotic-manipulation-course/blob/main/robotics/game/calibrate_board_map.py
• https://github.com/ripl/hs-robotic-manipulation-course/blob/main/robotics/game/board_position_map.json

Students should understand that the robot does not know the board automatically. The board map connects camera observations and robot positions to named Tic-Tac-Toe squares.

Integration Loop

The full robot game follows the same pattern students saw in Python games:

1. Read the current board.
2. Decide whether the game is over.
3. Choose a legal move.
4. Move the robot arm to place a piece.
5. Observe the board again.

Repository Code

• smart_game.py: integrated game flow.
• smart_player.py: decision-making player.
• robot_player.py: robot-controlled player.
• gamevision.py: board vision.
• game_ui.py: user interface.

Source folder:

https://github.com/ripl/hs-robotic-manipulation-course/tree/main/robotics/game