This week, we partook in lengthy discussions on our final product and decided that we should proceed with our backup plan of having Omar handle the integration and assembly of the robot since we could not get our iRobot to take in motion commands after weeks of debugging.
Andy worked on fine-tuning the color thresholding algorithm as well as experimenting with downsampling and morphological contours. In the video below it shows how a robot would move and detect objects it desires to collect. The numbers below each object signifies the “vertical” distance in the camera frame that the objects are away from the robot.
Hojun tried various methods to get the iRobot base to take in motion commands such as using the mini-Din serial pin with the Xavier and the laptop and using the Arduino on the DB25 port. After 3 weeks into it the group decided that using Omar’s backup robot would be better going forward since we are running out of time for integration and need to proceed straight away. Hojun has sent over the relevant parts regarding the robot arm and the gripper to Omar and will help Omar do integration. He will also create a revised CAD model.
Omar has asked for help from a robotics mentor of his from the past because integrating all the components in the short amount of time would be a challenge for a single person. He has begun integration of the gripping mechanism onto his robot.