2019 Fall UIUC ME461 Computer Control of Mechanical Systems course project
Kitchens are changing, incorporating more robotic processes to make life more simple and efficient. Some restaurants have tried completely automating their dishes. Bevier Cafe, located at the University of Illinois, challenged us to create a Waffle Robot, where students could effortlessly make their morning waffles and interact with robotics in the food industry.
The Waffle Robot has 5 modes:
1. Waits for a user
2. Asks a user to load the batter and spray the iron
3. Pours the batter, throws away the cup, and asks the user to close the lid
4. Rotates and cooks the waffle
5. Prompts the user to remove the waffle and waves goodbye
A button allows the user to indicate when they have finished performing prompted tasks. The button indicates if users would like to make a waffle, if users have sprayed the pan, if users have closed the lid, and when users have received the waffle.
The system uses three major components:
MSP 430
OrangePI
Arduino
The MSP430 microcontroller controls the state of the robot and sends this information over I2C to the OrangePI. It receives data from the OrangePI indicating when the waffle has finished cooking. The MSP430 also controls the rotation of the waffle iron.
The OrangePI processes camera data, receives the waffle state from the MSP430, computes inverse kinematics and sends the angular position commands to the Arduino. The number of green pixels is collected from the camera data to indicate if the waffle has finished cooking. Due to fluctuations in lighting, the green pixel count was averaged over 20 images.
This average is then compared to the previous average. If there is a large difference in the readings, this indicates the LED had turned on. If enough green pixels are counted (and the waffle has cooked for at least 90 seconds), the waffle has finished cooking and is rotated to its original position for removal.
Because the OrangePI has the most computational power, it is used to compute the inverse kinematics code and send the angular position data to the Arduino over I2C. Different commands are sent based on the waffle state received from the MSP430. The Arduino then sends these commands to the servomotors in the Sainsmart robot arm.
Lifting the waffle iron with a servo motor required too much torque for the servo motors available for the project. Therefore the tasks of lifting the waffle iron and removing the waffle are given to the user. In addition, the robot arm lifts 2 small cups of batter in order to minimize the torque.
Comments