The Nut Sorting Machine
The Nut Sorting Machine is a project that challenged each team to make the fastest yet accurate sorting machine possible. Each team is given the same sensor and objects to sort. The grading scheme is scaled based on the fastest yet accurate machine of all classes combined.
Constraints
-
Only sensors given in the tool kit can be used. (Given sensors include: distance sensor, ON/OFF switch, 10 DOF sensor, pressure sensor and temperature sensor).
-
The code for the machine cannot be hardcoded. Judges reserves the right to change the order of the sorting at anytime.
-
The machine must fit within a 1m x 1m square.
-
The nuts must be dropped into a cup to count as a successful sort.
Design
My machine is designed based on a wheel where the sensor senses the nut just before the sorting and makes the decision for the arm to move as the nut is moved towards it. The sensors I used were a distance sensor, ON/OFF switch and 10 DOF sensor.
The logic is as follows:
IF the switch is ON then the nut must be a big brass nut or big nylon nut. The ON/OFF switch is purposely set at the height where it will only trigger at the height of big nuts.
IF the magnetic sensor is triggered, then the nut must be a small steel nut or copper nut. The reading of the magnetic field is taken to determine which.
IF the distance sensor is triggered, then the nut must be a small brass nut, small steel nut, big brass nut or big nylon nut. The reading for the distance is taken since each material reflects the IR differently.
Reflection
My project achieved an 99% total grade as it was the second fastest sorting done in the course. Some issues I faced during this project is that the 10 DOF sensor runs with a slight lag every time is takes a reading. Eventually the lag will become so great that the sorting would not be accurate. This issue is due to the way the 10 DOF gets input and how it is cached in the Arduino. This issues is mitigated during my trials by repeatedly uploading the code to refresh its cache. For future trials, I would research a way to reduce the cache lag possibly through a code and also integrate more sensors.