Coffee Learning is a project that targets the incorrect recycling habits of many coffee drinkers, and how we can automate and simplify the process of recycling commercial coffee cups through experiential design and machine learning
Over 2 billion cups of coffee are consumed everyday, worldwide. Among this figure, are college and university students. Investigations into recycling habits at Sheridan college have revealed that a majority of the student body does not dispose of coffee cups correctly, even after Sheridan College implemented recycling measures across campus to better dispose of these cups in the appropriate methods.
How might we automate the recycling process to encourage recycling habits at Sheridan College and simplify the process overall?
This smart recycling bin analyzes the coffee cup components, being the cup, the sleeve, and the lid, and decides for the user, which bin it belongs in. The user places each component one by one on the middle platform, and the camera detects which bin to turn the platform towards.
Ideation for this problem began with primary research on what recycling solutions existed both inside of Sheridan College, and what is being used in large corporations today. Sheridan college has an organic recycling process in place to properly dispose of coffee cups (removing the plastic from the paper cups). However, this process is inhibited by the improper sorting that majority of students partake in. Sheridan encourages students to place the cups into the organics bin, and to not place the entirety of the cup, sleeve and lid in the garbage or recycling bins.
Through research, the team discovered a few different methods of recycling automation.
The target audience for this project is Young adults, aged 17 to 25, attending Sheridan College
After understanding the needs and pain points of Sheridan students, the team began brainstorming, researching, and sketching functional product designs. These designs needed to save the students time, and make it undoubtably clear what choice was correct and remove any necessary problem solving
Initial Sketches
Before deciding to incorporate machine learning, the team experimented with low fidelity prototypes and sketches. These involved custom lids with holes shaped like the desired coffee cup part, circular lids that would rotate based on detected parts, and a screen that would provided colour feedback (Red/Green) based on the users attempt/ anticipated action to further UX experimentation.
Majority of the time spent on this project was spent learning, teaching, and tweaking Google's Teachable Machine. The system needed to be taught what the different components that made up a coffee cup actually were.
The final concept identified 5 states that needed to be recognized by the system. These would include the Fully Assembled Coffee Cup, Cup, Sleeve, Lid, and Place Holder (Empty). Placeholder was necessary to return the system to a default and not in use state. The design also incorporated user feedback in the form of a red light when the Coffee Cup was Fully Assembled, and a green light following the automated movement, when a single component was placed on the platform.