TOWEL FOLDING BAXTER

About our Project

Our project is interesting because it takes automates a task that we are all familiar with - folding clothing. It also has an interesting component of human-robot interaction, an emerging field, with the human providing the piece of clothing and the robot performing the action. Ultimately, implementing every aspect of this project would simulate a complex task from beginning to end, automating a process that we are all familiar with because everyone does laundry on a regular basis.

Our project requires us to solve a number of interesting problems in order for it to function properly. We need to use sensing in order identify important waypoints in the clothing. This requires the use of several sensing techniques ranging from edge detection to corner detection to convolutional neural networks. Baxter would have to plan the paths in order to fold the clothes correctly without colliding with itself or the table. This of course would require the translation of the real-world coordinates of the towel into the frame of Baxter. Finally, it has to use actuation in order to execute all of the folds, all without damaging itself or the clothes.

To learn more about why our project is interesting and the real-world robotics applications, visit the Large Picture page.

Inspiration

The inspiration of this project comes from past research projects done by UC Berkeley and IEEE. We also wanted this to be practical for users and thought about people who may have a hard time with physical movement. We all have to do laundry week after week, while the folding process is repetitive and time consuming. It is particularly difficult for those with arthritis and other disabilities. With these robots surrounding us in class, we hope to utilize them to greatly improve the lives of individuals.