Date of Completion

2022

Document Type

Honors College Thesis

Department

Computer Science

Thesis Type

Honors College

First Advisor

Josh Bongard

Second Advisor

Nick Cheney

Keywords

Soft Robotics, Human-computer Interaction, Citizen Science, Crowd Source, Voxels, Twitch Plays Soft Robotics

Abstract

Scalability has been a long-standing open problem in the field of robotics and especially in the field of soft robotics where there has never been a technology that allows for crowdsourcing the creation and training of soft robots. The crowdsourced approach is very advantageous because it strives to maximize the use of available resources, cut expenses, diversify contributions, and promote transparency (Uhlmann et al., 2019). I posit that by using Voxcraft, Google Colab and Twitch we can not only create a way to crowdsource the creation of simulations of soft robots but also do it in an accessible, replicable, and scalable manner. I not only demonstrate that this can be done for the first time ever but also demonstrate that soft robots can be collectively created using this approach. The crowdsourced, low barrier to entry, replicability and scalability of this approach suggests that Twitch Plays Simulated Soft Robotics (TPSS) can be replicated and built upon in the future to not only get other user inputs such as fitness function or mutation rate to train robots but also scaled up to possibly crowdsource the crowdsourcing of soft robotics.

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.

Download

Share

COinS