Crowdsourcing Genetic Programming for Pathfinding

A crowdsourced fitness function to train a sonar-equipped robot to avoid obstacles and follow walls.

Similar projects worth following
  • Description
  • Details
  • Files 0
  • Components 0
  • Logs 0
  • Instructions 0
  • Discussion 0
This project was done for my Master's thesis in 2010, and largely ended up under the rug for ten years while the world caught up. It's actually still quite interesting now, especially since this was done with a lot of recycled equipment and literally zero budget.

This led to SEBeLS - State and Event Based Language System - and a genetic programming framework first written in Python (probably a poor choice for efficiency, mind you) to support it. Compiling it with cython, it was deemed "fast enough" - but we didn't have servers.

Enter crowdsourcing. With the help of certain people in the adult entertainment industry, we were able to seed a number of starting individuals per computer based on its capacity, and see what patterns emerged.

The thesis paper details the most interesting parts of the solution, as well as the SEBeLS language. We hope you enjoy.

Enjoy this project?



Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates