Research Background
We are amidst one of the greatest technology eras of all time. From cell phones, to smart phones, to laptops, and now jumping robots, technology has advanced in the past 20 years more than it ever has before. Right now however, is a unique time period for Robotics in this great technology boom, as technology's focus is beginning to shift to automation. More and more data is being collected, analyzed, and learned. Robotics is at the forefront of this shift, learning from mass amounts of data and through powerful computation how to complete tasks only humans could complete, or couldn't complete at all.
What is Sokoban?
Sokoban is a type of transport puzzle, where the player pushes boxes in warehouse, trying to get them to a storage location. It’s a strategy game that was created in 1981 by Hiroyuki Imabayashi. Sokoban can be studied using the theory of computational complexity, and the problem of solving Sokoban puzzles has been proven to be NP-hard. There is also a practical aspect to this game; solving Sokoban can be compared to the automated planning that needs to be done by a robot to move boxes in a warehouse. That is what we did in this project.
What does the project do?
The main goal of our project is to simulate the Sokoban game using a Baxter robot. We have two "game" modes - automatic and manual. In the automatic game, Baxter plays itself using a search and move algorithm that we gave it, and in the manual game, the user can play himself using the keyboard.
The project is made up of three parts:
1) Map building
2) Path planning
3) Moving.
Though a fairly simple skeleton, there were numerous interesting challenges along the way, most of which you will read about in our Conclusion. Overall, the project was enlightening, showing us that a seemingly simple problem has so much complexity. At the same, it’s not just a project to play for fun, but also a meaningful task that can be transferred to actual applications, such as express delivery and warehouse management.
The project is made up of three parts:
1) Map building
2) Path planning
3) Moving.
Though a fairly simple skeleton, there were numerous interesting challenges along the way, most of which you will read about in our Conclusion. Overall, the project was enlightening, showing us that a seemingly simple problem has so much complexity. At the same, it’s not just a project to play for fun, but also a meaningful task that can be transferred to actual applications, such as express delivery and warehouse management.