A Robot That Plays Rubik's Race, a Slide Puzzle Game

This paper presents a robotic system with a generalpurpose robotic arm that can play the game of Rubik's race autonomously. Two puzzle-solving algorithms suitable for online implementation are proposed and compared. One is an adaptation of the IDA* search with additive pattern database heuristics that worked well in solving <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$N^{2}-1$</tex> puzzle. Another is a randomized sequential path-planning method. The system is designed to handle the physical challenges of real-time computing and agility required for playing the game. The paper discusses the key parameters of the robotic arm, the challenges encountered and design modifications to make the sliding movement more robust. The system can be adapted to solve other sliding puzzle games and serves as an important step toward the development of a general-purpose physical game-playing robot capable of competing with humans. Because of its low-cost and multidisciplinary nature, it can serve as a great STEM education platform to engage K-12 students and educators.