Rubikon: Intelligent Tutoring for Rubik's Cube Learning Through AR-enabled Physical Task Reconfiguration

TL;DR

Rubikon introduces an AR-enabled intelligent tutoring system for Rubik's Cube learning that automatically reconfigures physical cube states to target unmastered knowledge components. By combining ArUco-based state tracking with cognitive-tutor design—task model, model tracing, and knowledge tracing—the system delivers real-time feedback and personalized practice. An empirical evaluation shows Rubikon yields higher post-test gains (about 25%) and broader, balanced practice across knowledge components, while reducing preparation effort compared to traditional video tutorials. The work demonstrates the feasibility and value of automated, AR-driven task reconfiguration for 3D physical tasks and outlines generalizable principles for MR-based tutoring beyond the Rubik's Cube.

Abstract

Learning to solve a Rubik's Cube requires the learners to repeatedly practice a skill component, e.g., identifying a misplaced square and putting it back. However, for 3D physical tasks such as this, generating sufficient repeated practice opportunities for learners can be challenging, in part because it is difficult for novices to reconfigure the physical object to specific states. We propose Rubikon, an intelligent tutoring system for learning to solve the Rubik's Cube. Rubikon reduces the necessity for repeated manual configurations of the Rubik's Cube without compromising the tactile experience of handling a physical cube. The foundational design of Rubikon is an AR setup, where learners manipulate a physical cube while seeing an AR-rendered cube on a display. Rubikon automatically generates configurations of the Rubik's Cube to target learners' weaknesses and help them exercise diverse knowledge components. In a between-subjects experiment, we showed that Rubikon learners scored 25% higher on a post-test compared to baselines.

Figures (10)

• Figure 1: Rubikon enables personalized learning through a closed-loop design. While the learner manipulates the cube, the Model Tracing module parses frames streamed from a webcam into cube states. Informed by a Task Model that defines the movement sequences learners need to perform, the Knowledge Tracing module analyzes the transitions between cube states to assess mastery. For knowledge components not yet mastered, the Task Generation module correspondingly samples cube configurations and renders them on the screen for deliberate practice.
• Figure 2: The Rubikon task model outlines 11 essential components for solving the first layer of the Rubik's Cube. Each component corresponds to a specific move sequence necessary to master distinct patterns. Difficulty levels are indicated by the number of *.
• Figure 3: Each knowledge component in Rubikon's task model has a unique pattern of configuration on the Cube. (a) show a specific configuration of the Rubik's Cube corresponding to the knowledge component Side. (b) shows that there are 8 configurations that follow this pattern, which all enable the learners to exercise the knowledge component Side (indicated by the white stars). (c) shows 4 configurations of the knowledge component Back Harder (indicated by the white stars).
• Figure 4: The Rubikon user interface has 5 key sections.
• Figure 5: (a) A physical Rubik's Cube with ArUco markers attached to all squares. (b) A fully-rendered Rubik's Cube on the display with AR rendering. (c) A fully-rendered Rubik's Cube with extended views.