Similarity-Aware Skill Reproduction based on Multi-Representational Learning from Demonstration
This work addresses a domain-specific problem for robotics by improving skill generalization in LfD, though it appears incremental as it builds on existing LfD methods with a novel similarity-aware approach.
The paper tackles the problem of inconsistent similarity in skill reproductions when generalizing learned skills over varying boundary conditions in Learning from Demonstration (LfD), proposing a multi-representational framework that improves generalization by selecting reproductions with the highest similarity values, validated in 3 simulated and 4 real-world experiments with a 6-DOF robot and evaluating 11 similarity metrics in 286 simulated experiments.
Learning from Demonstration (LfD) algorithms enable humans to teach new skills to robots through demonstrations. The learned skills can be robustly reproduced from the identical or near boundary conditions (e.g., initial point). However, when generalizing a learned skill over boundary conditions with higher variance, the similarity of the reproductions changes from one boundary condition to another, and a single LfD representation cannot preserve a consistent similarity across a generalization region. We propose a novel similarity-aware framework including multiple LfD representations and a similarity metric that can improve skill generalization by finding reproductions with the highest similarity values for a given boundary condition. Given a demonstration of the skill, our framework constructs a similarity region around a point of interest (e.g., initial point) by evaluating individual LfD representations using the similarity metric. Any point within this volume corresponds to a representation that reproduces the skill with the greatest similarity. We validate our multi-representational framework in three simulated and four sets of real-world experiments using a physical 6-DOF robot. We also evaluate 11 different similarity metrics and categorize them according to their biases in 286 simulated experiments.