Zero-Shot Generalization from Motion Demonstrations to New Tasks

2026-03-16

AI Summary

Existing dynamical systems approaches struggle to generalize from demonstrations of isolated tasks to novel tasks, lacking the ability to reuse knowledge across tasks. This work proposes a Gaussian graph structure that models the spatial distribution of movement primitives as graph nodes and integrates graph search with continuous control to design two compositional frameworks—Stitching and Chaining. For the first time, this approach enables zero-shot task generalization of dynamical systems within a shared workspace while preserving convergence guarantees. Experimental results demonstrate that the method effectively generalizes to unseen tasks in both simulation and real-world robotic platforms, significantly outperforming existing baselines.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Mixed Discrete/Continuous Planning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Learning motion policies from expert demonstrations is an essential paradigm in modern robotics. While end-to-end models aim for broad generalization, they require large datasets and computationally heavy inference. Conversely, learning dynamical systems (DS) provides fast, reactive, and provably stable control from very few demonstrations. However, existing DS learning methods typically model isolated tasks and struggle to reuse demonstrations for novel behaviors. In this work, we formalize the problem of combining isolated demonstrations within a shared workspace to enable generalization to unseen tasks. The Gaussian Graph is introduced, which reinterprets spatial components of learned motion primitives as discrete vertices with connections to one another. This formulation allows us to bridge continuous control with discrete graph search. We propose two frameworks leveraging this graph: Stitching, for constructing time-invariant DSs, and Chaining, giving a sequence-based DS for complex motions while retaining convergence guarantees. Simulations and real-robot experiments show that these methods successfully generalize to new tasks where baseline methods fail.

Problem

• Zero-Shot Generalization
• Motion Demonstrations
• Dynamical Systems
• Task Generalization
• Motion Primitives

Innovation

• Zero-Shot Generalization
• Gaussian Graph
• Dynamical Systems
• Motion Primitives
• Task Composition