Geometric Task Networks: Learning Efficient and Explainable Skill Coordination for Object Manipulation

TL;DR

This paper introduces Geometric Task Networks (GTNs), a hierarchical planning framework that learns from exhaustive planners to enable efficient, explainable, and generalizable skill coordination for complex object manipulation tasks.

Contribution

It presents a novel hierarchical and compositional planning framework that learns GTNs automatically, improving efficiency, performance, and transparency in robotic manipulation.

Findings

Significant improvement in offline learning efficiency

Enhanced online performance in manipulation tasks

Increased transparency of decision-making process

Abstract

Complex manipulation tasks can contain various execution branches of primitive skills in sequence or in parallel under different scenarios. Manual specifications of such branching conditions and associated skill parameters are not only error-prone due to corner cases but also quickly untraceable given a large number of objects and skills. On the other hand, learning from demonstration has increasingly shown to be an intuitive and effective way to program such skills for industrial robots. Parameterized skill representations allow generalization over new scenarios, which however makes the planning process much slower thus unsuitable for online applications. In this work, we propose a hierarchical and compositional planning framework that learns a Geometric Task Network (GTN) from exhaustive planners, without any manual inputs. A GTN is a goal-dependent task graph that encapsulates both the transition relations among skill representations and the geometric constraints underlying these transitions. This framework has shown to improve dramatically the offline learning efficiency, the online performance and the transparency of decision process, by leveraging the task-parameterized models. We demonstrate the approach on a 7-DoF robot arm both in simulation and on hardware solving various manipulation tasks.