Learning to Coordinate Controllers - Reinforcement Learning on a Control Basis

By Manfred Huber, R. Grupen. 1997

Subfields: Reinforcement Learning, Robotics & Control

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Reinforcement Learning

In Robotics & Control

Citations in the corpus

Learning Parameterized Skills (2012, background): “This framework effectively defines a control basis that can be used to construct hierarchical motor control plans and is strongly related to the options framework, though it was proposed first [28].”
Learning grounded finite-state representations from unstructured demonstrations (2015, background): “In the control basis framework (Huber and Grupen, 1997), a graph of admissible behaviors is automatically built based on the predicates and constraints of multiple hand-crafted controllers, allowing safe composite policies to be learned.”
Intrinsically motivated hierarchical manipulation (2008, background): “Previous work by Huber [12], [11] and Platt [17] addresses how to direct exploration using constraints on A and how such programs generalize to new contexts.”
Incremental Semantically Grounded Learning from Demonstration (2013, background): “In the control basis framework [8], a graph of admissible behaviors is automatically built based on the predicates and constraints of multiple hand-crafted controllers, allowing safe composite policies to be learned.”
Improving Grasp Skills Using Schema Structured Learning (2006, methodology): “One way to automatically learn the correct sequence of primitives is to encode the problem as a Markov Decision Process, and solve it using Reinforcement Learning [1], [2].”
Intrinsically Motivated Reinforcement Learning: A Promising Framework for Developmental Robot Learning (2005, background): “There are several examples of layering RL on top of a behavioral basis, e.g. ( Huber & Grupen 1997 ).”
A Survey of Collectives (2004, background): “Many problems ranging from controlling a robot’s gait to controlling a chemical plant to allocating constrained resource have been addressed with considerable success using RL [97, 114, 166, 186, 247].”
From Skills to Symbols: Learning Symbolic Representations for Abstract High-Level Planning (2018, background): “…of control so that an agent can learn and plan using higher-level skills, echoing and modeling similar approaches in robotics based on building and sequencing a collection of low-level motor controllers (Nilsson, 1984; Malcolm & Smithers, 1990; Brooks, 1991; Huber & Grupen, 1997; Gat, 1998).”