This paper presents a set of methods by which a learning agent can learn a sequence of increasingly abstract and powerful interfaces to control a robot whose sensorimotor apparatus and environment are initially unknown. The result of the learning is a rich hierarchical model of the robot's world (its sensorimotor apparatus and environment). The learning methods rely on generic properties of the robot's world such as almost-everywhere smooth effects of motor control signals on sensory features. At the lowest level of the hierarchy, the learning agent analyzes the effects of its motor control signals in order to define a new set of control signals, one for each of the robot's degrees of freedom. It uses a generate-and-test approach to define sensory features that capture important aspects of the environment. It uses linear regression to learn models that characterize context-dependent effects of the control signals on the learned features. It uses these models to define high-level control laws for finding and following paths defined using constraints on the learned features. The agent abstracts these control laws, which interact with the continuous environment, to a finite set of actions that implement discrete state transitions. At this point, the agent has abstracted the robot's continuous world to a finite-state
|
6327
|
Introduction to algorithms
– Cormen, Leiserson, et al.
- 1990
|
|
2268
|
A robust layered control system for a mobile robot
– Brooks
- 1986
|
|
1161
|
Robot Vision
– Horn
- 1986
|
|
1055
|
Self-organization and associative memory
– Kohonen
- 1984
|
|
984
|
Learning from Delayed Rewards
– Watkins
- 1989
|
|
442
|
Multivariate Analysis
– Mardia, Kent, et al.
- 1979
|
|
394
|
Learning regular sets from queries and counterexamples
– Angluin
- 1987
|
|
392
|
Integrated architectures for learning, planning, and reacting based on approximating dynamic programming
– Sutton
- 1990
|
|
390
|
Neuronlike Adaptive Elements That Can Solve Difficult Learning Control Problems
– Barto, Sutton, et al.
- 1983
|
|
362
|
A robot exploration and mapping strategy based on a semantic hierarchy of spatial representations
– Kuipers, Byun
- 1991
|
|
288
|
A simplified neuron model as a principal component analyzer
– Oja
- 1982
|
|
243
|
Integration of Representation into Goal-Driven Behavior-Based Robots
– Mataric
- 1992
|
|
213
|
Forward models: Supervised learning with a distal teacher
– Jordan, Rumelhart
- 1992
|
|
188
|
Modeling spatial knowledge
– Kuipers
- 1978
|
|
183
|
Complexity of automaton identification from given data
– Gold
- 1978
|
|
182
|
Reinforcement Learning for Robots Using Neural Networks
– Lin
- 1993
|
|
173
|
Cooperative computation of stereo disparity
– Marr, Poggio
- 1976
|
|
140
|
Topological mapping for mobile robots using a combination of sonar and vision sensing
– Kortenkamp, Weymouth
- 1994
|
|
139
|
Inference of finite automata using homing sequences
– Rivest, Schapire
- 1993
|
|
130
|
Made-up minds: A constructivist approach to artificial intelligence
– Drescher
- 1991
|
|
115
|
Robotic exploration as graph construction
– Dudek, Jenkin, et al.
- 1991
|
|
108
|
Neural Computation and Self-Organizing Maps: An Introduction
– Ritter, Martinetz, et al.
- 1992
|
|
89
|
On the complexity of minimum inference of regular sets
– Angluin
- 1978
|
|
84
|
A robust, qualitative method for robot spatial learning
– Kuipers, Byun
- 1988
|
|
80
|
Navigation and mapping in large-scale space
– Kuipers, Levitt
- 1988
|
|
76
|
Autonomous Learning from the Environment
– Shen
- 1993
|
|
72
|
Automatic Control Systems
– Kuo
- 1991
|
|
61
|
Navigating With a Rat Brain: A Neurobiologically-Inspired Model for Robot Spatial Representation
– Mataric
- 1991
|
|
59
|
Principles of Multivariate Analysis: A User’s Perspective
– Krzanowski
- 1988
|
|
58
|
Continual Learning in Reinforcement Environments
– Ring
- 1994
|
|
37
|
Learning to explore and build maps
– Pierce, Kuipers
- 1994
|
|
36
|
Inferring finite automata with stochastic output functions and an application to map learning
– Dean, Angluin, et al.
- 1992
|
|
32
|
A new approach to unsupervised learning in deterministic environments
– Rivest, Schapire
- 1987
|
|
26
|
Rule creation and rule learning through environmental exploration
– Shen, Simon
- 1989
|
|
26
|
Representing knowledge of large-scale space
– Kuipers
- 1977
|
|
25
|
The need for constructive induction
– Matheus
- 1991
|
|
25
|
Learning Multiple Goal Behavior via Task Decomposition and Dynamic Policy Merging
– Whitehead, Karlsson, et al.
- 1993
|
|
23
|
Learning from Delayed Reinforcement In a Complex Domain
– Chapman, Kaelbling
- 1991
|
|
22
|
On automated scientific theory formation: A case study using the AM program
– Lenat
- 1979
|
|
21
|
A simpli ed neuron model as a principal component analyzer
– Oja
- 1982
|
|
19
|
Reinforcement-Learning Connectionist Systems
– Williams
- 1987
|
|
18
|
Inference of nite automata using homing sequences
– Rivest, Schapire
- 1993
|
|
17
|
Complexity of automaton identi cation from given data
– Gold
- 1978
|
|
15
|
Functional transformations in AI discovery systems
– Shen
- 1990
|
|
14
|
Connectionist modeling and control of finite state environments
– Bachrach
- 1992
|
|
13
|
Evangelos Kokkevis, and Oded Maron. Inferring finite automata with stochastic output functions and an application to map learning
– Dean, Angluin, et al.
- 1992
|
|
12
|
Learning hill-climbing functions as a strategy for generating behaviors in a mobile robot
– Pierce, Kuipers
- 1991
|
|
10
|
Uncertainty in graph-based map learning
– Dean, Basye, et al.
- 1993
|
|
10
|
Learning a set of primitive actions with an uninterpreted sensorimotor apparatus
– Pierce
- 1991
|
|
4
|
Learning turn and travel actions with an uninterpreted sensorimotor apparatus
– Pierce
- 1991
|
|
