Robot Learning

Building a robot that learns to perform a task has been acknowledged as one of the major challenges facing artificial intelligence. Self-improving robots would relieve humans from much of the drudgery of programming and would potentially allow operation in environments that were changeable or only partially known. Progress towards this goal would also make fundamental contributions to artificial intelligence by furthering our understanding of how to successfully integrate disparate abilities such as perception, planning, learning and action.
Although its roots can be traced back to the late fifties, the area of robot learning has lately seen a resurgence of interest. The flurry of interest in robot learning has partly been fueled by exciting new work in the areas of reinforcement earning, behavior-based architectures, genetic algorithms, neural networks and the study of artificial life. Robot Learning gives an overview of some of the current research projects in robot learning being carried out at leading universities and research laboratories in the United States. The main research directions in robot learning covered in this book include: reinforcement learning, behavior-based architectures, neural networks, map learning, action models, navigation and guided exploration.

1 Introduction to Robot Learning.- 1 Motivation.- 2 The Robot Learning Problem.- 3 Background.- 4 Domains.- 5 Roadmap.- 2 Knowledge-based Training of Artificial Neural Networks for Autonomous Robot Driving.- 1 Introduction.- 2 Network Architecture.- 3 Network Training.- 4 Performance Improvement Using Transformations.- 5 Results and Comparison.- 6 Discussion.- 3 Learning Multiple Goal Behavior via Task Decomposition and Dynamic Policy Merging.- 1 Introduction.- 2 Basics of Reinforcement Learning.- 3 Multiple Goal Tasks.- 4 A Decomposition Approach.- 5 Search-based Merging.- 6 A Hybrid Architecture.- 7 Summary.- 4 Memory-based Reinforcement Learning:Converging with Less Data and Less Real Time.- 1 Introduction.- 2 Prioritized Sweeping.- 3 A Markov Prediction Experiment.- 4 Learning Control of Markov Decision Tasks.- 5 Experimental Results.- 6 Discussion.- 7 Conclusion.- 5 Rapid Task Learning for Real Robots.- 1 Introduction.- 2 Behavior-based Reinforcement Learning.- 3 Exploiting Local Spatial Structure.- 4 Using Action Models.- 5 Highly Structured Learning.- 6 Summary.- 6 The Semantic Hierarchy in Robot Learning.- 1 Introduction.- 2 The Cognitive Map and the Semantic Hierarchy.- 3 From Simulated Robot to Physical Robots.- 4 From Tabula Rasa to Cognitive Mapping.- 5 From Low-Speed to High-Speed Motion.- 6 Conclusions.- 7 Uncertainty In Graph-Based Map Learning.- 1 Introduction.- 2 Qualitative Navigation and Map Learning.- 3 Theoretical Development.- 4 Problem Classification.- 5 Summary of Results.- 6 Conclusions.- 8 Real Robots, Real Learning Problems.- 1 Introduction.- 2 Motivation.- 3 The Main Types of Learning.- 4 The Main Methods of Learning.- 5 Simulation.- 6 Conclusion.