For many years, building autonomous robots that can learn and gain new capabilities for accomplishing task(s) or adapt to changing environments, has been a major research area in AI, Robotics, and Cognitive Sciences.
Although Machine Learning methods have already been applied to some robotic problems, due to the difficulties of learning in natural environments e.g. partial observability, high-dimentionality, and time-complexity of learning process, few examles of successfull application of them is available for commercially existing robots.

Learning techniques can be applied to various stages of robot control form perception to action e.g. combining multi-modal information, attention control and specifying the relevant sensory information, decision making and high-level planing, action generation and motor control. The growing number of research programs on robot learning shows an ever increasing interest on this field and its application specially to allow robots to be accepted as a companion by human.

In this special issue of Journal of Advanced Robotic Systems on Robot Learning, we intend to outline recent progress in the application of machine learning techniques to robotics. Examples of topics of interest include, but are not limited to:

Bio-inspired robot learning
Evolutionary Robotics
Imitation and learning by demonstration
Reinforcement learning on robots
Incremental Lifelong learning and behavior shaping
Hierarchical task learning
Distributed multi-robot learning
Cooperative and Competitive learning and knowledge transfer between homogeneous and heteregenous robots
Statistical learning and probabilistic reasoning for robotic tasks
Locomotion learning and Environment adpatation
Software and control architecture for robot learning
Attention control and state abstraction
Cognitive robotics