The Humanoid Perspective: Adaptive peripersonal space representation and motor learning for a humanoid robot

In recent years, humanoid robotics have been the object of growing interest and investments in the academic and industrial worlds. And indeed, autonomous robots are announced to imminently enter our households. However, beyond the boastful declarations, many challenges still remain to be met on this way. This book focuses on one of the most fundamental ones, the question of action-oriented space representation and motor learning. Highlighting the main difficulties encountered by researchers, it shows how taking inspiration from the human space representation and motor control can lead to original robotics solutions. It suggests a way to enable a humanoid to learn how to control its limbs and to learn simple tasks from human demonstrations. In doing so, the book explores a fruitful but difficult dialogue between engineering and cognitive sciences, illustrating its potential benefits and pitfalls. This book is intended to people wishing to have a more precise idea of current advances in humanoid space representation and motor learning, as well as for scientists who may find it a useful source of inspiration for their own research.