Imitation learning focuses on three important issue:
(1)efficient motor learning
(2)the connection between action and perception
(3) modular motor control in the form of movement primitives
In order to overcome the need for tele-operation and manual hard-coding of every behavior, a learning approach is required. The most general approach is reinforcement learning.
At the beginning of the 1980s, the idea of imitation learning started to find increasing attention in the field of manipulator robotics as it seemed to be a promising route to automate the tedious manual programming of robots.
(1)symbolic approaches to imitation learning
Each state becomes a graph node and each action a directed link between two nodes. Symbolic reasoning could now unify different graphical representations for the same task by merging and deleting nodes such that the final program for imitating the demonstrated movement was achieved.
(2)inductive approaches to imitation learning
(3)imitation learning of novel behaviors
several research projects have focused on imitation as a method of creating novel behaviors. Three major approaches can be distinguished:
(a) learning a control policy directly: task-level learning.
(b)learning from demonstrated trajectories
(c)model-based imitation learning
Imitation learning focuses on three important issue: (1)efficient motor learning (2)the connection between action and perception (3) modular motor control in the form of movement primitives In order to overcome the need for tele-operation and manual hard-coding of every behavior, a learning approach is required. The most general approach is reinforcement learning.
At the beginning of the 1980s, the idea of imitation learning started to find increasing attention in the field of manipulator robotics as it seemed to be a promising route to automate the tedious manual programming of robots.
(1)symbolic approaches to imitation learning Each state becomes a graph node and each action a directed link between two nodes. Symbolic reasoning could now unify different graphical representations for the same task by merging and deleting nodes such that the final program for imitating the demonstrated movement was achieved.
(2)inductive approaches to imitation learning
(3)imitation learning of novel behaviors several research projects have focused on imitation as a method of creating novel behaviors. Three major approaches can be distinguished: (a) learning a control policy directly: task-level learning. (b)learning from demonstrated trajectories (c)model-based imitation learning