Cognitive robotics

Abstract

Cognitive robotics presents unique research challenges as it straddles the boundaries between two very different fields. Its name promises robots that exhibit cognitive behaviour and suggests that the research would be beneficial and interesting to researchers in both robotics and cognitive science. However, this is not generally the case. Many roboticists are disinterested in the developments of cognitive robots unless these robots are proven to be superior in terms of speed, efficiency and accuracy. Similarly, cognitive scientists are disinterested in the work of roboticists unless the robot’s behavior is fit to some known empirical data and the models implemented are cognitively plausible. These different requirements, each from their own field, are often at odds, and has hampered the development of cognitive robotics, causing researchers to use robots only as a platform for simulating cognitive ideas or to use cognitive science as a weak source of ideas for robot mapping. In this article, we argue that a better synthesis of ideas from both fields must be encouraged and that cognitive robotics should move beyond its artificial limitations and in the process better serve robotics and cognitive science.