Developmental robotics


Developmental robotics

Developmental Robotics (DevRob), sometimes called epigenetic robotics, is a methodology that uses metaphors from neural development and developmental psychology to develop the mind for autonomous robots. The focus is on a single or multiple robots going through stages of autonomous mental development (AMD). Researchers in this field study artificial emotions, self-motivation, and other methods of self-organization. The program that simulates the functions of genome to develop a robot's mental capabilities is called a developmental program.

Different from traditional machine learning, some major features of developmental robotics are:

  • Task-nonspecificity: Since it is difficult for the genome to predict what tasks the baby will learn and perform in his life, the developmental program is body-specific (species specific) but not task-specific.
  • Environmental openness: Due to the task-nonspecificity, AMD must deal with unknown and uncontrolled environments, including various human environments.
  • Raw sensors: AMD must directly deal with continuous raw signals from sensors (e.g., vision, audition and touch), since different tasks require different information in the sensors. Only raw signals have all.
  • Online processing: At each time instant, what the machine will sense next depends on what the machine does now.
  • Incremental processing: Acquired skills must be used to assist in the acquisition of new skills, as a form of scaffolding. This requires incremental processing.

DevRob is related to, but differs from, evolutionary robotics (ER). ER uses populations of robots that evolve over time, whereas DevRob is interested in how the organization of a single robot's control system develops through experience, over time.

DevRob is also related to work done in the domains of Robotics, Artificial Life.

Contents

History

Cresceptron (ICCV 1992, IJCV 1997) was the first published developmental learning method for detecting and recognizing a general object in a complex natural background based on a composite image view. It also segments the detected object from the complex natural background. Human-machine interactions through the sensory-end and the motor-end teach the Cresceptron, while the internal self-organization is fully autonomous.

The NSF/DARPA funded Workshop on Development and Learning was held April 5–7, 2000 at Michigan State University. It was the first international meeting devoted to computational understanding of mental development by robots and animals. The term "by" was used since the agents are active during development.

DevRob was explained in Weng et al. Autonomous mental development by robots and animals. Science 291:599-600, 2001. Its major uniqueness is the task nonspecificity concept of a new kind of program: developmental program (DP). A DP simulates the developmental functions of the "genome".

The first undergraduate courses in DevRob were offered at Bryn Mawr College and Swarthmore College in the Spring of 2003 by Douglas Blank and Lisa Meeden, respectively.

The first graduate course in DevRob was offered at Iowa State University by Alexander Stoytchev in the Fall of 2005.

Main Journals

Main Conferences

External links

Academic institutions and researchers in the field

Blogs and Other Links


Wikimedia Foundation. 2010.