Intelligent Systems And Their Societies

ATTILA

ATTILA was written by Rodney Brooks of the MIT Artificial Intelligence Laboratory. It was described in the SIGART journal of August '91.

This is a curious Intelligent System (IS) because it is so different from all the others described in this chapter. It is an artificial insect, 35 cm long and 30 cm wide, having six legs. It can move around and recharge its batteries. It has a Signetics 68070 principal processor and 10 secondary processors. It has 150 sensors and 23 actuators. It also uses an infrared range sensor and a CCD camera together with a 25 cm long, movable whisker.

Its main objective is to explore its environment and its learning consists in noting and storing landmarks. Once it has learned its environment, a person can indicate towards which landmark it should move, and it does so, choosing a safe, collision free, path. This is done by activating the landmark representation, which in turns activates near landmarks, although to a lower level.

ATTILA's movements are not coordinated by a central brain. Each actuator has its own response rules (called responses) that connect a sensation to a response. It uses "hormones" (the numerical value of a variable) that originate in the body or the brain, and which intensify or diminish the activity of a response rule. They also influence sensors and processors. Different hormones can be active at the same time and reinforce each other, influencing the activity of a certain leg. Several response rules are also active at the same time. There are some more general response rules such as standing up and sleeping, composed of many lower level response rules. All response rules are present at the start, wired or programmed. In this sense ATTILA is more similar to an insect than to a higher animal, which can learn its responses.

For continuos reading, like a book - continue here
Jump to the top of the ATTILA document /Artificial Intelligent Systems /e-book Contents.