Arnab Chakraborty's assembly language tutorial: page 7 of 11

cwave.eu5.org
Also see: http://www.angelfire.com/dragon/letstry
cwave04 at yahoo dot com

My Guestbook

Last updated on Wed Dec 22 10:24:57 IST 2010.

Home >

... tutorials >

... ... Assembly language

Artificial intelligence

A robot differs from other machines in that it has "intelligence". Like human intelligence a robot's artificial intelligence also comes in two flavours. Imagine playing a game of chess. You think hard for a clever move, then extend your hand, grab a chess piece and actually make the move. Any one can see how intelligence is required in the first stage: thinking of good move. But intelligence is also required in the second stage, where you use your hand. Here your brain is guiding your hand to reach the piece, then guiding your fingers to curl around it, and press it with just the right amount of force in order to lift it, etc. All the while your eyes are monitoring the movement to send feedback signals to the brain.

Similarly, a robot has two levels of AI:

Concrete: The kind of intelligence used in the second stage above,
Abstract: The kind of intelligence used in thinking about a clever move.

Clearly, a robot with only abstract AI is like a poet who has no tool to write down the verses that occur to him. Pretty useless! So even to experiment with abstract AI we need some concrete AI to start with. A brilliant technique exists to solve all concrete AI needs. The technique is so very powerful and general that there is not a single robot in the world that does not use it over and over again. This technique is called a state machine or deterministic finite automaton (DFA) to use a more impressive term.

Abstract AI, on the other hand, has no unique technique. It changes from robot to robot. We shall discuss a very fundamental technique of abstract AI called graph search. It is widely used, and forms the backbone of many other techniques.

In the next page we shall discuss concrete AI using state machines.