Some aspects of robots are influenced by nature. It is sensible therefore to study ‘real’ life.
Living systems generally:
- Have some built-in instincts
- But are able to learn - a feedback process
- Can communicate in some fashion
- And are able to reproduce
- And over many generations they evolve
Those who study robots utilise many of these concepts.
Most animals have some form of a brain. They are ‘intelligent’ to an extent, but different animals need different ‘intelligence’.
The human brain developed on African plains to enhance tracking and understanding of animals. Since then human brains have progressed greatly for other tasks - e.g. we play chess.
Most other animals have developed in order to survive, find food and avoid being eaten.
Whilst animals can learn to do better (and we will explain learning later), much of this is based on simple built-in behaviour - instincts - which can be achieved with very simple brains.
These simple brains have neurons which are simple processing elements which take information from sensors, say, and cause something to happen. Let’s illustrate this with a video showing three robots, so called Braitenberg vehicles, which take information from light sensors and so determine how their motors should turn.
The neurons are set so these vehicle’s instinctive behaviour is that they look for lights (they are called light seekers) or they avoid lights (they are called light phobes). Watch the video in the next Step to find out more.
© University of Reading