Robotics/Computer Control/Control Architectures

From Wikibooks, open books for an open world
< Robotics
Jump to: navigation, search

Robots need some form of artificial intelligence in order to be able to navigate around a room or to solve a problem. Such AI can range from simple, as in following particular rules when a specific input is received, through more complicated, as in fuzzy logic rules or simple neural nets, to complicated, as in state of the art machine learning methods.

How complicated you can make your AI depends on a few factors:

  • Processing power: How fast is your microprocessor, is it 8-, 16-, or 32-Bit, How many microprocessors do you use?
  • Available memory: Both temporary and "permanent" memory. Limited amounts of memory restrict how much data can be held at a time, e.g. a map of the surroundings, sensor history, movement history. Systems with much permanent storage and little ram are slower than systems with large amounts of ram.
  • Your knowledge: AI is a very broad subject. Parts of it are very math intensive and can be tricky to understand.