Hybrid PLC Mechatronics - Chapter 21 Artificial Intelligence

After a look at statistical forecasting methods, an introduction to relational AI is given. Relational AI is at the core of artificial intelligence programming. Introduced are early tests for machine thought processing as well as discussion of various types of machine intelligence. Questions of ethics and scope of use of AI are discussed as well. A problem using a search agent introduces the decision tree search algorithm. The problem introduced to searching a maze is useful throughout the chapter and becomes one of the few programs easily adapted to the PLC for implementation. More sophisticated search methods are also discussed although their implementation is not included. Probability is introduced as a means to choose a best path through a tree. The section on probability includes the development of Bayes’ Theorem and gives examples of its use. More sophisticated examples of problems solved by AI programs are given but not solved since they are not readily implemented in the programming of the standard PLC instruction set. Knowledge-based expert systems must also consider fuzzy logic and neural network systems. Fuzzy logic deals with fuzzy linguistic terms while neural networks can learn using supervised and unsupervised learning. Once considered part of the core of modern AI, fuzzy logic and neural networks are not as popular as at one time and are not considered as a first solution in most AI problems. Accompanying Video: PLC Series Chapter 21 - Artificial Intelligence (44:08) Hybrid Lab Text: Ch. 27 – The Maze Additional Chapters in Hybrid Lab Text Hybrid Lab Text: Ch. 28 – Festo Ch. 30 – Conveyors (see Additional Public Access to Materials below)