A review of artificial intelligence systems for site characterization

Site characterisation has become one of the most extensively developed areas of application for artificial intelligence systems in geotechnical engineering. A total number of 33 knowledge-based systems and 14 neural network systems have been reported in the literature. Systems have been developed for site investigation planning, interpreting ground conditions, classification of soil and rock and the interpretation of geotechnical parameters. Techniques used have involved a wide range of techniques: simple rule-based approaches; probabilistic methods; fuzzy sets; object-oriented programming; case-based systems. Many systems are only simple developmental prototypes, with much more work needed before they would prove useful in geotechnical practice. Nevertheless, they do demonstrate the potential of both knowledge-based systems and neural networks in this yield. Hybrid systems which combine different artificial intelligence techniques and other techniques (eg database systems, Geographical Information Systems) will probably prove to be the most useful type of system for site characterisation in the future.