A probabilistic object-oriented data model

Intelligent systems such as autonomous robots interacting with the real world encounter uncertainty in their knowledge about the world, due to incomplete information, data acquisition and measurement errors, and uncertainty about cause and effect. To operate in such an environment, uncertain information needs to be represented explicitly in their world model (database). We outline an object-oriented data model that describes uncertainty through the use of (Bayesian) probabilities. The model represents uncertainty with respect to values of attributes, and with respect to the class hierarchy.