Towards a general theory of action and time
Artificial Intelligence
Situation recognition: representation and algorithms
IJCAI'93 Proceedings of the 13th international joint conference on Artifical intelligence - Volume 1
Understanding causal descriptions of physical systems
AAAI'92 Proceedings of the tenth national conference on Artificial intelligence
Hi-index | 0.00 |
This paper presents Event Calculus, a model for representing the identifying characteristics of physical events in terms of changes in a scene and time-related combinations of other physical events. The model is used to construct a knowledge-based system for event recognition which forms a high-level description of changes in a scene, given a low-level description as input. Time-varying information is represented in the form of "GRAPHs", data structures which plot the elements of various domains against time. Several varieties of operations are presented which map GRAPHs into GRAPHs, and representations of physical events are formed as symbolic expressions involving these operations. The paper concludes with an overview of the event recognition system, as implemented in INTERLISP on a VAX 11/780, and an example of a session with this system.