Proceedings:
Knowledge Representation and Reasoning
Volume
Issue:
Proceedings of the AAAI Conference on Artificial Intelligence, 17
Track:
Reasoning about Actions and Time
Downloads:
Abstract:
High-level robot controllers in realistic domains typically deal with processes which operate concurrently, change the world continuously, and where the execution of actions is event-driven as in ``charge the batteries as soon as the voltage level is low''. While non-logic-based robot control languages are well suited to express such scenarios, they fare poorly when it comes to projecting, in a conspicuous way, how the world evolves when actions are executed. On the other hand, a logic-based control language like congolog, based on the situation calculus, is well-suited for the latter. However, it has problems expressing event-driven behavior. In this paper, we show how these problems can be overcome by first extending the situation calculus to support continuous change and event-driven behavior and then presenting ccgolog, a variant of congolog which is based on the extended situation calculus. One benefit of ccgolog is that it narrows the gap in expressiveness compared to non-logic-based control languages while preserving a semantically well-founded projection mechanism.
AAAI
Proceedings of the AAAI Conference on Artificial Intelligence, 17