A combined adaptive/fuzzy network-control and its application within the Munich open system architecture

Abstract

Urban traffic control systems mainly apply two approaches: an adaptive control which uses traffic and efficiency models to perform a numerical optimization of a pre-defined performance index; and a rule-based control in which an expert designs a specific knowledge base to detect traffic situations together with explicit control strategies to deal with them. Both methods have their particular benefits and weaknesses especially regarding the definition of user objectives. This article presents an approach to combine these diverse philosophies and describes its implementation for the new open system architecture of the Bavarian capital Munich.