Simulation of large and complex systems plays an increasingly important role with knowledge-based measurement and control. In real-time situations the utmost in computing speed is required. Parallel computations are necessary to answer the need for ultra-speed at compecting costs. An appreciable cost reduction can be obtained by a certain degree of specialization in the computer: the algorithm oriented processor. A very fruitful specialized parallel architecture is the linear processor array (LPA). The costs are at the minicomputer level. This paper describes the concept and the implementation of LPA's. As an illustrative application the minimum path problems for large vehicle traffic networks (e.g. 20000 nodes) are presented. In that problem use is made of a traffic-assignment algorithm in space-time and a variety of minimum path strategies are considered for their effectiveness in computing time. Real-time control is indeed feasible.
The 'linear processor array' (LPA) as an efficient tool for real-time traffic control in large networks
Das sogenannte lineare Prozessor-Array (LPA) als effizientes Werkzeug für eine Echtzeit-Verkehrsüberwachung in großen Netzen
1990
7 Seiten, 2 Bilder, 3 Quellen
Conference paper
English
ECHTZEITBETRIEB , DATENVERARBEITUNGSGESCHWINDIGKEIT , KOSTEN , PARALLELVERARBEITUNG , ARRAY-PROZESSOR , LINEARES SYSTEM , VERKEHRSLENKUNG , STRASSENVERKEHR , IMPLEMENTIERUNG , NETZSTRUKTUR (ELEKTRISCH) , KOMPLEXES SYSTEM , MESSEN , SIMULATION , EXPERTENSYSTEM , DIGITALE SIGNALVERARBEITUNG , RECHNERARCHITEKTUR
Integrated Memory Array Processor and Real-Time Vision system for Vehicle Control
British Library Conference Proceedings | 1994
|Neural Networks for Real-Time Traffic Signal Control
Online Contents | 2006
|Neural Networks for Real-Time Traffic Signal Control
IEEE | 2006
|Real-Time Detection of Traffic Signs on a Multi-Core Processor
British Library Conference Proceedings | 2008
|