Computational Technology Resources - CCP

X. Jin*, R.-L. Cheu*, D. Srinivasan+, K.-C. Ng$, Y.-L. Ng$ and K.-H. Lee$

*Department of Civil Engineering
+Department of Electrical Engineering, National University of Singapore
$Navigation Systems Department, CET Technologies Pte. Ltd., Singapore

doi:10.4203/ccp.61.9.3

purchase the full-text of this paper

X. Jin, R.-L. Cheu, D. Srinivasan, K.-C. Ng, Y.-L. Ng, K.-H. Lee, "Artificial Intelligence Techniques in the Development of Expressway Incident Detection Systems", in B. Kumar, B.H.V. Topping, (Editors), "Novel Design and Information Technology Applications for Civil and Structural Engineering", Civil-Comp Press, Edinburgh, UK, pp 235-242, 1999. doi:10.4203/ccp.61.9.3

Abstract

The applications of genetic algorithm (GA) and probabilistic neural network (PNN) in the development of expressway automated incident detection (AID) models are presented. To date, AID model development is impeded by insufficient training data, and lack of model adaptability when site condition changes. This paper investigates these two problems in the development framework, and proposes solutions based on artificial intelligence techniques. For the first problem, GA is used to calibrate the system parameters of a traffic simulation model against real traffic data. The method of using the calibrated model to simulate incident database is also presented. In the second problem, the site adaptation potential of an AID model that has been developed based on PNN network is investigated. A set of network adaptation procedures is proposed to improve the model performance when the trained PNN is transferred to a new site that has a different traffic condition. The skeleton of this study could serve as the basis for developing an adaptive AID model for future real-time applications.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £66 +P&P)

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Proceedings ISSN 1759-3433 CCP: 61 NOVEL DESIGN AND INFORMATION TECHNOLOGY APPLICATIONS FOR CIVIL AND STRUCTURAL ENGINEERING Edited by: B. Kumar and B.H.V. Topping Paper IX.3 Artificial Intelligence Techniques in the Development of Expressway Incident Detection Systems X. Jin, R.-L. Cheu, D. Srinivasan+, K.-C. Ng$, Y.-L. Ng$ and K.-H. Lee$ *Department of Civil Engineering +Department of Electrical Engineering, National University of Singapore $Navigation Systems Department, CET Technologies Pte. Ltd., Singapore doi:10.4203/ccp.61.9.3 purchase the full-text of this paper Full Bibliographic Reference for this paper X. Jin, R.-L. Cheu, D. Srinivasan, K.-C. Ng, Y.-L. Ng, K.-H. Lee, "Artificial Intelligence Techniques in the Development of Expressway Incident Detection Systems", in B. Kumar, B.H.V. Topping, (Editors), "Novel Design and Information Technology Applications for Civil and Structural Engineering", Civil-Comp Press, Edinburgh, UK, pp 235-242, 1999. doi:10.4203/ccp.61.9.3 Abstract The applications of genetic algorithm (GA) and probabilistic neural network (PNN) in the development of expressway automated incident detection (AID) models are presented. To date, AID model development is impeded by insufficient training data, and lack of model adaptability when site condition changes. This paper investigates these two problems in the development framework, and proposes solutions based on artificial intelligence techniques. For the first problem, GA is used to calibrate the system parameters of a traffic simulation model against real traffic data. The method of using the calibrated model to simulate incident database is also presented. In the second problem, the site adaptation potential of an AID model that has been developed based on PNN network is investigated. A set of network adaptation procedures is proposed to improve the model performance when the trained PNN is transferred to a new site that has a different traffic condition. The skeleton of this study could serve as the basis for developing an adaptive AID model for future real-time applications. purchase the full-text of this paper (price £20) go to the previous paper go to the next paper return to the table of contents return to the book description purchase this book (price £66 +P&P)
Back to top	©Civil-Comp Limited 2023 - terms & conditions