Development and adaptation of constructive probabilistic neural network in freeway incident detection |
| |
| Institution: | 1. Department of Civil Engineering, National University of Singapore, Block E1A, #07-03, 1 Engineering Drive 2, Singapore 117576, Singapore;2. Department of Electrical and Computer Engineering, National University of Singapore, Block E4, #05-48, 4 Engineering Drive 3, Singapore 117576, Singapore;1. Science Department, “Roma Tre” University, via della Vasca Navale 84, Rome, Italy;2. E.V.O. srl, Rome 00134, Italy;1. University of Novi Sad, Faculty of Technology Novi Sad, Department of Applied and Engineering Chemistry, Bulevar cara Lazara 1, 21000, Novi Sad, Serbia;2. Institute of Food Technology, University of Novi Sad, Bulevar cara Lazara 1, 21000 Novi Sad, Serbia;1. Department of Physics, Faculty of Sciences, University of Mohaghegh Ardabili, P.O. Box 179, Ardabil, Iran;2. Department of Engineering Sciences, Faculty of Advanced Technologies, University of Mohaghegh Ardabili, Namin, Iran;1. Department of Physics and Astronomy, Ghent University, Belgium;2. Department of General Economics, Ghent University, Belgium;3. University College Roosevelt, The Netherlands;4. Utrecht University School of Economics, The Netherlands |
| |
| Abstract: | This paper investigates the use of constructive probabilistic neural network (CPNN) in freeway incident detection, including model development and adaptation. The CPNN was structured based on mixture Gaussian model and trained by a dynamic decay adjustment algorithm. The model was first trained and evaluated on a simulated incident database in Singapore. The adaptation of CPNN on the I-880 freeway in California was then investigated in both on-line and off-line environments. This paper also compares the performance of the CPNN model with a basic probabilistic neural network (BPNN) model. The results show that CPNN has three main advantages over BPNN: (1) CPNN has clustering ability and therefore could achieve similarly good incident-detection performance with a much smaller network size; (2) each Gaussian component in CPNN has its own smoothing parameter that can be obtained by the dynamic decay adjustment algorithm with a few epochs of training; and (3) the CPNN adaptation methods have the ability to prune obsolete Gaussian components and therefore the size of the network is always within control. CPNN has shown to have better application potentials than BPNN in this research. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
