Urban traffic state detection based on support vector machine and multilayer perceptron
US9037519B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Oct 18, 2012 |
| Grant date | May 19, 2015 |
| Priority date | — |
| Expiry date | Jul 27, 2033 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06N20/20
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
A system and method that facilitates urban traffic state detection based on support vector machine (SVM) and multilayer perceptron (MLP) classifiers is provided. Moreover, the SVM and MLP classifiers are fused into a cascaded two-tier classifier that improves the accuracy of the traffic state classification. To further improve the accuracy, the cascaded two-tier classifier (e.g., MLP-SVM), a single SVM classifier and a single MLP classifier are fused to determine a final decision for a traffic state. In addition, fusion strategies are employed during training and implementation phases to compensate for data acquisition and classification errors caused by noise and/or outliers.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.