A5079955784- Traffic Prediction and Management Techniques
- Time Series Analysis and Forecasting
- Traffic control and management
- Advanced Computational Techniques and Applications
- Vehicle License Plate Recognition
- Advanced Sensor and Control Systems
- Smart Parking Systems Research
- Water Quality Monitoring and Analysis
- Simulation and Modeling Applications
- Industrial Technology and Control Systems
Bohai University
2018-2020
Jilin University
2013-2014
Short-time traffic flow prediction is necessary for advanced management system (ATMS) and traveler information (ATIS). In order to improve the effect of short-term prediction, this paper presents a multistep method based on similarity search time series. Firstly, landmark model used represent series data. Then input data are determined through searching similar Finally, echo state networks prediction. The performance proposed measured with expressway collected from loop detectors in...
With the rapid economic development, parking problems have become increasingly prominent due to city's development model and emergence of a large number private cars. Parking management departments around world focus on intelligent system in order solve problems, but most them are limited upgrading infrastructure. There is no effective solution from perspective fundamental cause problem. At present, it generally believed that guidance systems can effectively alleviate provide drivers traffic...
To increase the efficiency and precision of large-scale road network traffic flow prediction, a genetic algorithm-support vector machine (GA-SVM) model based on cloud computing is proposed in this paper, which analysis characteristics defects algorithm support machine. In environment, firstly, SVM parameters are optimized by parallel algorithm, then used to predict flow. On basis data Haizhu District Guangzhou City, was verified compared with serial GA-SVM MPI (message passing interface)....
In the process of predicting link travel time, difficultly handling high-dimensional and nonlinear data consistently poses problems to researchers. Because single least squares support vector machine cannot handle these issues, a prediction method based on wavelet package analysis square was proposed forecast time. The verification results obtained from VISSIM simulation show that time is better has definite application value.
To solve the problem of traffic state identification for regional road network which includes large amount data and difficult calculation, improve every link fuzzy C-means clustering algorithm, a method urban was proposed combined with genetic algorithm MapReduce parallel programming model cloud computing. A typical environment twelve intersections set up in VISSIM simulate software, verified by simulated data. Experiment results show that can accurately identify network, greatly efficiency...