The Short-time Traffic Flow Prediction at Ramp Junction Based on Wavelet Neural Network
- Resource Type
- Conference
- Authors
- Xiao, Jinjian; Xie, Yingna; Wen, Yubo
- Source
- 2021 IEEE 5th Advanced Information Technology, Electronic and Automation Control Conference (IAEAC) Advanced Information Technology, Electronic and Automation Control Conference (IAEAC), 2021 IEEE 5th. :664-667 Mar, 2021
- Subject
- Communication, Networking and Broadcast Technologies
Engineering Profession
Robotics and Control Systems
Signal Processing and Analysis
Training
Roads
Neurons
Time series analysis
Predictive models
Prediction algorithms
Junctions
ramp
traffic flow
wavelet neural network
prediction
- Language
- ISSN
- 2689-6621
In order to accurately analyze the impact of ramp traffic flow on main road traffic flow, it is necessary to predict short-time traffic flow at ramp junction. Based on the comprehensive analysis of the characteristics of short-time distribution of traffic flow at typical ramp junction, 15 minutes was selected as the traffic flow sampling interval, the traffic flow time series data chain is formed. And the structure of wavelet neural network is optimized and trained by means of the gradient method used to modify the network weights and wavelet basis functions. The short-time traffic flow prediction algorithm with high fitting accuracy and no repeated correction is formed. The prediction training, fitting and error correction of the model are carried out by using the 300 traffic flow time series obtained from road monitoring. The real-time prediction analysis shows that the prediction algorithm can improve the matching accuracy of short-time traffic flow.