基于回归-卡尔曼滤波组合模型的航道整治区域船舶交通流时空预测

Spatio-temporal prediction of ship traffic flow in waterway regulation area based on regression-Kalman filter combination model

为分析及预测工程不同阶段航道整治区域船舶交通流量变化,通过确立研究断面,获取交通流数据,对研究断面处的船舶交通流时空特性进行分析.基于现有数据,将回归模型预测值替代卡尔曼滤波模型中的状态转移值,建立回归-卡尔曼滤波组合模型,在卡尔曼滤波模型预测具有实时性的基础上,提高预测精度,并利用历史数据进行了预测.预测结果与实际数据对比验证了回归-卡尔曼滤波组合模型在航道整治区域船舶交通流时空预测方面的有效性与可靠性.

In order to analyze and predict the changes of ship traffic flow in different stages of waterway regulation area,by establishing the study section and obtaining traffic flow data,the spatio-temporal characteristics of ship traffic flow at the study section were analyzed.Based on the existing data,the state transition value in the Kalman filter model was replaced with the predicted value of the regression model to establish the combined regression Kalman filter model.On the basis of the real-time prediction of the Kalman filter model,the prediction accuracy was improved,and the historical data was used for prediction.By comparing the actual data and the prediction results,the validity and reliability of the regression Kalman filter combination model in the spatio-temporal prediction of ship traffic flow in waterway regulation area have been varified.