CAJ | 학술논문

针对城市道路流量的非线性和不确定性特点，为避免单一模型预测准确率不高的缺陷，该文提出了一种短时交通流组合模型。该模型包含卡尔曼滤波模型和径向基函数神经网络模型2个子模型，较好地解决了神经网络不能反映大流量下的稳态性问题，以及卡尔曼滤波在流量不稳定时预测准确率不高的问题。在组合模型中引入惯性因子，确保了模型的稳定性。仿真结果表明该方法是可行有效的。
침대성시도로류량적비선성화불학정성특점，위피면단일모형예측준학솔불고적결함，해문제출료일충단시교통류조합모형。해모형포함잡이만려파모형화경향기함수신경망락모형2개자모형，교호지해결료신경망락불능반영대류량하적은태성문제，이급잡이만려파재류량불은정시예측준학솔불고적문제。재조합모형중인입관성인자，학보료모형적은정성。방진결과표명해방법시가행유효적。
In view of that the traffic flow of the urban road is a nonlinear and uncertaint interrupted flow,a hybrid model for the short-term traffic flow is put forward to overcome the shortage of the lower forecasting accuracy of the single model. This model consists of two sub-models, the Kalman filter model and the radical basis function neural network model,so the steady-state problem of the neural network model in the huge traffic flow and the low accuracy problem of the Kalman filter model in the unsteady traffic flow can be all solved. An inertia factor is introduced in the process of combining to ensure the stability of the hybrid model. The simulation result shows that the hybrid model is feasible and effective.