Travelling Time Forecasting And Research In Route Guidance Algorithm Based On Simple Road Network Model

The rapid development of urban economy brings huge pressure for intelligenttransportation. Solving the traffic congestion, easing traffic pressure and reducingnegative influence caused by traffic congestion are urgent problems to be solved).Intelligent Transportation System(ITS) research gets more and more attention, and theresearch content is also gradually rich. Among all research contents, prediction in ITSplays the basic role and it is the data source of other techniques. At the same time,route guidance algorithm is the core technology of traffic flow guidance and vehiclenavigation. This paper does research on travel time prediction and route guidancealgorithms. The main contributions of this paper are as follows:①Making reference on the methods that kalman filter is used to forecast trafficparameters, in view of the characteristics of simplified network model, this paper putsforwards a single-step travel time prediction model based on kalman filter and designsa single-step travel time prediction algorithm. The time lag problem is solvedeffectively. Meanwhile the historical average data can update automatically. Thecombination of single-step travel time prediction model and existing simplifiednetwork model makes the predicted travel time more in line with the actual situation.②Based on single-step travel time prediction model, a kalman filter-basedmulti-step travel time prediction model and corresponding algorithm are proposed,which can forecast link travel time in about20minutes. In order to solve the problemthat there are no observation data in future, the HD (History Day) algorithm isproposed.③On the results of multi-step travel time prediction, in view of the existingalgorithms only considering a single data, the route navigation algorithm whichcombines real-time data, multi-step prediction data and historical data is proposedtaking advantages of the real-time data, prediction data and historical data. In themeantime, the core algorithm Dijkstra_pred is designed, which solved the problem oflonger travel time that the navigation route calculated by route guidance algorithmwhich considered only real-time travel time data takes.④Prediction model and the algorithm proposed in this paper are verified bysimulated data. Experiments show that:1) the travel time of single-step and multi-steptravel time prediction algorithms can eliminate the time lag effectively;2)The multi-step algorithm can forecast travel time reasonably within20minutes in future;3)The travel time calculated by the route guidance algorithm based on the real-timedata, the forecast data and historical data is superior to the travel time calculated by thealgorithm based on real-time data, and route calculated based on the algorithmproposed in this paper changes less than algorithm considering only real-time data.