Real-time Based On Multi-time Speed Prediction Research On Path Optimization

With the development of urbanization,the road traffic environment in recent years has become increasingly prominent,frequent accidents and environmental pollution have seriously affected people’s quality of life,intelligent transportation system is becoming more and more concerned.Path optimization is also a very important research direction.Considering the current demand for path optimization system,this paper studies the real-time dynamic path optimization system,and it based on multi-time velocity prediction from the historical speed value in traffic environment information.The main contents of this paper:Based on the speed prediction model of support vector machine,the current time velocity prediction model is put forward according to the historical time velocity value after pretreatment.For the forecast of the speed value of the future time,the speed value at a certain time predicted as the previous time is used as an input for predicting the next time speed value,and the prediction of the speed value at any future time is realized.It is found that the model has a large error in predicting the velocity value in the future,and then the back propagation algorithm is introduced to select the optimal weight to further integrate the predicted value with the historical real value,and ultimately achieve the purpose of reducing the error.Based on the dynamic path optimization model,choose the depth of the first search algorithm ideas,introduce refresh predictive factor,realize the construction of dynamic full path search model.And the dynamic path optimization system based on the speed prediction model is proposed,it satisfies the demand for accessibility and the shortest time.In order to prevent the congestion problem in multi-vehicle routing,this paper designs an anti-congestion model which has been proved by simulation to avoid the excessive overlapping of the planning path between multiple vehicles.The purpose of this paper is to provide a solution for dynamic path optimization research,to provide a valuable reference for further research by improving or proposing new algorithms and models,and has proved the rationality and validity of the research results of this paper.