Operation time forecasting method of buses on road segments of bus lanes

Abstract

The invention relates to an operation time forecasting method of buses on road segments of bus lanes. According to the operation time forecasting method, through the learning process that input vectors are mapped to a high-dimensional space through a kernel function and a regression function is solved in the high-dimensional space, a nonlinear problem is converted to a linear problem, and forecasting precision of travel time of buses on the bus lanes is obviously improved; in addition, basic kernel functions are subjected to weighted addition to obtain a combined kernel function, characteristics of different kernel functions are maintained, different types of data inputs can be better processed, a model has higher robustness, the establishment process of smart traffic is promoted, and the method has far-reaching influence on optimal operation of bus systems on the bus lanes and outgoing convenience of passengers, and therefore the method can be widely applied to the field of operation time forecasting of the buses on the road segments of the bus lanes.

Description

technical field [0001] The invention relates to a time prediction method, in particular to a method for predicting the running time of a bus section on a bus lane. Background technique [0002] The bus-only lane is an important road section where the road conditions permit and the traffic is congested. One or several lanes are delineated through signs and markings, and the passage of other vehicles is restricted all-weather or at different times for the exclusive use of buses. APTS (Advanced Public Transportation System, advanced public transportation system) and ATIS (Advanced Traveler Information System, advanced traveler information system) are the two core subsystems of ITS (Intelligent Transport System, intelligent transportation system), which accurately and reasonably predict For ATIS, the bus running time on the bus lane can provide effective travel information, so that passengers can reasonably arrange their own travel plans and improve the quality of bus service. F...

AI Technical Summary

Problems solved by technology

[0003] The running time of buses on bus lanes is affected by many random factors, such as weather, traffic conditions, and changes in passenger flow, etc. There is a very complicated relationship between the running status of buses and these factors, so it is difficult to apply mathematics model for accurate calibration and prediction

The existing neural network algorithm used to predict the running time of bus sections on bus-only lanes has problems such as structural determination, over-learning and under-learning, and local convergence, so that the accuracy of the neural network algorithm is not high. Method - The accuracy of SVM (Support Vector Machine, standard support vector machine) depends too much on the choice of kernel function, and there is no mature theory to guide it, resulting in poor robustness, which is often not obtained in practical applications. better forecast

Images