Based on Bayesian weighted combination intersection dynamic steering ratio estimation method

The present invention discloses a method based on Bayesian weighting intersection dynamic steering ratio combining estimation method utilizing intersection import and export channel detection to road traffic, designed to improve the Kalman filter, improved back propagation neural network and genetic algorithms are algorithms for solving three sub-intersection dynamic steering ratio and combined on the basis of historical data, considering the variation of amendments to the history and current estimated using the Bayesian formula and dynamically update the calibration weights will be three weighting algorithm result obtained seed to give the combination method to estimate dynamic steering ratio. For different traffic conditions, the existing method of estimating dynamic steering ratio advantage in terms of accuracy and efficiency, each with advantages and disadvantages, the present method can be embodied in various ways on the whole, to avoid local excessive deviation, with adaptable high accuracy, good stability, the best overall characteristics that can provide the basic data to support real-time traffic signal control system management and information services.