大型停车场空闲车位短时预测方法

Simulation of prediction for free parking spaces in large parking lots

为了提升智能交通系统性能及停车场利用率,针对大型停车场空闲车位短时预测进行了研究,提出了一种基于灰色理论、BP神经网络和马尔可夫链的组合预测方法以提高预测精度与时效性。该方法使用灰色理论处理数据,弱化其随机性,再通过人工神经网络训练得到数量预测结果,最后使用马尔可夫链消除系统产生的随机误差得到最终结果。实验表明,这种组合预测方法有效提高了预测精度,预测结果符合实际停车场数据变化规律,为驾驶员提前作出合理的停车场选择提供了可靠依据,能有效提高停车场车位利用率。

In order to improve the performance of intelligent transportation system and the utilization rate of parking lot,aiming at the short-term prediction of free parking space in large parking lot,this paper came up with a forecasting method based on the combination of grey theory,BP neural network and Markov chain.At beginning,it used the grey method to weaken randomness of data itself and combined with factors that influenced the quantity of free parking spots.Then it went through BP neural network training to get the short time free parking spots prediction.Finally,it used the Markov chain to eliminate the random error generated by the system and obtained the final result.According to the experiment results,the combined forecasting method can effectively improve the prediction accuracy.The result that matched the data flow of actual parking lot is able to provide reliable analysis for drivers to choose the right parking lot and improve efficiency of parking lots usage.