停车场泊位占有率预测方法评价
发布时间：2018-10-09 19:05
【摘要】：采用上海市五角场地区的停车泊位检测数据,分析了商业、办公和体育场3种不同类型停车场泊位占有率(parking occupancy rate,POR)的时变特征,并评价了ARIMA(autoregressive integrated moving average)、卡尔曼滤波和BP(back propagation)神经网络等3种常用方法在POR预测中的适用性.结果表明,ARIMA和BP神经网络的预测精度总体优于卡尔曼滤波,BP神经网络在商业和办公停车场的短时预测中有较好的精度;3种方法的预测精度均随预测时间步长的增加而逐渐降低;不同类型停车场的POR预测精度存在较大差异,工作日的预测精度一般高于非工作日,且模型具有较好的自适应性.
[Abstract]:Based on the parking berth detection data of Wujiuchang area in Shanghai, this paper analyzes the time-varying characteristics of parking berth occupancy rate (parking occupancy rate,POR) of three different types of parking lots: commercial, office and stadium. The applicability of ARIMA (autoregressive integrated moving average), Kalman filter and BP (back propagation) neural network in POR prediction is evaluated. The results show that the prediction accuracy of Arima and BP neural networks is better than that of Kalman filter BP neural network in the short-term prediction of commercial and office parking lots. The prediction accuracy of POR in different types of parking lots is quite different. The prediction accuracy of working day is generally higher than that of non-working day, and the model has good adaptability.
【作者单位】： 同济大学道路与交通工程教育部重点实验室;
【基金】：“十二五”国家科技支撑计划(2014BAG03B02)
【分类号】：U491.7
本文编号：2260440
[Abstract]:Based on the parking berth detection data of Wujiuchang area in Shanghai, this paper analyzes the time-varying characteristics of parking berth occupancy rate (parking occupancy rate,POR) of three different types of parking lots: commercial, office and stadium. The applicability of ARIMA (autoregressive integrated moving average), Kalman filter and BP (back propagation) neural network in POR prediction is evaluated. The results show that the prediction accuracy of Arima and BP neural networks is better than that of Kalman filter BP neural network in the short-term prediction of commercial and office parking lots. The prediction accuracy of POR in different types of parking lots is quite different. The prediction accuracy of working day is generally higher than that of non-working day, and the model has good adaptability.
【作者单位】： 同济大学道路与交通工程教育部重点实验室;
【基金】：“十二五”国家科技支撑计划(2014BAG03B02)
【分类号】：U491.7
【相似文献】
相关期刊论文 前3条
1 宋俊;许轶博;;基于神经网络模型与最短路径模型下的公交路线设计[J];硅谷;2013年07期
2 翁建玲;;基于神经网络模型的软基工后变形分析[J];山西建筑;2007年04期
3 ;[J];;年期
相关会议论文 前1条
1 赵志理;李炜;王虎平;刘波;;基于超强学习机的交通客流预测方法研究[A];第九届中国不确定系统年会、第五届中国智能计算大会、第十三届中国青年信息与管理学者大会论文集[C];2011年
相关硕士学位论文 前1条
1 付春芬;基于深度学习的安全带检测方法研究[D];华中科技大学;2015年
，本文编号：2260440
本文链接：https://www.wllwen.com/kejilunwen/daoluqiaoliang/2260440.html
