摘要
介绍短时交通流预测的背景和意义,将短时交通流预测的方法分为5类,包括基于统计分析的预测模型、非线性理论模型、基于仿真的预测模型、智能预测模型及混合预测模型。对这5类预测模型进行逐一介绍,并对其在算法复杂度、预测精度、计算时长、适用路段等方面进行分析。短时交通流预测研究领域今后可能的发展趋势是数据来源多样化、混沌理论和深度学习逐渐发展,组合预测模型更加多样,预测精度不断提高。
This study briefly describes the background and definition of short-term traffic flow prediction.Short-term traffic flow forecasting methods are divided into five categories:prediction models based on statistical analyses,non-linear theoretical models,prediction models based on simulations,intelligent prediction models,and hybrid prediction models.Each of these models is introduced and their complexity is analyzed in terms of the type of algorithm,prediction accuracy,calculation duration,and applicable road sections.Possible development trends in the field of short-term traffic flow prediction will include the diversification of data sources,gradual development of chaos theory and deep learning techniques,increase in combined predictive models,and continuous improvement in prediction accuracy.
作者
赵宏
翟冬梅
石朝辉
ZHAO Hong;ZHAI Dongmei;SHI Chaohui(Beijing Jiaotong University Software College,Beijing 100044)
出处
《都市快轨交通》
北大核心
2019年第4期50-54,共5页
Urban Rapid Rail Transit
基金
国家自然科学基金面上项目(51778047)
关键词
智能交通
短时交通流
预测
intelligent transportation
short-term traffic flow
prediction
