城市干道系统交通事故预测模型研究英文.pdfVIP

A ccident P redictio n M odels fo r U r ban Ar ter ial System ——M EN G Xia ng hai　CHEN T ianen et al 13 A ccident Pr ediction M odels for U r ban A rt erial Sy st em M EN G Xiang hai　CH EN T ianen　SHEN G Hongf ei　J IA N G M eili Har bin Institute of T echno lo gy , Har bin 150090 , China) : Abstract It relies g reatly on the accident prediction models to make eff ectiv e traff ic saf et y countermeasur es . T herefo re , by taking Har bin ur ba n ar ter ial netw or k co mposed o f 46 8 art er ial links and 163 at-g rade inter sect ions as a case , the bro ad g eomet ry and tr aff ic flow dat a of the netw o r k w ere collected , as w ell as 8 510 accident data occur r ed on the netw o rk dur ing 1999 t o 2004 . F ir st ly , t he char acteristics of t he t raff ic accident data w ere analy zed , and the r esults show that the accident dat a follo w the N eg ativ e Bino mial distr ibutio n . Secondly , links and inter sectio ns w er e classif ied accor ding t o t he clu ster analysis met ho d , and then the accident pr edict ion models that can be u sed to predict the accident frequencies occur r ed on each kind o f links and inter sectio ns w er e established . T hirdly , the quantit ativ e relatio nship betw een the accident index o f the links dur ing r u sh ho urs and the v/ c o f them w as discu ssed. T o tally , 24 pr edict ion mo dels w ere calibr ated . F inally , the prediction models w er e applied to a case study o n par tial r oad netw or k of