37 6 2011 6 Vol 37 No 6 JOURNAL OF BEIJING UNIVERSITY OF TECHNOLOGY Jun 2011 100124 2 BP BP BP U 491 A 0254-0037 2011 06-0882 - 06 1 70% BP 1 2 2 2 2009-10-25 50578003 2006BAG01A01 1984
6 883 2 1 3 3 5 3 v = 4 = 0 a v a a = { 0 1 E = m = 1 v E m E 0 05 0 01 2 p 0 001 0 05 0 01 p 2 2 4 Wlson 5 Hart 6 7-8 32% 42% 9 1 16 8 19 ~ 29 1 Table 1 Basc Informaton of Partcpants 1 2 19 ~ 29 8 25 3 1 1 21 ~ 27 8 24 7 30% ~ 42% 2 1 2 3 3 3 Sm World 4 8 3
884 2011 Fg 1 1 1 2 1 An example of benchmar heart rate curve Fg 2 An example of heart rate curve durng the experment a v v L Δv ΔS a L 3 Sm World Fg 3 Vsual effects durng evacuaton smulaton 30 Hz 2 3 2 3 Table 2 2 3 An example of data collected from smulator of test vehcle t /s v / m s - 1 a / m s - 2 470 10 5 30 1 14 470 20 5 34 1 15 t /s Table 3 An example of data collected from smulator of leadng vehcle ΔS /m v L / m s - 1 a L / m s - 2 470 10 22 00 21 87 13 56 1 46 470 20 22 00 22 77 13 63 2 03 4 BP 4 1 BP BP BP EXCEL 10 BP p 0 = p 0 1 p 0 2 p 0 n p = p 1 p 2 p n 2 p 0 p γ 0 p 0 p = 1 + ξ max max Δp 0 + η max max 1 + λ 1 Δp 0 + λ 2 Δp' 0 + ξ max max Δp' 0 Δp 0 + η max max Δp' 0 3
6 885 λ 1 λ 2 0λ 1 + λ 2 = 1 λ 1 λ 2 ξ η 0 5 ξ η λ 1 λ 2 ξ η λ 1 λ 2 Δp 0 = p 0 - p Δp' 0 = p' 0 - p 0 p' = p - p - 1 max max = 1 2 n = 2 m n m p 0 p γ = γ 0 p 0 p = 1 γ n n = 1 γ p 0 p v v L Δv ΔS a L 4 4 Table 4 4 The result of Gray-correlaton Analyss γ a a L γ a Δv γ a v L γ a v γ a ΔS 0 967 0 947 0 847 0 846 0 723 4 2 BP BP 4 11 BP 3 3 11 4 BP Fg 4 Archtecture of BP neural networ 11 BP 11 BP 1 3 a L Δv v L T 1 a BP 20 Hyperbolc tangent sgmod MATLAB7 7 tranlm Levenberg- Marquardt BP MSE 1 10-11 BP 3 5 BP 5 308 5 BP 6 R
886 2011 N R = ( 槡 a - a 槇 2 N ) R a a槇 BP N 5 Table 5 5 MATLAB An Example of Data Used wth MATLAB a L / m s - 2 v L / m s - 2 Δv / m s - 1 a / m s - 2 0 69 0 03-0 28 0 23 0 64 0 31-0 34 0 21 Fg 5 5 BP Tran result of BP Neural Networ Fg 6 6 BP Smulaton result of BP Neural Networ model R 17 8% 12 5 2 5 3 BP 1 J 2007 33 10 1070-2074 HU Hong LIU Xao-mng YANG Xao-uan Car-followng model featured wth factors reflectng emergency evacuaton stuaton J Journal of Bejng Unversty of Technology 2007 33 10 1070-2074 n Chnese 2 M 1998 533-534 3 J 1999 20 30-32 JI Hong-guang WANG Ha-mng CHEN Yao-zhong et al The plot study of a fatgue questonnare J Journal of Navy
6 887 Medcne 1999 20 30-32 n Chnese 4 HART S G HAUSER J R In flght applcaton of three plot worload measurement technques J Avaton Space and Envronmental Medcne 2004 58 402-410 5 WILSON G F DONNELL R D Measurement of operator worload wth the neuropsychologcal worload test battery J Advances n Psychology 1988 52 63-200 6 HART S G WICHENS C D An approach to systems ntegraton M New Yor Van Nostrand Renhold 1990 257-296 7 APPARIES R J RINIOLO T C PORGES S W A psychophy sologcal nvestgaton of the effects of drvng longercombnaton vehcles J Ergonomcs 1998 41 5 581 8 DEKKER J M SCHOUTEN E G Heart rate varablty from short electrocardographc encodngs predct mortalty from all causes n mddle aged and elderly men J Am J Epdermal 1997 45 1 899 9 J 2008 36 10 1372-2377 YAN Yng LIU Hao-xue GUO Zhong-yn Research on drvers Psychologcal and Physologcal Characterstcs n dfferent traffc flow condtons J Journal of Tongj Unversty Natural scence 2008 36 10 1372-2377 n Chnese 10 D 2005 PING Xue-lang Research on grey system theory and ts applcaton n data acquston and process of reverse engneerng D Nanjng College of Mechancal Engneerng and Appled Electroncs Technology of Nanjng Unversty of Aeronautcs and Astronautcs 2005 n Chnese 11 MATLAB2007 M 2007 3 12 WANG L LIU Xao-mng ZHONG Xao-mng A car-followng model based on artfcal neural networs n urban expressway sectons C Proceedng of 85th Transportaton Research Board Washongton D C Transportaton Presearch Board 2006 Car Followng Model Under Emergency Evacuaton Stuaton Based on the BP Artfcal Neural Networ XU Zh YANG Xao-uan ZHAO Xao-hua LI Lng-je College of Archtecture and Cvl Engneerng Bejng Unversty of Technology Bejng 100124 Chna Abstract Scenaro smulaton s employed to create a drvng envronment n the emergency evacuaton and the questonnare nvestgaton and the Mram ECG are used to verfy the valdty of the drvng envronment n the emergency evacuaton from subjectve and objectve aspects Gray-correlaton Analyss s undertaen to determne whch factors have more mpact on the acceleraton of the followng car Acceleraton of the followng car s set as the mother factor seres and the speed of the followng car speed of the front car speed dfference between vehcles headway between vehcles acceleraton of the front car are set as the sub-factor seres Fnally acceleraton of the front car speed dfference between vehcles speed of the followng car are determned as three most mportant factors BP neural networ s desgned and the three factors mentoned above are selected as nput varables and the acceleraton of the followng car s selected as output varable Smulaton of the BP neural networ wth data gathered from drvng smulator reveals that BP neural networ has a hgh precson n the predcton of the car-followng model Key words car followng model emergency evacuaton gray-correlaton BP Neural Networ