A study on traffic accident measures in municipal roads by using GIS

Transcription

1 icccbe 010 Nottigham Uiversity Press Proceedigs of the Iteratioal Coferece o Computig i Civil ad Buildig Egieerig W Tizai (Editor) A study o traffic accidet measures i muicipal roads by usig GIS Satoshi Fuamoto Bureau of Waterworks, Tokyo Metropolita Govermet, Japa Nguye Va Nham, Akiori Morimoto Dept. of Civil Egieerig, Utsuomiya Uiversity, Japa Hirotaka Koike Faculty of City Life, Utsuomiya Kyowa Uiversity, Japa Abstract Traffic Accidets are serious problems i all coutries. I Japa, thousads of people are killed due to traffic accidets every year. Therefore, the Govermet of Japa has submitted ad put ito operatio may measures to reduce traffic accidets. However, choosig effective measures is difficult because it ivolves may hazardous levels, various locatios of accidets, various traffic volumes, etc. I case of muicipal roads, the problem is more difficult tha arterial roads because of their wide rages of accidet circumstaces, lower traffic volumes, ad ukow hazardous levels. Therefore, this study focused o submittig idices of hazardous spots ad hazardous meshes as a methodology to choose measures appropriately for the scope of muicipal roads. To ivestigate the method, GIS is used to aalyze traffic accidets, ad Utsuomiya City was chose as a case study area. The Microcomputer Accidet Aalysis Package (MAAP) was used as a tool to do cluster aalysis ad desity aalysis. Cluster aalysis is a fuctio to determie hazardous accidet spots, ad desity aalysis is for determiig hazardous accidet meshes. As the result, may hazardous meshes are cocetrated i the city cetre. Furthermore, the importat thig is that we classified hazardous levels of traffic accidets ad divided the measures ito two groups called spot measures ad mesh measures for reducig traffic accidets o muicipal roads. Keywords: traffic accidet, muicipal road, measure, GIS, microcomputer accidet aalysis package 1 Itroductio 1.1 Backgroud ad Purpose Traffic accidets are serious problems i all coutries because it causes may deaths ad ijuries every year. Although govermets take various measures to solve the problem, traffic accidets are still persistet problems. This situatio is the same i may coutries icludig Japa. I Japa, accordig to the data take from 1965 to 007, the umber of traffic accidets, deaths ad ijuries icreased sharply from 1965 to 1971 due to the isufficiecy of traffic safety measures. I 008, the umber of traffic accidet deaths is reduced to 5,155 which are the lowest i half a cetury. However, the umber of ijured people exceeded oe millio for 7 years sice Besides, traffic accidet measures implemeted by the govermet are almost cocetrated o solvig problems at the accidet-proe places such as itersectios of atioal roads or prefectural roads, ad little attetio was paid to muicipal roads because of their ifrequet occurreces of accidets. However, although there were fewer accidets occurred o muicipal roads, they were

2 widely scattered everywhere o a large scale. Therefore, if all accidets o muicipal roads are summed up, the umber of traffic accidets is ot a small oe at all. I order to reduce traffic accidets o muicipal roads, the measures applied for atioal roads maybe iappropriate due to the differece betwee the eviromet of the atioal road ad the muicipal road. Therefore, the purpose of this study is to fid out hazardous positio of traffic accidet ad submit a method to choose measure appropriately ad effectively for the scope of muicipal road. 1. Review of related studies Hideki s study (Hideki et al., 000) made a aalysis o traffic accidet for muicipal roads from the viewpoit of sight distace of itersectio. The author also defied causes of accidets by usig variatio tree method). Besides, The Road Admiistratio Maagemet of Miistry of Lad, Ifrastructure, Trasport ad Tourism (MLIT) of Japa has defied accidet-proe places, ad measures are chose with the idicator called casualty rate calculated by the umber of accidets i a year ad traffic volume (See Equatio 1). Casualty rate = Number of deaths ad ijures Number of cars i oe km (case/cars) However, the traffic accidets o muicipal road have a characteristic that the occurrece rate is low, ad gatherig traffic volume data o all muicipal roads is very difficult. Therefore, we thik that ew idices for aalyzig traffic accidet o muicipal road as proposed i this study are very ecessary. Cocept of traffic accidet measure i case of muicipal road..1 Curret traffic accidet situatio of Utsuomiya city ad data used Number of accidets Number of deaths Utsuomiya trai route umber of occurreces umber of ijured people umber of dead people Fig.1 Process of traffic accidet i Utsuomiya km Utsuomiya statio Fig. Traffic accidets displayed o GIS There were 4, traffic accidets with 6 people dead ad 5,461 people ijured occurred i Utsuomiya city i 007. If comparig with the data of 006, umber of traffic accidets decreased by 178 (.9%), ad umber of ijured people decreased by 197 (.5%). But, umber of deaths icreased by 4 (1.5%). So, it may be said that traffic accidet becomes a serious problem i Utsuomiya city (See Figure 1).

3 Data used i this study is of the year 006 with,088 traffic accidets occurred o muicipal roads i Utsuomiya city. These accidets were displayed o GIS as Figure.. The tred of traffic accidet distributio i zoes I Utsuomiya city, occurrece desity of traffic accidets i zoes which was calculated shows that occurrece situatios of traffic accidets are differet by the lad use. Moreover, occurrece desity of traffic accidets i residetial zoe is very high. I order to kow distributio of accidets i detail, the etire Utsuomiya city area was divided ito meshes with size legths of 500m. Ad i these meshes, variace was used to determie the distributio types such as cocetratio type, scatter type, ad mix of both types. Each mesh was set up with x axis ad y axis (See Figure ). Coordiate of accidets i each mesh was examied. The, variace σ x ad σ y of x axis ad y axis correlatively were calculated (See Equatio ). ( xi x) 1 σ x = () ( yi y) 1 σ y = Fig. Calculatio of variace i meshes The, scatter graph was created by calculated variaces ad probability cocetratio ellipse (PCE) (90% cofidet limit). (See Figure 4) Fig.4 Distributio graph. Cocept of traffic accidet measures for Muicipal Roads Fig.5 Cocept of traffic accidet measures i Muicipal road Accordig to figure 4, traffic accidets o muicipal roads have differet distributios i each mesh. So, it was thought that uderstadig of distributio types of traffic accidets occurred o muicipal road is very essetial. I fact, i case of cocetratio type, safety will icrease higher if spot accidet measure is used. But i case of large scale with radom accidets, it was thought that mesh measure

4 will become more effective. I summary, if distributio types of accidets are determied, either spot measure or mesh measure will be chose to implemet to reduce traffic accidets effectively (See Figure 5). Ivestigatio of the classified method base o spot hazardousess ad mesh hazardousess..1 Evaluatio of spot hazardousess ad mesh hazardousess. As discussed above, we kew that traffic accidets occurred ad distributed everywhere o muicipal roads. Ad, we thik that accidets with ear distaces will become hazardous if they are grouped together. Therefore, i order to ivestigate the methods classified as i Sectio., we focused o evaluatio of spot hazardous level ad mesh Fig.6 Grouped traffic accidets hazardous level. About the evaluatio of spot hazardousess, the accidets was grouped as figure 6 by circles with radius r = 16.m that was calculated by usig the cocept of earest eighbour distace. Accordig to this, we ca fid out accidet-proe places ad defie them as Accidet Hazardous Spot (AHS). Next is about the evaluatio of mesh hazardousess. We divided area ito equal size meshes i which the weightig of accidet will be calculated, ad hazardous level of mesh will be evaluated accordig to the accidet weightig. So, meshes which have high hazardous levels will require immediate attetio to implemet couter-measures. We defied the mesh as Accidet Hazardous Mesh (AHM).. Weightig of traffic accidet I order to evaluate hazardous level i each spot ad mesh, weightig of traffic accidet was used. It is also called as Accidet Score (AS) i this study. Its equatio is show as follows (See Equatio ). α + α B + C = Accidet score () 1 A α A: Number of fatal accidets B: Number of seriously ijured accidets C: Number of slightly ijured accidets α 1 : Parameters ( α 1 =0.76 =0. =0.0) Parametersα 1 were determied by the rates after coutig amout of loss of each perso accordig to various cases such as death, after effect, ijury based o car isurace data (Geeral Isurace Associatio of Japa, 007).. Priority of measure iside the hazardous mesh After determiig AS for meshes ad spots as above, priority measures will be chose to implemet effectively based o distributio ratio (DR) of hazardous level couted as Equatio 5. (4) The process of choosig measure is show i the chart below.

5 Fig.7 Proces I case 1 with o AHS i the mesh, the ecessary measure will obviously be the mesh measure. But, i case with oe AHS, ad case with two or more AHS, spot measure or mesh measure will be prioritized accordig to the value of distributio ratio (DR) (Figure 8). 4 Determiig accidet distributio i Utsuomiya city 4.1 Determiatio of accidet distributio type by usig Accidet Aalysis Software Utsuomiya city was used as a case study for the aalysis. Traffic accidet distributio i Utsuomiya city will be determied by usig Microcomputer Accidet Aalysis Package (MAAP). The accidet aalysis software was provided by TRL Compay i Eglad. It helps aalyzig accidet causes ad submittig measures for reducig accidet by displayig hazardous levels optically. As a aalysis method, desity aalysis (mesh aalysis) ad cluster aalysis (spot aalysis) will be focused o. Desity aalysis divides a pre-defied geographical area ito equally dimesioed squares. MAAP will the cout the umber of accidets i each square ad thematically colour the map correspodigly. A list of the worst squares, i.e. the squares with the highest accidet couts are produced ad the accidets cotaied with-i the square ca be aalyzed with MAAP s other aalysis tools (TRL, 001) Cluster aalysis searches for collectios of geographically related accidets withi a defied geographical area. MAAP defies geographically related accidets as lyig withi a user defied distace of each other. Accidets belog to the same cluster as all other accidets geographically related to it, ad to all accidets geographically related these earest eighbours. Therefore it is possible to move from ay accidet i a cluster site to ay other accidet i a cluster site, movig accidet to-accidet whilst ever movig more tha the resolutio distace. MAAP couts the umber of accidets i each site. This iformatio is the thematically represeted o the map ad a list of the sites with the highest accidet couts produced (TRL, 001). Geerally speakig, cluster aalysis is used to evaluate spot accidet hazardousess, ad desity aalysis is used to evaluate mesh accidet hazardousess. 4. Desity aalysis result As the result of desity aalysis, traffic accidet hazardous meshes are cocetrated ad distributed i the city cetre. Hazardous level of cetral area is also higher compared with suburba area (See Figure 9). 4. Cluster aalysis result As the result of cluster aalysis, 97 spots of 0 grouped spots are hazardous ad displayed o GIS map as i Figure 10 which shows that almost all accidet hazardous spots are cocetrated i itersectio areas.

6 : Utsuomiya statio 5km Fig.9 Distributio of accidet hazardous mesh 4. Choosig accidet measures for each mesh : Accidet hazardous spot : Utsuomiya statio Fig.10 Distributio of accidet hazardous spot i the east side of Utsuomiya statio I order to choose a appropriate measure for each mesh, we applied the steps show i figure 7, ad used the easter area of Utsuomiya statio as a case for aalysis. Desity ad cluster aalysis results of the area were displayed o GIS map as Figure 11. The, the distributio ratio was computed ad show as i the Table 1. Accordig to the table, the meshes of No. ad No.7 which have high distributio ratio of spot will eed spot measure o a priority basis. Besides, the other meshes which have high distributio ratio of mesh will be asked for mesh measure to be implemeted effectively. 5 Coclusio The purpose of this study is to fid out traffic accidet hazardous places ad propose the method of choosig effective accidet couter-measure i the scope of muicipal roads. Therefore, the idices of spot ad mesh hazardousess to choose ad implemet measure effectively have bee submitted. Accordig to the idices, spot measure ad mesh measure have also bee submitted i this study. Furthermore, i practice, MAAP software was used to aalyze traffic accidet i Utsuomiya city. As the aalysis result displayed o GIS, the city cetre was kow as the hazardous area due to the cocetratio of traffic accidets. So, the determiatio of appropriate couter-measures for reducig accidets has bee called for. Refereces GENERAL INSURANCE ASSOCIATION OF JAPAN. Traffic accidets based o car isurace data, 007 HIDEKI FURUYA, HIDEKAKI KANO SHIMA, NOBUHISA MAKINO, ATSUSHI TERAMUKU. A cosideratio o actual occurrece coditios ad restraits of traffic accidet o muicipal road focused o itersectio ecouter accidets due to ocompliace of safety cofirmatio. 0 th Research Recital Thesis Report of Japa Society of Civil Egieers pp1-4, 000 TRL LIMITED, B.Hills, A.Morriso, S.Vadgama, G.Roberts. Applicatio Guide 8, Map User Guide, pp.10 1, 001