CROSSING BEHAVIOUR PEDESTRIANS OF PEDESTRIANS IN MIXED TRAFFIC CONDITIONS

Abstract

Pedestrians are important part of urban transportation system. They form single road user group, as all road users are pedestrians at some time or other. Pedestrian movement in urban environment are of two types; one interacting with traffic like crossing and other without interacting with traffic like walking through sidewalks, precincts etc. In these movements crossing is important as it increases the vulnerability of pedestrians to accidents. It was observed that the pedestrians cross at location convenient to them and at locations which would cause little delay to them (Garder, 1989; Hamed, 2001; Hollo et al., 1995; Sisiopiku and Akin, 2003). Enforcement of the rules by police is infrequent, and considered by the public to be unwarranted (Schonfeld and Musumeci, 2003). Engineering measures also tend to be resisted, with measures such as overpasses and underpasses (Hollo et al., 1995) and pedestrian barriers (Kopelias et aI., 2002) have little effect on illegal crossing behaviour. These violations of traffic laws by the victim were found to be one of the "predominant contributing factors" in all pedestrian accidents (Ashur et al., 2003). In India the situation is also not very different. In India, pedestrian related fatalities in the year 2008 accounted for about 11% of all motor vehicle related deaths (Accident Report -2008, MoRTH). Around 72 % of all pedestrian fatalities in 2003 occurred in urban areas (National Highway Traffic Safety Administration, 2004). The share of non-motorized users in traffic fatalities is 88% in Mumbai of which 80% are pedestrians and 8% are bicyclists (Mohan, 2004). Pedestrians, by far, constitute the most vulnerable road user group comprising high proportion of road accident fatalities, typically 40 — 50% in India and 70 — 90% in metropolitan cities like Mumbai and Kolkata. Over 54% of all traffic fatalities in Delhi are pedestrians (Grebert, 2008). In Bangalore, pedestrian fatalities account for nearly 54% of the total traffic fatalities. The share of children (<15 years), young men (16 to 45 years) and elderly in pedestrian fatalities are found to be 6%, 51% and 17% respectively. It was also found that fatalities in elderly women pedestrians are much greater than elderly male pedestrians. The pedestrian fatalities are higher in the outskirts of the city and about 58% of the pedestrian fatalities are due to heavy vehicles (buses, van, and trucks) with the share of two-wheelers being 19% (NIMHANS BISP report 2010). Studies also show that pedestrian fatalities are more in midblocks than at junctions and other locations (Mohan, 2009) While going through the literature it is found that several studies relating to pedestrian behaviour have been conducted abroad where uniform and disciplined traffic flow conditions prevail. The findings and recommendations of these studies cannot be applied to most of the developing nations where the nature of traffic is mixed or untidy. No detailed study is reported on crossing behaviour of pedestrian with respect to their age and gender, gap acceptance, etc. under mixed traffic conditions. Literature clearly lacks in pedestrian crossing studies under mixed traffic conditions. The Indian code of practices on pedestrian crossing also remains silent on various design aspects. The capacity analysis and Level of Service (LOS) criteria of a pedestrian crossing facility are not discussed in Indian Roads Congress (IRC) codes and those reported in literature do not consider the perception of pedestrians in defining the LOS. This has prompted the author to conduct a detailed crossing behaviour study in Indian conditions. Data were collected at 17 midblock locations in five cities of India (Chennai, Chandigarh, Coimbatore, New Delhi and Erode). Data were collected mainly by videography. Pedestrian crossing time, gap accepted, approach vehicle speed, classified traffic volume count and pedestrian arrivals were extracted from the videography. Then a pedestrian perception survey was conducted at two locations; one in Roorkee and another in Connaught place, New Delhi for the estimation of Level of Service (LOS). Data extracted from the videos were used for further analysis. Three types iv of analysis are conducted on pedestrian crossing. They are pedestrian flow analysis, vehicle interaction and risk analysis. Finally, a LOS service criteria is proposed incorporating the pedestrian perception. Flow analysis is divided into two parts; behavioural analysis and statistical analysis. Behavioural analysis shows that three types of crossing pattern exist in Indian conditions. These are single stage, rolling and two stages. Female and old pedestrians are found to adopt single stage crossing more. Pedestrians were found to follow zebra crossing more if obstructions like barrier is provided along the carriage way. Higher usage of zebra crossing is observed for male and female pedestrians in metropolitan cities. Crossing speed is generally higher than all earlier studies except that reported by Knoblauch et al. (1996) and Fitzpatrick et al. (2006). High average crossing speed is observed in Indian condition on road with two-way traffic. Female and old pedestrians speed was found to be lower than other groups. Pedestrian speed was found to decrease with increase in group size. Speed-density and flow density form exponential relations whereas flow-speed relation is logarithmic. Pedestrian arrivals were found to follow generally geometric distribution. Pedestrian-Vehicle interaction was studied by estimating the gap accepted, critical gap and waiting time. The average gap accepted in India was found to be higher than that reported in literature. Old and female pedestrians accepted large gaps. It was found that accepted gap varies with conflicting flow as inverse function and with crossing speed as power function. Critical gap was estimated from cumulative frequency distribution functions drawn for gap accepted and inverse of crossing time. Critical gap was found to increase with road width and decrease with traffic volume. Waiting time was generally low in mixed traffic conditions. Female pedestrians V waiting time is more than male pedestrians and varies with conflicting traffic as exponential function. Pedestrian risk was modelled by three parameters; accepted gap, safety margin and conflicts. With respect to gap accepted, female and old pedestrians risk was found to be high. Risk was also high when the approaching vehicle is a large vehicle like bus. With respect to safety margin, female and old pedestrians were found to adopt large safety margin and safety margin adopted by pedestrian is high when the approaching vehicle is bus. Based on the type of conflict, the risk is low due to absence of fatal accidents and low percentage of Type III conflicts (vehicle weaves) at all locations. Finally, a level of service based on cluster analysis of waiting time data is developed and further modified by pedestrian perception survey data. This is used to estimate the LOS at different locations. The average waiting time at all locations is below 23s, which shows that all the locations selected in the present study fall in LOS category above `C'. VI