The crossing speed and safety margin of pedestrians at signalized intersections

Transcription

1 Available online at ScienceDirect Transportation Research Procedia 22 (2017) th EURO Working Group on Transportation Meeting, EWGT2016, 5-7 September 2016, Istanbul, Turkey The crossing speed and safety margin of pedestrians at signalized intersections Pelin Onelcin a,*, Yalcin Alver a a Ege University, Civil Engineering Department, Izmir, Turkey Abstract Pedestrians are the vulnerable part of the road users worldwide. A significant number of pedestrian fatalities occur in road traffic accidents each year. Hence, to minimize the pedestrian-vehicle accidents it is important to optimize the signal timings according to pedestrians crossing speeds and delays at signalized intersections and to understand the pedestrians safety perception. This paper investigates the pedestrians crossing speeds, delays and gap perceptions at six signalized intersections in Izmir, Turkey. The crosswalk where the pedestrian density is high at each intersection was selected for observations. Each intersection was observed on weekdays during afternoon peak ( ) and evening peak ( ) hours using video cameras. In total, 2,694 pedestrian crossings were observed. The average crossing speed is found to be 1.31 m/s and the average 15 th percentile crossing speed is found to be 1.07 m/s. Existing delay models are compared with the observed delays and the results showed that none of the models give good results to explain the observed delays and a new model is needed. Speed limit and position*speed limit interaction revealed significant effects on safety margin The Authors. Published by Elsevier B.V. Peer-review under responsibility of the Scientific Committee of EWGT2016. Keywords: Pedestrians crossing speed; safety margin; signalized intersection 1. Introduction No matter which type of transportation mode is chosen by the people, walking is a part of that trip. Hence, each person is a pedestrian during some part of his/her trip. In this paper, the results of an observational study at * Corresponding author. Tel.: ; fax: address: onelcinpelin@gmail.com X 2017 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the Scientific Committee of EWGT /j.trpro

2 4 Pelin Onelcin et al. / Transportation Research Procedia 22 (2017) 3 12 signalized intersections are presented. Pedestrians walking speed, delay and safety margins are given and interpreted in terms of pedestrian safety. Pedestrians crossing speed varies for each country. The pedestrian walking speed is given as 1.2 m/s in Highway Capacity Manual (HCM) 2010 (Transportation Research Board, 2010). The number of elderly pedestrians affects the recommended walking speed. For more than 20% elderly pedestrians the crossing speed is given as 1.0 m/s. The Manual on Uniform Traffic Control Devices (MUTCD 2003) recommends a crossing speed of 1.21 m/s. In Turkey the design walking speed is given as 1.4 m/s by Turkish Standards Institution (TSI, 2012). This value is considerably high especially for elderly pedestrians. In Australia Bennett et al. (2001) found the average crossing speed to be 1.24 m/s. In Jordan, Tarawneh (2001) recommended an average 15 th percentile speed of 1.11 m/s. They found that pedestrians who crossed in groups had a higher crossing speed than individuals, the younger pedestrians had a higher crossing speed than the elderly pedestrians, and males walked faster than females. In India, Chandra and Bharti (2013) and in Canada, Montufar et al. (2007) also found that males walked faster than females. Gates et al. (2006) conducted a study in Wisconsin and observed that the individuals walked faster than the pedestrians who crossed in groups. They recommended a 1.2 m/s walking speed. At signalized intersection not only vehicles but also pedestrians experience delays. Pedestrian phases usually consist of green, clearance and red phases. During red phases pedestrians are not allowed to walk but are allowed to walk during green phase. On the other hand, during clearance phase if pedestrians are already on crosswalk they are required to speed up and if those who are not yet on crosswalk are not allowed to enter the crosswalk. Flashing red or flashing Don t Walk are usually used for clearance phases (Li et al., 2005). The signalized intersection investigated in this paper had green, red and all red signals for pedestrians. Pedestrian delay estimation is based on uniform arrival rates and fixed pedestrian timing in HCM 2010 (TRB, 2010). The model developed by Braun and Roddin (1978) assumed uniform pedestrian arrival rate, complete signal compliance, fixed cycle length, and no pedestrian actuation. Virkler (1998) indicated that there are times when pedestrians do not comply with signal rules to minimize their delays. Data was collected at 18 crosswalks in Brisbane, Australia. Pedestrians who crossed during green signal, pedestrians who entered the crosswalk during the clearance interval, pedestrians who entered the crosswalk during the red interval, and the delay of all pedestrians were observed. By modifying the equation developed by Braun and Roddin (1978) they used 69% of the clearance interval for entering the crosswalks. Li et al. (2005) developed a model for pedestrian delays in developing cities like Xi an, China. The authors noted that the models currently used in developed countries do not take into account the delay that pedestrians may encounter during the green signal. They found that the average delay of pedestrians during the green signal was 1.9 s. In Turkey, pedestrians experience delays during green signal as well. Pedestrians making illegal crossings either cross against the lights or away from the lights (Lange et al., 2011). Safety margins for pedestrians who crossed illegally (within the 25 m from the crosswalk) were measured in this study. Safety margin is defined as the difference between the time a pedestrian crosses the traffic and the time the next vehicle arrives at the crossing point by Chu and Baltes (2001). This definition is adopted for the present study. Age was found to be a predominant factor in the gap perception studies. Oxley et al. (1997) investigated the age related differences in crossing actions and found that elderly pedestrians made less safe choices compared to the young pedestrians despite they left longer gaps because the elderly pedestrians underestimated their crossing times. The study of Connelly et al. (1998) only investigated the gap thresholds of 48 school children whose ages ranged from five to twelve. About one third of the children took into account both distance and the speed of the oncoming vehicle. Oxley et al. (2005) conducted two experiments to investigate the effect of age in selecting safe time gaps in a simulated road-crossing task. Lobjois and Cavallo (2007) focused on the effect of age, vehicle speed, and time constraints on gap acceptance decisions. Their experiments are similar to the experiments of Oxley et al. (2005). Elderly participants accepted lower time gaps at higher speeds. The young-old group had a safety margin between 0 2 s while the old-old group had a negative safety margin of up to -10 s. A negative sign indicates underestimated time required crossing the road. Dommes et al. (2012) used simulator-based training for a group of 20 pedestrians whose age ranged from 65 to 83 to reveal if any training could change the perception of elderly pedestrians for selecting adequate gap for safe crossing. The results of the training showed that elderly participants adopted higher safety margins, made more safe decisions, and made fewer unsafe decisions on the post-tests than pre-test. Lobjois

3 Pelin Onelcin et al. / Transportation Research Procedia 22 (2017) et al. (2013) obtained higher safety margins at lower vehicle speeds. Safety margin was 1.81 s at 60 km/h speed while it was 2.29 s at 40 km/h vehicle speed. Liu and Tung (2014) found a lower safety margins for older pedestrians at their simulation based study. Koh and Wong (2014) studied the gap acceptance of non-compliant pedestrians who crossed during red light The average accepted time gap for near ends was 6.3 s and for far end crossing it was 5.2 s. In the present study each intersection was observed on weekdays during afternoon peak ( ) and evening peak ( ) hours using six video cameras. During a preliminary study it was observed that the major part of the illegal crossings were within the 25 m of the crosswalk. ANOVA analysis were conducted to reveal the factors (gender, age, group size, and items carrying) affecting the pedestrian walking speed and safety margin. The purposes of this research are to obtain the crossing speed of the pedestrians and compare the result with the given value by Turkish Standards Institution; to compare the observed delays with the delays computed by the commonly used equations in literature; and to find out the factors affecting the safety margin. 2. Methodology Six signalized intersections in Izmir, which is the third biggest city of Turkey with 4,113,072 inhabitants, were observed. Table 1 shows the observation sites and their characteristics. There are two three-legged intersections (Sirinyer and Karsiyaka) and four four-legged intersections (Cankaya, Sair Esref, Bostanli and Narlidere). At each intersection the densest crosswalk was chosen to be observed. Crosswalks are located on divided roadways and there are two lanes at each direction. Table 1. Characteristics of the observed intersections Location Crosswalk Geometry Signal Cycle Traffic Volume Number of Land Use (m) (s) (veh/h) observations Length Width Green Red Residential Commercial Through traffic Sirinyer /682 Karsiyaka /1390 Cankaya /750 Bostanli /461 Sair Esref /766 Narlidere /866 Data was collected at the signalized intersections with video recording technique. Two hours of recordings were conducted at each intersection during afternoon peak (12:30-13:30) and evening peak (17:00-18:00) hours in The posted speed limit at Bostanli intersection is 30 km/h whereas at other intersections it is 50 km/h. Pedestrians age was categorized in three groups; the young group (10-19 years old), the adult group (20-59 years old), and the senior group (over 60 years old). Group size was categorized in two groups; individuals, and pedestrians who walked within a group of two and more people. The delay model developed by Braun&Roddin is given in Equation 1. In this equation d is average pedestrian delay, C is cycle length, and g is green time. The model was developed under the assumptions of uniform pedestrian arrival rate, complete signal compliance, fixed cycle length, and no pedestrian actuation. d C g C (1) Another equation is given assuming that pedestrians receive no delay if they violate traffic signals. Broun&Roddin introduces a fraction factor F to describe pedestrians who arrive during non-green phases and comply with traffic signals. Equation 2 is as follows.

4 6 Pelin Onelcin et al. / Transportation Research Procedia 22 (2017) 3 12 d 0.5( C g) F C 2 (2) The delay model developed by Virkler is given in Equation 3. In this equation A is clearance time. d ( C ( G 0.69 A)) 2C 2 (3) At six signalized intersections a total of 444 (142 female and 302 male) pedestrian crossings within the 25 m from the crosswalk were observed. When the pedestrian stepped into lane if the lane was empty, it was not included in the safety margin calculations. Hence only the data of 334 (111 females and 223 males) pedestrians were analysed, that is the 75.23% of the pedestrians who crossed illegally. For safety margin measurements, only the data of the first lane crossings were evaluated. Pedestrians could be at either side of the road (at refuge, at sidewalk or at the parking lane). 3. Results 3.1. Crossing speed Crossing speed was computed by distance covered by the pedestrian divided by the crossing time. The crossing time is measured by the time difference between the starting time of the crossing (the pedestrian steps into the lane) and the ending time of the crossing (the pedestrian reaches the sidewalk) minus waiting time at the median since the crossing speed is zero while waiting. The total observed number of pedestrians crossing at the crosswalk is 2,694. Table 2 shows the average and the 15 th percentile crossing speeds computed at each observed intersection. The highest crossing speed is seen at Bostanli intersection. Here, the traffic volume is much lower compared to other intersections. At Karsiyaka intersection the lowest crossing speeds were observed. Table 2. The average and the15 th percentile crossing speeds at each observed intersection Gender Age Group size Items carrying Pedestrian crossing speed (m/s)/ Pedestrian 15 th percentile crossing speed (m/s) Cankaya Sirinyer Karsiyaka Bostanli S. Esref Narlidere Female 1.19/ / / / / /1.11 Male 1.25/ / / / / /1.04 < / / / / / / / / / / / /1.04 > / / / / / /0.92 Individual 1.27/ / / / / /1.04 Group of / / / / / /1.04 With items 1.22/ / / / / /0.93 Without 1.25/ / / / / /1.10 items Table 3 shows the average crossing speeds, 15 th percentile crossing speeds and standard deviations based on gender, items carrying, group size, and age. Males had a greater crossing speed than females. Pedestrians with items had a lower crossing speed than pedestrians without items. Pedestrians who moved in groups of two and more people had a lower crossing speed than individuals. The young pedestrians had the highest crossing speed followed by the adults and the senior pedestrians. The average crossing speed is found to be 1.31 m/s and the average 15 th percentile crossing speed is found to be 1.07 m/s. Compared with the recommended crossing speed of 1.4 m/s in Turkish Standards (TSI, 2012) the need of a revision is quite obvious.

5 Pelin Onelcin et al. / Transportation Research Procedia 22 (2017) Table 3. The average and the 15 th percentile crossing speed of all pedestrians Gender Age Group size Items carrying N Avg. Speed (m/s) Std. Dev. 15 th Percentile Speed (m/s) Female 1, Male 1, < , > Individual 1, Group of 2+ 1, With items Without items 1, Total 2, The following graphs show the crossing speed distribution of the pedestrians. In Figure 1 pedestrians crossing speed distribution based on gender is given. The 15 th percentile crossing speed is very close between females and males. The speed difference is clearly visible above 60 th percentile crossing speed. Figure 2 shows the pedestrians crossing speed distribution based on items carrying. In the graph it can be clearly seen that pedestrians without items walk faster than pedestrians with items. The 15 th percentile crossing speed is close for both groups. Figure 3 shows the pedestrians crossing speed distributions based on group size. Individuals significantly walked faster than pedestrians who moved in groups of two and more people. Finally, Figure 4 shows the crossing speed distribution based on age. The difference in crossing speeds between senior and adult pedestrians and senior and young pedestrians is more significant compared to adult and young pedestrians. The slowest group is the senior pedestrians. Percentage below given speed 1,0 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0 0,0 0,3 0,6 0,9 1,2 1,5 1,8 2,1 2,4 Walking speed (m/s) Males Females Percentage below given speed 1,0 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0 0 0,3 0,6 0,9 1,2 1,5 1,8 2,1 2,4 Walking speed (m/s) With items Without items Figure 1. Crossing speed distribution based on gender Figure 2. Crossing speed distribution based on items carrying

6 8 Pelin Onelcin et al. / Transportation Research Procedia 22 (2017) 3 12 Percentage below given speed 1,0 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0 0 0,3 0,6 0,9 1,2 1,5 1,8 2,1 2,4 Walking speed (m/s) Percentage below given speed 1,0 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0 0 0,3 0,6 0,9 1,2 1,5 1,8 2,1 2,4 Walking speed (m/s) Individuals Group of 2+ Young Adult Senior Figure 3. Crossing speed distribution based on group size Figure 4. Crossing speed distribution based on age ANOVA analyses were conducted to reveal the factors affecting the pedestrian walking speed. In Table 4 these factors are listed from the most significant to the least significant. Group size is the most significant factor with F(1,2693) = Group size and age have the same significance level (p<0,0001). The ANOVA analysis revealed a significant effect of items carrying F(1,2693)=11,955, p=0,001, gender F(1,2693)=5.605, p<0,05, and age*group size interaction F(1,2693)=3.203, p<0,05. Table 4. ANOVA analyses on crossing speed Factor F Significance level Group size 103,089 0,000 Age 51,060 0,000 Items carrying 11,955 0,001 Gender 5,605 0,018 Age*Group size 3,203 0, Pedestrian delay Pedestrians delays obtained from six signalized intersections are compared with the results of the equations presented in HCM2010, by Braun&Roddin, and by Virkler. Two different delays have been computed. The first one is the delay experienced by pedestrians who arrive during red phase and wait until the green phase to cross (Delay1). The second is the total delay both experienced during red and green phases (Delay2). Table 5 shows the results of the equations discussed above and the observed delays. The results of Virkler equation gave relatively better results compared with the observed delays. When Delay1 and Delay2 are compared it can be seen that Delay2 has lower values. This shows that illegal crossings (crossing during red phase) decreases the delays. Except for Karsiyaka intersection the highest delay is computed by HCM2010 equation. HCM 2010 only considers the effective green duration and the signal cycle. Braun& Roddin assumed that the pedestrians who obey the signal rules experience delays. Virkler added 69% of the clearance phase to green duration. The results of Virkler equation are lower than the results of HCM2010 equation. The results of HCM 2010, Braun&Roddin equations are compared with Delay1 and the results of Virkler equation is compared with Delay2. The following figures are created to visualize the relation between the equation results and the observations. Although there does not exist a significant relation between the observed delays and the computed delays, the highest R 2 value was obtained by Virkler equation among the given models. However, even in this model large discrepancies are clearly visible from the graph. This suggests that none of the models give good results to explain the observed delays.

7 Pelin Onelcin et al. / Transportation Research Procedia 22 (2017) Table 5. Observed and computed delays HCM 2010 Braun&Roddin Delay1 Virkler Delay2 Cankaya 22,05 12,44 21,05 12,27 16,34 Sirinyer 32,81 21,35 32,28 19,36 25,51 Karsiyaka 13,51 14,00 22,97 15,62 21,55 Bostanli 25,35 18,09 15,67 8,76 6,71 Çankaya 23,25 15,83 21,94 23,24 16,66 Narlidere 28 27,35 23,34 20,81 13,65 Figure 5 shows the relation between the observed delays (Delay1) and the delays computed with HCM 2010 method. The R 2 value is found to be Computed delay (s) y = 0,5404x + 11,799 R² = 0, Observed delay (s) Figure 5. Observed delays vs HCM2010 method Figure 6 shows the relation between the observed delays (Delay1) and the delays computed with Braun&Roddin method. The R 2 is lower compared to HCM2010 method. Computed delay (s) y = 0,3246x + 10,751 R² = 0, Observed delay (s) Figure 6. Observed delays vs Braun&Roddin method Figure 7 shows the relation between the observed delays (Delay2) and the delays computed with Virkler method. The R 2 is found to be

8 10 Pelin Onelcin et al. / Transportation Research Procedia 22 (2017) 3 12 Computed delay (s) y = 0,4103x + 9,809 R² = 0, Observed delay (s) Figure 7. Observed delays vs Virkler method The results showed that the above mentioned models are not suitable to reflect the local behavior of pedestrians in Izmir, Turkey. Researchers who confronted the same problem in different countries tried to make use of additional parameters to better reflect the pedestrian delay. Li et al (2005) developed a pedestrian delay model in China based on non-uniform arrival rate with the following parameters; average delay of pedestrians arriving during green phases, adjustment factor for non-uniform arrival rate, absolute value of the decreasing line and effective red time. Marisamynathan and Vedagiri (2013) showed that the existing models failed to provide necessary accuracy of pedestrian delay estimation at signalized intersections in India because those models did not consider all possible pedestrian crossing behaviors under mixed traffic conditions. For the purpose of developing a new delay model they introduced three components. The first component is average waiting time delay which is developed from HCM 2000 model. The second component is crossing time delay (walking speed is set to 1.34 m/s). And the third component is pedestrian-vehicular interaction delay time which is calculated from field data. For future research a new model may be developed including new components to the existing models or by modifying them for Turkey Safety margin Safety margin is the time that a vehicle needs to arrive to the point where the pedestrian crosses. The data of 334 pedestrians were analysed for safety margin. The safety margin and distance gap observations are discussed in detail in a previous study of Onelcin and Alver, (2015). Here, a brief summary of the analyses is given in order to present the safety perception. In Table 6 safety margin values based on gender, items carrying, group size and age are given. The minimum safety margin for females is found to be 1.45 s. The minimum safety margin is lower for males (1.19 s) compared to females. The mean safety margin is also found lower for males (7.43 s) than for females (8.25 s). The maximum safety margin is measured as s. No pedestrians crossings were observed within the 25 m from the crosswalk at intersections where the speed limit was 50 km/h. Table 6. Observed and computed delays N Mean Std. Deviation Minimum Maximum Gender Female 111 8, , Male 223 7, ,19 28,37 Age With items Without items Individuals Group size Group of Young 56 7,68 4, Items carrying Adult Senior Total 334 7, ,

9 Pelin Onelcin et al. / Transportation Research Procedia 22 (2017) Table 7 shows the ANOVA analyses based on safety margin. Both speed limit F(1,333)=4,997and position*speed interaction F(2,332)=4,499 had a significance level of p<0,05. Safety margin decreased as the speed limit increased. This is in line with the previous studies of Oxley et al. (2005), Lobjois and Cavallo (2007) and Lobjois et al. (2013). When the vehicle speed was over 30 km/ h pedestrians accepted lower time gaps to cross. Table 7. ANOVA analyses on safety margin. Factor F Significance level Speed limit ,026 Position*Speed limit , Conclusions Six intersections were observed to measure the average and the 15 th percentile crossing speeds, and to find the factors affecting the crossing speed and safety margin. Delay methods were compared with the observed delays to find the most relevant method that reflects the measured delays. The factors affecting the crossing speed were found to be the group size, age, items carrying, gender, and age*group size interaction. Individuals walked faster than pedestrians who crossed in groups. Males walked faster than females. Pedestrians who did not carry any items walked faster than pedestrians who carried items. Young pedestrians walked faster than adults and adults walked faster than senior pedestrians. The average crossing speed is found to be 1.31 m/s and the average 15 th percentile crossing speed is found to be 1.07 m/s. The recommended crossing speed is 1.4 m/s in Turkish Standards which is higher than the observed crossing speed. The crossing speed is used in signal timing arrangements. If the crossing speed is overestimated then pedestrians will not be able to complete their crossings within the given time. On the other hand if it is underestimated then the vehicles will have delays at the intersections. Thus, to apply the optimum signal timings the recommended crossing speed should reflect a realistic crossing speed of the pedestrians. Where the elderly pedestrians population is high a lower crossing speed should be considered. In this study the recommended 15 th percentile crossing speed is 0.93 m/s. TSI should revise the present crossing speed and recommend a more realistic crossing speed value. The delay method developed by Virkler gave relatively better results. Nevertheless, it is still incapable of reflecting the local behaviour of pedestrians. It was observed that pedestrians arriving during green phases also experienced delays. Further studies might be useful to develop a delay model that considers the pedestrian nonuniform arrivals and signal non-compliance for developing countries. The safety margin was measured using the data of 334 pedestrians who crossed within the 25 m from the crosswalk. The minimum and maximum safety margin was found to be 1.19 s and s, respectively. Age did not reveal a significant effect on safety margin in contrast to the previous studies (Connelly et al., 1998; Oxley et al., 1997, 2005; Lobjois and Cavallo, 2007; Dommes et al., 2012; Lobjois et al., 2013; Liu and Tung, 2014). Speed limit and position*speed limit interaction revealed significant effects on safety margin. Safety margin decreased as the speed limit increased. In literature researches have been made on pedestrian behaviour, however they generally focused solely on crossing speed, or on gap perception, and limited number of studies focused on pedestrian delays. In this study these three aspects of pedestrian behaviour were analyzed simultaneously. Although, observations were performed in different parts of the city to minimize the site-specific effects, expanding the study to different cities would provide evidence to confirm the findings of the present study. References Bennett, S., Felton, A., Akcelik, R., Pedestrian movement characteristics at signalized intersections, Proceedings of the 23rd Conference of Australian Institutes of Transport Research (CAITR 2001), Monash University, Melbourne, Australia, December Braun, R., Roddin, M., NCHRP Report 189: Quantifying the Benefits of Separating Pedestrians and Vehicles. Transportation Research Board, National Research Council, Washington, DC.

10 12 Pelin Onelcin et al. / Transportation Research Procedia 22 (2017) 3 12 Chandra, S., Bharti, A.K., Speed Distribution Curves for Pedestrians during Walking and Crossing. Procedia - Social and Behavioral Sciences 104, Chu, X., Baltes, M.R., Pedestrian Mid-block Crossing Difficulty. National Center for Transit Research (NCTR). Connelly, M. L., Conaglen, H. M., Parsonson, B. S., Isler, R. M., Child Pedestrians Crossing Gap Thresholds. Accident Analysis and Prevention, 30(4), Dommes, A., Cavallo, V., Vienne, F., Aillerieb, I., Age-related Differences in Street-Crossing Safety Before and After Training of Older Pedestrians. Accident Analysis and Prevention, 44(1), Gates, T.J., Noyce, D.A., Bill, A.R., Ee, N.V., Recommended Walking Speeds For Pedestrian Clearance Timing Based On Pedestrian Characteristics, Transportation Research Board Annual Meeting CD-ROM. Koh, P., Wong, Y., Gap Acceptance of Violators at Signalized Pedestrian Crossings. Accident Analysis and Prevention, 62, Lange, F., Haiduk, M., Schwarze, M., Eggert, F., The Dark Side of Stimulus Control Associations between Contradictory Stimulus Configurations and Pedestrians and Cyclists Illegal Street Crossing Behavior. Accident Analysis and Prevention, 43, Li, Q., Wang, Z., Yang, J., Wang, J., Pedestrian Delay Estimation at Signalized Intersections in Developing Cities. Transportation Research Part A 39, Liu, Y.C., Tung, Y.-C., Risk Analysis of Pedestrians Road-crossing Decisions: Effects of Age, Time Gap, Time of Day, and Vehicle Speed. Safety Science, 63, Lobjois, R., Cavallo, V., Age-related Differences in Street-crossing Decisions: The Effects of Vehicle Speed and Time Constraints on Gap Selection in an Estimation Task. Accident Analysis and Prevention, 39, Lobjois, R., Benguigui, N., Cavallo, V., The Effects of Age and Traffic Density on Street-crossing Behavior. Accident Analysis and Prevention, 53, Manual of Uniform Traffic Control Devices MUTCD, US Department of Transportation, Federal Highway Administration Manual of Traffic studies (1999) Institute of Transportation Engineers, US. Marisamynathan, S., Vedagiri, P., Modeling Pedestrian Delay at Signalized Intersection Crosswalks under Mixed Traffic Condition. 2nd Conference of Transportation Research Group of India. Procedia - Social and Behavioral Sciences,104, Onelcin, P., Alver, Y., Illegal Crossing Behavior of Pedestrians at Signalized Intersections: Factors Affecting The Gap Acceptance. Transportation Research Part F, 31, Oxley, J., Fildes, B., Ihsen, E., Charlton, J., Day, R., Differences in Traffic Judgements between Young and Old Adult Pedestrians. Accident Analysis and Prevention, 29(6), Oxley, J. A., Ihsen, E., Fildes, B. N., Charlton, J. L., Day, R. H., Crossing Roads Safely: An Experimental Study of Age Differences in Gap Selection by Pedestrians. Accident Analysis and Prevention, 37(5), Tarawneh, M.S., Evaluation of Pedestrian Speed in Jordan with Investigation of Some Contributing Factors. Journal of Safety Research, 32(2), Transportation Research Board, Highway Capacity Manual. National Research Council, Washington, DC. Turkish Standards Institute, Urban Roads - Design Criteria on Sidewalks and Pedestrian Areas (TS12174). Virkler, M., Pedestrian Compliance Effects on Signal Delay. Transportation Research Record 1636,