R4TLI Conference Proceedings 2017 
ISSN: 2513-2512 Paper Reference: R4TL (17) – C31 
 
- 189 - 
Importance of Pedestrian Start-Up Times at 
Signalised Crosswalks 
Sewmini Jayatilake 
University of Peradeniya, Sri Lanka 
Niroshana Premachandra 
Road Development Authority, Sri Lanka 
Vasantha Wickramasinghe 
University of Peradeniya, Sri Lanka 
1. Introduction 
The total time required for pedestrians to cross the road is the summation of the time 
required to traverse the crosswalk and the pedestrian start-up time. But most signal 
time calculations do not pay attention to start-up time. This may result in dangerous 
situations caused due to pedestrian-vehicle interactions.  
Pedestrian start-up time is the difference between the time at which the pedestrian 
stepped off the curb to cross the road and the time at which the pedestrian signal 
phase turns into green. If the pedestrian steps off the curb during red phase, then it 
is early start-up time. If the pedestrian steps off the curb during green phase, it is 
delayed start-up time. 
Pedestrian behaviour varies according to several factors. Knoblauch et al. (1996), 
have considered weather, age, gender, curb height etc. and data were categorised 
accordingly. But Golani and Damti (2007), have considered the behaviour of a 
group of pedestrians. In this research, age, gender, group size and the familiarity of 
the pedestrians to the signal phase were considered. When considering the results 
of Vujanic et al. (2014), 17.5% pedestrians have started crossing during red phase. 
The highest number of offenders of both genders is under the age of 30 and the 
majority is male. 
This paper shows the importance in considering start-up times of pedestrians at 
signalised crosswalks.  
2. Objectives 
The objectives are primarily to find the start-up times of pedestrians crossing at 
signalised crosswalks, and secondly to find the variation of start-up time by age, 
gender, group size and pedestrian’s familiarity with signal phase. 
 R4TLI Conference Proceedings 2017 
 Paper Reference: R4TL (17) – C31 
 
- 190 - 
3. Methodology 
Signalised crosswalks with time indication were selected at intersections, and at 
each location video surveying was done. Using the captured video data, images 
were extracted with an accuracy of 0.1s. In finding the start-up times, pedestrians at 
the front line who are waiting for the pedestrian’s signal phase to turn into green 
were considered. Start-up times were calculated using the extracted images while 
observing the age, gender and group size of each pedestrian. The next analysis of 
start-up time was done using ‘Easy fit’ software and start-up times were graphically 
interpreted. ‘Z test’ was carried out to check whether there is a significant difference 
between data gathered in two main cities in Sri Lanka (i.e., Kandy and Colombo). 
Kandy has recently installed pedestrian signals while Colombo is having a long 
history of signalized pedestrian crossings. Probability analysis was done for early 
and delayed start-up times under each factor separately. Finally, a matrix of 
probability values was created. 
4. Results 
A total of 194 pedestrian crossing phases with 1,420 pedestrians were captured and 
considered in Kandy, and 158 crossing phases with 672 pedestrians were captured 
and considered in Colombo. Next, probability curves were obtained separately for 
early and delayed start-up times in the two cities and standard deviations were found 
for each as shown in Table 1. 
Table 1: Mean and Standard Deviation 
 Total start-up time (s) Early start-up time (s) Delayed start-up time (s) 
 Mean Standard 
Deviation 
Mean Standard 
Deviation 
Mean Standard 
Deviation 
Total 0.05 
(0.69) 
1.43 
(1.36) 
-1.23 
(-0.95) 
1.21 
(0.74) 
0.93 
(1.39) 
0.69 
(0.82) 
Male -0.08 
(0.74) 
1.45 
(1.36) 
-1.27 
(-1.00) 
1.17 
(0.76) 
0.92 
(1.41) 
0.67 
(0.83) 
Female 0.20 
(0.61) 
1.40 
(1.32) 
-1.16 
(-0.87) 
1.28 
(0.70) 
0.94 
(1.36) 
0.70 
(0.67) 
Alone 0.02 
(0.67) 
1.44 
(1.39) 
-1.24 
(-0.99) 
1.22 
(0.76) 
0.92 
(1.40) 
0.68 
(0.85) 
Group 0.33 
(0.84) 
1.37 
(1.12) 
-1.06 
(-0.70) 
1.16 
(0.59) 
1.02 
(1.35) 
0.75 
(0.68) 
10-30 -0.02 
(0.51) 
1.43 
(1.27) 
-1.23 
(-0.90) 
1.27 
(0.81) 
0.89 
(1.21) 
0.61 
(0.76) 
30-60 0.12 
(0.77) 
1.38 
(1.39) 
-1.17 
(-0.99) 
1.13 
(0.71) 
0.94 
(1.49) 
0.71 
(0.84) 
Above60 -0.23 
(1.06) 
1.66 
(1.28) 
-1.52 
(-0.93) 
1.43 
(0.55) 
0.97 
(1.58) 
0.68 
(0.68) 
  
 R4TLI Conference Proceedings 2017 
 Paper Reference: R4TL (17) – C31 
 
- 191 - 
Legend  
1.0 - Kandy 
(1.0) - Colombo 
Using those curves, the cumulative probability curves were plotted separately as 
shown in Figures 1 to 3. Z statistic showed that the start-up times in two cities are 
significantly different. The probability matrix created separately for the two cities 
can be used to find the probability of an individual to different start-up time ranges 
when the age, gender & group size of that pedestrian is known. 
 
Figure 1: Cumulative probability variation for early start-up times of male & female 
pedestrians in Kandy 
 
 
Figure 2: Cumulative probability variation for early start-up times according to 
group size in Kandy 
0
10
20
30
40
50
60
70
80
90
100
-5 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
Cu
m
ul
at
iv
e 
Pr
ob
ab
ili
ty
 (%
)
Start-up time (s)
Early start-up time (Gender)
Male
0
10
20
30
40
50
60
70
80
90
100
-5 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
Cu
m
ul
at
iv
e 
Pr
ob
ab
ili
ty
 (%
)
Start-up time (s)
Early start-up time (Group size)
Alone
Group
 R4TLI Conference Proceedings 2017 
 Paper Reference: R4TL (17) – C31 
 
- 192 - 
 
Figure 3: Cumulative probability variation for early start-up times according to age 
groups in Kandy 
 
Table 2: Probability Matrix 
Start-up time Total Male Female Alone Group 10 to 30 30 to 60 Above 60 
-4.1 to -5 1.58 1.10 0.48 1.3 0.27 0.69 0.75 0.14 
-3.1 to -4 1.58 
(0.15) 
0.89 
(0.15) 
0.69 
(0) 
1.51 
(0.15) 
0.07 
(0) 
0.21 
(0.15) 
1.17 
(0) 
0.21 
(0) 
-2.1 to -3 4.8 
(2.98) 
3.16 
(1.93) 
1.65 
(1.04) 
4.46 
(2.83) 
0.34 
(0.15) 
1.1 
(1.34) 
2.81 
(1.64) 
0.89 
(0) 
-1.1 to -2 9.47 
(7.89) 
6.04 
(5.36) 
3.43 
(2.53) 
8.92 
(6.99) 
0.55 
(0.89) 
2.06 
(2.98) 
6.18 
(4.17) 
1.24 
(0.74) 
0 to -1 24.37 
(19.64) 
13.52 
(11.9) 
10.84 
(7.74) 
21.83 
(16.67) 
2.54 
(2.98) 
6.31 
(8.33) 
15.58 
(10.42) 
2.47 
(0.89) 
0.1 to 1 35.35 
(25.45) 
17.30 
(17.11) 
18.05 
(8.33) 
31.64 
(21.73) 
3.71 
(3.72) 
8.37 
(10.57) 
23.75 
(13.10) 
3.23 
(1.79) 
1.1 to 2 18.74 
(28.87) 
9.20 
(20.24) 
9.54 
(8.63) 
16.27 
(23.51) 
2.47 
(5.36) 
4.32 
(9.85) 
12.70 
(16.07) 
1.72 
(2.98) 
2.1 to 3 4.05 
(11.90) 
1.99 
(8.48) 
2.06 
(3.42) 
3.43 
(9.67) 
0.62 
(2.23) 
0.55 
(3.72) 
3.02 
(6.99) 
0.48 
(1.19) 
3.1 to 4 0.27 
(3.12) 
0.14 
(2.08) 
0.14 
(10.40) 
0.21 
(3.12) 
0.07 
(0) 
0 
(0.45) 
0.27 
(2.38) 
0 
(0.03) 
 
Legend 
2.0 - Pedestrians are not familiar with the signal phase 
(2.0) - Pedestrians are familiar with the signal phase 
0
10
20
30
40
50
60
70
80
90
100
-5 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
Cu
m
ul
at
iv
e 
Pr
ob
ab
ili
ty
 (%
)
Start-up time (s)
Early start-up time (Age groups)
10 to 30
age
30 to 60
age
above 60
age
 R4TLI Conference Proceedings 2017 
 Paper Reference: R4TL (17) – C31 
 
- 193 - 
5. Conclusion 
Pedestrians in Kandy have an early start-up when compared with pedestrians in 
Colombo. Therefore, pedestrians’ familiarity to signal phases affect start-up times. 
Male pedestrians have early start-up times than female pedestrians. Pedestrians who 
are alone have early start-up times than pedestrians who cross in groups. Start-up 
times show different variations within different age groups. Therefore, 
consideration of start-up times of pedestrians is important to improve the safety of 
pedestrians. 
6. References 
 R.L. Knoblauch, M.T. Pietrucha & M. Nitzburg. (1996) Field Studies of 
Pedestrians Walking Speed and Start-up Time. In Transportation Research 
Record 1538, Journal of the Transportation Research Board, National 
Research Council, Washington. D.C., pp27-38 
 A. Golani, H. Damti. Model for Estimating Crossing Times at High 
Occupancy Crosswalks. In Transportation Research Record No.2002, 
Journal of the Transportation Research Board, National Academies, 
Washington. D.C.,2007, pp125-130 
 M. Vujanic, D. Pesic, B. Antic & E. Smailovic. Pedestrian Risk at the 
Signalized Pedestrian Crossing Equipped with Countdown Display. 
International Journal for Traffic and Transport Engineering. Volume 4(1), 
pp52-61 
Keywords: start-up time, early start-up, delayed start-up