The t table value c) Effect of Side roads
was 2.776 for pedestrian accidents and 2.571 for all other The number of side roads is an important parameter which accidents (5% level of significance). Here in Table IV, t value affects the accident rate. In section 5&6 there are 6 side roads of each regression coefficient is greater than the t table value, it each. Here, accident rate is more.
means that all the independent variables contribute significantly
to the explanation in the dependent variable. In Table V, t value d) Effect of Speed
of some regression coefficients such as width of the road, Mismatch of speeds on road is a direct impact from the alignment of the road, number of side roads and speed are heterogeneous traffic. Slow moving vehicles become a barrier lower than the t table value; it means that all the independent to the movement of fast moving vehicles. Second problem is variables do not contribute to the explanation in the dependent the over speeding of certain private vehicles to reach there variable. Therefore it is clear that traffic volume has greater destination. This is most common in many places. Speed has a effect on the accidents than speed.
greater influence on fatal accidents. For a 5% increase of speed,
fatal accidents increase to 3.5%.
B. Results And Discussion
a) Effect of Width
e) Effect of Volume
Almost all sections have the same width. However, it will When volume increased by 2%, total accidents, rear end be appropriate to study the effect of one parameter on accidents accidents, fatal accidents, grievous accidents and minor
accidents increased by 1.3%, 1.2%, 15.6%, 5.5% and 4.6%
10. When width was 5% more, an increase in pedestrian
respectively.
volume and traffic volume by 5%, reduced accidents
by 0.7%.
f) Combined Effect of Width and Speed
When the width of the road increased by 5%, a speed of 5% C. Validation of the Model
increase will not cause any accidents. This shows that width has Validation of the model was done and root mean square major influence on accidents. In this case, total accidents and error between the observed and the expected values were found fatal accidents reduced by 3.8 % and 2.8 % respectively.
out. For the validation two road links were used, Pongumoodu
–Ulloor and Pongumoodu –Sreekaryam.
g) Combined Effect of Width and Volume
RMSE value was obtained as 0.379, 0.435, 0.316, 0.324,
When the volume of traffic increases, for reducing the 0.280, and 0.445 for total accident; rear-end accident, fatal number of accidents, the width of the road has to be increased accident, grievous accident, minor accident and pedestrian proportionally. When volume increased by 2%, width should be accident model respectively. Here RMSE value obtained is very
increased by 10% for 1.5% accident reduction.
low and hence the developed models are valid.
h) Factors Affecting Pedestrian Accidents
The model shows that there are many numbers of factors
V. SUMMARY AND CONCLUSIONS
which depends on pedestrian accidents.
The following conclusions are drawn from the analysis.
1. Presence of footpath reduces accidents to a great
When width increased by 5 %, total number of accident
extend.
reduced by 1.8%, rear end accident reduced by 1.8% and
2. When width increased by 5%, pedestrian accident
minor accident reduced by 1.5%.
reduced by 6.1%. Section 4 was a good example for Speed has a greater influence on fatal accidents. For a 5%
this. Section 4 has 8.8 meter width. There pedestrian
increase of speed, fatal accidents increase to 3.5%.
accidents were low compared to other sections.
When volume increased by 2%, total accidents, rear end
3. When road is in straight alignment, both drivers and
accidents, fatal accidents, grievous accidents and minor
pedestrians can see each other which results in
accidents increased by 1.3%, 1.2%, 15.6%, 5.5% and 4.6%
accident reduction. But here in section 1 & 2 even
respectively.
though alignment is straight, pedestrian collision are When volume increased to 2%, width should be increased
more. This is due to the high speed at those sections.
to 10% for 1.5% accident reduction.
When alignment is straight drivers have a good sight RMSE value was obtained as 0.379, 0.435, 0.316, 0.324,
distance. So they have a tendency to drive in greater
0.280, and 0.445 for total accident; rear-end accident, fatal
speed. This in turn results in collision. Sections 9&10
accident, grievous accident, minor accident and pedestrian
have curve and sharp alignment. There drivers and
accident model respectively. The lower RMSE value shows
pedestrians cannot see each other. In such case speed
that the developed model can predict the accidents
increase of a vehicle results in collision.
accurately.
4. Increase in side roads increases the accident rate. It is
clear from the data shown above.
5. When pedestrian volume increased by 18% accident
REFERENCES
rate increased by 0.8%.
6. When traffic volume increased by 5%, accident rate [1] M. Abdel-Aty, (2003). “Analysis of driver injury severity levels at increased by 0.33%.
multiple locations using ordered probit models.” J. Safety Res. , 34(5),
7. When speed increased by 10%, accident rate increased
597–603.
[2] M. Bedard, G.H. Guyatt, M.J. Stones, J.P. Hirdes, (2002). “The
by 0.3%.
Independent Contribution of Driver, Crash, and Vehicle Characteristics to
8. For 5% increase in traffic volume, pedestrian accident
Driver Fatalities.” Accident analysis and Prevention, Vol. 34, 717-727.