Pedestrian Crossing Speed Model Using Multiple Regression Analysis

1
Pedestrian Crossing Speed Model Using Multiple
Regression Analysis
2
Transportation
I. Introduction

• Is the system of moving of people and goods from
  one point to the other in any given time.

pedestrians
3
Transportation cycle
Introduction
4
Main Problem
Introduction

• The typical crosswalk signal is of fixed time.
• The time allotted is not actuated to the number
  of pedestrians.
• No apparent provisions for the elderly and
  disabled.

5
Objectives

• To enumerate and identify the different variables
  that may affect pedestrian speed.
• To formulate a model to predict the pedestrian
  speed in a crosswalk for any given cycle.
• To establish an algorithm for an alternative
  computation of pedestrian speed.

6
Conceptual Framework
7
Significance of the Study

• The result may help the Traffic Engineers for the
  proper allocation of time for crossing
  pedestrians.
• The output of this study is a call to Urban
  Planners of the importance of pedestrian flow to
  the transportation system.

8
IV. Methodology
9
Scope and Limitations
IV. Methodology

• Criteria for selecting crosswalk
• Crosswalks must be in the proximity of school or
  any educational establishments.
• Must have a functioning signal at the time of
  survey.
• All must be in one corridor.
• Must have existing median or refuge island.

10
Study Area
IV. Methodology

• Four (4) crosswalks have been chosen to be the
  study area for this research. These crosswalks
  are
• CCP
• StopShop
• Lacson
• Gilmore

11
V. Presentation, Analysis and Interpretation of
data
Existing Situation of the Crosswalk
12
TABLE 4-6 - Corresponding properties of crosswalk
13
Speed Profile
Cumulative frequency curve of crossing speeds
taken at all crosswalks on i and j direction
14
Speed Profile
Table 4.5 - 15th and 85th percentiles on each
direction of all crosswalks
15
Non Compliance Ratio
Non-Compliance Ratio for each crosswalk and all
sites combined.
16
Design Speed Profile
Column graph of length green time ratio per
crosswalk
17
Relationships between variables
18
Speed Density Relationship
Speed on i direction vs. total density (K)
Speed on j direction vs. total density (K)
19
Speed Density Relationship
Speed vs. density relationship along i direction
(ki)
Speed vs. density relationship along j direction
(kj)
20
Speed Length Relationship
21
Model Estimation
22
Model 1

• Ws f (length (Lc), density along i direction
  (ki), density along j direction (kj))

23
Model 1
Goodness of fit statistics (Ws)
Analysis of variance (Variable Ws)
24
Model 1
Equation of the model
Ws 0.725 0.027Lc - 0.308ki -0.052kj
25
Model 2

• Ws f (length (Lc), flow along i direction
  (qi), flow along j direction (qj))

26
Model 2
Goodness of fit statistics (Ws)
Analysis of variance (Variable Ws)
27
Model 2
Equation of the model
Ws 0.722 0.026Lc - 0.182qi - 0.009qj
28
Model 3

• Ws f (length (Lc), volume along i direction
  (ki), volume along j direction (kj), green time
  (G))

29
Model 3
Goodness of fit statistics (Ws)
Analysis of variance (Variable Ws)
30
Model 3
Equation of the model
Ws 0.691 0.025Lc - 0.005Wvi - 0.001Wvj
0.003G
31
Model Selection
After determining the possible variables and its
relation to speed, we select MODEL 1 as our final
model.
Ws 0.725 0.027Lc - 0.308ki -0.052kj
This model has the highest value of F. Even if
MODEL 3 obtained the highest value of R2, the
researchers would like to emphasize that in
multiple regression, the high value of F rather
than R2.
32
Conclusion and Recommendations
33
Findings

• There are a significant number of pedestrians who
  do not clearly understand the meaning conveyed by
  the pedestrian signalization. About 60 or 6 out
  of 10 pedestrians do not follow the pedestrian
  signals.
• The pedestrian speed profile in all four
  crosswalk is slightly lower than the recommended
  speed profile.
• The signal timing allocation in these crosswalks
  is not based the length of the crosswalk. Even if
  the crosswalk length is increased, the signal
  timing does not necessarily be increased.
  (Existing Case)

34
Conclusions

• To enumerate and identify the different variables
  that may affect pedestrian speed.
• Several factors may affect the speed of the
  pedestrians and the researchers came to
  conclusion that it is not merely by density. The
  dimension of the crosswalk affect the speed in
  terms that when the length of the crosswalk is
  lengthen, the tendency is to increase their speed
  as not to be caught up by the movement of
  traffic. In the course of our research, we
  conclude that the presented variables are not
  enough to explain the variation on speed.

35
Conclusions

• To formulate a model to predict the pedestrian
  speed in a crosswalk for any given cycle.
• The researchers had formulated a model to
  predict the pedestrian signal but due to its low
  coefficient of determination, the researchers
  conclude that the model may not predict the
  correct pedestrian speed located outside the
  study area.

36
Conclusions

• To establish an algorithm for an alternative
  computation of pedestrian speed.
• Due to the low coefficient of determination of
  the model, the researchers failed to present an
  alternative computation of pedestrian speed.

37
Recommendations

• Additional model for obstructions and wall
  avoidance of the microscopic pedestrian
  simulation model is suggested to perform a better
  capacity analysis.
• Integration of pedestrian flow in planning
  and design.
• Inclusion of pedestrians to the education
  system.
• Use of fully actuated signals for
  pedestrians and of countdown pedestrian timers to
  minimize vehicular conflicts and non compliancy.

38

• Improvement of the automatic video data
  collection toward the occlusion problem is highly
  recommended to enhance the performance of the
  system for higher pedestrian traffic density.
• Further studies on pedestrian flow which can
  include
• Larger study area that can include all metro
  manila.
• Analysis of both signalized and unsignalized
  crosswalks.
• Analysis for both one-way and two-way pedestrian
  lane.

39

• Inclusion of model variables such as
• delay
• pedestrian generators (schools, offices, railway
  station, malls)
• pavement markings
• dissipation time
• vehicular movement and speed
• pedestrian interaction
• queuing time
• Pedestrian analysis using
• Benefit cost cellular model
• Cellular automata model
• Magnetic force model
• Social Force model
• Queuing network model

40
THANK YOU!