Project Presentation� Title: Travel Demand Modelling of IIT-Jodhpur campus

• KANDUKURI VENKATA SUBRAHMANYA SRIMUKH BABU -M20CI003
• M.Tech -Civil and Infrastructure Engineering, specialization in Energy
• IIT-Jodhpur

1

• Supervisor
• Dr. Ranju Mohan
• Assistant Professor
• Department of Civil and Infrastructure Engineering
• IIT-Jodhpur

Abstract

2

Travel demand modelling is one of the key processes in understanding the traffic and people travel pattern of any study area, Travel demand modelling involves creating mathematical model for trip generation and using the trips produced a trip-distribution matrix is generated, modal-split analysis and route-assignment are done subsequently. The study area for this project is IIT-Jodhpur campus. Based on the results from these processes insights regarding total trips made, distribution of trips to different zone, traffic-flow, modal-split, average travel time, average cost of travel, parking analysis, identifying busy routes and methods to avoid congestion, important routes and important timings for providing E-Rickshaw service in campus. Also becomes instrumental in determining if new routes need to be laid in campus. Although travel demand modelling is applied regularly to major cities it is very rarely applied on universities. Even though campuses produce lot of trips and the travel pattern inside the campus is very different from conventional city travel trips, very less amount of research work is done in this area. By applying the 4-stage travel demand model to the university, the project aims at providing information to the management so that it can provide good travel experience and transportation facilities to its population.

Key words: Campus travel demand modelling, 4 stage travel demand modelling, Trip-generation, Uniform growth trip distribution, Modal split analysis, All or none trip assignment model.

Contents

Introduction

Methodology

Literature Review

IIT-Jodhpur Campus study

Data collection

Trip Generation

Trip Distribution

Modal Split

Trip Assignment

Conclusions and Suggestions

References

​

3

Introduction and Literature review

4

• Travel demand modelling is an important process in understanding the flow of traffic and people travel pattern of any region or institute or of a study area. It involves 4 steps namely trip generation, trip distribution, modal analysis, and trip assignment
• Until recently this technique is mostly used for studying travel pattern in metropolitan cities and that of busy city zones.
• Universities with large campuses and huge population often generate a greater number of trips [1] and have issues of congestion [10], issues of parking on regular basis [5] and also few lanes in the campus gets very busy at times of a day and it becomes difficult for vehicles like cars and university buses to travel along such roads [10].
• Often university’s traffic act like a town-ship or a high-end educational area’s traffic containing lot of schools, offices, coaching centers and with lot of pedestrian traffic as well as that of individual vehicles.
• There are very less research articles available on university Travel demand modelling [1], and that too from INDIAN university the number is quite low. This proves that there is certain amount of gap in research in the area of campus travel demand modelling [4].
• Though there is this much travel is happening the amount of research done in this area to provide good transportation model for better Intra campus transportation facilities are very less.
• And this with lot of is one of the important processes to understand the traffic flow of any area.

Methodology

5

1. In this project first we tried to understand the campus and its residents in detail by building an entire campus model in Arc-GIS .
2. After having understood the different types of reasons for trip production a travel survey form is circulated and data related to travel is obtained.
3. Trip generation model is created and total number of trips are calculated.
4. In the next phase we have created a Origin-destination matrix for sample population and by using uniform growth factor model we created Origin-destination matrix for the entire population.
5. Modal split analysis for the entire population is calculated.
6. In the final step we have done trip assignment by using Arc-GIS.
7. All the 4 stages of transportation modelling are implemented for IIT-Jodhpur campus, results and conclusions are drawn.

​

​

Campus study

Data collection

Trip generation model

Trip distribution

Modal split analysis

Trip assignment

Conclusions and suggestions

Analytics

Figure 1

IIT-Jodhpur campus study

6

• IIT-Jodhpur has one of INDIA’s biggest campuses with an area of around 852 acres.
• IIT-Jodhpur has a student strength of around 2800 and faculty strength of around 200. Around 120 office staff and around 400 people who are family members of Staff of IIT-Jodhpur.
• IIT-Jodhpur also have lot of Non-academic working staff like security personal, hostel guards, campus doctors, medical staff, etc.

Figure 2: Campus master plan

Figure 3: IIT-J population

Table 1: IIT-J population over years

IIT-Jodhpur campus study

7

Geo-Referencing and digitizing the campus

• Initially campus image is taken and geo-referenced with co-ordinates of 1984 N 43 and all the building and locations are plotted in ArcGIS software by origin-destination details.
• This georeferencing has also helped in creating Network .
• The main aim of using ArcGIS is to properly locate all the building and also assigning them with exact co-ordinates and creating model is to use this model for measuring distances between any two locations in campus

Table 2 co-ordinates

Table 3 co-ordinates

IIT-Jodhpur campus study

8

2.3 Digitizing Campus Buildings, Infrastructure and Roads

1. In this project ArcGIS software is used to digitize the campus building, and Infrastructure.
2. Also, the entire road network is built in ArcGIS with 2-way movement on road with no one-way roads.
3. Also, all the junctions facilitate movement into any other road irrespective of left or right turn.

Figure 4:Digitized Campus model

IIT-Jodhpur campus study

9

Zones of campus:

Zone1: the entrance gate of the campus, security guards watch tower comes under this zone.

Zone2: Type-C faculty housing, Community center and other facilities nearby.

Zone3: This zone is a cluster of Main office building, Library, Lecture Hall and shop.

Zone4: student hostels as well as mess, and also sports facility and Gymnasiums.

Zone5: Type-B faculty housing and also laundry service in berms is also considered under this zone

Zone6: This zone houses the following departments-Computer science and mathematics, Bio and chemical departments, Humanities and social sciences.

Zone7: This zone is comparatively more populous than zone6 as it consists of more departments. Civil and Infrastructure, Mechanical, Metallurgy, Electrical, Physics, Material science, School of management and Entrepreneurship. It also has the campus health center.

Zoning of campus:

Zoning is a process in which the entire study area is divided into smaller regions called zones to study and understand the travel pattern that occurs in the study-area.

Figure 5: Campus zoning

Data collection

10

Data collection process is one of the key steps to obtain good results. “Quality of data is more important than the quantity of the data”.

Amount of Data required:

As the people of same categories more or less follows same kind of travel pattern it is considered that very large amount of survey data is not very much necessary for the modelling.

Direct and Indirect data:

• Here getting travel data directly from user or traveler is considered as direct data in this project
• The data about the security guards and watchmen travel which is obtained from the respective office i.e., not directly from the traveler is treated as indirect data.
• As few of the watchmen and security guards are not well educated obtaining indirect data has helped in maintaining the sanctity of the travel data.
• Also their travel pattern is fixed so collection of indirect data is best in this scenario

Figure 6: Travel survey form

Data collection

11

Avoiding sampling Bias:

1. It is very important to avoid sampling bias which happens to be taking survey from only one particular set or kind of population either more or less.
2. To avoid this, we tried our best to reach out to the different types of user persona present in the campus. So, to avoid that bias error in the data,
3. The survey data that is obtained is from sample population is proportionate to the actual population of the data

Groups of people in campus:

By observing as well as by conducting field studies and survey the following groups are predominantly present in the campus:

1. Students B. Tech and M. Tech
2. PhD scholars
3. Teaching Faculty
4. Non-teaching Staff
5. Security Guards for all academic, office and staff residential buildings
6. Watch men for Student hostels
7. Others

Figure 7: Actual population percentage distribution vs Sample population distribution percentage

Trip generation

12

Trip Generation

Travel demand modelling starts with trip generation. Because trip generation involves in finding or predicting the total number of trips made or attracted from different zones or from the entire study area.

Trip Production and Trip Attraction:

• If any trip consists one end of the trip either origin or destination it is called as Trip-Production.
• If both the ends of a trip do not consist of house or residential zone then it is called Trip-Attractions.

Types of Trips:

Trips are classified based on the purpose also the involvement of house or living zone

• Home based trips: These trips have one end as house or residential zone.
• Non-Home-Based trips: These trips do not have any end as house or residential zone

Figure 8: Home and Non-home trips

Trip generation

13

Home based trips in campus

In campus for students Hostels are considered as homes and for faculty and staff their respective type-B and type-C housing are considered as homes. So, any trip made by students involving hostels, for faculty either of type b or c is involved either in origin or destination then those trips are considered as home-based trips.

Difference between campus and city travel

• The difference between the population of campus and city is that in campus the population can be easily segregated into groups
• The number of different groups formed is very less.
• Also, the best part is all the members of a group will have almost similar travel pattern and that makes the results more accurate.

​

Factors that can Influence Trip Generation in Campus

Trip generation inside campus is different when compared to the trips generated outside campus. Trips in towns and cities happen for various reasons like for earning money, to work, to school. To cinema hall etc.

But in Campus it happens due to number of

trips made to sports facility, trips made to department, trips made to library, trips made to auditorium, trips made to lecture hall, mode choice, trips made to campus-dhobi, trips made to mess, trips made to hostel, trips made by security personnel, trips made by watch men all these factors decide the number of trips produced inside the campus.

Trip generation

14

• Fixed Travel Pattern:

When compared to students, PhD scholars, teaching and non-teaching faculty the trips made by Security personnel and Watchmen is very much restricted. Once they take charge of particular location, they do the duty for 8 hours. Also, after their duty they go back to their homes or rest rooms. So, their movements are very less and can be tracked more accurately.

1. Travel pattern of Security Guards inside the campus:

There are around 180 security guards in campus. These guards work in 8-hour shifts, so in a day there are 3 timings namely 6am-2pm, 2pm-10pm,10p6am. So a total of around 360 trips happen.

2. Travel Pattern of Hostel Watchman inside the campus:

There are around 42 Hostel watchmen in the campus. These guards work in 8-hour shifts, so in a day there are 3 timings namely 6am-2pm, 2pm-10pm,10pm-6am. During each slot around 14 security guards are posted at hostels. During the shift change around 28 guards will make travel in campus. Therefore, a total of 84 trips are possible. Around 30 guards have 2 wheelers.

3. Travel Pattern for Dhobi Service:

Dhobi service can be availed by the students twice a week. Also, the service on specific days is provided

• Flexible Travel Pattern:

When it comes to students, they are free to move anywhere in the campus and also, they don’t have much constraints apart from classes. For example, a student has 2 classes and a lab on a particular day, he needs to be present at lecture hall for 2 hours and at department for 1 hour and these trips are mandatory and, in a way, can be taken as fixed travel but after that the student can go to sports facility, his room, to shop, to dhobi etc. places and it depends on his wish which is difficult to predict without the data from the student so we will do direct survey with student.

​

Trip generation

• Regression Analysis for Total population Sample
• Y= 0.8978+ (0.2413X1) + (0.1587X2) + (0.6002X3) + (0.5406X4) + (0.6843X5) + (0.6738X6) + (0.5204X7) + (0.8426X8)

​

​

15

Table 4: Coefficients of the regression equation

Trip generation

16

Understanding the table

• The multiple R value obtained is 0.795 which can be considered as decent value w.r.t correlation.
• Adjusted R square value is around 0.61 that is acceptable.
• The significance F value is very small and this indicates our model is good.
• The t-values for almost all the coefficients are greater than 2 and a max value of 7.2.
• The p-values for almost all the coefficient are of order E-05 and least value is of 1.39 E-11

Table 5,6,7: regression equation analysis

17

Trip generation

• Predicting the number of trips that can be made in a day by an individual:

• The average number of trips made by the total population according to the sample size is 7.14 trips/person.
• In campus there are around 3086 people are there so total number of trips that are occurring in campus in a day is 22,034 trips. 180 security persons create around 360 trips and around 42 watchmen generate around 84 trips, so a total of 444 trips need to be added.
• Therefore, total number of trips = 22,478 trips

​

Trip generation

18

Trips made to lecture hall �

Lecture hall is located in zone-3 and the following results are the number of trips made by students, Ph.D. students and faculty members to lecture hall in a day

Figure 9: Comparison of trips made by students, Ph.D. students and faculty to lecture hall

Trip generation

19

Usage of Dhobi service �

Dhobi service center is located in zone-5 and the following results are the usage of dhobi service by students, Ph.D. students and faculty members

Figure 10: Comparison of usage of dhobi service by students, Ph.D. students and faculty

Trip generation

20

Trips made to department�

There are 2 zones for departments i.e. zone6 for department cluster-1 and zone7 for department cluster2 is located in zone-3 and the following results are the number of trips made by students, Ph.D. students and faculty members to respective department in a day

Figure 11: Comparison of trips made by students, Ph.D. students and faculty to department

Trip generation

21

Figure 12: Number of students having classes during working hours in a week

Table 8: Number of students having classes during working hours in a week

Number of students having classes during working hours in a week:

​

In this we tried to find out how many students may be travelling to classes during different hours on different days of a week

Assumptions made:

​

1. It is assumed that all the B.Tech courses will have same student enrollment for each class. Similar for M.Tech/MBA/ other programs as well.
2. We assumed no student drops from a course in between
3. Almost all the classes happen offline
4. Almost all the students of the class attend the lecture.

​

Trip distribution

• Trip distribution:
• Trip distribution is a process in which all the trips that are made in the study zone obtained from trip generation are distributed among different zones. It forms the 2^nd step in travel demand modelling. In this step all the trips are assessed and the matrix pertained to the origin and destination locations along with the trips is made. By this matrix we get the idea of how many trips are made for each zone. Trip distribution plays a crucial role in understanding the demand of travel between two zones.

​

22

Table 9: Trip distribution matrix for the sample population

Trip distribution

• Uniform growth factor model:
• In this model a growth factor is calculated and with that growth factor entire sample distribution matrix is multiplied to obtain the entire population distribution matrix. Growth factor is obtained by dividing the total number of trips that are made by the trips that are made by the population sample.
• T_ij = f * t_ij
• Here T_ij is the total number of trips that are made by population obtained from trip generation analysis, f is growth factor and t_ij is the trips made by the sample population

​

​

23

The growth factor is 8.3153

Table 10: Trip distribution matrix for the total population excluding security guards trips

Table 11: Trip distribution matrix for the total population

Trip distribution

24

Reason for selecting Uniform growth factor model:

There are numerous methods to distribute trips and for creating a origin destination matrix for entire population of study area but uniform growth model is selected because

1.It preserves the travel pattern of study area.

2. It gives good results when there is more uniformity in the travel nature of population of the study area

3. It is comparatively easy to implement and to understand.

​

Desire line diagram

All the distributed trips from different zones can be represented in a diagram called as Desire line diagram

​

Figure 13: Desire Line Diagram

Modal split

• Modal choice Modal choice is the third step in travel demand modelling. In this step the mode of transport used by the people travelling is captured. By modal choice we can be able to understand the preferred mode of transport of people in that region. Actually, Modal choice plays key role in transportation planning of a region. By Modal choice a better decision can be taken in choosing a particular type of public transport services in a region. It also aids in determining the type of roads that need to be laid based on the type of vehicles used in the region
• Mode choice pattern in campus: Upon field study as well from the data obtained most of the motorized vehicles are of faculty and staff. The number of motorized vehicles used by students is comparatively very less. Also, most students prefer to either walk or travel by cycle. Based on this it is understood that there is a pattern present in modal choice also, so it is better to use growth factor modelling here i.e., the sample size’s modal choice is preserved by the population’s modal choice and with this assumption the modal analysis for the entire population is done.

​

25

Figure 14: Comparison of mode choice split of students, Ph.D. students and faculty

Table12: Modal split table

Trip assignment

26

Trip or traffic assignment is a process in which the traffic that is flowing in the study area is assigned to different routes to understand the travel pattern of the study area and also how much usage of different roads is happening and also the traffic volume of different roads can be observed.

All or Nothing trip assignment:

In this type of trip assignment, the roads users choose a path that has minimum travel cost or minimum travel time. Based on this if there are n number of routes between A to B but among all the n routes the whole traffic moves in a single route i.e., the route which has least travel time or least travel cost.

Reasons for selection All or nothing trip assignment:

In campus there are very fewer alternate routes and those alternate routes present are also comparatively long in distances. Also, the campus is designed in such a way that people are encouraged to travel by Bicycle or by feet so the short distant routes are created in the heart of campus to easily connect all important locations. To facilitate travel of heavy-duty vehicles like construction trucks, water vans, load carrying lorries wide outer roads are built so that they need not enter the heart of campus and can travel bypass. Also, from field observations it is understood that almost all the student, faculty and staff population use short distant routes for their day-to-day commute. Based on all these it is decided All or nothing trips is best for campus trip assignment.

Figure 15: All or Nothing Trip assignment

Trip assignment

27

Most used Travel routes in the Campus

All the important routes that are used for travel are plotted in the network model, using least distance algorithm the routes are marked in Arc-GIS and in practice (via field observation) also those same routes are being used mostly by the maximum population of the campus

Table 13: roads used for travelling between different zones

Figure 16: Trip assignment routes

Trip assignment

28

Algorithm of the Arc-GIS in finding shortest routes:

In the network model, first roads have to laid by digitizing the real roads which are visible on the campus tiff image, then all the roads are to be designated as single network unit. The entire campus needs to be geo-referenced such that any point on the map will have co-ordinates. Now if an origin and destination locations are marked then the algorithm moves along the road networks which are constructed and will travel in the route which is having least distance based on the origin and destination co-ordinates

Travel cost matrix: This matrix provides the cost of travel between two zones for 1 trip

Table 14: Travel cost matrix

Figure 17: shortest route plotted in Arc-GIS

Conclusions and suggestions

29

Re-routing the motorized traffic:

• As of now all the pedestrians, cyclists, and all motorized vehicle travelers are using the same roads to travel, and due to this the congestion may increase if the population of campus and vehicle ownership numbers increase.

​

• Also it may lead to discomfort for both non-motorized users and motorized users to travel along same road as their travel behavior is different.

​

• So if different routes are used by motorized and non-motorized vehicle travelers then it would be reducing the crowd on existing roads as well as the travel time will be reduced

Figure 18: Current route vs Proposed route for motorized vehicles

Conclusions and suggestions

30

Hostel and home allocation based on department and nature of travel:

It is important to understand the nature of mode of transport of groups of population and their work nature and department to allocate residence is campus.

For a person if the distance between office and work is very less the chances of him using cycle or walk to the office is very high.

In other case if the distance is high the chance of using motorized vehicle is high.

So persons who are willing to cycle to work should be given preference to stay in Type-B housing facility.

And among student population Bachelors and masters students have less cycle usage so they should be given hostels near to lecture hall.

Figure 19: Hostels nearer to and far away from lecture hall

Conclusions and suggestions

31

Building cycle renting stands in campus

​

In recent time after the introduction of E-rickshaw in campus more number of students are using it to travel to far-away distance and the trips made by E-rickshaw is increasing.

​

So if cycle renting system is in place, trips of E-rickshaw can be reduced and also the locations which are more favorable for this rental stand is also suggested.

Figure 20: Bicycle rental locations analysis

Conclusions and suggestions

32

Parking Locations analysis:

Dedicated Parking zones

There are certain places in Campus that are dedicated for parking and these are facilities that are established with roofs and have markings indicating individual spaces. Generally, these zones are built near faculty housing B, C and also near the office blocks on the main entrance road.

Parking zones formed out of demand:

These are the places which attract lot of vehicles and so are formed into parking zones. Generally, these zones are formed near offices, departments and near lecture hall buildings

Time table design based on population capacity and to reduce parking congestion:

Creating time table for classes based on the number of students enrolling for slots so as to not over populate or under populate the roads during those slots. This looks a bit difficult task but with help of certain AI mapping tools this can be achieved in near future.

E-rickshaw services in faculty areas along with timings that match with class timings

E-Rickshaw service are more needed in the faculty housing zone, especially in type-C housing as it is far away. Also, the E-rickshaw should be scheduled to make trips from Type-c to departments based on the time-table of faculty who can’t drive to work

Figure 21: Parking locations analysis

Analytics

33

Average travel distance= Average trip distance * number of trips

= 799.5*7.14=5708 meters or 5.708 Kms in a day

Average travel time = Average travel distance/ average speed

Table 15: Zone to zone matrix

Table 16: Travel time matrix

Table 17: Travel cost matrix

Analytics

34

Table 18: Travel cost matrix

Figure 22: Travel distance network model for all buildings

References

35

[1]. Garikapati, Venu & You, Daehyun & Pendyala, Ram & Patel, Tushar & Kottommannil, Jiji & Sussman, Aaron. (2016). Design, Development, and Implementation of a University Travel Demand Modeling Framework. Transportation Research Record: Journal of the Transportation Research Board. 2563. 105-113. 10.3141/2563-15.

[2]. Dhulipala, Sowjanya & Tahlyan, Divyakant & Chalumuri, Ravi Sekhar. (2014). Travel Demand Modelling for a Metropolitan City.

[3]. Shekhawat, Arun. (2016). TRAVEL DEMAND FORECASTING - A Case study of sitapura industrial area.

[4]. Vechione, Matthew & Paudel, Sohil & Gurbuz, Okan. (2021). Trip Distribution Patterns on a University Campus: A Smarter Travel Demand Forecasting Approach. 1-7. 10.1109/ISC253183.2021.9562860.

[5]. Schoeman, Ilse. (2014). Land use and transportation integration within the greater area of the North West University (Potchefstroom Campus), South Africa: problems, prospects and solutions. 607-618. 10.2495/UT140501.

[6]. Sanghai, Sanket. (2014). Development of Trip Generation model using Activity based Approach. International Journal of Civil, Structural, Environmental and Infrastructure Engineering Research and Development (IJCSEIERD). 4. 61-78.

​

[7]. Berki, Zsolt & Monigl, Janos. (2017). Trip generation and distribution modelling in Budapest. Transportation Research Procedia. 27. 172-179. 10.1016/j.trpro.2017.12.023.

[8]. Chu, Zhaoming & Cheng, Lin & Chen, Hui. (2012). A Review of Activity-Based Travel Demand Modeling. 48-59. 10.1061/9780784412442.006.

[9]. Chang, Justin & Jung, Dongjae & Kim, Jaekyung & Kang, Taeseok. (2014). Comparative analysis of trip generation models: Results using home-based work trips in the Seoul metropolitan area. Transportation Letters The International Journal of Transportation Research. 10.1179/1942787514Y.0000000011.

[10]. Kaplan, David & Clapper, Thomas. (2007). Traffic Congestion on a University Campus: A Consideration of Unconventional Remedies to Nontraditional Transportation Patterns. Planning for Higher Education.

[11]. Santos, Georgina & Maoh, Hanna & Potoglou, Dimitris & Brunn, Thomas. (2013). Factors Influencing Modal Split of Commuting Journeys in Medium-size European Cities. Journal of Transport Geography. 30. 127-137. 10.1016/j.jtrangeo.2013.04.005.