Traffic Study of Individual Travel Behaviour in Tiruchirapalli City Using GPS

ABSTRACT
GPS data helps in finding the accurate trip rate information like Travel time, Delay data, Congestion Index, Acceleration Noise, Vehicle speed and Trip ends. In this study, an algorithm is developed for identifying the trip ends of a vehicle by taking dwell time, heading change and speed as the criteria obtained from the GPS unit installed in the vehicle. This trip end data is layered onto the city maps for locating the trip ends and route travelled by the vehicle. Further the Delay, Congestion Index and Acceleration noise for each trip has been calculated. True start and end points of each trip were identified in advance in a trip log completed by participants so that the accuracy of each automated trip division method is measured and compared. A Heuristic model, which combines data collected through GPS and the road network, performs best, correctly identifying 94% of trip ends.

1. INTRODUCTION
Infrastructure development is the theme of the day. In India, the economic growth is greatly influenced by rapid development of Infrastructure. Transportation Infrastructure is a vital component not only as part of the Infrastructure but also as an efficient field to boost the growth of it. In this context there is a need to develop innovative methods which can outcome the conventional methods in terms of time taken, amount of labour and accuracy for Transportation demand modeling and transportation planning.

Trip rate, the number of trips made by household or an individual in a period of time (usually an hour or a day), is an essential measure in travel behaviour research and transportation demand modeling. It is widely used in planning for trip generation, emission modeling, as well as other related evaluations. Conventional survey methods, such as home interview, roadside interview, mail and phone surveys that collect travel information based on participant recall, are limited, especially in capturing short trips or trip chains. Short trips are frequently omitted by survey participants especially at the end of the day. Congestion analysis, the data for this involve the measurement of accurate speed, time and lot of spatial information where and when the delay has occurred. This data collection also poses similar problems as above.

As mentioned above there is a need for fast and effective means of survey methods for Transportation planning. This study is carried out to develop a Geographical Information System in which Global Positioning System is used as a tool, which can automatically do the trip rate analysis and congestion analysis, for the trips along the major roads of Tiruchiraplli City .

2. GPS APPLICATIONS IN TRANSPORTATION ENGINEERING
GPS can give real-time reliable data on position, speed of the vehicle in which it is installed, provided the quality of the instrument used. This data obtained from the GPS can be processed and utilized for various studies in Transportation Engineering. The various applications Include Trip Rate Analysis, Travel Time and Delay Studies, Automatic Vehicle Identification, Dynamic Route Guidance System, Advanced Travel Information System, Travel Behaviour analysis.

Several studies have been carried out in these areas, Stopher 2003 et.al used dwell time as the parameter to identify the trip ends from the GPS data and identified the trip ends. Jianhe Du and Lisa Aultman-Hall 2007et.al proposed a model which combines dwell time distance and direction change from the GPS data for the successful identification of trip ends. Michael A.P. Taylor , Jeremy E. Woolley, Rocco Zito 2000 et.al carried out congestion studies by integrating GPS and GIS.

3. OBJECTIVES OF THE STUDY

1. To develop an algorithm that identifies the trip ends using dwell time.
2. To develop maps for the identified trip ends and further identifying the additional trip ends occurred due to the heading change.
3. To find the delays occurred in the Journey time.
4. To find the Congestion indices like Congestion Index and Acceleration Noise for the trips that occurred in the major arterials of the city.

4. STUDY AREA
Tiruchirapalli, situated on the banks of the river Cauvery is the fourth largest city in Tamil Nadu, It has a population of 7, 46,062 according to 2001 census. Tiruchirapalli is well connected by road network, it is connected by National Highways NH-45, NH-65 and NH-210. Tiruchirapalli is the hub of Southern Railway's operation to connect this central part of Tamil Nadu to various parts of India, notably regions in Kerala, Andhra Pradesh, West Bengal, Maharashtra, Karnataka, Delhi, and Madhya Pradesh. Tiruchirapalli has an international airport about seven kilometers from the city, which operates flights to Indian cities, territories, and neighbouring countries including Malaysia, Sri Lanka, Singapore, and the Gulf.

5. METHODOLOGY
Four major steps involved in the project are

1. Data collection
2. Data Processing
3. Trip Rate Analysis
4. Congestion Analysis

Data Collection
This study focuses on finding the method for processing the large volumes of GPS data and uses it for travel behavior analysis. The method used is passive data collection from GPS; GPS is installed in a probe vehicle, which is a taxi cab and it requires no intervention from the driver it automatically records the spatial data and it is collected later for data analysis. The in-vehicle manual logs are also provided to the driver to compare the results that are obtained from the trip identification method.

Equipments and Software Used
The equipments used for carrying out GPS survey are:

1. HI -305 III GPS Receiver
2. HP iPAQ Pocket PC

The software used for the data collection from GPS receiver in to the palm top PC is Arc Pad 6.0.1

The HI -305 III GPS receiver is a very smart hinge design, can folded from 00 to 1800 accommodate the best satellite receiving angle. We can adjust the HI -305 III in best angle with our PDA so that the external antenna will not be needed. We can fix the device with receiver in the dashboard of the car. Also, we don't need to face the mobile device display to the sky for better receiving angle. The HP- iPAQ is a pocket PC used for loading the necessary data. With DGPS collections even sub-meter accuracies can also be achieved Fig.1 shows the Pocket PC connected with GPS receiver.


Fig. 1. HI -305 III GPS Receiver with Pocket PC attached

DATA PROCESSING
The GPS data will give continuous Latitude, Longitude, Time and several other data for every two second interval. The data is stored in Track log data file, this file is processed using the Microsoft Excel for further analysis. The processing is explained in the following steps.

Distance Measurement
The distance between the two consecutive Latitude Longitude points is obtained by using the formula:
d = R arccos(sin(j1) * sin(j2) + cos(j1) * cos(j2) * cos(l1-l2))

Where R is the radius of the earth and d is the distance

(l1, j1) and (12, j2) are the two consecutive latitudes and longitudes
This formula can be given in Microsoft Excel as shown below, for getting the distance between the two consecutive GPS points, where the first columns of the sheet must be latitudes and longitudes. The distance in miles is converted into meters by multiplying with 1609.44. After this the cumulative distance is calculated in meters.

Distance=(ACOS(((SIN(A1*PI()/180))*(SIN(A2*PI()/180)))+((COS(A1*PI()/180))*(COS (A2*PI()/180))*(COS((ABS(B2-B1))*PI()/180))))*180/PI())*69

Identification of Trip Ends
A potential trip end is flagged off by following the two criteria which from the research of Stopher et al. (2003)

1. A potential trip end was flagged if the elapsed time was equal to or greater than 120 seconds, the difference in successive latitude and longitude values was less than .0000510 and at the same time the speed must be zero.
2. When the heading changes within 15 2-s interval GPS points were between 1800 + 2 a trip end was marked.

An algorithm is developed to find the trip ends using the above criteria. The input file is a text file which is taken from Microsoft Excel and processed. The output gives the Latitude, Longitude and the time where trip end has occurred. After obtaining trip end, Latitude and Longitude of it is mapped onto the Tiruchirapalli city map in the Google Earth as shown in the Fig. 2. Then the route of the trip is also mapped onto the city map in Google Earth.


Fig. 2. Plotting of Trip End on Google Earth Imagery

Congestion
From the processed data the following parameters are calculated for congestion analysis.
Delay: A commonly accepted definition of delay is system delay (d) defined as the excess travel time above the minimum (free flow) travel time to traverse a network element. If T is actual time T0 is the free flow time, then the system delay is

D= T- T ₀

Congestion Index: The level of delay as defined by the system delay can be expressed in terms of a dimensionless Congestion Index (CI), as described by Richardson and Taylor (1978).

CI = (T- T ₀) / T ₀

Acceleration Noise: Acceleration Noise (AN) is a parameter derived from acceleration pattern contained within speed-time profile for the journey and provides a measure of the quality of traffic flow and thus of the level of congestion.

AN is given by the equation:



Where T = Total time
Δv_i = Change in the velocity
Δt_i = Change in time

5. ANALYSIS AND RESULTS

Trip Rate Analysis Trip ends are identified by processing the data as mentioned in the previous discussion. The GPS data collected for one week from the cab is processed and the trip ends for all the days are identified. The starting and ending time of the trips and distance travelled are also calculated, for example trip for a day are shown in Table 1 and these are presented on the maps using Google –Earth as shown in Fig. 3.

Table 1. Trip details for the day 26 -10-07

From	To	Start Time	End Time	Total travel Time	Length
Indian Bank Stop	Toll gate	10:50:12	10:54:38	0:04:26	2383.823
Toll gate	Mananrpuram	11:10:46	11:16:12	0:05:26	2691.396
Mananrpuram	Mannarpuram Junction	11:19:16	11:19:59	0:00:43	135.6311
Mannarpuram Junction	Indian Bank Stop	12:06:46	12:08:53	0:02:07	1273.791
Indian Bank Stop	Tarannalur	12:58:21	13:09:14	0:10:53	4765.389

Discussion

Identification of Trips

1. The trip ends identified by the algorithm considering 120s criteria includes some times the stopped delays during signals, like signal near Madurai junction which has the cycle time more than 120s. This false trip ends are eliminated by mapping the trip ends.
2. Some false trip ends are identified, if the vehicle stops for longer period. This can be rectified by drawing the graphs between Speed and Time for every 1000 GPS points.

Fig. 3. Trip details for the day 26 -10-07 on Google - Earth

Comparison of Trip ends
To verify the results obtained from the method used for identification of trip ends from GPS data and manual trip ends entered by the driver of the probe vehicle in the data log sheet provided with him. It is observed that on most of the days of the survey, the trip ends obtained from the algorithm are more than the manual trip ends. This indicates the trip reporting negligence of the driver, small trips are not entered into the data sheets by them. On 22-11-07 trip ends identified by the algorithm are less than those entered manually, because of the data loss in GPS data collection.

Congestion Analysis
The major trips identified in Trip rate analysis during the one week survey are considered for the congestion analysis. The congestion parameters Delay, Congestion Index, Acceleration are calculated for the above nine trips which are obtained from the probe vehicle. These results are consolidated and shown in Table 2 below

Discussion
Delays
Delays occurred were mainly due to the stops at the signaled intersections except on 23 -11-07. Signaled delays occurred at the following intersections

1. Mannarpuram Junction
2. Madurai Junction
   
   Table . 2 Congestion Analysis for Major Trips

   Date	From	To	Delay	Congestion Index	Acceleration Noise
   21-11-07	Railway station	Indian bank Stop	00:1:42	0.508	0.343
   21-11-07	Indian Bank Stop	Near Income Tax Office Stop	00:08:53	1.349	0.250
   21-11-07	Indian Bank Stop	K.K. Nagar Stop	00:00:38	0.144	0.035
   23-10-07	Chevrolet Show room	Indian Bank Stop	00:35:38	2.548	0.0107
   3-11-07	Indian Bank Stop	Trichy distilleries	00:07:49	1.098	0.084
   23-11-07	Trichy distilleries	Campaign Higher School Secondary School	00:15:56	2.01	0.018
   26-11-07	Toll Gate	Mannarpuram Stop	00:00:55	0.203	0.191
   26-11-07	Indian Bank Stop	Toll gate	00:01:11	0.395	0.228
   26-11-07	Indian bank Stop	Tarannalur	00:03:53	0.55	0.233

   
   
   
3. Toll Gate Junction
4. Head post office Junction

Hence these signals should be properly designed for coordination without causing delay to traffic. Congestion delays occurred mainly on 23-11-07 where the vehicle has traveled mostly on the NH-45.

Speed
The average speed of the probe vehicle for the trips except on 23-11-07 has been 28 Kmph. On 23-11-07 it is very low due to more number of congestion and signal delays, all the trips occurred on that day have very low speeds on NH-45 stretch between Palpanni to Toll gate.

Hence it is observed that the highway stretch between Palpannai to Toll Gate is accounting for traffic congestion. This is due to the construction work going on the stretch and also due to the large volume of traffic on this road.

6. CONCLUSION
In this study the need for automatic data collection system for transportation demand modeling and planning is discussed. The advantages of the Global Positioning for collecting continuous data required for transportation modeling and planning are utilized, it is observed that GPS gives more accurate and reliable data for trip rate analysis than the conventional methods like trip reporting by the drivers. The short trips that are not listed by the driver are identified by the method used for trip end identification using GPS. In this study it is observed that from the continuous speed and time data obtained from the GPS data, the travel behavior of the probe vehicle is studied and routes which have high congestion during the survey are identified. The congestion analysis has been carried out for the selected trips from the one week survey which occurred on main arterials of Tiruchirapalli city.

REFERENCES

1. Jianhe Du, Lisa AultmanHall (2007) “Increasing the Accuracy of Trip Rate Information from Passive Multi – Day GPS Travel Datasets: Automatic Trip End Identification Issues”, Transportation Research Part A 41.



2. Yasuo Asakura, Takamasa Iyro (2007) “Analysis of Tourist Behaviour Based on the Tracking Data Collected Using a Mobile Communication Instrument”, Transportation Research Part A 41.



3. Yasuo Asakura, Eiji Hato (2004) “Tracking of Individual Travel Behavior Using Mobile Communication Instruments”, Transportation Research Part C 12.



4. Sutti Tantiyanugulchai and Robert L. Bertini (2003) “Arterial Performance Measurement Using Transit Buses as Probe Vehicles”, 2003 IEEE.



5. M. L Kulkarni, Vijay Singh Chowdary (2003) “Global Positioning System: A Useful Tool for Intelligent Vehicle – Highway Systems”, Department of Civil Engineering IIT- Bombay.



6. Michael A.P. Taylor.Jermey E. Woolley, Rocco (2000) “Integration of the Global Positioning System and Geographical Information Systems for Traffic Congestion Studies”, Transportation Research Part C 8.



7. Nobuaki Ohmori, Yasunori Muromachi, Noboru Harata and Katustoshi Ohta (2000) “Analysis of Day-Day Variations of Travel Time Using GPS and GIS”, Institute of Environmental Studies, The University of Tokyo.