Академический Документы
Профессиональный Документы
Культура Документы
by
ACKNOWLEDGEMENTS
I take this opportunity to express my regards, indebtedness and profound sense of gratitude to my supervisor Prof. Geetam Tiwari and Dr.A.K.Swamy for their inspiring guidance, constant encouragement and ever cooperating attitude, which enable me to undertake the present work. I appreciate their understanding, untiring enthusiasm and the great care they took in bringing up the work in the present form. My foremost thanks are due to my parents for their encouragement, support, love and affection and moral boosting, which kept me going throughout the duration of the work. I sincerely thank Dr.Manika Agarwal(DIMTS) for providing the data used in this project and TRIPP (Transportation Research and Injury Prevention Programme), especially Mr.Rahul Goel, Research Scholar, TRIPP, for providing all the necessary data and help regarding the work.
August, 2013
CERTIFICATE
This is to certify that the thesis title Time Series Analysis of Vehicular Delay in Delhi is a bonafide record of work done by Mehvesh Mushtaq for partial fulfillment of the requirement for the degree of Master of Technology in Transportation Engineering, Department of Civil Engineering, Indian Institute of Technology (IIT) Delhi, New Delhi, India. She has fulfilled the requirements for the submission of this thesis, which to the best of my knowledge has reached the required standard. This thesis was carried out under my supervision and guidance and has not been submitted elsewhere for the award of any other degree.
(Dr. Geetam Tiwari) Professor Department of Civil Engineering, Indian Institute of Technology Delhi, New Delhi, India
Dr.A.K.Swamy Assistant Professor Department of Civil Engineering, Indian Institute of Technology Delhi, New Delhi, India
ABSTRACT
Speed Studies can be temporal or spatial i.e., studying speed variations over time or over space respectively. The objective of this project is to study temporal variation of Bus speed over various bus routes of Delhi and identify bottlenecks in traffic in time and space. It also aims to divide a route into segments, each segment being a part of road between Stopping points like Bus Stops, Intersections(3 ways,4 ways), and Roundabouts. The mean speed over each segment is calculated for all hourly time slots during which buses ply on the route for all days of the week. Thus an hourly speed profile for all sections of the route is available for all times of bus movement.
TABLE OF CONTENTS:
Chapter1. 1 1.1 1.2 1.3 1.4 1.5 Introduction Definition of Time Series... .7 Autocorrelation... 8 Correlograms...8 Box-Jenkins Models (Forecasting). 8 Why Time Series.8
Chapter2. 2 2.1 2.2 Literature Review Purpose of Literature review. 10 Literature review.. .10
Chapter 3. 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 Data Collection and Analysis Description of Available Data...23 Treatment of dataset1 23 Stationarity..24 Methodology of Project (for dataset1). .25 Results of Application of ADF test on data ..26 Interpretations of Results. .27 Data Set 2 ..28
3.7.1 Route 108 Up.31 3.7.2 Route 108 Down. ..36 3.7.3 Route 185 Up. ...41 3.7.4 Route 185 Down.....49 3.7.5 Route 411 Up.. ..54 3.7.6 Route 411 Down... 59 Chapter 4. 4. Conclusions .. ...69 4.1 4.2 Definition of Bottleneck. ..70 Scope for further studies. ..72
References....80
4
LIST OF FIGURES: Figure 1: Route 108 Up.31 Figure 2: Route 108 Down.36 Figure 3: Route 185 Up.41 Figure 4: Route 185 Down 49 Figure 5: Route 411 Up.54 Figure 6: Route 411 Down 59 Figure 7: Comparison chart of mean speed for all routes.. ...69 Figure 8: Comparison of mean speeds for different time slots .71
LIST OF TABLES: Table1. Critical values for DF and ADF tests...26 Table2. Compiled Results of ADF test on Dataset 1. ...27 Table3. Route Characteristics for Route 108 30 Table4. Segments for Route analysis 108 Up ...32 Table5.Conclusions for Route 108 Up..34 Table6. Segments for Route 108 Down 36 Table7.Conclusions for Route 108 Down 38 Table8. Route Characteristics for Route 185 39 Table9.Segments for Route 185 Up. .42 Table10.Conclusions for Route 185 Up.44 Table11. Segments for Route 185 Down ..49 Table12.Conclusions for Route 185 Down ...51 Table13. Segments for Route 411 Up ...55 Table14.Conclusions for Route 411 Up56 Table15. Segments for Route 411 Down ..60 Table16.Conclusions for Route 411 Down ...62 Table17. Comparison of mean speed for different time slots.... ...70 Table 18. Bottleneck speed ...71 Table 19. Comparison chart of Mean speeds over various Routes.. .73
CHAPTER - 1 INTRODUCTION
Speed Studies can be temporal or spatial i.e., studying speed variations over time and over space respectively. The objective of this project is to study temporal variation of Bus speed over various bus routes of Delhi and identify bottlenecks in traffic in time and space. It also aims to divide a route into segments, each segment being a part of road between Stopping points like Bus Stops, Intersections (Three ways,Four ways), and Roundabouts. The mean speed over each segment is calculated for all hourly time slots during which buses ply on the route for all days of the week. Thus an hourly speed profile for all sections of the route is available for all times of bus movement. The data used for this study is GPS (Global Positioning System) Data which provides the location of a Particular bus after regular intervals of time (in the data used for this project it is 10 secs approx.).This is ideal for studying the data as a time series as a Time-series is essentially an ordered sequence of values of a variable at equally spaced time intervals.
Introduction to Time-Series
1.1. Definition of Time Series: An ordered sequence of values of a variable at equally spaced time intervals. Time series analysis accounts for the fact that data points taken over time may have an internal structure (such as autocorrelation, trend or seasonal variation) that should be accounted for. 1.1.1. Applications: The usage of time series models is twofold: a) Obtain an understanding of the underlying forces and structure that produced the observed data. b) Fit a model and proceed to forecasting, monitoring or even feedback and feed forward control.
7
1.1.2. Types of time series: I) Continuous vs. Discrete: Continuous observations made continuously in time; Discrete observations made only at certain times. II) Stationary vs. Non-stationary: Stationary Data that fluctuate around a constant value; Non-stationary A series having parameters of the cycle (i.e., length, amplitude or phase) change over time. III) Deterministic vs. Stochastic: Deterministic time series This data can be predicted exactly; Stochastic time series Data are only partly determined by past values and future values have to be described with a probability distribution. This is the case for most, if not all, natural time series. So many factors involved in a natural system that we cannot possibly correctly apply all of them. 1.2 Autocorrelation: A series of data may have observations that are not independent of one another. To find out if autocorrelation exist, Autocorrelation Coefficients measure correlations between observations a certain distance apart. 1.3 Correlograms: The autocorrelation coefficient r(k) can then be plotted against the lag (k) to develop a correlogram. This will give us a visual look at a range of correlation coefficients at relevant time lags so that significant values may be seen. 1.4 Box-Jenkins Models (Forecasting): Box and Jenkins developed the AutoRegressive Integrative Moving Average (ARIMA) model which combined the AutoRegressive (AR) and Moving Average (MA) models developed earlier with a differencing factor that removes in trend in the data. 1.5 Why Time Series? What we need from a modeling technique or a data-analysis tool is an ability to respond quickly, provide simple forecasting techniques and ability to provide accurate detailed local forecasts.
Limitations of traditional Complex Model Systems and Model Packages: a) Data collection and preparation is an enormous task (because of behavioral and socio-economic variables). b) c) Results are less accurate than trend extrapolation or experts judgment Forecasting errors: As much as 90% for a 7-year forecast Average error for 7 yr forecast 30% Average error for 3 yr forecast 20% (Horowitz and Enslie, 1978) d) Techniques for short-range planning are simpler but still inaccurate.
Also, The response of the most popular of these techniques (decomposition, exponential smoothing, moving average),to significant traffic changes is inadequate, hence they cannot predict traffic volume or other such variables with accuracy (Holmesland (1979)). Time Series analysis: Time series has recently become a more attractive tool for traffic engineers. Traditionally, traffic engineers do not explicitly assume that successive events are correlated and usually consider events in the time domain to vary randomly around a trend line. Autocorrelation was also ignored because the calculation and adjustment required for it, was tedious. However, recently developed computer software makes this a much easier and very inexpensive process. In summary, time-series analysis is an attractive tool for analysis because: a) we have exhausted most of the possibilities within the old set of forecasting techniques, b) c) many of these existing techniques are not giving us good solutions, we have the tools to extend our work into consideration of autocorrelated events.
2.1.
The aim of the literature review is to summarize the major work done in the study of travel time variation. It includes study of travel time variation, modeling of travel time, travel time prediction.
2.2.
Literature Review
Paper no.1 Title: Analysis of travel time variation over multiple sections of Hanshin Expressway in Japan The paper classifies sources of uncertainty in travel time into the categories: demand-side factors ( like traffic volume), supply-side factors (like traffic accidents) and external effects (like rainfall intensity).It also classifies seven sources of events that cause travel time variation: Traffic-influence events( traffic incidents, work zones, weather),traffic demand events(fluctuations in normal traffic, special events),physical highway features(traffic control devices and bottlenecks).The Seemingly Unrelated Regression Equations (SURE) model was used by the authors, as opposed to traditional models like Multiple Linear Regression (MLR) model as the latter fail to estimate the error correlation across various equations, also called contemporaneous error correlation. The papers novelty is also in that it considers the effect of uncertainties on travel-time variation across multiple sections. Methodology: The study assumes a linear relationship between the travel-time and the factors affecting the travel-time variation. The route is divided into 3 sections based on on-ramp and off-ramp criteria. The sections are considered dependent and hence, the error covariance across the equation is not zero. Since it is believed that there could be several unobserved characteristics of the uncertainties among various sections that will affect the travel-time variation, therefore the error terms can be
10
correlated across sections. Therefore, the regression equations are estimated jointly as a set of Seemingly Unrelated Regression Equations. Travel-time estimation for the study area: Using the collected vehicle detector data, spot speed for every 500mt interval was estimated. Corresponding travel-times were estimated from these. Path travel time for the three sections were estimated using time-slice method, which was found to be more suitable for offline application rather than online application when speed varies over time. Time-slice method was also found to provide better results than the instantaneous method. Travel time statistical parameters like mean, median, standard deviation, probability, cumulative distribution and standard deviation were calculated.MLR analysis was carried out to understand the influence of all the incidents on travel-time variation. The residual error obtained by this model was used for estimating the error covariance matrix. Using the error covariance matrix, SURE model coefficients were estimated. Conclusions: The Standard Error (SE) obtained using the SURE model for the three sections was lower than the MLR model. The model coefficients obtained by this method were found to be more appropriate than those obtained from the MLR model. The coefficients estimated by the MLR model underestimate the travel time as compared to the SURE model. Except for free-flow situations, the results obtained by the independent models have over-estimated the travel time under the influence of correlation among various sections due to traffic-volume (demand-side factor), traffic-accident (supply-side factor) and rainfall (external factor). Paper no: 2 Title: Bus Arrival Time Prediction Using Artificial Neural Network Model. Aim: The aim of this work was to develop and apply a model to predict bus arrival time using AVL (Automatic Vehicle Location) data. The data considered are traffic congestion and dwell time data. Methodology: A historical data based model, regression models and an artificial neural network model were used. The difference between the predicted and observed arrival times was used to qualify accuracy.
11
AVL data was collected in Houston, Texas over 6 months in 2000(from June to November) by Houston Metro buses equipped with DGPS (Differential Global Positioning System) receiver at 5 second interval. The test bed was Route 60 which was highly congested in the morning and afternoon peaks; it had two corridors, a downtown and a north area corridor, and only the south-bound direction was studied. The input variables were arrival time, dwell time, and schedule adherence. The time periods were weekday peak, weekday nonpeak, weekday evening, weekend. It was found that the variability of dwell time is larger than that of arrival time. Models: Historical Data Based Model: Link travel time between transit stops is calculated. It includes stopped delay at intersections but does not include dwell times. Arrival times are calculated at transit stops. Regression Models: Five multiple linear regression specifications were tested in this research after analyzing stepwise regression and correlation coefficient. Dwell time was not used to develop regression models since it was not important statistically. Artificial Neural Network Models: ANNs emulate the learning process of the human brain. They are calibrated in two steps: training, and testing. Out of 13 different training functions, Levernberg-Marquardt optimization algorithm was chosen as the training function. The ANN architecture used had three layers: an input layer, a hidden layer, an output layer. The weights and parameters associated with the hidden layer were identified during the calibration process. Fifteen different hidden neurons were tested and the best number of neurons was selected for each ANN model based on the concept of minimizing the prediction error. The prediction results from the fifteen different neurons were not significantly different from each other. The back propagation algorithm and the Hyperbolic Tangent Sigmoid transfer function were used in the model development. After testing fourteen different learning functions, a Perceptron Weight and Bias learning function was used. The average MAPE (mean absolute percentage error) for these fourteen functions was not significantly different.
12
Model Evaluation: The MAPE was used as a MOE (measure of effectiveness) in this work. It was found that clustering data led to smaller MAPE in Historic data based model and regression models. However, the clustering results in poorer results than the non-clustering option in the artificial neural network models. It is, therefore, hypothesized that ANN as a universal function approximator, was able to identify the non-linear relationships associated with different clusters. However, there may not have been enough observations to adequately fit the functions. The lowest MAPE of the Historical model of downtown area was for the weekday peak. It is proposed that congestion reduces the variability in travel times and this makes the historical model more accurate for this time period. The use of Real-time schedule adherence did not improve the results much and hence, it was proposed that there is a non-linear relationship between arrival time and schedule adherence. The ANN has the lowest MAPE as compared to the Historic model and the MLR (Multiple Linear Regression) model. It was proposed that the use of historic data (representing congestion) coupled with real-time schedule adherence data (representing real-time congestion and demand inputs) resulted in better performance of the ANN model. Conclusions: This paper describes the results of three bus travel time prediction algorithms which were calibrated and tested on a transit route in Houston, Texas. The input to the models consisted of historic data (i.e., link travel time and dwell time) and real-time schedule adherence data. It was found that the Artificial Neural Network models (used without clustering of the data) performed considerably better than either a historic data based model or MLR models. It was hypothesized that ANN was able to identify the complex non-linear relationship between travel-time and the independent variables and this led to the superior results. Paper no.3 Title: Using bus Travel Time Data to Estimate Travel Times on Urban Corridors. Aim: This study determines whether transit vehicles/buses can be used as probe vehicles for collecting travel time data for automobiles on urban corridors. It analyses the nature of information collected by the buses and develops formulas to covert the travel time of a bus to that of an automobile. Data on bus and automobile
13
travel time on various sections of arterials in the northern part of New Castle County, Delaware was used for this purpose. Methodology: The tasks involved are: a) to measure the travel time of the bus and the automobile for the same section, b) to analyze the characteristics of the components of the Bus Travel Time (BTT) and there variability, c) to develop a model that converts the travel time of the bus to the average travel time of the automobile, d) to verify the model by the collected data. The procedure is to convert the BTT to the ATT (Automobile Travel Time) that is expected before the next BTT data are updated. The predicted travel time is assumed to be equal to the estimate obtained from the last available BTT. It is assumed that the predicted travel time would be closer to the actual travel time if the data is collected at shorter intervals, which depends on the frequency of buses (or measurement intervals). The required accuracy of the predictions was debated, since a higher accuracy complicates the measurement plans and the procedure of conversion (BTT to ATT).On the other hand, a much lower accuracy of prediction may render the predictions useless. Assigning a monetary value to the travel time and value assigned to differences between predicted and actual travel time, the tolerable error of estimate was found to be 10% to 15% of the actual travel time. The distance over which travel time was estimated worked out to 4.6km (assuming a 55km/hr speed and a travel time of 5min). The difference of BTT from the average travel time of the stream is a random variable. Buses typically, travel in the rightmost lane of the urban corridors and this induces a bias in the travel time of buses. However, despite the sources of randomness and bias in the difference between ATT and BTT, buses run on heavily travelled urban corridors (at a high frequency during peak hours), follow traffic rules and observe speed limits. These characteristics make them attractive as probe vehicles. Postulating that the differences between ATT and BTT arise because of the following factors: the stopping time of the bus at bus stops, the time lost by the bus because of repeated accelerations and decelerations from and to a stop, basic difference between the operating abilities of the bus and the automobile, adherence
14
(by the bus and the automobile) to the posted speed limits, the tendency of the bus to use the right lane; a simple predictive equation treating the actual running time of the bus as an independent variable is formed. The equation was changed repeatedly taking into account various factors: statistical importance of calibration constants, insight provided by the calibration constants into the relation between the variables, effort to make the model as calibration free as possible. Results: Five models were developed for the five sites and the results are presented in equations (i) and (ii): =
+ 0.14
(i)
+ (0.18 )
(ii)
For more frequently congested roads Using this result, we can predict the average travel time of the automobile from the data on the BTT and the general characteristics of the road section. Although five sites may not be enough to develop a rule of thumb, such a rule may be developed after the study of many more arterial sections. The use of AVL equipped buses as a data source is promising because the measurement function is already available by default and the task of prediction can be performed with minimum manual intervention. Paper No.4 : Title: Chaotic analysis of traffic time series. Authors: Pengjian Shang, Xuewei Li, Santi Kamae Input Variables: Speed, volume, occupancy collected every 20s. Aim: Paper applies non-linear time series modelling techniques to analyse the traffic data collected from Beijing Xizhimen Methodology: Raw data screened for errors, aggregated into 2min data, average speed, average volume, total occupancy calculated. Draw curves for correlation
15
function v/s r, range of scaling region, from this plot the chaotic nature of traffic time series is known, the slope of the line in the scaling region is the correlation dimension. Correlation dimension v/s Embedding dimension is plotted. Phase space is reconstructed using "method of delays. The slope values corresponding to the largest Lyapunov exponent were obtained after the least-squares line fit for the average speed time series and was found to be 0.25 (deviation +- 0.02). Results: Saturation of correlation dimension beyond a certain embedding dimension value is an indication of the presence of deterministic dynamics, the finite and low correlation dimension is an of the existence of deterministic dynamics. Positive value of Lyapunov Exponent is a strong indicator of chaos. Conclusions: Traffic time series is deterministic and can be modelled using phase space techniques. the predicting length of the traffic time series should be about 8min. Paper No.5 Title: Use of the Box and Jenkins Time Series Technique in Traffic forecasting Authors: Nancy L. Nihan and K Jello O. Holmesland, Department of Civil Engineering, University of Washington, Seattle, U.S.A. Input Variables: Average Weekday Volume (AWD) (1968-1977) Aim: To show the short-range accuracy of the simplest possible model, to investigate the accuracy of the Box-Jenkins technique for short-range forecasting (12-month forecasting period). Methodology: 2 steps were followed: 1) data fitting,2) model selection. After examining several models and conducting many statistical tests, ARIMA (Autoregressive Integrated Moving Average) was finally choosen. Two types of forecasts - a simple forecast and an adaptive forecast were made. All errors were found to be around 5% or less. Results: It was found that it is possible to fit an ARIMA model as well as a multiplicative model to traffic data from the highway under consideration, using the Box and Jenkins technique. The ARIMA model selected was only two percent away from the measured values at the end of a twelve-month forecast, ARIMA is,
16
therefore, highly accurate and easy to use after it has been estimated. It requires less data input, and is flexible. It can accommodate more than one interval in complex time series models, It can be used to relate two or more time series, and changes taking place in a time series can be detected very soon, so it can be used as an early warning system.
Paper No. 6. Title: A multivariate state space approach for urban traffic flow modeling and prediction. Authors: Anthony Stathopoulos, Matthew G. Karlaftis, Department of
Transportation Planning and Engineering, School of Civil Engineering, National Technical University of Athens. Input Variables: 3-min volume measurements from urban arterial streets near downtown Athens, flow (volume) and occupancy data used to estimate speed and travel time. Aim: Developing flexible and explicitly multivariate time-series state space models using core urban area loop detector data, to model and predict flow at an urban signalized arterial. Methodology: Data from 144 loop locations, 5 sequential (multivariate setting) detectors along an important 3-lane per direction signalized arterial on the periphery of the core area of the city (Alexandras Avenue) are chosen for further analysis. Time series is tested for stationarity using the augmented Dickey Fuller (ADF) test. Determination of basic autoregressive and cross-correlation characteristics of the time series for the various loop locations, State space models were developed for both the pooled data (data from all time periods combined), and the data from the various periods separately. The models developed were flexible for an autoregressive and a moving average order of up to three lags. 70% of the data was used for model development and 30% for testing. Results: Predictions obtained from the state space models are superior to those obtained from the ARIMA models; in one of the loops, the state space model yields
17
a mean absolute percent error (MAPE) of 12% compared to a 20% MAPE value from the ARIMA model; MAPE values reported here are rather high when compared to values reported in other studies. Conclusions: Despite its potential usefulness, traffic flow in signalized urban arterials cannot be predicted, at least in the short-run, with as much accuracy as flow in urban freeways. The results of the models developed clearly suggest that, at least in the case of Athens, different specifications are appropriate for different time periods. Further, it also appears that the use of multivariate state space models is relevant in the urban roadway system. Paper No.7. Title: A time-series analysis of public transit ridership in Portland,Oregon,19711982. Authors: Michael Kyte, James Stoner, Jonathan Cryer. Input Variables: level of transit service available, (2) relative costs of travel by transit and by automobile, (3) the size of the travel market, and (4) other factors such as gasoline shortages, weather, etc. Aim: Comprehensive analysis of public transit usage in Portland, Oregon, from 1971 through 1982 using time-series analysis, The impacts of the 81 service changes and 5 fare changes implemented have been analyzed at both at the system and route levels using transfer function and intervention time-series models, the effects of auto travel costs and the local economy are included. Methodology: A) Model development: Transfer function model is chosen and Impact Analysis and Intervention analysis are conducted, B) forecast procedure was accomplished by not using the final one year of data (July 1981-June 1982) for the system data and fall 1981-spring 1982 for the route-level data), estimating the models, and then making forecasts for 12 months or four quarters ahead. The forecasts were then compared with the actual ridership data. C) Three levels of data aggregation were used: system level, sector level, and route level. Three classes of time-series models were developed including univariate transfer function, intervention and simultaneous equations transfer function.
18
Results: The effects of service-level and fare changes on transit ridership are not instantaneous but are delayed and distributed over specific periods of time. The existence of these lag structures is expected from consumer behavior theory. The effect of fare changes can be measured for up to three months after their implementation. Gasoline price and employment level changes do have immediate effects with no discernable lag structures. Feedback relationships were identified between transit ridership, fare, service level, and gasoline price. For example, gasoline price changes affect future transit fare changes. The service-level and fare elasticities computed for the system, the six sectors, and the 26 routes were in the range reported by previous studies. Models were generally consistent, in terms of lag structure and elasticities, between the three data aggregation levels. However, some variables are inherently more effective at one level than another. The system models had mean absolute percent errors (MAPE) of less than five percent. Paper No.8. Title: Modeling and Forecasting Vehicular Traffic Flow as a Seasonal ARIMA Process: Theoretical Basis and Empirical Results. Authors: Billy M. Williams and Lester A. Hoel. Aim: To present a case for acceptance of a specific time series formulationthe seasonal autoregressive moving average processas the appropriate parametric model for a specific type of ITS (Intelligent Transportation System) forecast: shortterm traffic condition forecasts at a fixed location in the network, based only on previous observations at the forecast location. Methodology: A) theoretical justification for the application of ARIMA models: assertion that a weekly seasonal difference will yield a stationary transformation of discrete time traffic condition data series, coupled with the Wold decomposition theorem, B) Hypothesis: properly fitted seasonal ARIMA models will provide accurate traffic condition forecasts, C) Testing of hypothesis through empirical results, where correlation analysis is shown as the basis for assessing the stationarity of series transformations using a first weekly difference; presentation of the modelfitting results and a discussion of the heuristic benchmarks used to assess the predictive performance of the fitted seasonal ARIMA models.
19
Results: One-step seasonal ARIMA predictions consistently outperformed heuristic forecast benchmarks. Assertions and findings presented in this paper directly contradict a statement in Kirby et al. 1997, namely that extending simple ARIMA models to include seasonal and other effects, in practice... did not have a substantial impact on the results. Theoretical foundation for seasonal ARIMA modeling negates any theoretical motivation to investigate high level nonlinear mapping approaches, such as neural networks. This assertion is supported by comparison to actual neural network forecasting results with a common data set. Paper No.9 Title: Multivariate Short-Term Traffic Flow Forecasting Using Time-Series Analysis. Authors: Bidisha Ghosh, Biswajit Basu and Margaret OMahony. Input Variables: traffic flow, number of maneuvers, time. Aim: Introducing a different class of time-series models called structural time-series model (STM) (in its multivariate form) to develop a parsimonious and computationally simple multivariate short-term traffic condition forecasting algorithm. Methodology: A) A "seemingly unrelated time-series equation (SUTSE) Model, which is also a multi-inputmulti-output short-term traffic flow forecasting model, where the number of input intersections is more than number of output intersections is choosen. B) The proposed multivariate SUTSE traffic flow forecasting methodology is applied to a congested urban transportation network at the city center of Dublin. A network of ten intersections within the transport network is chosen for this purpose. C)The univariate traffic flow observations obtained over each 15-min interval from the inductive loop detectors situated at these ten intersections and their nearest available upstream junctions are modeled using the proposed multivariate traffic flow model. D) The cross-correlational structure of the ten chosen junctions is verified. E) Traffic flow time series is reduced to stationary form by 'differencing, second-order stationarity of the time-series data sets used in the MST model are checked by plotting the autocorrelation functions (ACFs) of the data sets, To ensure stationarity, seasonal differencing has been performed on all the modeled traffic flow
20
time-series data sets. All of the ten series of traffic flow observations are modeled using homogeneous SUTSE models. Results: Hyperparameter estimates and the plot of the seasonal component show that the seasonality is deterministic in nature. The trend component is stochastic and depicts the within-day local fluctuations in the data. The trend component varies about a zero mean value, validating the assumption that there is no slope component latent within the traffic flow data set. The SUTSE model is computationally more efficient than some of the other existing multivariate short-term traffic flow forecasting methodologies.The checking for stationarity conditions is not critical to the development of the model. The developed SUTSE model can separately trace the evolution of each individual component (trend, seasonality, etc.) of the traffic flow data over time. Consequently, the deterministic nature of the seasonal component of the traffic volume observations from junctions at urban signalized arterials has been established.MST model can additionally include the effect of changes in traffic conditions at one or more immediate upstream junctions to improve the predictions at the downstream output junction. Paper No.10. Title: Travel Time Prediction using a Seasonal Autoregressive Integrated Moving Average Time Series Model. Authors: Angshuman Guin. Input Variables: Volume, average speed and lane-occupancy data. Aim: Investigating the possibility of extending the Box and Jenkins technique to develop a Seasonal ARIMA (also sometimes referred to as SARIMA) prediction model for travel times. Methodology: Average speed converted into travel time, several weeks travel time data is plotted over time, plots are superimposed and the periodicity is detected, The ACF( Autocorrelation Function) plot of the raw travel time data for all weekdays is plotted, the ACF plot of the single lag (15-minute interval) differenced travel time data is plotted, the ACF plots for a dataset with just the Mondays of the consecutive weeks is plotted, ACF of Raw Monday Travel Time Data (10 days) is plotted to check for stationarity, the first difference of the series is plotted but it does not yield
21
any stationarity, the autocorrelation plots indicate that a seasonal differencing at weekly level would generate a stationary series. A multiplicative seasonal autoregressive integrated moving average process of period s, with regular and seasonal AR (Autoregressive) orders p and P, regular and seasonal MA orders q and Q, and regular and seasonal differences ,is referred to as an ARIMA (p,d,q)(P,D,Q)s model. Conclusions: Travel times have strong weekly seasonality, such seasonality is to be expected at higher levels of aggregation and not in the system level data at which the detectors record the data, weekly periodicity can be successfully used in a predictive model for segment travel times, this model is expected to provide effective travel time forecasts irrespective of whether the travel time estimates are based on point detection data, probe vehicle data or Automatic Vehicle Identification (AVI) data.
Relevance of Literature review: Paper no.3 , Using bus Travel Time Data to Estimate Travel Times on Urban Corridors which is included in the literature review is particularly useful because it considers the use of data collected from Buses fitted with AVL (Automated Vehicle Location) equipment for estimating travel time for motor vehicles on the same routes. The data being used in this project has been collected in a similar manner from buses fitted with GPS equipment. Hence, the models/methodology used in the papers can be used to see how the data available with us can be used to estimate travel times for all vehicles on the routes covered.
22
3.1
Data Set 1: The data has been collected over a period of 1 week for Route 419(BRT corridor) The data has been collected using GPS enabled AVL ( Automated Vehicle Location) project implemented by DIMTS (Delhi Integrated Multi-Modal Transit System Ltd) that record certain parameters at every 10 seconds for every bus. The parameters recorded are: latitude, longitude, timestamp, speed, distance travelled by object, user_id( registration of the bus). The available data covers the 24 hour schedule of the bus. The bus route generally varies from 9.30 am in the morning to 7pm in the evenings. Since the buses return do circular routes, the data for the return journey is not recorded by the GPS. Dataset 2: The GPS data over various bus routes of Delhi collected from 1-1-2013 to 31-12013. The parameters recorded are: latitude, longitude, timestamp, speed, distance travelled by object, user_id( registration of the bus). The data is already separated into UP and DOWN trips for all routes.
3.2
Treatment of Dataset 1:
The data available was present in the form of Latitude and Longitude and time stamp, hence the first step was the conversion of these to distance and speed values. Using the spherical law of cosines formula (equation (iii)) distance is calculated from latitude, longitude and R (radius of earth);
23
Spherical law of cosines: = (acos sin 1 . sin 2 + cos 1 . cos 2 . cos (iii)
3.3
Stationarity:
A stationary time series is one whose statistical properties such as mean, variance, autocorrelation, etc. are all constant over time. Most statistical forecasting methods are based on the assumption that the time series can be rendered approximately stationary (i.e., "stationarized") through the use of mathematical transformations. A stationarized series is relatively easy to predict: you simply predict that its statistical properties will be the same in the future as they have been in the past. Another reason for trying to stationarize a time series is to be able to obtain meaningful sample statistics such as means, variances, and correlations with other variables. Such statistics are useful as descriptors of future behavior only if the series is stationary. For example, if the series is consistently increasing over time, the sample mean and variance will grow with the size of the sample, and they will always underestimate the mean and variance in future periods. And if the mean and variance of a series are not well-defined, then neither are its correlations with other variables. For this reason it is important to be cautious about trying to extrapolate regression models fitted to non-stationary data. First Difference: The first difference of a time series is the series of changes from one period to the next. If Y(t) denotes the value of the time series Y at period t, then the first difference of Y at period t is equal to Y(t)-Y(t-1). In Statgraphics, the first difference of Y is expressed as DIFF(Y). If the first difference of Y is stationary and also completely random (not autocorrelated), then Y is described by a random walk model: each value is a random step away from the previous value. If the first difference of Y is stationary but not completely random--i.e., if its value at period t is autocorrelated with its value at earlier periods--then a more sophisticated forecasting model such as exponential smoothing or ARIMA may be appropriate. (Note: if DIFF(Y) is stationary and random, this indicates that a random walk model is
24
appropriate for the original series Y, not that a random walk model should be fitted to DIFF(Y). Fitting a random walk model to Y is logically equivalent to fitting a mean (constant-only) model to DIFF(Y)). Test for Stationarity: If the series has a stable long-run trend and tends to revert to the trend line following a disturbance, it may be possible to stationarize it by de-trending (e.g., by fitting a trend line and subtracting it out prior to fitting a model, or else by including the time index as an independent variable in a regression or ARIMA model), perhaps in conjunction with logging or deflating. stationary. Such a series is said to be trend-
series stationary, in which case it may be necessary to transform it into a series of period-to-period and/or season-to-season differences. If the mean, variance, and autocorrelations of the original series are not constant in time, even after detrending, perhaps the statistics of the changes in the series between periods or between seasons will be constant. Such a series is said to be difference-
stationary. (Sometimes it can be hard to tell the difference between a series that is trend-stationary and one that is difference-stationary, and a so-called unit root test may be used to get a more definitive answer.
3.4
1)
2)
Trend Analysis
Definition of Non-Stationarity: A Unit root test tests whether a time series variable is non-stationary using an autoregressive model. A well-known test that is valid in large samples is the augmented DickeyFuller test. The optimal finite sample tests for a unit root in autoregressive models were developed by John Denis Sargan and Alok Bhargava. These tests use the existence of a unit root as the null hypothesis.
25
ADF Test: An Augmented DickeyFuller test (ADF) is an augmented version of the Dickey Fuller test for a larger and more complicated set of time series models. The augmented DickeyFuller (ADF) statistic, used in the test, is a negative number. The more negative it is, the stronger the rejection of the hypothesis that there is a unit roots at some level of confidence.
3.5
The ADF test was applied on the data for each run (the bus performs 4 runs in a day), and the results obtained were compiled in Table 2 and were compared to the critical values for DF (Dickey-Fuller) and ADF (Augmented Dickey-Fuller) tests provided in Table 1.The ADF values were greater than the critical values for the test.
Table 1: Critical values for DF and ADF Tests (Fuller, 1976, p373) Significance level C.V. for constant but no trend C.V. for constant and trend 10% -2.57 -3.12 5% -2.86 -3.41 1% -3.43 -3.96
26
Table 2. Compiled Results of ADF test on Dataset 1: S.No . Type ADF test statistic 1% -3.96 5% -3.41 10% -3.12 Durbin Watson Statistic
1st run,1st difference, trend only, lag 10 -9.558 -3.448 -2.869 -2.570 2.003
3.6
Interpretation of results:
Since the ADF value is always greater than the critical values, hence the null hypothesis is rejected. The series is stationary.
27
Interpreting the DurbinWatson statistic: Since d (DurbinWatson statistic) is approximately equal to 2(1 r), where r is the sample autocorrelation of the residuals, d = 2 indicates no autocorrelation. The value of d always lies between 0 and 4. If the DurbinWatson statistic is substantially less than 2, there is evidence of positive serial correlation. As a rough rule of thumb, if DurbinWatson is less than 1.0, there may be cause for alarm. Small values of d indicate successive error terms are, on average, close in value to one another, or positively correlated. If d > 2 successive error terms are, on average, much different in value to one another, i.e., negatively correlated. In regressions, this can imply an underestimation of the level of statistical significance. The DurbinWatson statistic approaches 2 in most of the tests, suggesting that autocorrelation does not exist.
3.7
Dataset-2
Description of data: The data provided by DIMTS was 1 month (1/1/13 to 31/1/13) data for different bus routes of Delhi. Methodology: The bus routes were divided into segments, using Bus stops, Intersections (3 way/4way) and Roundabouts as Segment ends. The mean speed for each segment was calculated and these were tabulated to see trends over time. The speeds are tabulated as Results according to Day of week and according to Hourly Time Slot (Appendix A of CD). A graph of each table is added at the end. The Bottlenecks exposed via these graphs (segments having low speeds) are tabulated in Conclusions tables. The range of speed for each Time slot and Day for each route is also provided in the Conclusions Table. Procedure: The data was provided by DIMTS as Combined Data file per route. This was divided into separate files (using MATLAB software) for separate days and for separate buses. Speed calculations were carried out on these to determine the instantaneous
28
speed along the routes. The routes were divided into segments according to occurrence of bus-stop/traffic light/intersection and the mean speed was calculated for each segment. The results for a particular time slot for many buses for a particular segment were averaged to give an estimate of mean speed at a particular segment of a route at a particular time (1 hour time slot).These were aggregated according to day of week. By this, a speed profile of every day of the week was obtained for all the routes studied. Following are the Route Data, Segment ends and Conclusions for each route:
29
Table 3: Route Characteristics for Route 108 Route No. 108 Depot:- Low Floor RUNNING TIME:- 64 Minutes Departure Time Hari Nagar. Clock Tower 0536 0552 0608 0616 0624 0640 0656 0712 0800 0816 0832 0840 0848 0904 0920 0936 1048 1104 1120 1128 1136 1152 1208 1224 1400 1416 1432 1440 1448 1504 1520 1536 1624 1640 1656 1704 1712 1728 1744 1800 1848 1928 1944 1952 2000 2016 2032 2048 0648 0704 0720 0728 0736 0752 0808 0824 0904 0928 0944 0952 1000 1016 1056 1112 Nehru Vihar 1200 1216 1232 1240 1248 1304 1320 1336 1512 1528 1544 1552 1600 1616 1632 1648 1736 1752 1808 1816 1824 1840 1920 1936 2024 2040 2056 2104 2112 2128 2144 2200
30
3.7.1
The route is divided into segments, the segments being separated by Bus Stopping points like Bus stops, Intersections (3 way/4 way, Roundabout)( Table 4).The Bottlenecks identified for different time slots and days are summarized in Table 5.
31
Table 4: Segments of Route for analysis: 1 Nehru vihar Road 2 Four way 3 Nehru vihar crossing 4 Police station timarpur 5 Balak ram hospital 6 Timar pur 7 Timarpur water tank 8 North mall road 9 Mall road 10 International students hostel 11 Four way 12 Khalsa college 13 Patel chest(4 way) 14 Sri ram college 15 Daulat ram college 16 Maurice nagar 17 Roop nagar 18 Roop nagar 19 Kamla nagar 20 Nangia park(roundabout) 21 Intersection 22 Leela vati vidya mandir 23 Gulabi bagh crossing 24 Shastri nagar 25 Gulabi bagh 26 Four way 27 DDA flats sarai basti 28 Shiv mandir 29 Inderlok 30 Intersection 31 Zakhira road
32
28.71015 77.22426 28.70975 77.22698 28.70885 77.22599 28.70686 77.22412 28.70505 77.22322 28.70061 77.22125 28.69746 77.2206
28.69511 77.20993 28.69175 77.20827 28.68906 77.20697 28.68805 77.20646 28.68676 77.20586 28.68462 77.20369 28.68413 77.20250 28.68320 77.20108 28.67991 77.19585 28.67772 77.18900 28.67712 77.18904 28.67516 77.18870 28.67261 77.18622 28.67097 77.18412 28.67007 77.18349 28.67097 77.17615 28.67160 77.17313 28.67206 77.16894 28.67079 77.16655 28.66876 77.16523
32 Zakhira 33 Roundabout 34 DCM chemicals 35 Campa cola 36 ESI dispensary 37 Intersection 38 Moti nagar market 39 Three way 40 Roundabout 41 F block 42 Kirti nagar ps 43 Furniture market 44 Saraswati garden 45 Wood market 46 Man sarovar garden 47 Four way 48 Mayapuri depot 49 Govt. press mayapuri 50 Maya puri metal forging 51 LIG flats 52 Swarag ashram 53 Beri-wala bagh 54 Round-about 55 DDU hospital 56 Hari nagar clock tower
28.66759 77.16421 28.66680 77.16387 28.66521 77.15986 28.66199 77.15211 28.66081 77.14885 28.65961 77.14603 28.65733 77.14178 28.65456 77.13680 28.65286 77.13842 28.65150 77.14076 28.64952 77.14382 28.64764 77.14236 28.64600 77.14028 28.64279 77.13713 28.63815 77.13241 28.63754 77.13065 28.63670 77.12868 28.63432 77.12777 28.63087 77.12479 28.63077 77.11937 28.63088 77.11519 28.63103 77.11247 28.63160 77.11163 28.62824 77.11103 28.62469 77.11039
33
Slot/Day: Sundays
Mondays
Tuesdays
Wednesdays
1) Four way after International students hostel to Khalsa College (5.37kmph) 2) Harinagar clock tower(1.37kmph)
Thursdays
1) Leelawati vidya mandir to Gulabi bagh crossing (4.44kmph) 2) Four way after man sarovar garden to Mayapuri depot (5.75kmph)
8.67kmph to 32.50kmph
Fridays
8.42kmph to 35.01kmph
Saturdays
1) Four way after Mansarovar garden to Mayapuri depot(5.96kmph) 2) Harinagar clock tower(5.15kmph)
9.26kmph to 34.90kmph
8am-9am
1) Four way after Mansarovar garden to Mayapuri depot(5.14kmph) 2) Harinagar clock tower(4.5kmph)
10.83kmph to 37.20kmph
9am-10am
10am-11am
11am12noon
1) Four way after Mansarovar garden to Mayapuri depot(5.68kmph) 2) Harinagar clock tower (5.41kmph)
12noon-1pm
34
9.79kmph to
32.46kmph 1pm-2pm 1) Khalsa College to Patel chock (5.37kmph) 2) Nangia Park to Intersection(5.57kmph) 2pm-3pm 1) Four way after Mansarovar garden to Mayapuri depot(4.91kmph) 3pm-4pm 1) Kamlanagar to Nangia park(5.91kmph) 8.76kmph to 32.69kmph 8.54kmph to 30.56kmph 8.43kmph to 32.55kmph 4pm-5pm 1) Nehru vihar road to Four way(1.11kmph) 2) Kamlanagar to Nangia park( roundabout) 3) Four way after Mansarovar garden to Mayapuri depot(4.91kmph) 5pm-6pm 1) Nehru vihar road to Four way(1.58kmph) 2) Kamlanagar to Nangia park(5.10kmph) 3) Four way after Mansarovar garden to Mayapuri depot(5.96kmph) 6pm-7pm 1) Nehru vihar road to Four way(1.42kmph) 2) Kamlanagar to Nangia park(4.27kmph) 7pm-8pm 1) Nehru Vihar to Four way(0.93kmph) 2) Kamlanagar to Nangia park(4.25kmph) 3) Leelawati vidya mandir to Gilabi bagh crossing(4.44kmph) 4) Four way after Mansarovar garden to Mayapuri depot(5.75kmph) 7.58kmph to 31.36kmph 7.82kmph to 34.38kmph 7.28kmph to 33.45kmph 8.06kmph to 34.15kmph
35
Fig2: Route 108 Down Table 6. Segments for Route 108 Down: Segment S.No. from/to -> 1 DDU hospital 2 Roundabout 3 Beri wala bagh 4 Swarg asharam 5 Swarg asharam 6 LIG flats (Four - way) 7 Ram singh 28.62992 77.12359 Latitude Longitude S.No Segment from/to -> 28.62805 28.63170 28.63147 28.63097 28.63101 28.63091 77.11088 77.11136 77.11202 77.11485 77.11510 77.11946 27 Inderlok 28 Four way 29 Shiv mandir 30 Shastri nagar E block 31 Four way 32 Gulabi bagh Shastri nagar A 33 block
36
28.67284
77.18622
marg Maya puri 8 metal forging 28.63064 77.12427 34 Gulabi bagh crossing Swami narayan 9 Govt press 10 Four way Man sarovar 11 garden 12 Chuna bhati 13 Wood market Saraswati 14 garden Furniture 15 market 16 Kirti nagar F block kirti 17 nagar 18 Rounabout 19 Kirti nagar Moti nagar 20 market 21 Moti nagar 28.65721 28.65882 77.14111 77.14419 46 Khalsa college 47 Three way International students 22 ESI dispensary 23 Campa cola 28.66047 28.66273 77.14731 77.15335 48 hostel 49 Entrance to side road Lucknow road govt 24 DCM chemicals 25 Roundabout 26 Zakhira 28.66521 28.66649 28.66766 77.15943 77.16335 77.16407 50 school 51 Lucknow road 52 Balak ram hospital 28.69577 28.69686 28.70521 77.21624 77.21667 77.22322 28.69625 28.69532 77.21190 77.21427 28.69446 28.6965 77.20951 77.21056 28.65138 28.65245 28.65483 77.14079 77.13908 77.13681 43 Sri ram college 44 Patel chest 45 Patel chest 28.68893 28.69144 28.69228 77.20678 77.20796 77.20845 28.6478 28.6497 77.14225 77.14374 28.64608 77.14004 28.63843 28.63911 28.6426 77.13249 77.13343 77.13673 37 Roundabout 38 Four way 39 Kamla nagar Roop nagar(Bus Stop 40 befpre roundabout) Roop nagar(Bue Stop 41 after roundabout) 42 Maurice nagar 28.6847 28.68629 77.20352 77.20549 28.68409 77.20208 28.67927 28.68142 28.68239 77.19385 77.19805 77.19947 28.63512 28.63743 77.12792 77.12977 35 marg(Three way) 36 Leelawati mandir 28.67552 28.67706 77.18872 77.18881 28.67489 77.18827
37
Table 7. Conclusions for Route 108 Down: Slot/Day: Sundays Bottleneck: 1) Govt. press to Four way 2) Roundabout after Leelawati mandir to following Four - way(<5kmph) Mondays 1) Four - way after Govt. press to Man sarovar garden (6.85kmph) 2) Four - way to Kamlanagar (6.61kmph) Tuesdays 1) Four - way to Man sarovar garden(6.27kmph) 2) Four - way to Kamlanagar(6.98kmph) Wednesdays 1) Moti nagar to ESI dispensary(5.36kmph) 2) Kamlanagar to Roopnagar( before roundabout)(6.8kmph) Thursdays 1) Four - way to Man sarovar garden (6kmph) 2) Four - way to Kamlanagar (6.73kmph) Fridays 1) Four - way to Kamlanagar(6.61kmph) 8.59kmph to 34.16kmph 9.14kmph to 34.97kmph Saturdays 1) Motinagar to ESI dispensary (4.96kmph) 2) Four - way to Kamlanagar(5.9kmph) 8am-9am 1) Four - way to Kamlanagar(7.9kmph) 2) Patel Chest to Khalsa College(7.45kmph) 9am-10am Four - way to Man sarovar garden(7.3kmph) Govt press to Four way (4.26kmph) 10.68kmph to 32.60kmph 12.70kmph to 34.50kmph 10.80kmph to 34.30kmph 10am-11am 10.74kmph to 33.03kmph 11am12noon 1) Four - way to Kamlanagar (6.78kmph) 2) International students hostel to enter side road(6.9kmph) 12noon-1pm 1) Roundabout after Leelawati mandir to Four - way (6.21kmph) 2) Four - way to Kamlanagar (6.61kmph) 1pm-2pm 1) Roundabout after Leelawati mandir to Four - way 10.32kmph to 9.40kmph to 32.30kmph 8.89kmph to 34.20kmph 9.20kmph and 33.60kmph 9.10kmph to 33.40kmph 9.20kmph to 33.60kmph Mean speed: 10.30kmph to 31.10kmph
38
(4.52kmph) 2) Four - way to Kamlanagar (5.9kmph) 2pm-3pm 1) Four - way to Man sarovar garden (6kmph) 2) Four - way to Kamlanagar (5.61kmph) 3pm-4pm 1)Roundabout to Four - way (4.24kmph) 2) Four - way to Kamlanagar(6.26kmph) 4pm-5pm 1) Roundabout to Four - way(6.67kmph) 2) Four - way to Kamlanagar(5.65kmph) 5pm-6pm 1) Four - way to Mansarovar garden (6.8kmph) 2) Roundabout to Four - way (4.25kmph) 6pm-7pm 1) Four - way to Kamlanagar(6.14kmph) 2) Kamlanagar to Roopnagar(6.16kmph) 7pm-8pm 1) Four - way to Mansarovar garden (6.27kmph) 2) Motinagar to ESI dispensary (4.9kmph) 3) Roundabout after Leelawati mandir to Four - way (5.2kmph)
34.09kmph
9.44kmph to 30.80kmph 9.92kmph to 31.28kmph 11.10kmph to 32.25kmph 9.89kmph to 31.80kmph 8.80kmph to 31.50kmph 7.11kmph to 35.60kmph
Route: 185
The route lies between the ends Nathupura and Kendriya Terminal. Route characteristics are summarized in Table 8. Table 8: Route Characteristics for Route 185 Route No. 185 Depot:- Standard Floor Running Time:-90 Mintues Nathu Pura to Kendriya Tr., Nathu Pura to I.S.B.T. 60 Minutes Departure Time Nathu Pura 0500 0605 0631 0710 1511 1524 1550 1611 0605 0710 0736 0815
39
0749 0800 0815 0828 0841 0946 0950 1010 1104 1121 1151 1230 1314 1340 1353 1406 1419 1458
1620 1629 1642 1655 1708 1721 1800 1905 1918 1938 1944 1950 2005 2036 2100 2115 2128 2200
0905 0933 1020 1051 1115 1209 1230 1255 1335 1419 1445 1458 1511
1747 1813 1905 2010 2023 2045 2049 2110 2141 2205 2220 2233 2305
40
41
Table 9: Segments for 185 Up Segment from/toS.No. > Kendriya 1 terminal 2 Roundabout Kendriya 3 terminal Kendriya 4 terminal 5 NDPO 6 Roundabout Gurudwara 7 bangle sahib 8 YMCA 9 Jantar Mantar 28.62546 28.62623 28.62790 77.20938 77.21243 77.21596 41 Mall Road International 42 Students Hostel 43 GTB nagar Three way(Mahatma 10 Palika Kendra 11 Regal Cinema Three way(Panchkurian road joins inner 12 circle) 13 Shivaji Park New Delhi 14 Railway Station 15 Roundabout 16 JL Nehru Marg 28.64145 28.64241 28.64207 77.22606 77.22680 77.22812 48 Daka Village Permanand 49 Crossing 50 Radio Colony 28.70870 28.71123 77.20452 77.20436 28.70710 77.20461 28.63423 28.63857 77.21697 77.22388 46 T.B.Hospital 47 Gandhi Ashram 28.70052 28.70466 77.20508 77.20510 28.62872 28.63073 77.21652 77.21742 44 gandhi marg) 45 Camp Chock 28.69874 28.69928 77.20483 77.20482 28.69612 28.69828 77.21182 77.20605 28.69348 77.21887 28.62167 28.62559 28.62653 77.20623 77.20649 77.20747 38 Old Secretariat 39 Khyber Pass Three - way(north 40 mall road) 28.69318 77.21982 28.68407 28.69017 77.22209 77.22118 28.61951 77.20612 28.61744 28.61739 77.20421 77.20555 Latitude Longitude Segment from/to> Civil lines metro 35 station 36 IP College Three way(Mahatma 37 Gandhi road) 28.68127 77.22274 28.67641 28.68005 77.22499 77.22356 Latitude Longitude
42
Zakir Hussain 17 College Three way(jawahar lal 18 nehru marg) 19 Asaf Ali College Delhi Nagar 20 Nigam 21 Hotel Broadway 22 Darya Ganj 23 Subhash Park 24 Three - way 25 Jama Masjid 26 Red Fort Four - way( shyam prasad 27 mukherjee marg) 28 GPO 29 GGS University 30 ISBT Three - way(lala hardev sahai 31 marg) Three - way (ISBT) to Shyam 32 nath marg 33 Ludlow Castle Three way(Shyam nath 34 marg) 28.67304 77.22568 28.66881 28.67211 77.22659 77.22593 66 Amrit Vihar 67 Nathupura 28.76413 28.76894 77.18379 77.18060 28.66845 77.22660 65 Kaushik Enclave 28.76080 77.18918 28.65958 28.66180 28.66534 28.66840 77.23631 77.23479 77.23012 77.22757 Sarvodaya 61 Vidyalay Burari 62 Burari Village 63 Burai Ghari 64 Laxmi vihar 28.74897 28.75331 28.75755 28.76002 77.19841 77.19881 77.19560 77.19093 28.64122 28.64086 28.64319 28.64899 28.64971 28.65124 28.65365 77.23520 77.23838 77.24036 77.23957 77.23905 77.23790 77.23639 54 Nirankari Sarovar Four - way(outer 55 ring road) Transport 56 Authority 57 Jharoda Diary 58 St. Nagar 59 Bengali Colony 60 Francis School 28.73074 28.73491 28.73828 28.73871 28.74477 77.19837 77.19745 77.19744 77.19756 77.19802 28.72808 77.19787 28.72736 77.19783 28.64074 28.64202 77.23079 77.23281 CB Raman IIT 52 Colony Nirankari Sarovar 53 Burari Crossing 28.72703 77.19780 28.72103 77.19988 28.64116 77.23001 51 Nirankar Colony 28.71458 77.20402
43
Table 10. Conclusions about Route 185 UP: Slot/Day: Sundays Bottleneck: 1) Transport Authority to Four - way(5.36kmph) 2) Intersection after DTC ambedkar terminal to Delhi gate(6.08kmph) 3) Kendriya Terminal(8.44kmph) Mondays 1) Nathupora to Amrit vihar colony(1.01kmph) 2) Kaushik enclave to Laxmi vihar(5.09kmph) 3) Transport Authority to Four way(4.10kmph) 4) Intersection(Jawaharlal Nehru marg) to Delhi Gate(5.94kmph) 5) Three - way(outer circle,Barakhamba road) to Three way(outer circle,Kasturba Gandhi marg)(5.59kmph) 6) Kendriya terminal Bus Stop to Kendriya terminal (5.01kmph) Tuesdays 1) Nathupora to Amrit Vihar Colony(3.10kmph) 2) Four - way after Transport Authority to Burari Crossing(4.64kmph) 3) ISBT to Yamuna Bazaar(4.02kmph) 4) Intersection(Jawahar lal Nehru marg) to Delhi Gate(3.56kmph) 5) Delhi Gate to LNJP hospital(5.37kmph) 6) Three - way(outer circle,Barakhamba road) to Three way(outer circle,Kasturba bagndhi marg)(3.04kmph) 7) St.Columbus School to Kendriya Terminal (3.72kmph) Wednesdays 1) Nathupora to Amrit Vihar Colony(5.59kmph) 2) Transport Authority to Four way(3.07kmph) 3) Intersection(Jawahar lal nehru marg) to Delhi Gate(4.96kmph) 4) Kendriya Terminal (6.0kmph) Thursdays 1) Transport Authority to Four - way(3.77kmph) 2) Intersection(Jawahar lal Nehru marg) to Delhi
44
Mean speed:
10.08kmph to 36.87kmph
8.66kmph to 36.09kmph
8.82kmph to 39.84kmph
9.21kmph to 41.20kmph
7.23kmph to 37.76kmph
gate(4.67kmph) 3) Three - way(outer circle,barakhamba road) to Three way(outer circle,Kasturba Gandhi marg)(6.05kmph) 4) Kendriya terminal(0kmph) Fridays 1) Nathupora to Amrit Vihar Colony(6.26kmph) 2) Transport Authority to Four - way(4.09kmph) 3) Intersection(Jawahar lal Nehru marg to Delhi gate(5.14kmph) 4) Four - way( maharaja ranjit singh marg) to Statesman house(6.41kmph) 5) Three - way(outer circle,barakhamba road) to Three way(outer circle,kasturba Gandhi marg)(7.64kmph) 6) Three - way( near Janpath) to Three - way(outer circle,Sansad Marg)(6.24kmph) 7) Kendriya Terminal(4.43kmph) Saturdays 1) Nathupora to Amrit Vihar Colony(4.87kmph) 2) Transport Authority to Four - way(4.05kmph) 3) Intersection(Jawahar lal Nehru marg) to Delhi Gate(6.83kmph) 4) Statesman house to Three - way(outer circle,Barakhamba road)(3.91kmph) 5) Kendriya terminal(3.49kmph) 8am-9am 1) New Delhi Railway station to Roundabout(7kmph) 2) Subhash Park to Three - way(7.63kmph) 3) Four way(outer ring road) to Transport Authority(0.44kmph) 4) Sarvodaya Vidyalaya Burari to Burari Village(5.6kmph) 9am-10am 1) New Delhi Railway Station to Roundabout(6.98kmph) 2) Three - way after Subhash Park to Jama Masjid(6.35kmph) 3) GPO to GCS university(6.72kmph) 4) Four - way(outer ring road) to Transport
45
8.82kmph to 39.06kmph
8.87kmph to 36.94kmph
5.00kmph to 39.30kmph
7.79kmph to 34.15kmph
Authority(4.6kmph) 5) Transport Authority to Jharoda Diary(3.41kmph) 10am-11am 1) Kendriya Terminal to Roundabout(1.53kmph) 2) Three - way after Subhash park to Jama Masjid(7.12kmph) 11am12noon 1) Kendriya Terminal to Roundabout(1.16kmph) 2) Subhash Park to Three - way(2.26kmph) 3) International Students Hostel to GTB nagar(7.55kmph) 4) Permanand Crossing to Radio Colony(5.57kmph) 5) Four -way(outer ring road) to Transport Authority(3.48kmph) 12noon-1pm 1) New Delhi Railway Station to Roundabout(5.19kmph) 2) Subhash Park to Three - way(2.53kmph) 3) Four - way(shyam Prasad Mukherjee marg) to GPO(5.25kmph) 4) Permanand Crossing to Radio Colony(2.62kmph) 5) Four - way(outer ring road) to Transport Authority(3.28kmph) 6) Nathupura(0kmph) 1pm-2pm 1) Darya Ganj to Subhash Park(2.16kmph) 2) Jama Masjid to Red Fort(5.34kmph) 3) Camp Chock to TB Hospital(4.81kmph) 4) Permanand Crossing to Radio Colony(3.70kmph) 5) Four - way(outer ring road) to Transport Authority(3.37kmph) 6) Nathupora(0kmph) 2pm-3pm 1) Subhash Park to Three - way(3.02kmph) 2) Jama Masjid to Red Fort(4.32kmph) 3) Red Fort to Four Four way( Ram Prasad Mukjarjee Marg)(4.35kmph) 4) Four - way(Shyam Prasad Mukherjee Marg) to GPO(4.22kmph) 5) GTB nagar to Three - way(Mahatma Gandhi Marg)(4 3.20kmph to 36.09kmph 3.63kmph to 34.78kmph 4.14kmph to 38.32kmph 2.29kmph to 35.38kmph 1.53kmph to 37.55kmph
46
.41kmph) 6) Camp Chock to T.B Hospital(3.49kmph) 7) T.B Hospital to Gandhi Ashram(5.17kmph) 8) Four - way(outer ring road) to Transport Authority(3.81kmph) 9) Nathupora(0kmph) 3pm-4pm 1) New Delhi Railway Station to Four - way(6.93kmph) 2) Hotel Broadway to Darya Ganj(5.48kmph) 3) Subhash park to Three - way (4.09kmph) 4) Four - way(Shyam Prasad Mukherjee marg) to GPO (6.72kmph) 5) Four - way(outer ring road) to Transport Authority(3.32kmph) 6) Nathupora(0kmph) 4pm-5pm 1) Hotel Broadway to Darya Ganj(4.35kmph) 2) Darya Ganj to Subhash park(4.68kmph) 3) Subhash park to Three - way(4.90kmph) 4) Red fort to 4way(shyam Prasad mukherjee marg)(4.67kmph) 5) Four - way(Shyam Prasad mukherjee marg)to GPO(3.72kmph) 6) Permanand Crossing to Radio crossing(6.88kmph) 7) Four - way(outer ring road) to Transport Authority(3.27kmph) 8) Nathupora(0kmph) 5pm-6pm 1) Regal cinema to Three - way(4.71kmph) 2) New Delhi railway station to roundabout(7.50kmph) 3) Hotel Broadway to Daryaganj(6.97kmph) 4) Darya ganj to Subhash park(5.93kmph) 5) Subhash park to Three - way(6.18kmph) 6) Red fort to Four - way(4.07kmph) 7) Four - way to GPO(3.13kmph) 8) Four - way(outer ring road) to Transport 4.37kmph to 33.20kmph 4.70kmph to 36.09kmph 5.29kmph to 33.47kmph
47
Authority(4.43kmph) 6pm-7pm 1) Three - way(Jawahar lal Nehru marg) to Asif ali college(2.93kmph) 2) Subhash park to Three - way(3.83kmph) 3) Four - way(shyam Prasad mukherjee marg) to GPO(3.85kmph) 4) Four way(outer ring road) to Transport Authority(4.43kmph) 5) Nathupora(0kmph) 7pm-8pm 1) International Students Hostel to GTB nagar(7.40kmph) 2) Four - way(outer ring road) to Transport Authority(4.47kmph) 3) Bengali Colony to Francis School(5.82kmph) 6.01 kmph to 32.12kmph 5.26kmph to 34.07kmph
48
Table 11. Segments for Route 185 Down S.No. Segment 1 Nathupura Amrit vihar 2 colony 3 Kaushik enclave 4 Laxmi vihar 5 Burari ganj 28.76477 28.76090 28.76000 28.75791 77.18364 77.18938 77.19114 77.19536 28 Intersection 29 Shanti van 30 Raj ghat 31 Intersection DTC ambedkar 6 Burari village 28.75522 77.19871 32 terminal 28.63965 77.24382 28.65108 28.64967 28.64174 28.64003 77.24570 77.24563 77.24696 77.24691 Latitude Longitude 28.76885 77.18087 Segment 27 Merging traffic Latitude Longitude 28.66067 77.24522
49
Intersection(jawahar 7 Burari xing Sarvodaya 8 vidtalay burari 9 Francis school 28.74862 28.74470 77.19853 77.19812 34 Delhi gate 35 LNJP hospital Intersection/Jawaharlal 10 Bengali colony 28.73870 77.19756 36 nehru marg Four way(maharaja 11 Jharoda diary Transport 12 authority 28.73319 77.19803 38 Statesman house Three way( outer circle ,barakhamba 13 Four way 28.72879 77.19812 39 road) Three way(outer circle,Kasturba gandhi 14 Burari crossing Three way15 intersection 28.72235 77.21614 28.72821 77.20028 40 marg) Three way(near 41 janpath) Three way(Outer 16 Gandhi vihar Gopal pur 17 crossing 28.71920 77.22386 43 Hanuman mandir Gurudwara bangla 18 Three way 19 PWD office 20 Majni ka tilla 21 Three way 22 Matkaf house 23 Three way 24 ISBT 25 Yamuna bazaar 26 Three way 28.70923 28.70264 28.69767 28.69682 28.68344 28.67991 28.67118 28.66481 28.66215 77.22732 77.22773 77.22732 77.22729 77.22920 77.22937 77.23141 77.23538 77.23935 44 sahib 45 St.columbus school 46 Kendriya terminal 47 Kendriya terminal 48 Roundabout 49 Kendriya terminal 50 Roundabout 51 Kendriya terminal bs 52 Kendriya terminal
50
28.75441
77.19908
28.64008
77.24113
28.63998 28.64019
77.23874 77.2356
28.64037
77.23215
28.73477
77.19763
28.62907
77.22666
28.63030
77.22394
28.63102
77.22269
28.63028
77.22139
28.6299
77.21970
28.72150
77.21874
42 circle,Sansad marg)
28.63025
77.21816
28.63003
77.21364
Table 12. Conclusions about 185 Down: Slot/Day: Sundays Bottleneck: 1) Transport Authority to Four way( 5.36kmph) Mean speed: 10.08kmph to 36.87kmph Mondays 1) Nathupora to Amrit vihar colony(1.01kmph) 2) Kaushik enclave to Laxmi vihar(5.09kmph) 3) Transport Authority to Four way(4.10kmph) 4) Intersection(Jawaharlal Nehru marg) to Delhi Gate(4.90kmph) Tuesdays 1) Nathupura(3.10kmph) 2) Four way after transport authority to Burari crossing(4.60kmph) 3) ISBT to Yamuna bazaar(4.02kmph) 4) Intersection(Jawahar lal Nehru marg) to Delhi gate(3.50kmph) 5) Three way(outer circle,barakhamba road) to Three way(Kasturba Gandhi marg)(3.04kmph) 6) St. Columbus school to Kendriya terminal(3.70kmph) Wednesdays 1) Nathupora to Amrit Vihar colony(5.5kmph) 2) Transport authority to Four way(4.87kmph) 3) Intersection( Jawaharlal Nehru marg)to Delhi Gate(4.96kmph) Thursdays 1) Transport authority to 4 way(3.77kmph) 2) Intersection( Jawahar lal Nehru marg) to Delhi gate( 4.67kmph) 3) Kendriya terminal (0 kmph) Fridays 1) Transport Aurhority to Four way(4.09kmph) 2) Intersection( Jawahar lal Nehru marg) to Delhi gate(5.14kmph) 3) Kendriya Terminal(4.43kmph) 8.82kmph to 39.06kmph 7.23kmph to 37.76kmph 9.20kmph to 41.20kmph 8.82 kmph to 39.80kmph 8.6kmph to 36.00 kmph
51
Saturdays
1) Nathupura to Amrit Vihar colony(4.80kmph) 2) Transport Authority to Four way(4.05kmph) 3) Statesman house to Three way(outer circle,barakhamba road)(3.90kmph) 4) Kendriya Terminal(3.40kmph) 8.8kmph to 36.90kmph
8am-9am
5.48kmph to 27.22kmph
9am-10am
10.72kmph to 34.28kmph
10am-11am
1)
2) Intersection after DTC Ambedkar nagar Terminal to LNJP hospital 11am12noon 1) Transport Authority to Four way(4.05kmph) 2) Intersection after DTC ambedkar terminal to Delhi Gate(4.90kmph) 12noon1pm 1) Kaushik enclave to Laxmi vihar(5.09kmph) 2) Transport Authority to Four way(3.07kmph) 3) ISBT to Yamuna Bazaar(3.65kmph) 4) Intersection after DTC Ambedkar nagar Terminal to Delhi Gate(4.9kmph) 5) Three way(outer circle,barakhamba road) to Three way(outer circle,Kasturba Gandhi marg)(5.33kmph) 6) Kedriya Terminal (6.00kmph) 1pm-2pm 1) Transport Authority to Four way( 5.47kmph) 2) Intersection after DTC Ambedkar Terminal to Delhi Gate(5.47kmph) 2pm-3pm 1) Transport Authority to Four way(6.24kmph) 2) Intersection after DTC Ambedkar terminal to Delhi Gate(5.22kmph) 3) Three way(outer circle,barakhamba road) to Three way(outer circle,kasturba Gandhi marg)(2.33kmph) 3pm-4pm 1) Nathupura to Amrit Vihar colony(1.01kmph) 2) ISBT to Yamuna bazaar(3.38kmph)
52
8.87kmph to 36.90kmph
9.26kmph to 37.07kmph
6.16kmph to 35.50kmph
8.44kmph to
3) Delhi Gate to LNJP hospital(5.37kmph) 4) Three way(outer circle,barakhamba road) to Three way(outer circle,kasturba Gandhi marg) 5) St.Columbus school to Kendriya Terminal(3.72kmph) 4pm-5pm 1) Nathupura to Amrit Vihar colony(4.8kmph) 2) ISBT to Yamuna Bazaar(4.02kmph) 3) Intersection after DTC Ambedkar Terminal to Delhi Gate(4.50kmph) 4) Three way( outer circle,barakhamba road to Three way,outer circle,kasturba Gandhi marg) 5) Kendriya Terminal bs to Kendriya Terminal(5.01kmph) 5pm-6pm 1) Statesman house to Three way(outer circle,barakhamba road)(4.60kmph) 2) Kendriya Terminal(5.12kmph) 6pm-7pm 1) Intersection after DTC Ambedkar terminal to Delhi Gate(5.14kmph) 2) Kendriya terminal(3.40kmph) 7pm-8pm 1) Intersection after DTC Ambedkar terminal to Delhi Gate(3.56kmph) 2) Statesman House to Three way(outer circle,Barakhamba road)(3.91kmph)
40.58kmph
9.3kmph to 41.20kmph
9.35kmph to 34.09kmph
8.77kmph to 32.60kmph
10.76kmph to 39.06kmph
53
54
Table 13. List of Bus-stops/Traffic Lights used for segmentation: Bus Stop/Traffic Light 1 Nityanand marg Latitude 28.66843 Longitude BS/TL Latitude 28.56672 Longitude 77.25294
2 3
28.66873 28.66945
28.56451 28.56231
77.25026 77.2515
isbt
28.66921
28.56205
77.25565
5 6 7 8
77.23025 39 kailash 77.23503 40 SNP depot 77.23616 41 Modi mill 77.23694 42 Three way Modi mill
Jama masjid
77.23875 43 crossing 77.23956 44 Modi mill 77.24040 45 Modi mill 77.24095 46 NSIC
10 Subhash park 11 Darya ganj 12 Delhi gate Ambedkar 13 stadium terminal 14 Three way
28.63977 28.63960
28.54931 28.54811
77.26257 77.26289
28.61590
28.52929
77.27123
19 Three way
28.61325
28.52859
77.27041
20 Pragati maidan
28.61141
77.24606 54 extension
28.52550
77.26079
55
77.24032 55 Three way 77.24017 56 Tara apartments 77.24012 57 Tughlaqabad Guru ravidas
24 Intersection
28.60735
28.51984
77.25456
25 Intersection 26 Sundernagar Delhi public 27 school 28 Roundabout Dargah hazrat 29 nizammudins Hazrat 30 nizammudins 31 Bhogal 32 Ashram 33 Four way
28.60585 28.60216
28.51947 28.51721
77.25439 77.25357
28.59872 28.59361
28.51566 28.51201
77.25302 77.25195
28.59139
28.51193
77.24962
77.24628 64 mandir 77.25135 65 Vayusena bad 77.25683 66 Tigri 77.25808 67 Devli crossing Ambedkarnagar
34 Nehru nagar
28.56864
77.25413 68 terminal
28.51507
77.22894
Table 14. Conclusions for 411 Up. Slot/Day: 9am-10am Bottleneck: <3kmph 1) Maharana Pratab ISBT bus stop to GCS University bus stop 2) Three way to IG stadium 3) Ambedkar nagar terminal 10am-11am <3kmph 1) ISBT to Kashmiri gate 2) Maharana pratab ISBT bus station to GCS 1.43kmph to 10.10kmph Mean speed: 2.12kmph to 12.60kmph
56
university bus station 3) Four way to Redfort 4) Subhash park to Darya ganj 5) Three way to National Stadium 6) Ambedkarnagar terminal 11am12noon <3kmph 1) Four way to Red fort bus station 2) Ambedkarnagar terminal 12noon1pm <3kmph 1) Maharana Pratab ISBT to GCS university 2) Subhash park bus station to Darya ganj bus station 3) Modi mill crossing to Modi mill 4) Devli crossing to Ambedkar nagar terminal 1pm-2pm <3kmph 1) Maharana Pratab ISBT to GCS university 2) Subhash park bus station to Darya ganj bus station 3) Three way to National stadium bus station 4) Ambedkarnagar terminal 2pm-3pm 5) Maharana Pratab ISBT to GCS university 6) Subhash park bus station to Darya ganj bus station 1) Devli crossing to Ambedkarnagar terminal 3pm-4pm <3kmph 1) Maharanapratab ISBT bus station to GCS university 2) Four way to Redfort bus station 3) Three way to National stadium bus station 4) Ambedkarnagar terminal. 4pm-5pm <3kmph 1) Maharana pratab ISBT bus station to GCS university 2) Four way to Red fort 3) Subhash park to Darya ganj 4) Three way to National Stadium 5) Intersection after Zoo to Intersection
57
2.53kmph to 10.11kmph
1.78kmph to 9.76kmph
1.96kmph to 10.82kmph
6) Ambedkarnagar terminal 5pm-6pm <3kmph 1) Maharana pratab ISBT bus station to GCS university 2) Intersection after Zoo to Intersection 3) Ambedkarnagar terminal 6pm-7pm <3kmph 1) Four way to Red fort bus station 2) Subhash park bus station to Draya ganj bus station 3) Tughlaqabad extension bus station to Three way 4) Ambedkarnagar terminal 7pm-8pm <3kmph 1) Ashram to Four way 2) Ambedkarnagar terminal 1.68kmph to 4.79kmph 2.24 kmph to 10.00kmph 1.38 kmph to 10.07 kmph
58
59
Table 15. List of Bus-stops/Traffic Lights used for segmentation: Bus stop/Traffic Light 1 Ambedkar nagar terminal 2 3 4 5 6 RPS Colony Devli crossing Vayusena bad Tigri Sri satyanarayan mandir 7 Sangam vihar 28.51206 77.24850 41 Dargah hazrat nizammundin 8 9 Hamdard nagar Three way 28.51205 77.24939 28.51193 77.25141 28.51515 77.25272 42 Roundabout 43 Golf club 44 Four way 28.59273 77.24379 28.60129 77.23794 28.60474 77.23568 28.59169 77.24462 28.51522 77.22873 28.51410 77.23375 28.51394 77.23458 28.51291 77.23890 28.51234 77.24317 36 Four way 28.57220 77.25770 28.57371 77.25729 28.58191 77.25166 28.58477 77.24957 28.58932 77.24631 28.51568 77.22654 35 Nehru nagar 28.56897 77.25382 Latitude Longitude Bus Stop/Traffic Light latitude Longitude
10 Apollo pharmacy 11 Hamdard nagar 12 Guru ravidas mandir 13 Tara apartments 14 Tughlaqabad 15 Three way
28.52257 77.25546
28.61811 77.24022
28.52542 77.26200
50 Four way
28.62503 77.24040
28.5271
77.26714
28.62765 77.24126
28.52872 77.27035
52 Bal bhavan
28.62899 77.23823
60
22 Govind puri metro station 23 Kalkaji mandir 24 Kalkaji temple/3 way 25 NSIC 26 Modi mill 27 SNP depot
28.54511 77.26398
56 Delhi gate
28.63740 77.24097
57
58 Darya ganj
59 Subhash park 60 Jama masjid 61 Red fort 62 Presentation convent senior secondary school
29 B Block Esat of 28.56195 77.25649 Kailash 30 Garhi village 31 Three way 32 Lajpat nagar crossing 33 Three way 34 PGDAV College 28.56517 77.25000 28.56724 77.25288 28.56216 77.25158 28.56239 77.25096 28.56436 77.25011
63 GPO
28.66178 77.23481
67 Nityanand marg
28.66815 77.22502
61
Table 16. Conclusions for 411 Down: Slot/Day: Sundays Bottleneck: 1) Ambedkarnagar terminal to Devil crossing(6pm7pm)(<3kmph) 2) 3) 4) 5) Mondays 1) Hamdard nagar to Three way(4.50kmph) Darya Ganj to Subhash Park(5.42kmph) Four way to ISBT(4.63kmph) Nityanand marg(7.03kmph) Ambedkar nagar Terminal to RPS colony(5.25kmph) 2) 3) 4) 5) 6) Hamdard nagar to Three way(5.15kmph) Three way to PGDAV(6.24kmph) Nehru nagar to Four way(5.86kmph) ITPO to Pragati Maidan Gate no5(5.35kmph) Four way to Ram charan agarwal chock(6pm7pm)(2.60kmph) 7) 8) 9) Four way Daryaganj to Daryaganj(6.23kmph) Subhash Park to Jama Masjid(5.60kmph) Red fort to Presentation convent SS school(4pm5pm)(1.90kmph) 10) Four way after CGS University to ISBT(4.70kmph) Tuesdays 1) 2) 3) 4) Ambedkar nagar terminal to RPS colony(3.50kmph) Hamdard nagar to Three way(4.80kmph) Tughlaqabad to Three way(5.70kmph) Lajpat nagar crossing to Three way(5pm6pm)(2.58kmph) 5) 6) Four way after golf club to Three way(3.30kmph) Pragati Maidan Metro station to Four way(3.50kmph) 7) Four way(after Pragati maidan metro) to Ram Charan Agarwal Chock(3.06kmph) 8) 9) Four way Daryaganj to Daryaganj(4.34kmph) Daryaganj to Subhash park(5.90kmph)
62
6.60kmph to 37.90kmph
6.47kmph to 37.01kmph
10) Subhash Park to Jama Masjid(3.50kmph) 11) Jama Masjid to Red fort(5.30kmph) Wednesdays 1) 2) 3) 4) 5) Ambedkar nagar terminal to RPS Colony(4.12kmph) 5.91kmph to Hamdard nagar to Three way(6.18kmph) C lal chock to Kalkaji depot(7.44kmph) Garhi village to Three way(3.75kmph) Three way(after garhi village)to Lajpat nagar crossing(6.10kmph) 6) Three way (after Lajpat nagar crossing) to PGDAV (5.64kmph) 7) 8) 9) Four way (after Golf club) to Three way(1.70kmph) ITPO to Pragati maidan Gate no.5(5.74kmph) Pragati Maidan metro station to Four way(5.56kmph) 10) Four way to Ram Charan Agarwal Chock(2.11kmph) 11) Four way Darya Ganj to Darya ganj (6.62kmph) 12) Presentation Convent SS school to GPO(6.28kmph) 13) Nityanand marg (2.76kmph) Thursdays 1) 2) 3) 4) 5) 6) 7) 8) Ambedkar nagar terminal to RPS Colony(3.93kmph) 7.30 kmph to RPS Colony to Devli crossing(4.26kmph) Devli crossing to Vayusena bad(5.75kmph) Hamdard nagar to Three way(4.80kmph) C lal chock to Kalkaji depot(5.80kmph) Nehru nagar to Four way(6.50kmph) Four way(after Golf club) to Three way(6.60kmph) Four way after Pragati maidan metro station to Ram charan agarwal chock(4.90kmph) 9) Bal Bhavan to Connector road(5.55kmph) 36.20 kmph 34.60kmph.
10) Four way Darya ganj to Darya ganj(6.60kmph) 11) Subhash park to Jama masjid (8.42kmph) 12) Presentation convent SS school to GPO(6.30kmph) 13) Four way (after CGS university) to ISBT(7.50kmph)
63
14) ISBT to Nityanand marg(6.95kmph) Fridays 1) 2) 3) 4) 5) Ambedkar nagar terminal to RPS Colony(3.26kmph) 7.80kmph to Hamdard nagar to Three way(6.40kmph) Tughlaqabad to Three way(6.02kmph) C lal chock to Kalkaji depot(8.44kmph) Three way(after Garhi Village) to Lajpat Nagar crossing(7.40kmph) 6) Three way(after Lajpat nagar crossing) to PGDAV college(6.11kmph) 7) 8) Nehru nagar to Four way(6.89kmph) Four way (after pragati maidan metro station) to Ram charan agarwal chock(1.73kmph) 9) Four way Darya ganj to Darya ganj(5.90kmph) 37.85kmph
10) Subhash park to Jama masjid(7.32kmph) 11) Presentation convent SS school to GPO(4.99kmph) Saturdays 1) 2) 3) 4) 5) Ambedkar nagar terminal to RPS Colony(4.80kmph) 5.99kmph to Hamdard nagar to Three way(4.62kmph) Tughlaqabad to Three way(7.40kmph) C lal chock to Kalka ji depot(6.80kmph) C block east of Kailash to B block east of kailash(4.55kmph) 6) 7) Garhi village to Three way(2.77kmph) Three way (after Lajpat nagar crossing) to PGDAV college(4.74kmph) 8) 9) Nehru nagar to Four way(4.32kmph) ITPO to Pragati Maidan Gate no.5(8.05kmph) 36.62kmph
10) Four way(after Pragati maidan metro station) to Ram charan agarwal chock(4.63kmph) 11) Four way darya ganj to Daryaganj(6.99kmph) 12) Presentation Convent SS school to GPO(5.63kmph) 13) CGS university to Four way(5.39kmph) 14) Nityanand marg (6.44kmph) 8am-9am <6kmph: 5.46kmph to
64
1)
39.50kmph.
2) 3) 4) 5)
Hamdardnagar to Three way(4.80kmph) Tughlaqabad to Three way(6.70kmph) Nehrunagar to Four way(5.80kmph) Four way after Golf club to next Four way(3.32kmph)
6)
7) 8) 9am-10am 1)
Four way to Ram charan agarwal chock(6.97kmph) Bal bhavan to Connector road (5.55kmph) Ambedkarnagar terminal to Devli crossing(3.17kmph) 6.82kmph to 37.90kmph
2) 3) 4)
Hamdard nagar to Three way(4.62kmph) Tughlaqabad to Three way(6.38kmph) Three way after Garhi village to Lajpat nagar crossing(7.43kmph)
5) 6)
Four way after Golf club to Three way(6.62kmph) Four way after Pragati maidan metro station to Ram charan agarwal chock(5.74kmph)
7) 10am-11am 1)
Presentation Convent SS school to GPO(7.54kmph) Ambedkarnagar terminal to Devli crossing(6.30kmph) 8kmph to 37.75kmph
2) 3) 4)
Hamdard nagar to Three way(6.66kmph) Tughlaqabad to Three way(7.77kmph) Three way after Garhi village to Lajpat nagar crossing(8.41kmph)
5)
6) 7) 8) 9)
Subhash Park to Jama Masjid(5.70kmph) Presentation Convent SS school to GPO(6.30kmph) CGS university to Four way(5.49kmph) Nityanand marg(7.20kmph)
65
11am12noon
1) 2) 3) 4)
Ambedkarnagar to RPS colony(3.01kmph) Hamdard nagar to Three way(7.48kmph) Tughlaqabad to Three way(7.87kmph) Three way afer Lajpat nagar crossing to PGDAV college(7.54kmph)
6.93kmph to 37.04kmph
5)
Four way after Pragati maidan metro station to Ram charan Agarwal chock(7.68kmph)
6)
7) 12noon-1pm 1) 2) 3) 4)
Presentation Convent SS school to GPO(8.13kmph) Ambedkarnagar terminal to RPS Colony(3.01kmph) Hamdard nagar to Three way(7.88kmph) C lala chock to Kalkaji depot(7.63kmph) Three way(after Garhi village) to Lajpat nagar crossing(7.45kmph) 8.06kmph to 37.42kmph
5)
6)
Four way(after Pragati maidan metro station) to Ram charan agarwal chock(5.94kmph)
7) 1pm-2pm 1) 2) 3) 4) 5) 6) 7) 8) 2pm-3pm 1) 2) 3) 4) 5)
Darya ganj to Subhash Park(5.49kmph) Ambedkar nagar terminal to RPS Colony(6.7kmph) Tughlaqabad to Three way(7.23kmph) C lal chock to Kalka ji depot(5.87kmph) Garhi village to Three way(1.15kmph) Lajpat nagar crossing to Three way(5.24kmph) Three way to PGDAV college(4.17kmph) Nehrunagar to Four way(6.57kmph) ITPO to Pragati maidan gate no.5(5.35kmph) Ambedkar nagar terminal to RPS Colony(3.26kmph) 4.71kmph to Hamdard nagar to Three way(6.09kmph) Tughlaqabad to Three way(6.51kmph) Garhi village to Three way(3.50kmph) Three way to Lajpat nagar crossing(5.83kmph) 37.50kmph 7.9kmph to 37.3kmph
66
6) 3pm-4pm 1) 2) 3) 4) 5) 6)
Presentation convent SS school to GPO(4.90kmph) Ambedkarnagar terminal to RPS Colony(6.13kmph) Three way to Lajpat nagar crossing(6.33kmph) Three way to PGDAV college(4.74kmph) PGDAV college to Nehrunagar(6.58kmph) Four way to Ram charan agarwal chock(4.63kmph) Four way Daryaganj to Daryaganj bus stop(6.19kmph) 1.70kmph to 37.70kmph
7) 8) 4pm-5pm 1)
Subhash park to Jama masjid(6.73kmph) Four way to ISBT(4.63kmph) Ambedkarnagar terminal to Devli crossing(3.50kmph) 5.61kmph to 39.03kmph
2) 3) 4) 5) 6) 7) 8) 9)
C lal chock to Kalkaji depot(6.80kmph) Garhi village to Three way(3.98kmph) Three way to Lajpat nagar crossing(5.99kmph) Lajpat nagar crossing to Three way(6.63kmph) Three way to PGDAV college(4.90kmph) Four way to Ramcharan Agarwal chock(5.60kmph) Four way Darya ganj to Darya ganj(5.90kmph) Red fort to Presentation Convent SS school(1.92kmph)
10) Nityanand marg(6.44kmph) 5pm-6pm 1) 2) 3) 4) RPS colony to Devli crossing(4.26kmph) Devli crossing to Vayusena bad(5.75kmph) Himgiri apartments to Kalkaji(3.80kmph) C block east of kailash to B block east of kailash(4.55kmph) 5) 6) 7) 8) 9) Garhi village to Three way(2.77kmph) Lajpat nagar crossing to Three way(2.58kmph) Three way to PGDAV college(4.90kmph) Four way to Ram charan agarwal chock(2.59kmph) Presentation convent SS School to GPO(5.63kmph) 4.26kmph to 34.53kmph
67
6pm-7pm
1)
2.60kmph to 35.80kmph
2)
Devli crossing to Vayusena bad(4.60kmph) Four way after Golf club to Three way(1.79kmph)
3)
4) 5) 6) 7) 7pm-8pm 1) 2) 3) 4) 5) 6)
Four way to Ram charan agarwal chock(1.73kmph) Four way Daryaganj to Daryaganj(6.24kmph) Presentation convent SS school to GPO(6.09kmph) Nityanand marg(3.78kmph) Four way to Ram charan agarwal chock(2.11kmph) Four way Daryaganj to Daryaganj(4.34kmph) Daryaganj to Subhash park(5.94kmph) Subhash park to Jama masjid(3.58kmph) Four way to ISBT(4.70kmph) ISBT to Nityanand marg(2.76kmph) 5.70kmph to 43.40kmph
68
CHAPTER 4 CONCLUSIONS
The mean speeds for all routes are tabulated in Table 19 and Fig 7 is a plot of these values. From these mean speed values obtained the following observations have been made: The mean speed for Route 108 Up varies between 11.30kmph to 32.00 kmph. For Route 108 Down it varies from 11.16kmph to 31.00 kmph. For Route 185 Up, it varies from 8.17kmph to 32.00 kmph. For Route 185 Down, it varies from 12.43kmph to 44.90 kmph.
50 45 40 35 Speed (kmph) 30 25 20 15 108 up 108down 185 down 185 up 411 up 411 down
10
5 0 1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 Segment No.
For Route 411 Up, it varies from 2.01kmph to 9.61kmph. For Route 411 Down, it varies from 6.63kmph to 35.87kmph.
69
For most routes the mean speed varies between 10kmph to 35 kmph, with Route 185 Down being the exception on the higher side(upto 45kmph) and Route 411 Up being the exception on the lower side(<10kmph). A comparison of mean speed for all the routes over different time slots(Table 17,Fig 8) reveals that for all routes the mean speeds are higher is the morning slot (8am-10am) and comparatively lower in the evening slot (5pm-7pm). Route 411 Up brings up the rear end of the speed values, with mean values as low as 5kmph.
4.1
Definition of Bottleneck:
The speed value that defines whether a segment classifies as bottleneck or not, is different for different routes since the mean speed is different for different routes. The values of bottlenecks used for different routes is provided in Table 18.The lowest value of speed for a particular segment is noted and values falling in the lowest 5kmph-speed-band are identified as bottlenecks (except for Route 411 Up, where the minimum mean speed is 2kmph and the upper limit of 3kmph is taken as the identifier for bottlenecks).
Table 17: Comparison of mean speed over whole routes for different time slots 108 Up 8am-9am 9am-10am 10am-11am 11am-12noon 12noon-1pm 1pm-2pm 108 Up 22.61 20.38 20.42 20.02 19.47 19.64 108 Down 23.68 21.38 21.29 20.65 20.08 19.98 185 up 25.66 23.60 23.21 20.96 20.52 19.68 185 down 18.81 19.85 20.18 19.56 20.90 19.45 5.94 5.75 5.96 5.85 5.73 411 up 411 down 23.07 21.64 20.76 20.24 20.18 20.31
70
Fig8: Comparison of mean speed over whole routes for different time slots
Table 18: Bottleneck speed Route 108 Up 108 Down 185 Up 185 Down 411 Up Maximum mean speed 32.00 31.00 32.00 44.90 7.61 Minimum mean speed 11.30 11.16 8.17 12.42 2.01 Bottleneck upperlimit 6.90 7.45 8.44 5.94 3.00
71
411 Down
35.87
6.63
8.42
4.2
The bottlenecks identified in this project could be examined and the reason behind formation of the bottleneck could be found. Measures to reduce bottleneck could then be suggested. Routes with maximum obtained mean speeds lower than expected speed could be examined for reallocation of traffic sources to reduce traffic flow and thus, increase mean speeds. If a scale factor for change of speed over months were to be developed, the available mean speeds over segments could be used to predict future speed on same segment for same time slot.
72
108 Up Segment Nehru Vihar 1 road 2 Four way Nehru vihar 3 crossing Police station 4 timarpur Balak ram 5 hospital
Table 19. COMPARISON CHART OF MEAN SPEEDS OVER VARIOUS ROUTES (all speed values in kmph) 108 Down 185 up 185 down 411 up 411 down Mean values Segment Mean values Segment Mean values Segment Mean values Segment Mean values Segment Ambedkar nagar 4.38 terminal 4.64 RPS Colony 8.14 Devli crossing Vayusena 2.67 bad 5.14 Tigri Sri satyanarayan 4.66 mandir Mean values
DDU 22.36 hospital 17.43 Roundabout Beri wala 18.81 bagh Swarg 20.64 Asharam Swarg 21.18 asharam
Kendriya 18.48 terminal Roundabo 20.58 ut Kendriya 22.19 terminal Kendriya 30.87 terminal 21.86 NDPO Roundabo 21.90 ut Gurudwar a bangle 29.58 sahib
13.35 Nathupura Amrit vihar 26.34 colony Kaushik 25.60 enclave 25.85 Laxmi vihar 26.94 Burari ganj Burari 19.07 village
Nityanand 13.24 marg 13.45 ISBT 13.45 Kashmiri gate Maharana 14.11 pratab isbt GCS 12.44 university
6 Timar pur Timarpur 7 water tank North mall 8 road 9 Mall road International Students 10 hostel 11 Four way
18.02 LIG flats (Four way) Ram singh 23.05 marg Maya puri metal 19.51 forging Government 14.43 press
13.36 GPO
20.52
24.45 YMCA Jantar 14.64 Mantar Palika 11.16 Kendra Regal 21.70 Cinema
20.20 Burari xing Sarvodaya vidtalay 27.54 burari Francis 26.92 school Bengali 16.78 colony Jharoda 17.77 diary
2.96 Sangam vihar Hamdard 4.86 nagar 3.98 Three way Apollo 2.83 pharmacy Hamdard 4.88 nagar
30.82
10.17 23.15
28.73 26.39
73
Khalsa 12 college Patel chest(four 13 way) Sri ram 14 college Daulat ram 15 college Maurice 16 nagar
18.88 Chuna bhati Wood 13.70 market Saraswati 25.57 garden Furniture 19.80 market 13.93 Kirti nagar F block kirti 28.46 nagar
Three way(Panc hkurian road joins inner 23.50 circle) Shivaji 24.75 Park New Delhi Railway 21.46 Station Roundabo 27.56 ut JL Nehru 19.71 Marg Zakir Hussain 19.11 College Three way(Jawa har lal nehru 21.09 marg) Asaf Ali 15.21 College Delhi Nagar 14.15 Nigam Hotel 11.99 Broadway
19.54 Four way Burari 18.58 crossing Three wayintersectio 22.03 n Gandhi 19.10 vihar Gopal pur 29.96 crossing
27.14
4.98
30.90
4.43 Tughlaqabad
10.84
5.94
12.95 22.77
17 Roop nagar
IP depot
35.88
20.62 Rounabout 11.33 Kirti nagar Moti nagar 17.55 market 16.65 Moti nagar
21.08 Three way 17.77 PWD office Majni ka 20.91 tilla 16.81 Three way
Pragati power 18.26 station 19.97 Three way Pragati maidan 19.32 Three way
33.00 20.55
13.68 18.23
74
ESI 18.21 dispensary 17.35 Campa cola DCM 20.81 chemicals 30.07 Roundabout 17.12 Zakhira
Govind puri 5.48 metro station Kalkaji 5.29 mandir Kalkaji temple/Thre 4.09 e way 6.68 NSIC 7.48 Modi mill
19.62 34.93
27.38 Three way Jama 31.42 Masjid 19.88 Red Fort Four way( shyam prasad mukherje 16.83 e marg) 17.43 GPO GGS 17.85 University
Delhi public 24.01 school Roundabout Dargah hazrat 26.20 nizammudins Hazrat 28.16 nizammudins
6.96 SNP depot C Block East 8.36 of Kailash B Block Esat 5.12 of Kailash
20.33 25.20
29 Inderlok
30 Intersection
17.98
31 Zakhira road
32 Zakhira
20.11 ISBT Three way(Lala hardev sahai 23.08 marg) Three way(ISBT) to Shyam 17.57 nath marg
26.91
26.12 Bhogal
10.40
30.97 Ashram
20.47
33 Roundabout
DCM 34 chemicals
35 Campa cola
Ludlow 25.91 Castle Three way(Shya m nath 27.36 marg) Civil lines metro 19.64 station
Intersectio 27.91 n
10.62
23.70 Delhi gate lNJP 20.10 hospital Intersectio n/Jawaharl al nehru 27.54 marg Four way(mahar aja ranjit singh 29.63 marg) Statesman 28.17 house Three way( outer circle ,barakham 23.51 ba road) Three way(outer circle,Kastu rba gandhi 32.73 marg) Three way(near 18.62 janpath) 76
19.53
15.49
36 ESI dispensry
12.30 Roundabout
19.49 IP College Three way(Maha tma gandhi 14.82 road) Old Secretaria 11.93 t
4.56
Four way
28.65
7.35 Ashram
25.41
6.52 Bhogal
18.94
39 Three way
15.98 Kamla nagar Roop nagar(bus stop before 21.13 roundbout) Roop nagar(bus 19.24 stop after
8.84 Jangpura
15.80
40 Roundabout
41 F block
20.58
33.66
roundabt) Internatio nal Students 18.67 Hostel 21.21 GTB nagar Three way(Maha tma gandhi 27.98 marg) Camp 16.21 Chock T.B.Hospit 28.92 al Gandhi 20.26 Ashram Daka 11.65 Village Three way(Outer circle,Sansa 13.46 d marg) Hanuman 16.01 mandir
26.95 32.47
Gurudwara bangla 28.13 sahib St.columbu 8.18 s school Kendriya 13.96 terminal Kendriya 26.91 terminal Roundabou 21.74 t
5.76 Four way 5.60 Three way 7.13 ITPO Pragati Maidan Gate 9.31 no5 Supreme 7.29 Court Pragati Maidan Metro 7.42 Station
23.56
26.03
Enter side 21.30 road Lucknow road government 14.22 school Lucknow 18.06 road
18.00
51 LIG flats
9.10
16.86
Swarag 52 ashram
32.15 24.41
CB Raman 25.60 IIT Colony Nirankari Sarovar Burari Crossing Nirankari Sarovar Four way(outer ring road) Transport Authority Jharoda Diary St. Nagar Bengali Colony Francis School Sarvodaya Vidyalay Burari Burari Village Burai Ghari Laxmi vihar Kaushik Enclave
44.79 Intersection
19.58
29.36 27.79
Connector 5.65 road Express 4.21 building Shaheed bhagat singh 4.93 park 8.82 Delhi gate Four way 5.10 darya ganj 9.41 Darya ganj 6.31 Subhash park 8.29 Jama masjid
22.36 23.19
18.64 12.02
Three way Tara apartments Tughlaqabad Guru ravidas ashram Guru ravidas mandir Hamdard nagar Apollo pahmacty
61
22.38
62 63 64 65
9.62 Red fort Presentation convent SS 4.47 school 5.07 6.89 7.61 GPO CGS university Four way
15.46
66 67
25.03 12.57
Tigri
3.85
16.45 10.97
79
References: 1. Chakroborty, P., Kikuchi, S.:Using Bus Travel Time Data to Estimate Travel Times on Urban Corridors,in: Transportation Research Record: Journal of the Transportation Research Board,No.1870,2004,p 18. 2. Jeong,R.,Rilett,L.R.:Bus Arrival Time Prediction Using Artificial Neural Network Model, in IEEE Intelligent Transportation Systems Conference,WeB2, 2004,p988. 3. Sekhar,C.R.,Iryo ,T.,Asakura,Y .:Analysis of travel-time variation over multiple sections of Hanshin expressway in Japan,in:CURRENT SCIENCE, Vol. 102, No. 10, 2012, p1. 4. Guin,A., Travel Time Prediction using a Seasonal Autoregressive Integrated Moving Average Time Series Model,in:IEEE Intelligent Transportation Systems Conference, MC3.3, 2006,p 493. 5. Ghosh,B. ,Basu,B.,OMahony,M.:Multivariate Short-Term Traffic Flow Forecasting Using TimeSeries Analysis,in: IEEE Transactions on Intelligent Transportation Systems ,v10 no.2, 2009, p246.
6. Williams,B.M., Hoel, L.A.:Modeling and Forecasting Vehicular Traffic Flow as a Seasonal ARIMA Process:Theoretical Basis and Empirical Results,in: Journal of Transportation Engineering,v 129, 2003,p664. 7. Kyte, M., Stoner,J.,Cryera ,J.:Time-series analysis of public Portland,Oregon,1971-1982,in:Transpn Res.-A,v 22A, 1988,p345. transit ridership in
8. Stathopoulos A.,Matthew, G.:A multivariate state space approach for urban traffic flow modeling and prediction,in: Transportation Research, Part C 11,2003,p121. 9. Nihan,N.L.,Holmesland,K.O.:Use of the Box and Jenkins Time Series Technique in Traffic Forecasting,in:Transportation, v 9,1980,p125.
10. Box, G.E.P.,Jenkins,G.M.,Reinsel,G.C.: Time Series Analysis, Forecasting & Control ,in: Pearson Education,Inc. 2009.
80