Lelitha Devi Vanajakshi

Estimating delays at different road segments and intersections is a primary step in any traffic management system. Delay along a path can be estimated using sensors such as Global Positioning System (GPS) sensors, Camera sensors, Wi-Fi sensors and On-Board Diagnostics (OBD) data of the vehicle. The former three techniques have been used for some time now, but the latter one has not been explored much, especially for estimating delay. The accuracy of methodologies using OBD, Wi-Fi and Camera was calculated using data from GPS as ground truth. Results obtained showed the performance of all these methods to be comparable. Choice of the right sensor may depend on various other factors such as cost, weather factors, accuracy, range, etc. Sometimes more than one sensor needs to be used for the desired results. This paper explores three techniques in more detail and summarizes the advantages and disadvantages of each one of them in various scenarios.

Motorised three wheeler is one of the most popular paratransit vehicles in India, due to its small size, cost-effectiveness, manoeuvrability, and availability. It provides the most flexible, rider-friendly, quick travel even in traffic-choked streets and narrow roads, exhibiting entirely different characteristics in terms of travel speed, and trip lengths compared to other paratransit modes like taxicabs. Reported studies on the behaviour of paratransit mainly concentrated on taxicabs and hence there is a need to analyse the travel behaviours of motorised three wheelers. In this regard, the present study aims to understand and characterise the travel patterns of motorised three wheelers. In addition, travel time prediction is an inevitable aspect for demand-responsive paratransit services like motorised three wheelers, taxicabs etc. It helps both the drivers and passengers to make smart choices about the routes by avoiding congested streets and to have information about the pickup and arrival time. The present study proposes a methodology using Support Vector Regression (SVR) to predict the travel time of motorised three wheelers by incorporating the trip characteristics under heterogenous lane less traffic conditions. The performance of the proposed method showed a clear improvement when compared with a Median based prediction methodology that was reported to be working well for travel time prediction problems.

Providing accurate and reliable travel time information to travellers is essential to improve the quality of public transit systems. With the availability of the latest technologies, it has become possible to collect a large amount of traffic data to analyze and understand these systems better. Traffic in India is characterized by lack of lane discipline and the presence of vehicles of varying static and dynamic characteristics, which makes prediction of bus travel time especially challenging. The aim of this study is to identify both a prediction algorithm that can handle high variability and suitable inputs or regressors to be used. Earlier studies performed offline manual grouping considering the patterns observed, which leads to limitations for automated field implementations. The present study explores the use of data-driven approaches, primarily clustering, to address the challenges for the prediction of bus travel time trends. Discrete wavelet transform (DWT) was used to extract trends from the travel time measurements. Three popular clustering algorithms—k-means, hierarchical, and self organizing maps (SOM)—were used to identify patterns. Travel time trends were then predicted by searching for similar cluster patterns within the historical database using pattern sequence-based forecasting (PSF). A comparison of the performance of these algorithms was carried out based on prediction errors. The clustering +prediction framework developed was also compared with the case when no clustering was done on the regressor dataset.

Even though new infrastructure is being developed to meet demand, increased urbanization and vehicle ownership have increased the congestion levels in Indian cities. Attracting more travelers to public transport is an option to reduce congestion but still remains a challenge, mainly because of the uncertainty of service. A reliable and accurate system for predicting vehicle arrival can help make public transportation more attractive. An accurate prediction method should be used to provide reliable information to passengers, and accuracy depends on the input data used. Therefore, identifying the optimum inputs and incorporating them in the prediction method become important. The optimum number of inputs required for best prediction performance was identified with an analytical approach. A model-based algorithm motivated by the Kalman filter was used to predict bus travel time with the use of GPS data. A case study was conducted on two selected bus routes in the city of Chennai, India, to evaluate the prediction accuracy of the proposed method. Results obtained from the algorithm were promising and showed the prediction accuracy to be ±5 min for a prediction window of 30 min in 92% of instances. The predicted travel time can be used to provide realtime bus arrival information to the public through various media, including web pages, mobile applications, and display boards.

In developing countries like India, number of people using public transport for everyday commute is large. Trip planner is a tool which helps commuters to plan their travel beforehand. In case of public transportation systems, trip with the minimum travel time is often of interest. A trip planner solves the time dependent shortest path problem (TDSPP) in a multimodal transport network to optimize one or more criteria like travel time, the number of transfers, etc. One subclass of this is the departure time planner. It suggests the optimal departure time from origin such that travel time to reach the destination will be minimum. This paper presents development of multimodal departure time planner using General Transit Feed Specification (GTFS) data. A citywide public transportation network is constructed with bus, metro and walking as modes of transport. Nodes represent transit stops and edges represent transportation services available in between nodes. The schedule corresponding to every mode is incorporated in the network. Origin, destination, and the latest arrival time at the destination for the trip are inputs from the user. A schedule-based algorithm is implemented which runs backward in time to calculate optimal labels at every node of the network. The results produced by the trip planner are found to be promising in terms of accuracy and feasibility.

Intersections account for most road accidents and delay in a road network. The intersection's efficiency and safety concerns can be addressed by collecting vehicle platoon size, queue length, and delay data and implementing a red light violation detection technique (RLVDS) to reduce accidents. It is challenging to collect this traffic information in heterogeneous and less lane disciplined traffic, a scenario often observed in developing countries such as India. Traditional sensors such as inductive loop, infrared, radar, or magnetic sensors and image processing solutions do not work well under these conditions. Hence the current work presents a set of robust techniques developed for heterogeneous traffic. First, the platoons were detected using foreground extraction, connected component analysis, and a density-based clustering algorithm. Then, the queue length was extracted using a progressive block processing technique. Separately the RLVD was performed using corner point tracking and user-defined detection zones.

Road traffic congestion has become a global worry in recent years. In many countries congestion is a major factor, causing noticeable loss to both economy and time. The rapid increase in vehicle ownership accompanied by slow growth of infrastructure has resulted in space constraints in almost all major cities in India. To mitigate this issue, authorities have shifted to more sustainable management solutions like Intelligent Transport System (ITS). Advanced Public Transportation System (APTS) is an important area in ITS which could considerably offset the growing ownership of private vehicles as public transport holds a noticeable mode share in several major cities in India. Getting access to real-time information about public transport would certainly attract more users. In this regard, this work aims at developing a reliable structure for predicting arrival/travel time of various public transport systems under heterogeneous traffic conditions existing in India. The data used for the study is collected from three cities-Surat, Mysore, and Chennai. The data is analyzed across space and time to extract patterns which are further utilized in prediction models. The models examined in this paper are κ-NN classifier, Kalman Filter and Auto-Regressive Integrated Moving Average (ARIMA) techniques. The performance of each model is evaluated and compared to understand which methods are suitable for different cities with varying characteristics.

This study explores the use of On-Board Diagnostics (OBD) data in the analysis and prediction of vehicle dynamics. Though various data sources are available for traffic data collection, these conventional approaches may not work for the complex traffic system in India, with its heterogeneity and lack of lane discipline. On-board units such as GPS and OBD are some devices, which perform independent of the traffic conditions. This study focuses on the use of OBD data along with GPS data for individual vehicle trajectory prediction. A machine learning tool, namely Long–Short-Term Memory (LSTM) model is employed and the prediction of speed and bearing for the next 1 s is done. Results obtained showed the OBD as a potential source of data that can be used for various real-time and offline applications.

This study evaluates and compares two major solutions used to mitigate the problem of delay of public transport buses at signalized intersections in urban arterials. The solutions considered are to implement bus signal priority (BSP) at intersections and to provide dedicated bus lanes (DBL). Comparison of performance is done in terms of travel time. Study site selected is a 5 km stretch in the IT corridor in Chennai for the DBL and with four intersections in that corridor for the BSP. Only the southbound movement is considered for the analysis. To evaluate the selected strategies, the study site was simulated in VISSIM, which is a micro-simulation software. Road, traffic, and control details from field were used as input, and a calibrated network was created. Using that, four scenarios were studied: 1. base condition with fixed time signals and no dedicated lane, 2. fixed time signal with DBL, 3. BSP without dedicated lane, and 4. BSP and DBL. Conditional Green Extension and Red Truncation priority strategies using Visual Vehicle Actuated Programming (VisVAP) from VISSIM were used for the BSP implementation. For dedicated bus lane condition, the mode preference was altered for the southbound movement links and the bus movement was allowed only in the left-most lane, which was kept as DBL. Results showed the bus signal priority having maximum impact in terms of reduction in total travel time. In the case of dedicated bus lane condition, travel time reduced effectively for bus mode, but with an increase in travel time of other modes.

Analysis of queue discharge characteristics at signalized intersections is a primary component of traffic signal analysis and design. On the basis of previous studies, mainly conducted in homogeneous traffic conditions, the discharge headway is assumed to be high at the start of green for the first few vehicles, mainly because of start-up lost times, and is also assumed to reach the minimum value by the fourth or fifth vehicle in the queue. The minimum headway is expected to continue until the end of the queue. However, this may not be the case under heterogeneous traffic conditions, such as those in India, which has the additional problem of lacking lane discipline. Most of the signals in India include a countdown timer that indicates the time left for the signal phase, which is also expected to affect queue discharge characteristics. This paper presents insights gained on queue discharge characteristics at signalized intersections under heterogeneous traffic conditions and on the effect of a countdown timer on the headway distribution. The analysis was carried out using data collected from two intersections, one with a timer and one without, in Chennai, India, through the use of a videographic technique. The data collected are classified into three discharge regimes: start-queue, mid-queue, and end-queue. Linear regression models are used to assess the impact of vehicle types on queue discharge characteristics. The results indicate that the accepted headway distribution is followed when there is no timer. However, with the presence of a timer, there is a clear change in the trend for reduced start-up lost time and end lost time.