Lelitha Vanajakshi

Support Vector Machine Technique for the Short Term Prediction of Travel Time

A vast majority of urban transportation systems in North America are equipped with traffic surveillance systems that provide real time traffic information to traffic management centers. The information from these are processed and provided back to the travelers in real time. However, the travelers are interested to know not only the current traffic information, but also the future traffic conditions predicted based on the real time data. These predicted values inform the drivers on what they can expect when they make the trip. Travel time is one of the most popular variables which the users are interested to know. Travelers make decisions to bypass congested segments of the network, to change departure time or destination etc., based on this information. Hence it is important that the predicted values be as accurate as possible. A number of different forecasting methods have been proposed for travel time forecasting including historic method, real-time method, time series analysis, and artificial neural networks (ANN). This paper examines the use of a machine learning technique, namely support vector machines (SVM), for the short-term prediction of travel time. While other machine learning techniques, such as ANN, have been extensively studied, the reported applications of SVM in the field of transportation engineering are very few. A comparison of the performance of SVM with ANN, real time, and historic approach is carried out. Data from the TransGuide Traffic Management center in San Antonio, Texas, USA is used for the analysis. From the results it was found that SVM is a viable alternative for short-term prediction problems when the amount of data is less or noisy in nature.

A Multiple Inductive Loop Vehicle Detection System for Heterogeneous and Lane-Less Traffic

This paper presents a novel inductive loop sensor that can detect vehicles under a heterogeneous and less-lane-disciplined traffic and thus can be used to support a traffic control management system in optimizing the best use of existing roads. The loop sensor proposed in this paper detects large (e.g., bus) as well as small (e.g., bicycle) vehicles occupying any available space in the roadway, which is the main requirement for sensing heterogeneous and lane-less traffic. To accomplish the sensing of large as well as small vehicles, a multiple loop system with a new inductive loop sensor structure is proposed. The proposed sensor structure not only senses and segregates the vehicle type as bicycle, motor cycle, scooter, car, and bus but also enables accurate counting of the number of vehicles even in a mixed traffic flow condition. A prototype of the multiple loop sensing system has been developed and tested. Field tests indicate that the prototype successfully detected all types of vehicles and counted, correctly, the number of each type of vehicles. Thus, the suitability of the proposed sensor system for any type of traffic has been established.

Data Fusion-Based Traffic Density Estimation and Prediction

Traffic congestion has become a major challenge in recent years in many countries of the world. One way to alleviate congestion is to manage the traffic efficiently by applying intelligent transportation systems (ITS). One set of ITS technologies helps in diverting vehicles from congested parts of the network to alternate routes having less congestion. Congestion is often measured by traffic density, which is the number of vehicles per unit stretch of the roadway. Density, being a spatial characteristic, is difficult to measure in the field. Also, the general approach of estimating density from location-based measures may not capture the spatial variation in density. To capture the spatial variation better, density can be estimated using both location-based and spatial data sources using a data fusion approach. The present study uses a Kalman filter to fuse spatial and location-based data for the estimation of traffic density. Subsequently, the estimated data are utilized for predicting density to future time intervals using a time-series regression model. The models were estimated and validated using both field and simulated data. Both estimation and prediction models performed well, despite the challenges arising from heterogeneous traffic flow conditions prevalent in India.

Estimation of bus travel time incorporating dwell time for APTS applications

Congestion has become a serious problem in the context of urban transport around the world. As more and more vehicles are being introduced into the urban streets every year, the mode share of the public transportation sector is declining at an alarming rate. Particularly in developing countries, more people have moved to personalized mode since it is becoming easily affordable and the quality of service offered by the public transit is not improving. To attract more people, the public transit should provide a high level of quality service to the passengers. One way of achieving this is by using Advanced Public Transport Systems (APTS) applications such as providing accurate real-time bus arrival information to the passengers which will improve the service reliability of the public transit. Travel time prediction has been a well-renowned topic of research for years. However, studies which were model based and incorporating dwell times at bus stops explicitly for heterogeneous traffic conditions are limited. The present study tries to explicitly incorporate the bus stop delays associated with the total travel times of the buses under heterogeneous traffic conditions. This will help in obtaining a reliable algorithm which can be adopted for bus arrival time prediction under Indian conditions.

Effect of Phase Countdown Timers on Queue Discharge Characteristics Under Heterogeneous Traffic Conditions

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.

Prediction of traffic density for congestion analysis under Indian traffic conditions

Traffic congestion is a serious problem which traffic engineers all over the world are trying to solve. Congestion increases the uncertainty in travel times leading to human stress and unsafe traffic situations. Better management of traffic through Intelligent Transportation Systems (ITS) applications, especially by predicting the congestion on various roads and informing the travelers regarding the same is one possible solution. Accurate and quick prediction is one of the important factors on which the reliability of such a system depends. If one is able to predict congestion on a roadway, then the travelers can be warned of the same either pre-trip or enroute so that they can take well informed travel decisions. The number of vehicles in a given stretch of a roadway (usually referred to as “traffic density”) is one of the most commonly used congestion indicator. Also, the travelers in general will be more interested to know what they can expect when they make the trip in future rather than the present scenario. This makes the short term prediction to future time intervals important. In this study, some of the reported techniques for density prediction under homogeneous traffic conditions are attempted under heterogeneous traffic conditions in order to determine their feasibility under the Indian traffic scenario.

An Efficient Multiple-Loop Sensor Configuration Applicable for Undisciplined Traffic

This paper presents an effective multiple-inductive-loop pattern suitable for heterogeneous and less lane-disciplined traffic and its performance evaluation. Vehicle detection system based on conventional inductive loops works well only for lane-based and homogeneous traffic. A multiple-loop system for sensing vehicles in a heterogeneous and less lane-disciplined condition has been reported recently. The scheme proposed in this paper employs a new configuration, where all the loops are connected in series, which considerably reduces the system complexity and improves reliability. Each loop has a unique resonance frequency and the excitation source given to the loops is programmed to have frequency components covering all the loop resonance frequencies. When a vehicle goes over a loop, the corresponding inductance and resonance frequency will change. The shift in frequency or its effect in any/every loop can be simultaneously monitored, and the vehicles can be detected and identified as a bicycle, a motorcycle, a car, a bus, etc., based on the signature. Another advantage of this scheme is that the loops are in parallel resonance; hence, the power drawn from the source will be minimal. A prototype multiple-loop system has been built and tested based on the proposed scheme. The developed system detected, classified, and counted vehicles accurately. Moreover, the system also computes and provides the speed of the vehicle detected using a single set of multiple loops. The accuracy of the speed measurement has been compared with actual values and found to be accurate and can be used for real-time intelligent transportation system (ITS) applications under heterogeneous and less lane-disciplined (e.g., Indian) conditions.

Impact of Signal Timing Information on Safety and Efficiency of Signalized Intersections

Signalized intersections are provided in traffic networks to improve the safety and efficiency of vehicular and pedestrian movement. Various measures under education, enforcement, and engineering headings being attempted to improve the safety and efficiency of operations at signalized intersections. Provision of signal countdown timer, a timer showing the remaining red and green time in a phase, is one such measure and is commonly adopted in India. However, studies on the effects of a countdown timer under Indian traffic conditions are very scarce. Traffic heterogeneity and lack of lane discipline make the transferability of models developed in other countries (with more organized traffic) infeasible. The present study is an attempt to analyze the changes in queue-discharge characteristics and red-light violations (RLVs) under Indian traffic conditions due to the presence of a timer. A before-and-after analysis was carried out using the data collected from a selected intersection in Chennai, India. The analysis was carried out for different vehicle types in the presence and absence of timers separately for the start and end of red/green. Results showed that the information provided at the start of green (end of red) enhances efficiency, the start-up lost time is reduced, and there is an increase in RLVs. Two-wheelers present at the start of the queue are found to be the category that is mostly affected by this information. However, the information provided at the end of green (start of red) was found to reduce the incidence of RLVs. In the presence of information, it was found that the propensity of RLV (proportion of cycles having RLV) decreased from 59 to 31% at the end of green (start of red) and increased from 12 to 75% at the start of green (end of red) with a statistically significant drop in the headways (indicating an increased efficiency). Also, in the presence of information, the intensity of RLVs (mean RLVs per RLV cycle) for both the start of red and end of red decreased from 3.32 to 2.30 vehicles and 8.52 to 5.60 vehicles, respectively. The impacts varied based on vehicle type with major impacts on two-wheelers. The queue-discharge models show a significant change in trend, implying a need to update the signal timings when timers are installed. These results also bring to light the trade-off between safety and efficiency and the choices drivers make in the presence of phase-change information. These trade-offs should be carefully considered as the technology advances and drivers are provided with more and more information. For example, with the advent of intellidrive technology (vehicle-to-infrastructure communications), the extent of information provided to drivers should be tailored to achieve system optimality. The results of studies such as this one can help in decision making.

A multiple loop vehicle detection system for heterogeneous and lane-less traffic

This paper presents a novel inductive loop sensor which detects large (e.g., bus) as well as small (e.g., bicycle) vehicles and help a traffic control management system in optimizing the best use of existing roads. To accomplish the sensing of large as well as a small vehicle, a multiple loop inductive sensor system is proposed. The proposed sensor structure not only senses and segregates the vehicle type as bicycle or motor cycle or car or bus but also enables accurate counting of the number of vehicles that too in a mixed traffic flow condition. A prototype of the multiple loop sensing system has been developed using a virtual instrumentation scheme and tested. Field tests indicate that the prototype successfully detected all types of vehicles and counted, correctly, the number of each type of vehicles. Thus the suitability of the proposed multi loop sensor system for any type of traffic has been established.

A model based approach to predict stream travel time using public transit as probes

Travel time is one of the most preferred traffic information by a wide variety of travelers. Travel time information provided through variable message signs at the roadside could be viewed as a traffic management strategy designed to encourage drivers to take an alternate route. At the same time, it could also be viewed as a traveler information service designed to ensure that the driver has the best available information based on which they can make travel decisions. In an Intelligent Transportation Systems (ITS) context, both the Advanced Traveler Information Systems (ATIS) and the Advance Traffic Management Systems (ATMS) rely on accurate travel time prediction along arterials or freeways. In India, currently there is no permanent system of active test vehicles or license plate matching techniques to measure stream travel time in urban arterials. However, the public transit vehicles are being equipped with Global Positioning System (GPS) devices in major metropolitan cities of India for providing the bus arrival time information at bus stops. However, equipping private vehicles with GPS to enable the stream travel time measurement is difficult due to the requirement of public participation. The use of the GPS equipped buses as probe vehicles and estimating the stream travel time is a possible solution to this problem. The use of public transit as probes for travel time estimation offers advantages like frequent trips during peak hours, wide range network coverage, etc. However, the travel time characteristics of public transit buses are influenced by the transit characteristics like frequent acceleration, deceleration and stops due to bus stops besides their physical characteristics. Also, the sample size of public transit is less when compared to the total vehicle population. Thus mapping the bus travel time to stream travel time is a real challenge and this difficulty is more complex in traffic conditions like in India with its heterogeneity and lack of lane discipline. As a pilot study, a model based approach using the Kalman filtering technique to predict stream travel time from public transit is carried out in the present study. Since it is only a pilot study, only twowheeled vehicles have been considered as they constitute a major proportion in the study area. The prediction scheme is corroborated using field data collected by carrying GPS units in two-wheelers traveling along with the buses under consideration. The travel time estimates from the model were compared with the manually observed travel times and the results are encouraging.