As one of the kernel components of ITS technology, Travel Time Estimation (TTE) has been a high-interest topic in highway operation and management for years. Out of numerous vehicle detection technologies being applied in this project, intrusive loop detector, as the representative of spot measurement devices, is the most common. The ultimate goal of this dissertation is to seek a TTE approach based primarily on spot speed measurement and capable of successfully performing in a certain accuracy range under various traffic conditions. The provision of real-time traffic information could offer significant benefits for commuters looking to make optimum travel decisions. The proposed research effort attempts to characterize typical variability in traffic conditions using traffic volume data obtained from loop detectors on I-66 Virginia during a 3-month period. The detectors logged time-mean speed, volume, and occupancy measurements for each station and lane combination. Using these data, the study examines the spatiotemporal link and path flow variability of weekdays and weekends. The generation of path flows is made through the use of a synthetic maximum likelihood approach. Statistical Analysis of Variance (ANOVA) tests are performed on the data. The results demonstrate that in terms of link flows and total traffic demand, Mondays and Fridays are similar to core weekdays (Tuesdays, Wednesdays, and Thursdays). In terms of path flows, Fridays appear to be different from core weekdays. A common procedure for estimating roadway travel times is to use either queuing theory or shockwave analysis procedures. However, a number of studies have claimed that deterministic queuing theory and shock-wave analysis are fundamentally different, producing different delay estimates for solving bottleneck problems. Chapter 5 demonstrates the consistency in the delay estimates that are derived from both queuing theory and shock-wave analysis and highlights the common errors that are made in the literature with regards to shock-wave analysis delay estimation. Furthermore, Chapter 5 demonstrates that the area between the demand and capacity curves can represent the total delay or the total vehicle-hours of travel if the two curves are spatially offset and queuing theory has its advantages on this because of its simplicity. As the established relationship between time-mean and space-mean speed is suitable for estimating time-mean speeds from space-mean speeds in most cases, it is also desired to estimate the space-mean speeds from time-mean speeds. Consequently, Chapter 6 develops a new formulation that utilizes the variance of the time-mean speed as opposed to the variance of the space-mean speed for the estimation of space-mean speeds. This demonstrates that the space-mean speeds are estimated within a margin of error of 0 to 1 percent. Furthermore, it develops a relationship between the space- and time-mean speed variance and between the space-mean speed and the spatial travel-time variance. In addition, the paper demonstrates that both the Hall and Persaud and the Dailey formulations for estimating traffic stream speed from single loop detectors are valid. However, the differences in the derivations are attributed to the fact that the Hall and Persaud formulation computes the space-mean speed (harmonic mean) while the Dailey formulation computes the time-mean speed (arithmetic mean). Chapter 7 focuses on freeway Travel Time Estimation (TTE) algorithms that are based on spot speed measurements. Several TTE approaches are introduced including a traffic dynamics TTE algorithm that is documented in literature. This traffic dynamics algorithm is analyzed, highlighting some of its drawbacks, followed by some proposed corrections to the traffic dynamics formulation. The proposed approach estimates traffic stream density from occupancy measurements, as opposed to flow measurements, at the onset of congestion. Next, the study validates the proposed model using field data from I-880 and simulated data. Comparison of five different TTE algorithms is conducted. The comparison demonstrates that the proposed approach is superior to the TTE traffic dynamics approach. Particularly, a multi-link simulation network is built to test spot-speed-measurement TTE performance on multi links, as well as the data smoothing technique's effect on TTE accuracy. Findings further prove advantages of utilizing space-mean speed in TTE rather than time-mean speed. In summary, a feasible TTE procedure that is adaptive to various traffic conditions has been established. Since each approach would under-/over-estimate travel time depending on the concrete traffic condition, different models will be selected to ensure TTE's accuracy window. This approach has broad applications because it is based on popular loop detectors. / Ph. D.

Travel time information is important for Advanced Traveler Information Systems (ATIS) applications. People traveling on urban freeways are interested in knowing how long it will take them to reach their destinations, particularly under congested conditions. Though many advances have been made in the field of traffic engineering and ITS applications, there is a lack of practical travel time estimation procedures for ATIS applications. Automatic Vehicle Identification (AVI) and Geographic Information System (GPS) technologies can be used to directly estimate travel times, but they are not yet economically viable and not widely deployed in urban areas. Hence, data from loop detectors or other point estimators of traffic flow variables are predominantly used for travel time estimation. Most point detectors can provide this data efficiently. Some attempts have been made in the past to estimate travel times from point estimates of traffic variables, but they are not comprehensive and are valid for only particular cases of freeway conditions. Moreover, most of these methods are statistical and thus limited to the type of situations for which they were developed and are not of much general use. The purpose of current research is to develop a comprehensive system for travel time estimation on urban freeways for ATIS applications. The system is based on point estimates of traffic variables obtained from detectors. The output required from the detectors is flow and occupancy aggregated for a short time interval of 5 minutes. The system for travel time estimation is based on the traffic flow theory rather than statistical methods. The travel times calculated using this system are compared with the results of FHWA simulation package TSIS 5.0 and the estimation system is found to give reasonable and comparable results when compared with TSIS results. / Master of Science

Emergency services has a vital function in society, and except saving lifes a functioning emergency service system provides the inhabitants of any give society with a sence of feeling secure. Because of the delicate nature of the services provided there is always an interest in improvement with regards to the performance of the system. In order to have a good system there are a variety of models that can be used as decision making support. An important component in many of these models are the travel time of an emergency vehicle. In In this study the focus lies in travel time estimation for the emergency services and how it could be estimated by using a neural network, called a deep learning process in this report. The data used in the report is map matched GPS points that have been collected by the emergency services in two counties in Sweden, Östergötland and Västergötland. The map matched data has then been matched with NVDB, which is the the national road database, adding an extra layer of information, such as roadlink geometry, number of roundabouts etc. To find the most important features to use as input in the developed model a Pearson and Spearman correlation test was performed. Even if these two tests do not capture all possible relations between features they still give an indication of what features that can be included. The deep learning process developed within this study uses route length, average weighted speed limit, resource category, and road width. It is trained with 75% of the data leaving the remaining 25% for testing of the model. The DLP gives a mean absolute error of 51.39 when trained and 59.21 seconds when presented with new data. This in comparison a simpler model which calculates the travel time by dividing the route length with the weighted averag speed limt, which gives a mean absolute error of 227.48 seconds. According to the error metrics used in order to evaluate the models the DLP performs better than the current model. However there is a dimension of complexity with the DLP which makes it sort of a black box where something goes in and out comes an estimated travel time. If the aim is to have a more comprehensive model, then the current model has its benefits over a DLP. However the potential that lies in using a DLP is entruiging, and with a more in depth analysis of features and how to classify these in combination with more data there may be room for developing more complex DLPs.

Travel time is one of the most important inputs in many Intelligent Transport Systems (ITS). As a result, this information needs to be accurate and dynamic in both spatial and temporal dimensions. For the estimation of travel time, data from fixed sensors such as Inductive Loop Detectors (ILD) and cameras have been widely used since the 1960 s. However, data from fixed sensors may not be sufficiently reliable to estimate travel time due to a combination of limited coverage and low quality data resulting from the high cost of implementing and operating these systems. Such issues are particularly critical in the context of Less Developed Countries, where traffic levels and associated problems are increasing even more rapidly than in Europe and North America, and where there are no pre-existing traffic monitoring systems in place. As a consequence, recent developments have focused on utilising moving sensors (i.e. probe vehicles and/or people equipped with GPS: for instance, navigation and route guidance devices, mobile phones and smartphones) to provide accurate speed, positioning and timing data to estimate travel time. However, data from GPS also have errors, especially for positioning fixes in urban areas. Therefore, map-matching techniques are generally applied to match raw positioning data onto the correct road segments so as to reliably estimate link travel time. This is challenging because most current map-matching methods are suitable for high frequency GPS positioning data (e.g. data with 1 second interval) and may not be appropriate for low frequency data (e.g. data with 30 or 60 second intervals). Yet, many moving sensors only retain low frequency data so as to reduce the cost of data storage and transmission. The accuracy of travel time estimation using data from moving sensors also depends on a range of other factors, for instance vehicle fleet sample size (i.e. proportion of vehicles equipped with GPS); coverage of links (i.e. proportion of links on which GPS-equipped vehicles travel); GPS data sampling frequency (e.g. 3, 6, 30, 60 seconds) and time window length (e.g. 5, 10 and 15 minutes). Existing methods of estimating travel time from GPS data are not capable of simultaneously taking into account the issues related to uncertainties associated with GPS and spatial road network data; low sampling frequency; low density vehicle coverage on some roads on the network; time window length; and vehicle fleet sample size. Accordingly this research is based on the development and application of a methodology which uses GPS data to reliably estimate travel time in real-time while considering the factors including vehicle fleet sample size, data sampling frequency and time window length in the estimation process. Specifically, the purpose of this thesis was to first determine the accurate location of a vehicle travelling on a road link by applying a map-matching algorithm at a range of sampling frequencies to reduce the potential errors associated with GPS and digital road maps, for example where vehicles are sometimes assigned to the wrong road links. Secondly, four different methods have been developed to estimate link travel time based on map-matched GPS positions and speed data from low frequency data sets in three time windows lengths (i.e. 5, 10 and 15 minutes). These are based on vehicle speeds, speed limits, link distances and average speeds; initially only within the given link but subsequently in the adjacent links too. More specifically, the final method draws on weighted link travel times associated with the given and adjacent links in both spatial and temporal dimensions to estimate link travel time for the given link. GPS data from Interstate I-880 (California, USA) for a total of 73 vehicles over 6 hours were obtained from the UC-Berkeley s Mobile Century Project. The original GPS dataset which was broadcast on a 3 second sampling frequency has been extracted at different sampling frequencies such as 6, 30, 60 and 120 seconds so as to evaluate the performance of each travel time estimation method at low sampling frequencies. The results were then validated against reference travel time data collected from 4,126 vehicles by high resolution video cameras, and these indicate that factors such as vehicle sample size, data sampling frequency, vehicle coverage on the links and time window length all influence the accuracy of link travel time estimation.

Road freight transport travel time estimation is an important task in fleet management and traffic planning. Goods often must be delivered in a predefined time window and any deviation may lead to serious consequences. It is possible to improve travel time estimation by considering more factors that may affect it. In this thesis work we identify factors that may affect travel time, find possible sources of information about them, propose a model for estimating travel time of heavy goods vehicles, and verify this model on real data. As results, the experiments showed that considering time related and weather related factors, it is possible to improve accuracy of travel time estimation. Also, it was shown that the influence of a particular factor on travel time depended on the considered road segment. Furthermore, it was shown that different data mining algorithms should be applied for different road segments in order to get the best estimation.

Providing transportation system operators and travelers with accurate travel time information allows them to make more informed decisions, yielding benefits for individual travelers and for the entire transportation system. Most existing advanced traveler information systems (ATIS) and advanced traffic management systems (ATMS) use instantaneous travel time values estimated based on the current measurements, assuming that traffic conditions remain constant in the near future. For more effective applications, it has been proposed that ATIS and ATMS should use travel times predicted for short-term future conditions rather than instantaneous travel times measured or estimated for current conditions. This dissertation research investigates short-term freeway travel time prediction using Dynamic Neural Networks (DNN) based on traffic detector data collected by radar traffic detectors installed along a freeway corridor. DNN comprises a class of neural networks that are particularly suitable for predicting variables like travel time, but has not been adequately investigated for this purpose. Before this investigation, it was necessary to identifying methods for data imputation to account for missing data usually encountered when collecting data using traffic detectors. It was also necessary to identify a method to estimate the travel time on the freeway corridor based on data collected using point traffic detectors. A new travel time estimation method referred to as the Piecewise Constant Acceleration Based (PCAB) method was developed and compared with other methods reported in the literatures. The results show that one of the simple travel time estimation methods (the average speed method) can work as well as the PCAB method, and both of them out-perform other methods. This study also compared the travel time prediction performance of three different DNN topologies with different memory setups. The results show that one DNN topology (the time-delay neural networks) out-performs the other two DNN topologies for the investigated prediction problem. This topology also performs slightly better than the simple multilayer perceptron (MLP) neural network topology that has been used in a number of previous studies for travel time prediction.

The accurate and reliable estimation of travel time based on point detector data is needed to support Intelligent Transportation System (ITS) applications. It has been found that the quality of travel time estimation is a function of the method used in the estimation and varies for different traffic conditions. In this study, two hybrid on-line travel time estimation models, and their corresponding off-line methods, were developed to achieve better estimation performance under various traffic conditions, including recurrent congestion and incidents. The first model combines the Mid-Point method, which is a speed-based method, with a traffic flow-based method. The second model integrates two speed-based methods: the Mid-Point method and the Minimum Speed method. In both models, the switch between travel time estimation methods is based on the congestion level and queue status automatically identified by clustering analysis. During incident conditions with rapidly changing queue lengths, shock wave analysis-based refinements are applied for on-line estimation to capture the fast queue propagation and recovery. Travel time estimates obtained from existing speed-based methods, traffic flow-based methods, and the models developed were tested using both simulation and real-world data. The results indicate that all tested methods performed at an acceptable level during periods of low congestion. However, their performances vary with an increase in congestion. Comparisons with other estimation methods also show that the developed hybrid models perform well in all cases. Further comparisons between the on-line and off-line travel time estimation methods reveal that off-line methods perform significantly better only during fast-changing congested conditions, such as during incidents. The impacts of major influential factors on the performance of travel time estimation, including data preprocessing procedures, detector errors, detector spacing, frequency of travel time updates to traveler information devices, travel time link length, and posted travel time range, were investigated in this study. The results show that these factors have more significant impacts on the estimation accuracy and reliability under congested conditions than during uncongested conditions. For the incident conditions, the estimation quality improves with the use of a short rolling period for data smoothing, more accurate detector data, and frequent travel time updates.

No description available.

Travel time estimation can be used in strategical distribution of ambulances and ambulance stations. A more accurate travel time estimation can lead to a better distribution of these ambulance sites. External factors such as weather and traffic conditions can affect the travel time from a starting location to a destination. In this work, we investigate how the SOS Alarm dataset of ambulance trips data and the machine learning model Gradient Boosted Decision Trees can be used to estimate travel time, and how these estimationscan be improved by incorporating aforementioned conditions when predicting travel time. Results showed that reasonable performance can be achieved for a subset of data where the precise origin and destination is known compared to a subset where the precise origin is unknown, and that traffic conditions could improve model performance on a subset of data containing trips only for a single route. Including weather represented as individual weather parameters did not, however, lead to enhanced performance.

Traveler information systems play a significant role in most travelers’ daily trips. These systems assist travelers in choosing the best routes to reach their destinations and possibly select suitable departure times and modes for their trips. Connected Vehicle (CV) technologies are now in the pilot program stage. Vehicle-to-Infrastructure (V2I) communications will be an important source of data for traffic agencies. If this data is processed properly, then agencies will be able to better determine traffic conditions, allowing them to take proper countermeasures to remedy transportation system problems under different conditions. This research focuses on developing methods to assess the potential of utilizing CV data to support the traveler information system data collection process. The results from the assessment can be used to establish a timeline indicating when an agency can stop investing, at least partially, in traditional technologies, and instead rely on CV technologies for traveler information system support. This research utilizes real-world vehicle trajectory data collected under the Next Generation Simulation (NGSIM) program and simulation modeling to emulate the use of connected vehicle data to support the traveler information system. NGSIM datasets collected from an arterial segment and a freeway segment are used in this research. Microscopic simulation modeling is also used to generate required trajectory data, allowing further analysis, which is not possible using NGSIM data. The first step is to predict the market penetration of connected vehicles in future years. This estimated market penetration is then used for the evaluation of the effectiveness of CV-based data for travel time and volume estimation, which are two important inputs for the traveler information system. The travel times are estimated at different market penetrations of CV. The quality of the estimation is assessed by investigating the accuracy and reliability with different CV deployment scenarios. The quality of volume estimates is also assessed using the same data with different future scenarios of CV deployment and partial or no detector data. Such assessment supports the identification of a timeline indicating when CV data can be used to support the traveler information system.