Traffic incidents are non-recurring events that can cause a temporary reduction in roadway capacity. They have been recognized as a major contributor to traffic congestion on our national highway systems. To alleviate their impacts on capacity, automatic incident detection (AID) has been applied as an incident management strategy to reduce the total incident duration. AID relies on an algorithm to identify the occurrence of incidents by analyzing real-time traffic data collected from surveillance detectors. Significant research has been performed to develop AID algorithms for incident detection on freeways; however, similar research on major arterial streets remains largely at the initial stage of development and testing. This dissertation research aims to identify design strategies for the deployment of an Artificial Neural Network (ANN) based AID algorithm for major arterial streets. A section of the US-1 corridor in Miami-Dade County, Florida was coded in the CORSIM microscopic simulation model to generate data for both model calibration and validation. To better capture the relationship between the traffic data and the corresponding incident status, Discrete Wavelet Transform (DWT) and data normalization were applied to the simulated data. Multiple ANN models were then developed for different detector configurations, historical data usage, and the selection of traffic flow parameters. To assess the performance of different design alternatives, the model outputs were compared based on both detection rate (DR) and false alarm rate (FAR). The results show that the best models were able to achieve a high DR of between 90% and 95%, a mean time to detect (MTTD) of 55-85 seconds, and a FAR below 4%. The results also show that a detector configuration including only the mid-block and upstream detectors performs almost as well as one that also includes a downstream detector. In addition, DWT was found to be able to improve model performance, and the use of historical data from previous time cycles improved the detection rate. Speed was found to have the most significant impact on the detection rate, while volume was found to contribute the least. The results from this research provide useful insights on the design of AID for arterial street applications.

<p> The McMaster Incident Detection Algorithm <MacAlg> automatically detects incidents en the Burlington Skyway for the Burlington Freeway Traffic Managment System <FTXS>. This paper describes the calibration, testing and evaluation of functions of northbound stations 1 through 6. The testing and evaluation of the two weekly data sets is illustrated and discussed. Some of the resulting functions are recommended to the Burlingtion FTNS to evaluate how well the MacAlg detects incidents. This research compliments the work: of Persaud, Hall and Hall (1989), who are developing and testing the logic of the MacAlg. The results of this paper contribute information to the further development and testing of the MacAlg's logic. </p> / Thesis / Bachelor of Arts (BA)

This report describes an assessment of using dedicated short-range communications(DSRC) technology to perform travel time monitoring and automated incident detectionon a segment of rural freeway. The assessment used the CORSIM traffic simulation toolto simulate traffic and incidents on a segment of rural freeway. Output data from thesimulation was subjected to post-processing to produce the "probe and beacon" data thatwould be produced by a DSRC-based system. An incident detection algorithm wasdeveloped, which used a travel time threshold and a counter. Travel times exceeding thethreshold incremented the counter, while travel times below the threshold decrementedthe counter (unless it was at zero). An alarm was generated when the counter reached apre-selected level. This algorithm was tested on selected data files, and the results wereused to identify the "best" values of the threshold and counter alarm level. Using these"best" values, the algorithm was then applied to the "probe and beacon" data todetermine how quickly the system could detect various traffic incidents. The analysisshowed that the system could provide rapid and reliable detection of incidents.During the simulation and analysis, several parameters were varied to observe theirimpacts on the system performance. These parameters included traffic volume, incidentseverity, percentage of vehicles with transponders, spacing of roadside readers, andlocation of the incident relative to the next downstream reader. Each parameter proved tohave a significant effect on the detection time, and the observed impacts were consistentwith logical expectations. In general, the time to detect an incident was reduced inresponse to (1) an increase in traffic volume, (2) an increase in incident severity, (3) anincrease in transponder population, (4) a reduction in reader spacing, and (5) a reductionin distance from incident location to next downstream reader.Preliminary estimates were developed of the costs associated with implementing aDSRC-based traffic monitoring system. The relationship between system cost andsystem performance was explored and illustrated.Recommendations were developed and presented. These included further analysis basedon traffic simulations, followed by a limited field deployment to validate the analysisresults.

This thesis presents a new testbed for Automatic Incident Detection (AID) systems that uses real-time traffic video and data feeds from the Ministry of Transportation, Ontario (MTO) COMPASS Advanced Traffic Management System (ATMS). This new testbed, termed the AID Comparison and Analysis Tool (AID CAAT), consists largely of a data warehouse storing a significant amount of traffic video, the corresponding traffic data and an accurate log of incident start/end times. An evaluation was conducted whereby the AID CAAT was used to calibrate, and then analyze the performance of four AID systems: California Algorithm 8, McMaster Algorithm, the Genetic Adaptive Incident Detection (GAID) Algorithm and the Citilog - VisioPAD. The traditional measures of effectiveness (MOE) were initially used for this evaluation: detection rate (DR), false alarm rate (FAR), and mean time to detection (MTTD). However, an in-depth analysis of the test results (facilitated by the AID CAAT) revealed the need for two additional MOEs: False Normal Rate and Nuisance Rate. The justification and sample calculations for these new MOEs are also provided. This evaluation shows the considerable advantages of the AID CAAT, and also suggests the strengths and weaknesses of the AID systems tested. / Thesis / Master of Applied Science (MASc)

Automatic Incident Detection Algorithms (AIDA) have been part of freeway management system software from the beginnings of ITS deployment. These algorithms introduce the capability of detecting incidents on freeways using traffic operations data. Over the years, several approaches to incident detection have been studied and tested. However, the size and scope of the urban transportation networks under direct monitoring by transportation management centers are growing at a faster rate than are staffing levels and center resources. This has entailed a renewed emphasis on the need for reliability and accuracy of AIDA functionality. This study investigates a new approach to incident detection that promises a significant improvement in operational performance. This algorithm is formulated on the premise that the current conditions facilitate the prediction of future traffic conditions, and deviations of observations from the predictions beyond a calibrated level of tolerance indicate the occurrence of incidents. This algorithm is specifically designed for easy implementation and calibration at any site. Offline tests with data from the Georgia-Navigator system indicate that this algorithm realizes a substantial improvement over the conventional incident detection algorithms. This algorithm not only achieves a low rate of false alarms but also ensures a high detection rate.

Recurrent congestion in urban traffic networks, especially on arterials, is a growing problem. Non-recurrent congestion, mainly due to incidents, only aggravates the problem. Any solution requires monitoring of the network, for which many developing countries, such as Turkey, do not have the traditional surveillance systems on arterials mainly due to high costs. An alternative solution is the utilization of Global Positioning System (GPS) technology, which is increasingly used in traffic monitoring. It is easy and cheap to obtain the GPS track information,even in real-time, from a probe-vehicle or a fleet of vehicles / and spatial variation of speed and travel time of the vehicle(s) in a network can be determined. GPS-based data, especially with only one probe-vehicle, would not provide information on the concurrent states of upstream and downstream traffic, needed to define the state of traffic in a network. To overcome this obstacle, a methodology based on statistical analysis of archival traffic conditions obtained through different sources is proposed to analyze traffic fluctuations and identify daily traffic pattern. As a result, bottleneck and resulting queues can be detected on a corridor. Thus, it enables detection of recurrent congestion and queues that may result from incidents. The proposed methodology is tested on a corridor the roadway between METU and Kizilay of in&ouml / n&uuml / Boulevard. The results show that the methodology can effectively identify bottleneck locations on the corridor and also an incident observed during the data collection is detected correctly by the proposed algorithm.

The early observation and elimination of non-recurring incidents is a crucial task in trafficmanagement. The performance of the conventional incident detection methods (trafficcameras and other sensory technologies) is limited and there are still challenges inobtaining an accurate picture of the traffic conditions in real time. During the last decade,the technical development of mobile platforms and the growing online connectivity made itpossible to obtain traffic information from social media and applications based on spatialcrowdsourcing. Utilizing the benefits of crowdsourcing, traffic authorities can receiveinformation about a more comprehensive number of incidents and can monitor areaswhich are not covered by the conventional incident detection systems. The crowdsourcedtraffic data can provide supplementary information for incidents already reported throughother sources and it can contribute to earlier detection of incidents, which can lead tofaster response and clearance time. Furthermore, spatial crowdsourcing can help to detectincident types, which are not collected systematically yet (e.g. potholes, traffic light faults,missing road signs). However, before exploiting crowdsourced traffic data in trafficmanagement, numerous challenges need to be resolved, such as verification of the incidentreports, predicting the severity of the crowdsourced incidents and integration with trafficdata obtained from other sources.During this thesis, the possibilities and challenges of utilizing spatial crowdsourcingtechnologies to detect non-recurring incidents were examined in form of a case study.Traffic incident alerts obtained from Waze, a navigation application using the concept ofcrowdsourcing, were analyzed and compared with officially verified incident reports inStockholm. The thesis provides insight into the spatial and temporal characteristics of theWaze data. Moreover, a method to identify related Waze alerts and to determine matchingincident reports from different sources is presented. The results showed that the number ofreported incidents in Waze is 4,5 times higher than the number of registered incidents bythe Swedish authorities. Furthermore, 27,5 % of the incidents could have been detectedfaster by using the traffic alerts from Waze. In addition, the severity of Waze alerts isexamined depending on the attributes of the alerts.

The early observation and elimination of non-recurring incidents is a crucial task in traffic management. The performance of the conventional incident detection methods (traffic cameras and other sensory technologies) is limited and there are still challenges in obtaining an accurate picture of the traffic conditions in real time. During the last decade, the technical development of mobile platforms and the growing online connectivity made it possible to obtain traffic information from social media and applications based on spatial crowdsourcing. Utilizing the benefits of crowdsourcing, traffic authorities can receive information about a more comprehensive number of incidents and can monitor areas which are not covered by the conventional incident detection systems. The crowdsourced traffic data can provide supplementary information for incidents already reported through other sources and it can contribute to earlier detection of incidents, which can lead to faster response and clearance time. Furthermore, spatial crowdsourcing can help to detect incident types, which are not collected systematically yet (e.g. potholes, traffic light faults, missing road signs). However, before exploiting crowdsourced traffic data in traffic management, numerous challenges need to be resolved, such as verification of the incident reports, predicting the severity of the crowdsourced incidents and integration with traffic data obtained from other sources. During this thesis, the possibilities and challenges of utilizing spatial crowdsourcing technologies to detect non-recurring incidents were examined in form of a case study. Traffic incident alerts obtained from Waze, a navigation application using the concept of crowdsourcing, were analyzed and compared with officially verified incident reports in Stockholm. The thesis provides insight into the spatial and temporal characteristics of the Waze data. Moreover, a method to identify related Waze alerts and to determine matching incident reports from different sources is presented. The results showed that the number of reported incidents in Waze is 4,5 times higher than the number of registered incidents by the Swedish authorities. Furthermore, 27,5 % of the incidents could have been detected faster by using the traffic alerts from Waze. In addition, the severity of Waze alerts is examined depending on the attributes of the alerts.

The McMaster incident detection algorithm is being tested on-line within the Burlington freeway traffic management system (FTMS) as an alternative to the existing California-type algorithm currently in place. This paper represents the most recent and comprehensive evaluation of the McMaster algorithm's performance to date. In the past, the algorithm has been tested using single lane detectors for the northbound lanes only. This evaluation uses data from lanes 1 and 2 for each of the 13 northbound and 13 southbound detector stations. The data was collected during a 60-day period beginning on November 15, 1990 and ending January 13, 1991. Detection rate, mean time-lag to detection and false alarm rate are used to evaluate the performance of the algorithm. As well, those factors such as winter precipitation, which influenced the performance of the algorithm are also examined. To improve the algorithm's detection rate and lower its false alarm rate, it is reccomended that the persistence check used to declare an incident be increased by 30-seconds from 2 to 3 periods. / Thesis / Candidate in Philosophy

The decision to select the best Intelligent Transportation System (ITS) technologies from available options has always been a challenging task. The availability of connected vehicle/automated vehicle (CV/AV) technologies in the near future is expected to add to the complexity of the ITS investment decision-making process. The goal of this research is to develop a multi-criteria decision-making analysis (MCDA) framework to support traffic agencies’ decision-making process with consideration of CV/AV technologies. The decision to select between technology alternatives is based on identified performance measures and criteria, and constraints associated with each technology. Methods inspired by the literature were developed for incident/bottleneck detection and back-of-queue (BOQ) estimation and warning based on connected vehicle (CV) technologies. The mobility benefits of incident/bottleneck detection with different technologies were assessed using microscopic simulation. The performance of technology alternatives was assessed using simulated CV and traffic detector data in a microscopic simulation environment to be used in the proposed MCDA method for the purpose of alternative selection. In addition to assessing performance measures, there are a number of constraints and risks that need to be assessed in the alternative selection process. Traditional alternative analyses based on deterministic return on investment analysis are unable to capture the risks and uncertainties associated with the investment problem. This research utilizes a combination of a stochastic return on investment and a multi-criteria decision analysis method referred to as the Analytical Hierarchy Process (AHP) to select between ITS deployment alternatives considering emerging technologies. The approach is applied to an ITS investment case study to support freeway bottleneck management. The results of this dissertation indicate that utilizing CV data for freeway segments is significantly more cost-effective than using point detectors in detecting incidents and providing travel time estimates one year after CV technology becomes mandatory for all new vehicles and for corridors with moderate to heavy traffic. However, for corridors with light, there is a probability of CV deployment not being effective in the first few years due to low measurement reliability of travel times and high latency of incident detection, associated with smaller sample sizes of the collected data.