ABSTRACT. We consider a situation where agents strategically decide which congested system to use. For example, commuters who choose a road or customers who decide which service system to seek service from. We use a model of parallel M/M/1 queues where strategic customers, who are interested in maximizing their own welfare, decide which queue to join without observing the system's state upon arrival. It is well known that due to externalities, selfish customers tend to overload the attractive servers, which dramatically affects social welfare. We show that when the servers are strategic and interested in maximizing their own revenue, there is a decentralized remedy. By letting the servers charge for priority while competing with each other, the equilibrium priority prices are such that the resulting customer-equilibrium routing strategy is socially optimal.

ABSTRACT. Autonomous vehicle (AV) technologies are rapidly evolving with the vision of having self-driving cars moving safely with no human input. However, it is clear that at least in the near and foreseeable future, AVs will not be able to resolve all road incidents and that in some situations remote human assistance will be required. However, remote driving is not trivial and introduces many challenges stemming mostly from the physical disconnect of the remote operator. In order to highlight these challenges and understand how to better design AV teleoperation interfaces, we conducted several observations of AV teleoperation sessions as well as in-depth interviews with 14 experts. Based on these interviews, we provide an investigation and analysis of the major AV teleoperation challenges. We follow this by providing design guidelines for the development of future teleoperation interfaces for assistance and driving of AVs.

ABSTRACT. In recent years there has been significant progress in the development of autonomous vehicles. However, much of the research revolves around safety impacts and their effect on traffic conditions in an urban environment. In this research, we examine the likely impact of autonomous vehicles on public transport service in the 300 to 700 km medium-haul market, which was assigned by the EU in 2010 as one of the targets for vehicle trip reductions through investments in transport infrastructure. Current data reveals that the private car remains a significant player in the medium-haul due to flexibility and door-to-door accessibility despite the additional time required and fatigue involved. However, the autonomous vehicle reduces these disadvantages, creating a robust alternative to the private car and increasing competition in the public transport market. To investigate the possible impact of the autonomous vehicle in the medium-haul market, we develop an applied game-theoretic formulation that analyses the potential transport equilibria outcome accounting for four profit-maximizing competitors, including high-speed railway, airlines and autonomous vehicle companies, offering single passenger or ride-sharing options. The supply-side decision variables for a high-speed railway company and airline include fares and frequencies, whilst the autonomous car company’s decision variables are fares and fleet size. We model the demand-side using a logit market share function that accounts for the transport providers' decisions, the passengers’ value of time characteristics and origin-destination paths. We apply the formulation to a case study of the Madrid-Barcelona corridor in Spain, using anonymized mobile phone records which are combined with data from the Google Maps API in order to calibrate the model. Preliminary results suggest that autonomous vehicles will directly impact high-speed railway fares and airfares, potentially undoing any gains achieved since the publication of the 2010 EU White Paper.

ABSTRACT. We present a model for policy optimization of the online booking of mobile personnel over a multiday horizon with a different cutoff for each day, where the goal is to maximize the expected ratio of accepted requests at steady-state. This model fits the practice of many service providers who allow booking of time slots over a horizon of multiple days and use availability control of the demand. Since the planning horizon is indefinite and the service horizon of each day overlaps the horizon of subsequent days, the objective is defined in terms of the steady-state performance. The interactions with the customers are performed in a single step: The system offers an assortment of time slots covering the next few days, and the user either chooses one of them or abandons the system. Upon the arrival of a service request, the provider estimates the opportunity cost of serving the request at each of the available time slots. We model this cost as a linear function and a Cobb–Douglas function of features that concisely represent the current system state. The assortment of time slots following each service request is constructed by maximizing the expected net gain from the assortment. The parameters of the opportunity cost functions are fitted using a simulation framework. The proposed method is benchmarked based on randomly generated datasets in various demand scenarios and geographies. The method is shown to outperform more straightforward baseline policies significantly.

ABSTRACT. Over the past decade, a large body of research has developed on walking and walkability, in part in response to increasing concerns over people’s health, climate change, livability, and social cohesion. Literature shows that some walkability indicators have a greater influence on walkability rates than others. These indicators include density, street network connectivity, land use mix, and open spaces. On the other hand, the literature identifies additional variables that may affect walking patterns. These variables refer to walking indicators related to people and their sociodemographic characteristics. Such variables include age, gender, household income, and average education level. It is not clear what is the exact nature of the relationship between neighborhood walkability and walking trips since logically a higher walking frequency can be triggered in different ways. A range of methods has been used to measure walkability and, subsequently, to analyze the relationship between walkability and walking. Most studies that analyze this relationship rely on calculating walkability scores using existing walkability index. Another common method is using basic regression analysis that is more suitable to small scale uncomplex data sets. This research aims to add to the literature by applying a big data approach and to assess to what extent big data approaches can successfully predict walking levels based on large data sets. Also, to identify the set of (walkability) parameters that most strongly shape walking in urban areas. The data used is part of the Tel-Aviv Metropolitan travel survey, enriched by GPS tracking and app-based reporting. The combination of some of the original parameters that were delivered by the data, and new parameters that were created using data mining methods allows a broad range of big data analyses, which may assist to reach this research`s goals. This analysis is done using the Random Forest classification algorithm in machine learning.

ABSTRACT. In recent years, lack of parking spaces has become a major concern in dense cities. To overcome this issue, several companies are offering Automated Parking Systems (APS), which enable higher land resources utilization then in conventional parking lots. In an APS, users park their cars in designated entry/exit rooms and later on pick them up from these rooms. The movement of the cars to/from parking spots is performed automatically by robots. This study focuses on a particular APS design in which cars are carried by shuttles, moving on a single horizontal rail in each floor and moving vertically using two elevators located at both ends of the rails, referred to as Double-Elevator Shuttle based APS (DES-APS). The operation of DES-APS introduces several challenges, including, parking spot allocation, job allocation to shuttles and shuttle routing, which consider the electro-mechanics limitation of the system. We define the DES-APS operational problem and propose a heuristic framework to solve it. This work focuses on a fundamental component of the framework, the scheduling of shuttle operations. Given that the routing and sequencing decisions have already been taken, a schedule which consider shuttle conflicts, floors’ shuttle capacity and drivers maximal waiting time is needed. We represent the scheduling sub problem by an alternative graph and propose a heuristic search procedure for the optimal schedule that utilizes information from the alternative graph. We then developed a MILP model, which used as a benchmark for small instances, to validate the suggested heuristic. The scheduling sub-problem heuristic will later be used as a subroutine in a LNS framework designed to search for promising solutions of the DES-APS operational problem.

ABSTRACT. Autonomous vehicles have been rapidly progressing towards full autonomy using fixed driving styles, which may differ from individual passenger preferences. Violating these preferences may lead to passenger discomfort or anxiety. Despite significant and rapid technological advancements, autonomous vehicles (AVs) face a crisis of trust and people are hesitant to ride in them: according to a 2019 AAA survey, 60–80% of drivers interviewed, expressed their reluctance to ride in a fully AV. trust, safety, and control issues are likely the underlying concerns behind their reluctance to ride in these vehicles. A technology that pretends to take the user’s feelings into account may promote a sense of comfort in them, which may contribute to the universal acceptance of such technology. in the first part of this study, We studied data from previous experiments, where passenger responses to different driving situations in a physical autonomous vehicle were recorded. Galvanic skin response (GSR) and heart rate (HR), which collected from 20 participants. Our results show how the magnitude of the driving kinematic features such as velocity, acceleration, and curvature can be used to predict the passenger’s GSR and HR responses, which are related to stress and anxiety. in the second part of this study, we designed an autonomous driving algorithm that performs driving tasks, while considering the passenger’s stress responses using deep reinforcement learning. using predictions of passengers’ GSR and HR our algorithm is punished when performing maneuvers that raise the stress of the passengers, and get rewarded while maintaining a stable and comfortable driving experience. this Autonomous driving is currently testing on an AV simulator.

ABSTRACT. מחקרים שונים הראו שנהגים סובלים מבעיות בריאות ועקה (דחק; מתח), בעקבות משתנים שונים בלתי נשלטים, כאשר נהגים מקצועיים (כגון נהגי אוטובוס, תובלה, מוניות וכו') הנוהגים במצטבר מעל 37,000 שעות במהלך חייהם, חשופים באופן רב יותר לבעיה. על מנת להימנע מגורמי הדחק, שתוצאותיהם בעיות בריאות, אנו חוקרים את המאפיינים המשפיעים על הנהג באמצעות מודל למידת מכונה. המטרה היא לבדוק האם יש קשר בין המדדים הפיזיולוגיים של הנהג לבין מאפיינים במסלול הנסיעה, כגון: משך הנסיעה, מספר התחנות בקו, כמות הנוסעים, עומס התנועה ועוד. לצורך איסוף הנתונים הפיזיולוגיים, נעשה שימוש בהתקן לביש מדגם E4 של היצרן האיטלקי Empatica. נאספו הנתונים הבאים: דופק, טמפרטורת עור, מוליכות אלקטרו-עורית (EDA), תנועות יד ב-3 צירים. בנוסף, התקבלו ממפעיל התחבורה הציבורית המידע הבא: נתוני מסלול הנסיעה, סוג כלי הרכב, מספר נוסעים עולים בכל תחנה, זמני נסיעה בפועל, מיקום גאוגרפי רציף ועוד. הנתונים נאספו במשך כשנה תוך דגש על מגוון משתתפים מכלל אוכלוסיית נהגי האוטובוסים. בנוסף, שולב אמצעי תיעוד ויזואלי מסוג מצלמת דרך מתקדמת, שבאמצעותה תועדה סביבת הרכב החיצונית במהלך הנסיעה של הנבדק. הנתונים נותחו באמצעות כלי למידת מכונה לבחינת הקשרים בין המדדים הפיזיולוגיים למאפייני הנסיעה. לצורך הניתוח השתמשנו בכלי כריית מידע באמצעות למידת מכונה. תוצאות ראשונות של הניתוח באמצעות Orange data mining softwareהראו: 1) ניתן לזהות באמצעות מודל Random Forest את הנהג ברמת דיוק של כ-84% על בסיס הנתונים הפיזיולוגיים, 2) לזהות באמצעות מודל Random Forest האם הנהג נמצא בתחנה, נמצא בנסיעה או בהפסקה ברמת דיוק של כ-97% על בסיס הנתונים הפיזיולוגיים, 3) באמצעות מודל Multi-Linear Regression לזהות ברמת דיוק של R^2=0.89 כמה זמן הנהג נמצא בנסיעה.

ABSTRACT. מרבית היוממים נוסעים בגפם וגורמים לחוסר יעילות תחבורתית וכלכלית כתוצאה מגודש, עודף נסועה ואובדן תוצר. אלה בתורם יוצרים זיהום, מטרדים ופגיעה במשאבים. שיתוף בין מספר יוממים שחולקים נסיעה יכול להוות כלי חשוב בסל של פתרונות מקיימים להתמודדות עם הבעיה, שכן הוא מפחית בצורה משמעותית את כמות כלי הרכב בדרכים בשעות שיא הביקוש. שיתוף נסיעות בהתארגנות עצמית (על ידי משתמשים בעבור משתמשים), ללא כוונת רווח, מצליח, בתנאים מסוימים, למשוך משתתפים רבים מקרב המשתייכים למעמד הביניים, שלרוב אינם מגיבים ליוזמות פורמליות ומסחריות. אולם, חרף הפוטנציאל הגלום בה, התופעה נתפסת במחקר ובפרקטיקה כאוסף של תופעות לוקליות ואנקדוטאליות שאינן ניתנות לשכפול - ובהתאם, הספרות הרלוונטית היא דלה. המחקר המוצג מניח כי ניתן להתייחס לתופעה כבעלת דפוסים חזרתיים, הניתנים להשפעה חיצונית. כדי להתמודד עם פער הידע בנושא, פותחה מסגרת מושגית לאפיון שיטתי של התופעה המורכבת, המאפשרת הבנת הדינמיקה שלה ואת התנאים הנדרשים להצלחתה. המסגרת מושתתת על סינתזה של ידע המתועד בספרות המדעית והאפורה ועל שילוב בין שלוש תיאוריות תומכות: חדשנות "מלמטה", כלכלה שיתופית וניתוח רב-שכבתי. לצורך בחינת התקפות והמועילות של המסגרת נערכה בחינה השוואתית של שמונה מקרי בוחן מסביב לעולם לשם חשיפת קטגוריות נוספות וביסוס המסגרת. החקירה נעשתה באמצעות שילוב של שיטות מחקר איכותניות וכמותניות, שכללו, בין השאר, ניתוח לאורך זמן של מסרים מילוליים בין משתתפים בפלטפורמות דיגיטליות (בעיקר אפליקציית ווטסאפ ופייסבוק), ראיונות מובנים למחצה עם גורמים רלוונטיים, תצפיות, וניתוח של חומר מתועד. ניתוח מקרי הבוחן באמצעות המסגרת שפותחה אפשר לחשוף מספר דגמים, דפוסי התפתחות והקשרים של מיזמים אלו, בנוגע להיבטים כגון הטריגרים למיזמים; מאפייני היזמים; התנאים הגיאוגרפיים, התחבורתיים, הכלכליים והחברתיים התומכים; זירת השיתוף; התלות בטכנולוגיה; שלבי הצמיחה המובחנים; הספים הקריטיים; הרגישות לאירועים חיצוניים; והתלות בהקשר המקומי ובתמיכת הרשויות. הבנת התופעה, על תכונותיה ודפוסיה החוזרים, מאפשרת הכללה וחיזוי, שעל בסיסם ניתן לתכנן התערבות אפקטיבית, לא יקרה יחסית וניתנת למימוש מהיר, שתעודד מיזמי שיתוף חדשים ותתמוך במיזמים קיימים לאורך מחזור חייהם.

ABSTRACT. עם העלייה הצפויה באוכלוסייה ובמספר כלי הרכב בעולם, התחבורה המקיימת תופסת מקום רב יותר בסל הפתרונות האפשריים להקטנת הצפיפות בדרכים. בהקשר זה מוזכרות נסיעות שיתופיות כאחד הפתרונות לבעיות התחבורה. יש הטוענים כי נסיעות שיתופיות מביאות תועלות סביבתיות, חברתיות וכלכליות באמצעות הפחתת מספר כלי רכב ממנועים, הפחתה במזהמים, בעלות הנסיעות ובעומסי התנועה. נסיעה שיתופית היא סוג של נסיעה בה פרטים חולקים את רכבם לצורך נסיעה משותפת וחולקים עם הנוסעים הנוספים את עלויות הדלק, החניה וכבישי האגרה. זמינות של אינטרנט והתפתחות טכנולוגיות ומודלים בכלכלת שיתוף פותחים אפשרויות חדשות לשירותי שיתוף תחבורה. בחירת אמצעי התחבורה היא בחירה מודעת הקשורה בעמדותיו של האדם ותפיסתו את צרכיו ואת רצונותיו, התנהגות מורכבת מכוונות התנהגותיות וכוונות מושפעות באופן ישיר מיחס ועמדות כלפי הפעולות האישיות. כוונות הן חלק מגורמי המוטיבציה שמשפיעים על התנהגות הפרט והן האינדיקציה למוכנותו של האדם לנסות, ומוכנותו להשקיע בכדי לבצע התנהגות מסוימת. עמדות כלפי אמצעי נסיעה מורכבות מ-3 רכיבים: הרכיב הסמלי, הרכיב האינסטרומנטלי והרכיב הרגשי. הרכיב הסמלי מתייחס לדרך בה אדם מביע את הזהות החברתית והאישית שלו. הרכיב האינסטרומנטלי מורכב מיתרונות אמצעי הנסיעה והרכיב הרגשי מתייחס לרגשות ולתחושות האדם - כנוסע או כנהג. הרכיבים הסמלי והרגשי מהווים חלק חשוב בבחירת הרכב הפרטי כאמצעי הנסיעה המועדף. בחלק הראשון של המחקר ניבדקו המאפיינים הפסיכולוגיים-חברתיים של המשתמשים הפוטנציאליים בנסיעות שיתופיות באזור מטרופולין ת"א. בחלק זה רואיינו 20 אנשים והוא נותח באמצעות grounded theory. נמצא שנהגים מתייחסים לרכבם כטריטוריה פרטית בדומה לביתם, דבר זה מהווה חסם משמעותי להכנסת נוסעים לרכב ומעבר לרכב של אחר. כמו כן, לנוחות הנהג והרגליו יש השפעה על ההחלטה האם להשתמש בנסיעות שיתופיות. בחלק השני נבדקה ההשפעה של מגפת הקורונה על ההחלטה להשתמש בנסיעות שיתופיות – האם מגפה זאת השפיעה ותשפיע בעתיד על החלטת נהגים להשתמש בנסיעות שיתופיות. הופץ שאלון המחולק ל-3 תקופות: מגפת הקורונה, פוסט קורונה והעתיד הרחוק. על השאלון ענו 227 נשאלים וכעת הוא מנותח על פי תאוריית ההתנהגות המתוכננת של אייזן.

ABSTRACT. The COVID-19 crisis has forced many people to work from home. Our study aims to assess the impact of the pandemic on telecommuting and commuting behavior in the post-pandemic era. We implemented a cross-country survey in Israel and Czechia addressing revealed preferences regarding work habits before and during the pandemic and stated intentions regarding work patterns after the pandemic. Two models were used for the data analysis, one addressing factors that affect the increased/decreased teleworking trend and the other addressing factors that affect the frequency of commuting. The results reveal that most respondents (over 60%) in both countries will maintain the same telecommuting / out-of-home work balance. About 19% of respondents expressed their intention to reduce the number of commuting days, while 6% stated they will increase out-of-home days. Consequently, a moderate reduction of 6-8% in the number of commuting trips is expected in the post-pandemic era. The anticipated decrease in commuting days is accompanied by an increase in teleworking: from 10% to 14% among those who work more than 20 hours a week and a drop in the rate of those who telework five hours or less a week. Self-employment, travel-time to work, working solely on-premise during lockdown, and personal preferences regarding telework versus working away-from-home were found to significantly contribute to a decrease in commuting days and to an increase in teleworking. An interesting finding is the high probability of increased teleworking among people who teleworked for the first time or increased their teleworking during lockdown. This indicates that the pandemic-related teleworking experience has enabled some people to view working from home as a viable option. Although, overall, the change in working habits does not seem dramatic, our results suggest that hybrid schemes for combining on-premise and telework are expected to be adopted by some sectors.

ABSTRACT. בשנים האחרונות, מהפיכת ׳האינטרנט של הדברים׳ מגיעה גם אל עולם הרכב. אלו הופכים טכנולוגיים יותר, ממוחשבים יותר ומקושרים יותר, במטרה לספק חווית הנסיעה בטוחה ומעשירה. אחד האתגרים המשמעותיים של הליך המיחשוב, הוספת טכנולוגיות ויכולות קישוריות הוא תג המחיר המתווסף בדמות העלייה באיומי הסייבר. ככל שהרכב דומה למחשב, הוא חשוף לאותן פגיעויות שהיו עד עתה עולמם של המחשבים – ובכלל זה תקיפות סייבר המסוגלות לפגוע ברכב. בשני מחקרים השונים, האחד של ד״ר חיים ויסמונסקי ומאי הר-שי והשני של גדי פרל, עמיתי מחקר במרכז פדרמן להגנת הסייבר באוניברסיטה העברית, נבחנו האתגרים המרכזיים בתחום, ושאלת מהלכי רגולציה הנחוצים להגנת סייבר בתחבורה חכמה בשוק הישראלי. בהמשך לכך, עו"ד ורד זליכה, עמיתת מחקר במרכז, גיבשה דו"ח מסכם עם מסקנות והמלצות לפעילות עתידית בישראל המיועדות בין היתר, לרגולטורים ולשוק הביטוח, על בסיס שני המחקרים ובתוספת מחקר עדכני וכן התייעצויות עם מומחים מהתחום וממערך הסייבר. הדו״ח, למיטב ידיעתנו ראשון מסוגו, מהווה בחינה כוללת של המצב הקיים בשוק הישראלי. הדו״ח כולל טקסונומיה של התחום והגדרת גבולותיו, לרבות מיפוי שוק התחבורה החכמה החדש וכן השחקנים החדשים בתחום – החל מבעלי הרכב הפרטיים, היצרנים הקלאסיים, וכן היווצרות מעמד נוסף וגדול יותר של יצרני רכיבי משני, אשר מוסיפים רכבים לרכב בין אם בשלבי הייצור, ובין אם לאחר מכן (After Market). הדו״ח בוחן בין היתר, מתי ניתן להסתמך על הרגולציה הקיימת, ומתי נדרשת אסדרה חדשה המותאמת לטכנולוגיה המתפתחת. נבחן הצורך ברגולציה רכה, הפועלת באמצעות יצירת תמריצים, וכן כללים מחייבים על מנת לדאוג לשלום הרכב ונוסעיו. נבחנו שאלות הנוגעות ליכולתו של הרגולטור הישראלי להשפיע על שוק שהוא בינלאומי באופיו, זאת בצד הצורך של מדינת ישראל לשמור על מרכזיות לפיתוח טכנולוגיות חדשות בתחום. לבסוף, אף נבחן מעמדן של חברות הביטוח, בשאלה האם יש צורך להרחיב את הפוליסות הקיימות לביטוח רכב על מנת שיכללו סיכוני סייבר, או שמא אלה כלולים כבר כעת- וכן את יכולתן של חברות הביטוח להתאים את עצמן למציאות המשתנה.

ABSTRACT. One of the most difficult road problems in the world is the lack of communication between drivers, which causes impatience and the increase of the extreme response during driving. This phenomenon, in addition to the tremendous load of vehicles as well as the time constraints that exist in modern society, results sharp and irresponsible reactions, follow with accidents and even deaths. The purpose of the system is to produce a broader symbolic language than the current signals (right / left, stop, and reverse), thereby increasing understanding and patience by providing information to the other driver about the driver's intentions. In emergencies the driver will be able to ask for help from other drivers in the area, like inform them about medical emergency or technical vehicle problems. The planned communication system that has been implemented prototype addresses two aspects: one is visual and has a range of up to 100 meters at day or night and the other is designed for an internal screen within the vehicle, which will transmit to all vehicles in the immediate vicinity. The product was built as a prototype for two remote controlled vehicles with all vehicles and their components connected to the local network created in its immediate vicinity. The network was built in tree topology to allow easy and hierarchical access to the connection destination. The mood choice is made via a touch screen which will be installed next to the driver and transmit and receive information from the vehicle's surroundings. In new vehicles, touchscreen devices and communication capability are installed so the system implementation can be easily implemented without the need for structural changes in vehicles. The system is intended to be tested in real-life situations and in cooperation with the National Road Safety Authority (Harelabad).

ABSTRACT. In traffic networks, determining the traffic signal plan for each intersection is crucial for obtaining a reasonable level of service. In this research a new method is developed that automatically designs complex actuated traffic signal plans. The method uses optimization algorithm combined with mesoscopic traffic simulation model to design and evaluate optimal traffic signal plans for each intersection in the traffic network. In return, the need of human intervention in the design process is reduced. Unlike available tools, the proposed method can simultaneously determine the optimal logical structure, optimal parameter values for the signal plan, optimal phase composition and sequence, and detectors placement including extension detectors, demand detectors, and pedestrian pushbuttons. Initially, a single intersection in the city of Haifa which was designed by experienced traffic engineers and operates by actuated traffic signal plan, was studied as a case study. The automatic design conducted in this research was done through both desiging signal plans with a fixed cycle length and with a variable cycle length. The results showed that both signal plans proposed outperformed the base signal plans. Therefore, the developed automated design method has been further extended to handle complex intersections. Such as intersections with complex geometry and those that include bus rapid transit (BRT) movements. Moreover, it can apply advanced mechanisms such as transit signal priority, green time compensation, and others. Lately another intersection in the city of Haifa was chosen to be tested by the improved version of the automatic design method. The initial results show that this approach has the potential to efficiently design signal plans without human intervention, which can minimize time, cost, and design effort. And it can also help uncover problems in the design that may otherwise not be detected, such as queue spillbacks and blockage between the main and storage lanes.

ABSTRACT. Delay imposed on buses is witnessed at isolated signalized intersections, as part of the delay is created due to bus and car interactions at the signalized approaches. Inexpensive solutions that do not involve new infrastructure are the most desired. Towards reducing person delay, pre-signal lights are installed at the upstream of the intersection. Pre-signal is a control tool that can manage the interaction and space conflict between cars and buses at the mixed lanes, located after the pre-signals, approaching the intersection. This research presents an optimal pre-signal control policy that determines signal settings based on minimization of total person-time spent. The developed model is compromised of a simplified signalized intersection with one approach, where two pre-signal lights are located, one for bus lane and one for car lane. A continuous-time model is first developed to describe the queue dynamics for the mixed lanes after the pre-signals and the bus and car lanes before the pre-signals. The optimal solutions are derived analytically via Pontryagin Maximum Principle, and the optimal pre-signal control policy is determined and synthesized. The derived analytical solutions are verified through the use of the optimal control software, DIDO, which uses an advanced optimization method and solves complex dynamic optimization problems without prior knowledge of the analytical computations or the solution. Several examples are used to test and analyze the derived solutions. Furthermore, a discrete-event model is developed and is currently being implemented in Aimsun where a microsimulation is to be conducted for further analysis of the optimization control proposed.

ABSTRACT. Highway Work Zones (HWZs) are associated with significant adverse impacts on safety, mobility, and work costs. The objective of this research is to reduce the negative impacts of HWZs. This was achieved by quantifying the impacts of HWZs on safety, mobility, and work costs. And developing an optimization model to minimize the total costs associated with HWZs, as well as a multi-objective optimization considering safety, mobility, and work costs. Quantifying the HWZs impacts and developing the optimization model was done by controlling several decision variables that are allocated into three categories: (1) site geometric design, such as lane and shoulder width and work zone length; (2) work management, such as determination of the number and timing of work hours; and (3) Temporary Traffic Control (TTC) measures, such as the use of dynamic speed displays or presence of flaggers. The developed model implements a location-based schedule within the cost evaluation. A genetic algorithm is used to determine a set of optimal scheduling and decision variables. The performance of the model is demonstrated in a case study. The results reveal that crash costs, which were often ignored or only included indirectly in previous works, are a substantial cost component. Their explicit inclusion in the optimization process significantly affects the total cost and the optimal operations of the HWZ. Furthermore, the inclusion of a location-based schedule in the model is instrumental and affects the optimal solution since all HWZ cost components are affected by the work processes and project duration. Moreover, considering the effects of TTC on the optimized function has a substantial influence on the total cost. The model can support transportation agencies and local authorities in mitigating the adverse impacts associated with HWZs.

ABSTRACT. Nairobi's endemic traffic congestion has severely impacted the city's resilience causing substantial economic losses, runaway urban sprawl, and transportation disadvantages primarily to its vulnerable populations in informal settlements. The Bus Rapid Transit (BRT) system has evolved as a mobility response expected to address these challenges. Its adoption has faced delays in implementation and opposition from existing stakeholders in the urban public transport sector. It espouses transport reform and is part of the investments from the recent flow of global capital into Kenya. However, these have focused on large-scale infrastructural developments that favor road-based investments to the improvement of existing public transportation and Non-Motorised Transportation (NMT) systems. Further, despite the devolution of the national government's urban planning functions to county governments through The New 2010 Constitution, the project's implementation and future are in jeopardy due to the reduced capacity of local authorities to coordinate BRT development because of the pre-existing problematic institutional and governance framework within the city's transportation planning. The project presents the challenge of transportation reform in an urban context that is dominated by informal transit trying to adapt to a world-class formal system (BRT) and the stakeholders involved, especially the existing transport operators. It also presents the neoliberal contradiction in the large and distorting role of external actors - donors, BRT advocates, and governments to create an international market for bus operations

ABSTRACT. Understanding the governance structure of public transport systems can help us understand the reasons services are well/poorly integrated, fare structure that encourages/discourages ridership, and service coverage, among other outcomes, all with important equity impacts. Unfortunately, the transportation literature has lacked consistent, regional measures of fragmentation of public transportation systems, impeding our ability to understand its impact on system performance indicators. We use data from four U.S. metropolitan regions to measure fragmentation based on providers' governance zones. We discuss the development of the measure, and apply it to four pilot regions in the United States. We then draw links between government systems, local and state funding boundaries, indicating avenues for future research into service outcomes and opportunities for similar research in Israel.

ABSTRACT. Public Transport (PT) has entered an era where automated vehicles can potentially turn Demand Responsive Transit (DRT) into the leading transportation mode. The research goal is to assess this future regarding Shared Automated Vehicle (SAVs) service in the Jerusalem Metropolitan Area (JMA). According to a recent publication (Schaller, 2021), most of the ride-hailing users in the US were former PT users, and AVs are unlikely to change auto users' resistance to shared services. A possible solution to force car users to move to shared services is to impose congestion and/or parking prices. The following study evaluates the consequences of a new PT-focused transport policy in the JMA that includes parking and congestion pricing and introduces a new SAV transportation mode as a possible game-changer of the existing equilibrium between PT and private car use. The Investigated effects of the SAV fleet operation parameters: size and location of the service area, fleet size are examined. We compare SAV service priorities; (1) Full SAV service – any ride-sharing request within the whole JMA area is served and (2) a Core-Periphery SAV service where either origin or destination of the request should be in the Jerusalem city center. The study's overall goal is to propose a balanced set of carrot-and-stick measures that could enforce sustainable modal shifts that can reduce congestion during the planned development of the JMA PT network between 2020 and 2040.

ABSTRACT. The age of automated vehicles (AVs) opens new opportunities to improve public transportation services, particularly for low-density suburbs. However, the characteristics of automated vehicles may adversely affect the transport equilibria by decreasing demand and service levels due to reduced public transport frequency. In this research, we aim to maximize the positive externalities due to AVs, such as mobility for under-served populations, whilst minimizing potential negative externalities such as congestion to ensure an equitable and inclusive transportation system. We develop a stylized suburban neighborhood model connecting to a central business district to create a two-stage game and compare the outcomes over multiple potential market structures for the AV operator. We look at a market demonstrating perfect competition, a monopoly operator market, and duopoly market setups. We utilize a real-world case study from Jerusalem, Israel, to compare the market structure results to an optimal social welfare outcome and derive the passenger subsidies and congestion fees needed to replicate the social optimum. The methodology and findings will aid policymakers in developing evidence-based regulatory tools to incentivize behavior and generate a positive transport equilibria outcome.

ABSTRACT. Ride-hailing services have become in recent years a prevalent transportation service that facilitates mobility in urban areas. A ride-hailing service system encompasses several research problems, among them is the operational assignment of on-demand ride requests to vehicles in real-time. The literature suggests various approaches to address similar systems, mainly optimizing the system efficiency, but recent studies pointed out that these systems are likely to cause geographical unfairness among passengers. Such unfairness implies that requests whose origin or destination are far from centralized locations, tend to suffer from excessive service rejections. In this work, we suggest an assignment mechanism that aims to overcome this phenomenon. We formulate the Online Ride-hailing Problem with Fairness that aims to maximize both efficiency and fairness in the system, while achieving an adequate balance between them. We find good assignment solutions from a mixed-integer linear programming (MILP) formulation that are applicable to the online setting. Then we develop a new policy to assign ride requests to vehicles, leveraging data-driven techniques that make use of good assignment solutions, found by the MILP. With a simulation study, we examine the performance of our new approach relative to commonly used dispatching rules, using synthetic random data suggested in the literature, which represent a real city layout and movement. Our results demonstrate that a good system performance is achieved using this approach.

ABSTRACT. Semi-autonomous transportation services use the autonomous capabilities of the vehicles partially, so that they adhere to current regulations. The multi-layered personal transit system is a special design in which convoys composed of human-driven lead vehicles and autonomous cabins provide station-to-station transportation. The cabins can detach from the convoys, led by human driven vehicles, and travel autonomously at restricted areas, typically in the proximity of stations. The assignment of passengers to cabins and their routing are typically determined dynamically. However, for ease of the operational planning, previous approaches applied fixed circular routes for the lead vehicles that which is the head of the convoys In this work, we introduce for each circular route a potential shortcut, with the aim of determining dynamically if and when a lead vehicle should take a shortcut from his given route. We formulate the decision problem as a Markov Decision Process in which the state represents the number of cabins at each station, an action represents the routes of the lead vehicles and the penalty represents the number of cabins waiting to be served. We develop an event-based simulation framework to represent the system’s dynamics and employ Reinforcement Learning method that aim to suggest for each state of the system how to spread the lead vehicles into the routes to reduce travel time. The obtained policies are compared to fixed route plans and simple-rule dynamic policies.

ABSTRACT. The Dial-a-Ride Problem (DARP) consists of defining routes and schedules for a fleet of vehicles serving multiple transportation requests within a service area. In this work, we define and study a new variant of the DARP which considers both transfers and walking, namely, the DARP with Transfers and Walking (DARPTW). In particular, passengers are allowed to transfer multiple times and their itineraries may include several walking segments. The goal of the DARPTW is to minimize a bi-objective function consisting of the total distance covered by the vehicles and the total excess time of the passengers, while considering limitations on the number of transfers performed by each user and the total distance walked. Introducing transfers and walking presents several opportunities. Multiple transfers may allow balancing the vehicle loads and reducing the service area covered by each vehicle, by decomposing itineraries to separate service segments that would be served by different vehicles. Walking may assist in reducing unnecessary vehicle detours to extreme regions of the service area. Additionally, walking may facilitate significant shortcuts that cannot be achieved due to driving. Nevertheless, these opportunities generate a challenging problem to solve. Specifically, the DARPTW generalizes the DARP and therefore is also NP-Hard. We devise an efficient algorithm for the scheduling sub-problem, which minimizes the total travel time of the passengers. The algorithm determines the feasibility of given routing plans and applies fast heuristics to construct good schedules. We implement the algorithm within an Adaptive Large Neighborhood Search framework in search of promising solutions of the DARPTW. Numerical experiments are conducted using real-world data obtained from Bubble-Dan in Tel Aviv. Preliminary results over thousands of scheduling sub-problem instances demonstrate that our heuristic algorithm finds the optimal schedule in more than 90% of the cases and that the entire framework produces high quality solutions.

ABSTRACT. Dynamic routing to achieve efficient and fair allocation The ability to make fast routing and allocation decisions in an uncertain environment has a great importance for numerous urban logistic and smart transportation applications. For example: food banks’ activities that collect an (unknown in advance) supply of food from suppliers and distribute it to welfare agencies. Due to a time constraint, only a subset of the suppliers and recipients may be visited in any order, and the routing and allocation decisions are made in real time based on supply information that arrives dynamically. We formulate a dynamic and stochastic vehicle routing problem combined with resource allocation decisions, in which information about supply quantities arrives gradually over time. In the problem we consider, the objective is delivering as much goods as possible while obtaining equitable allocations to the agents. The resulting model suffers from the three causes of curse of dimensionality (state space, action space and value computation of any state and action). Hence, we develop a heuristic approach that copes with all the three causes using a novel implementation of approximate dynamic programming. Specifically, we suggest to augment a Neural Network that approximates the future value given any state and action (decision). We develop upper bounds and benchmark solutions, which are used to evaluate the performance of our method. Towards that, we conduct a numerical experiment over 11 to 80 sites instances that are based on a dataset provided to us by the Berlin foodbank, as well as on randomly generated datasets. Our preliminary results confirm that the heuristic obtains fast high-quality solutions, which are on average about 10% better than the benchmark solutions.

ABSTRACT. Conditionally automated vehicles are expected to offer safer and faster travel than conventional vehicles. However, one of the issues this technology raises is the concern of behavioral adaptation, i.e., the driver’s behavioral changes due to adopting the autonomous vehicle and integrating it into his daily routine. Many studies have examined how drivers behave in an autonomous vehicle for the first time and for a relatively short length. These drivers didn’t have the time to adapt their behavior to the technology and examine the best conduct within the vehicle. Therefore, it will be essential to study driver behavior over multiple drives to understand behavioral changes better. This study aims to understand how drivers change their behavior due to exposure to autonomous vehicles by time and number, how their risk perception will change over time, and the relation between these changes. In a within-subjects pilot study, four participants drove four times each in a fixed simulator in a total of twelve different scenarios over three weeks (n=4). The subjects drove in each scenario twice while connected to an eye-tracking system. Results show that the participants generally drove more carefully in the second drive of each scenario in terms of velocity (standard deviation and absolute value) and the driving wheel’s standard deviation. This fact may hint that drivers adapt their behavior to conditionally automated driving and better assess their abilities in this type of driving as time progresses. In addition, subjects’ mental workload seems to decrease among almost all drivers. Results may also indicate a positive correlation between risk perception and trust in automation, hinting that subjects tend to judge the automation system more positively after riskier scenarios. This pilot provides preliminary evidence of behavioral adaptation drivers undergo when dealing with conditionally autonomous vehicles, to be investigated more extensively in the future.

ABSTRACT. Bikesharing is an affordable mode of transportation and a potential tool to reduce car usage in cities. However, in many cities, bikesharing seems to be used mostly by affluent populations. Indego, Philadelphia’s bikeshare, embraced the promotion of equity as part of its primary goals. While previous measures were not adequate for that cause, Indego decided to integrate e-bikes into its system to promote usage among current non-users. In this study, I examine how the integration of e-bikes influences Indego’s usage in disadvantaged areas. For that purpose, I combined official publicly available data using spatial analysis methods. Furthermore, I used random forest and spatial negative binomial regression to examine factors influencing shared bicycle and e-bike usage in Philadelphia. The findings show that e-bikes increase the overall usage of Indego, specifically in disadvantaged areas. In these regions, the users use shared e-bikes for commute, leisure, and other utilitarian purposes, while in the rest of the city, users use e-bikes mainly for commuting. I conclude that the integration of e-bikes was successful in promoting bikesharing usage in disadvantaged areas.

ABSTRACT. ההתפתחות הטכנולוגית המואצת בעשורים האחרונים והיישומים המתקדמים של טכנולוגיות מידע מציבים בפני האדם בן זמננו שאלות רבות על מהות הטכנולוגיה ועל הקשרים השונים בינה לבין הקיום האנושי. בהרצאה אבקש להצביע על התרומה האפשרית של נקודת המבט האונטולוגית להבנה והטמעה של טכנולוגיה מתקדמת בכלל ומערכות תחבורה חכמה בפרט, על בסיס טיעוניו של ברונו לאטור בספרו משנת 1993. לאטור בוחן במחקר אתנוגרפי ניסיון לפתח מערכת תחבורה ציבורית חכמה בפריז בשם "אראמיס" שתוכננה כ-"תחבורה אישית מהירה". הפרוייקט נועד לאפשר הסעת נוסעים מנקודה לנקודה באופן אוטונומי על פי בחירת הנוסע וללא עצירות ביניים, ובכך לשמש כתחליף קרוב לרכב הפרטי מחד, ומאידך לשמור על היתרונות של התחבורה הציבורית בחיסכון בהשקעה בתשתיות ובהפחתת זיהום ועומס במרכזי הערים. לאחר כ15 שנות פיתוח בהן הושקעו סכומים גבוהים ונערכו ניסויי היתכנות שהראו תוצאות חיוביות, הוחלט בשנת 1987 על סגירת הפרוייקט. לאטור מבקש לבחון בספרו מה היו הסיבות לכישלון ומה ניתן ללמוד מכך על הטמעה מוצלחת של טכנולוגיה חדשנית. הוא בוחן את הפרוייקט מנקודת מבט אונטולוגית וטוען כי פרוייקט טכנולוגי מורכב לא רק מרכיבים טכניים והנדסיים אלא מבוסס על רשת של משמעויות ופרשנויות הניתנות לו על ידי השחקנים השונים המעורבים בפרוייקט. לאטור מראה כי בהתאם למספר השחקנים המעורבים ולשלבי ההתפתחות השונים, כך נוצרו עשרות "סיפורים" שונים על מהותו וייעודו של הפרוייקט. המסקנה המרכזית היא כי על אף ההיתכנות הטכנולוגית והאינטרס הציבורי, חוסר היכולת של השחקנים השונים להגיע להסכמה על מהותו של הפרוייקט הובילה לכך שהוא נשאר כרעיון בלבד ולא התגבש לכדי אובייקט מציאותי. הוא רואה את הפרוייקט כדוגמא הממחישה את האופן בו רעיונות של טכנולוגיה חדשנית מתממשים או נכשלים. נראה כי על אף ההבדלים בטכנולוגיה ובאתגרי התחבורה, אימוץ נקודת המבט האונטולוגית יכול לתרום להבנה טובה יותר של אופן ההתפתחות של רעיונות טכנולוגיים של תחבורה חכמה בת זמננו, ולקדם בכך את המחקר התיאורטי והיישומי בנושא.

ABSTRACT. אוטובוסים עירוניים הם מחוללים משמעותיים של זיהום אוויר, רעש ופליטה של גז"ח, וזאת במיוחד ככל שצי האוטובוסים הוא מיושן יותר. מדינות בעולם וביניהן ישראל, משקיעות כספים רבים בהצטיידות באוטובוסים 'ירוקים' יותר, ביניהם: אוטובוסי דיזל יורו 6 (העומדים בתקני פליטה מחמירים יותר מבעבר), אוטובוסים הנעים על גט"ד (גז טבעי דחוס), אוטובוסים היברידים (דיזל וחשמל) ואוטובוסים חשמליים. כל אחת מחלופות, כרוכה בעלות רכישה והפעלה שונות, כמו-גם, בתועלות בריאותיות-סביבתיות שהיא מייצרת. מחקרים בעולם שבחנו אילו חלופות הן מיטביות ביחס של עלות מול תועלת, מצאו כי הדבר אינו מובן מאליו, ותלוי מאוד בתמהיל הדלקים של המדינה לייצור חשמל כמו גם בהיבטים מקומיים נוספים. מטרת העל של המחקר שאציג הייתה להשוות בין הכדאיות הכלכלית וסביבתית של הנעות חלופיות לאוטובוסים בישראל (הנעה חשמלית בטכנולוגיית סוללות; היברידית (דיזל-חשמל); גט"ד (CNG); ודיזל חדש (יורו 6), בשילוב מטרות המשנה הבאות: 1. כימות של העלויות הישירות והתועלות החיצוניות העיקריות המתקבלות מכל חלופה באמצעות בחינה משווה של מחזור חיים סביבתי (Life Cycle Analysis) וכלכלי (Life Cycle Cost)מותאם לישראל (משלב הייצור/שימוש/סוף שימוש של האוטובוס). 2.הצגת הזדמנויות מול אתגרים שונים הכרוכים עם כל חלופה 3.המלצה על צעדים שיאפשרו למקסם את התועלות הפוטנציאליות של כל חלופה. המחקר כלל גם ניטור משווה של הרעש הנגרם מכל חלופה ובחינה של מגוון סוגיות נוספות שטרם ניתנה עליהם את הדעת מבחינת מדיניות. מתוצאות המחקר עולה כי, הדבר הנחוץ ביותר כיום הוא להקדים את ההחלפה של צירי האוטובוסים הישנים בכל אחת מהחלופות שנבחנו, ולעשות זאת, בהקדם האפשרי ביותר. החלופה החשמלית שהיא כיום היקרה והסביבתית ביותר, יכולה לייצר תועלת גבוהה במיוחד בישראל, אך תחת תנאים מסוימים שמהם נגזרים צעדי תכנון ומדיניות בהתאם. בנוסף, חלופה זו, כרוכה עם מספר אתגרים משמעותיים ('פילים בחדר'), שעד היום נראה כי לרוב התעלמו מהם, והמחקר מנסה לראשונה להציג אותם ולהציע דרכי התמודדות.

ABSTRACT. Power and efficiency relation of the combined electrochemical and combustion processes was analyzed. Combustion is a spontaneous process that results in a high thermal energy of the combustion products, which is then utilized for mechanical work production in an expansion process. On the other hand, electrochemical reaction is an electrically controlled process, in which fuel energy is utilized directly for electrical work generation. A device called fuel cell (FC) implements the realistic irreversible electrochemical process. In this process, apart from work, heat is generated also and can be recovered into additional work by mechanical expansion. The reported analysis uses the principles of finite-time and finite-speed thermodynamics (FTT and FST) to relate between power and efficiency. Novikov in 1957 and Curzon-Ahlborn in 1975 proposed independently those principles for heat engines. A further development for other types of cycles was performed by Hoffman et al. in 1997. Sayyaadi et al. proposed an integrated analysis of FST and FTT for internal combustion engine (ICE) modelling. By applying FST and FST, our study shows that the combined cycle has the potential of a significant power gain at much higher efficiencies and beyond 70%. These efficiency and power are not achievable, if FC and ICE are operated alone. Power-efficiency curves of the hybrid cycle are obtained for various conditions and components properties. The results are of high importance for transportation propulsion since specific power affects the travel range and the powertrain weight and size.

ABSTRACT. This research aims at understanding the effect of aerodynamic forces on the dynamic stability of racecars moving at very high speeds. This was done in the context of the recent Indy Autonomous Challenge (IAC), the first competition of autonomous race cars competing on the famous Indianapolis Motor Speedway (the home track of indy500). Unique to race cars is the use of spoilers to increase the downforce between the tires and ground, which improves the vehicle's dynamic stability, i.e. no-slip, along a curved path. In this work, we investigated the effect of the aerodynamic forces on the vehicle speed limit, above which the vehicle might become dynamically unstable. Considering the aerodynamic forces in the system dynamics allows computing the trajectory (race line) along the track. The optimal race line can be used to control the racecar while pushing through its limits, and yet, be safe from different slip and roll-over effects of motion. It is shown that the aerodynamic forces, which are speed-dependent, affect the vehicle speed limit as a function of curvature, coefficient of friction, vehicle mass, and the lift coefficient, in a single expression. It is shown that for some combinations of these parameters, the difference between the current speed and the limit speed grows bigger as we increase the vehicle speed, which in turn increases the vehicle dynamic stability. For other combinations, increasing the vehicle speed would diminish the gap to the speed limit and thus reduce the vehicle stability at high speeds. This is counterintuitive since it is widely accepted that the higher the speed the better the grip. These effects of the aerodynamic forces on the time-optimal velocity profile along a given path are demonstrated in several examples.