دانلود کتاب تجزیه و تحلیل داده ها و یادگیری ماشین برای مدیریت یکپارچه راهرو

Mitigating traffic congestion and improving safety are the important cornerstones of transportation for smart cities. An INRIX study [59] found that in 2017, traffic congestion cost nearly $305 billion and caused Americans to lose 97 hours per person annually in gridlock. This costs theU.S. $87 billion annually in lost time. Many drivers are frustrated due to long (but potentially preventable) delays at intersections. The Governors Highway Safety Administration (GHSA) found that pedestrian deaths have steadily increased with 6590 deaths in 2019 an estimated 60 percent increase over 2009. However, traffic management and control are a big challenge in all major urban areas today, especially growing cities. This is because there are limitations on traffic capacity and throughput in most urban areas. Hence, it becomes essential to utilize the existing road infrastructure and networks as efficiently as possible [148, 19]. Traffic intersections, specifically signalized intersections and traffic controllers are an elementary component of all the road networks and are used to solve the problem of traffic conflicts by implementing a basic version of time division multiplexing [148, 149]. However, this comes with its own set of challenges. The operation of signalized intersections represents a specialized facet of traffic management, demanding distinct expertise and resources. The role of signal operations in improving efficiency is frequently undervalued. Even with sufficient resources, including budget and experienced staff, effectively allocating these resources can pose challenges in the absence of comprehensive knowledge regarding the operational performance of various locations. Addressing these challenges requires a thorough understanding of traffic patterns throughout the transportation network including intersections, road segments between intersections, and even limited-access facilities that comprise the overall transportation network. Current performance measures typically compare travel-times before and after retiming efforts for validation. However, incorporating additional performance metrics, e.g. arrival on green percentage and travel time reliability measures, will provide greater utility supporting additional decisions on prioritizing which corridor of signals need retiming the most, and handling additional, varying patterns as demand fluctuates across times of the day, days of the week, seasons, and special events. One of the key components of traffic management today is the extensive use of Advanced Traffic Management Systems (ATMS). ATMS systems are good at gathering displaying, and reporting raw data from wide variety of disparate systems and sensors that cannot communicate with each other. The next level of abstraction which enables traffic engineers of today to better manage the road networks is Automatic Traffic Signal Performance Measures (ATSPM) Systems. These draw upon the vast body of recent research on the quantification of the performance of signalized intersections, enabling both better management of signalized intersections and manual optimization of intersections and key corridors. ATSPM utilizes high resolution data from signals and automatically converts the data into performance measures that allow a practitioner to understand the performance of any given intersection and the nature of the problem at said intersection (if any). Furthermore, the availability of high resolution (10 Hz) controller logs opens an even broader range of possibilities. This data, in many cases, is available with small latency (a few minutes) making it amenable to real-time decision making and addressing bottlenecks. However, this plethora of information without proper decision-making tools adds a burden on transportation professionals. This is challenging even for small cities comprising of a few hundred traffic intersections. There is a need for a system that provides corridor-level and city-level information in a succinct and actionable form. The integration of integrated corridor management (ICM) systems, data, and algorithms is crucial for creating effective traffic management systems and is the focus of this book. We first discuss some general concepts on the traditional approaches to traffic management. We next explore the role of data-driven decision support systems, automation, and machine learning in optimizing traffic flow and improving transportation systems, and finally provide an outline for the rest of the book.