Yasmine Waleed, Rola wafi, Celiusty Naguib, Hagar Maged , Amal Khaled.
Traffic congestion is unavoidable in large and growing cities across the world. Life will be easier if there is an application to guide you when to go out tomorrow to reach your meeting on time, or say to you to avoid a certain route because it will be clogged due to heavy rains next Tuesday. Traffic Prediction system will make life simpler specially here in Egypt. There are predictable factors that cause traffic like weather and traffic crashes . Our project’s goal is to predict the traffic congestion before months or days using traffic data-set. Traffic prediction can help in guiding the best route to take and manage traffic congestion. The algorithm used in this system is Long-Short Term Memory”LSTM” and “Support Vector Regression Machine “SVR”.
1.1 Purpose of this document
The purpose of this Software Requirements document is to summarize the specific topic and goals of a Traffic Prediction Mobile application .It’s usually placed in the beginning to provide the reader a clear vision of what the document will be about. The document will show the requirements and the type of system which the developers will be developing and all the important information that will be used to write the system. The document will explain the time plan for the mobile application to be ready and will explain the requirements of the products, who is working on the specific requirements. The audience for this document are stakeholders such as automotive industry, the developers of this software and the graduation projects committee at Misr International University, Faculty of Computer Science.
1.2 Scope of this document
The Traffic Prediction Mobile application is being developed to support transportation safety, save time for drivers, and help engineers to improve transportation design.• The Document will explain more or make the system overview more clear by discussing similar systems, system description, functional requirements, Data Design, Preliminary Object oriented domain analysis, operational scenarios, project plan, and appendices.
1.3 System Overview
The data set is loaded from PEMS. it contains 6 columns which are: lane 1 flow , lane 2 flow , lane 3 flow, lane 4 flow, flow(vehicle per 5 minutes), lane points. First 4 columns ( lane 1 flow , lane 2 flow , lane 3 flow and lane 4 flow are representing the number of cars in each lane. the column (flow) is showing the sum number of cars in the 4 lanes per 5 minutes. Then we processed these data using normalization to adjust the data and change the numeric variable in the data set to a typical scale, without misshaping contrasts in the range of value. Then we applied long short term memory model. LSTM is a deep learning model we used it to predict the traffic congestion using the features in the data set with the higher accuracy that we can achieve. Then we integrated the system with google API. The output of this system is traffic prediction and suggesting another route if the route is crowded with showing the reason of the traffic. All of this system is applied using flutter mobile application using google API for the maps.
1.4 System Scope
The proposed system is designed to predict the traffic that may be caused by the events, weather and specific occasions within the hours, days, months in the current year and display its location and the surrounded streets which might affect traffic congestion flow and show the routs that should be avoided.
• Be easy and fast way to open your navigation.
• The system should predict the traffic from the user input data. These data are destination,
date and time.
• The system will help the driver to show the fastest route to avoid traffic congestion, and reach
the drop point on time.