LSTM network based reinforcement learning for vehicle traffic optimization to improve public guarded intelligent transportation system

Abstract

An Intelligent Transportation System (ITS) and the Internet of newlineThings (IoT) provide superior results for effectively solving unpleasant newlinecomplex transportation challenges. Predicting the vehicle traffic, vehicle newlinecrashes, vehicle demand, vehicle location, vehicle communication, and travel newlinesafety are the critical problems in today s transportation systems. newlineThe proposed research optimizes vehicle traffic by incorporating newlinereroute recommendations, higher utilization of public transportation, and newlineprovides intelligent health assistance for onboard vehicle drivers. newlineAdditionally, this research focuses on resolving complex transportation issues newlinesuch as vehicle s live location, exact vehicle count information on each route newlineand onboard vehicle vacant seat information. Further, vehicle communication newlinehelps to increase the proposed system s efficiency and avoids communication newlinedelay or traffic information loss to the registered users and the cloud server. newlineFirstly, the magnetic sensor is utilized to identify the type of newlinevehicle, and the exact vehicle count in traffic is calculated based on the newlinevehicle s magnetic flux. The sensed vehicle information is updated to the newlinecloud server using cluster vehicle communication. In the cloud server, an newlineintelligent agent utilizes reinforcement learning to comprehend real-time newlinetraffic flow collected from multiple routes to anticipate legitimate and newlineefficient route recommendations to the registered users, thereby minimizing newlinethe traffic congestion. newline