Deep learning algorithms and mechanisms in navigation for vehicular crowd managementsystems in real-time for smart transportation

Abstract

Accurate predictions of vehicle mobility and density are necessary for a wide range of mobile applications, including VANETs, crowdsourcing, participatory sensing, network provisioning, and shared transportation. The difficulty of forecasting is exacerbated by the scarcity and scale of vehicular mobility data. Crowd management and navigation analysis of vehicular networks that make use of deep learning techniques are the focus of this study. Multihop path based edge computing is used to analyze vehicular network navigation, and a markov spatio reinforcement neural network is used to manage vehicular crowds. The number of vehicles in the network and its navigation analysis are the basis for the experimental analysis. Throughput, data transmission rate, latency, network traffic analysis, and scalability are the parameters analyzed.proposed technique attained data transmission rate of 94%, latency of 62%, throughput of 61%, network traffic analysis of 59%, scalability of 63%.