Abstract
Modern urbanization is demanding smarter technologies to improve a variety of applications in intelligent transportation systems. Mobile crowdsourcing enabling automatic sensing tasks constitutes an excellent mean to complement existing technologies. In this paper, we exploit the high amount of data that can be collected by on-board and infrastructure-based sensors to evaluate traffic network statuses and improve the navigation of vehicles in urban areas. The objective is to design real-time route planning algorithms that determine fastest trajectories for both single and multiple destinations, in a real-time manner based on the frequent data inputs. We first formulate the routing problems as integer linear programs (ILPs) and then, design iterative approaches levels to iteratively solve the ILPs while considering updated traffic data. Afterwards, lower complexity sub-optimal graph-based algorithms are designed to solve the real-time routing problems. Unlike traditional navigation solutions, the proposed approaches update the vehicle trajectory after a certain period characterized by timely correlated data. Uncertainty and erroneous data inputs are also considered to determine fastest and least risky trajectory. Our results show that crowdsourcing-based real-time navigation outperforms outperform traditional navigation solutions by selecting less congested roads and avoiding blocked streets.