No abstract

The goal to achieve accurate and ubiquitous localization is the driving force for location based services in vehicular ad hoc networks (VANETs). In urban areas, global positioning system (GPS) and in-vehicle navigation sensors (e.g. odometers) suffer from prolonged outages and unsustainable error accumulation, respectively. The need for precise vehicle localization remains paramount, and cooperative vehicle localization based on ranging techniques are being exploited to this end. This paper presents a novel cooperative localization scheme that utilizes round trip time (RTT) for inter-vehicle distance calculation, integrated with inertial sensor measurements to update the position of not only the vehicle to be localized, but its neighbors as well. We adopted the extended Kalman filter (EKF), to limit the effect of errors in both the sensors and the neighbors' positions, in computing the new location. In comparison to the existing cooperative localization techniques, our proposed cooperative scheme does not depend on GPS updates for the neighbors' positions thus making it far more suitable in urban canyons and tunnels. In addition, our scheme considers updating the neighbors' positions using their current inertial sensor measurements resulting in; better position estimation. The scheme is implemented and tested using the network simulator 3 (ns-3), vehicle traces are generated using SUMO and error models are introduced to the sensors and initial positions for different velocities and densities. Results show that our scheme outperforms the inertial navigation systems (INS) technology typically used in environments where GPS fails.

No abstract