Localization for V2X communication with noisy distance measurement

• Published in: International journal of intelligent networks Vol. 4; pp. 355 - 360
• Main Authors: Javed, Iram, Tang, Xianlun, Saleem, Muhammad Asim, Javed, Ashir, Zia, Muhammad Azam, Shoukat, Ijaz Ali
• Format: Journal Article
• Language: English
• Published: KeAi Communications Co., Ltd 2023
• Subjects: C-Shaped; Extended Kalman filter; Format; Mobile anchor; V2X
• ISSN: 2666-6030; 2666-6030
• DOI: 10.1016/j.ijin.2023.11.007

Mobile sensor network localization is a growing research topic after IEEE 802.15.4 specified the procedure of low-rate wireless personal area networks (LR-WPANs), which further helps localize vehicles in the automobile industry. This paper presents a new localization scheme based on flying anchors deployed in vehicular infrastructure. The mobile anchor nodes follow a random C-shaped trajectory. A global positioning system (GPS) is attached to each anchor node, transmitting beacons with ID and location to all other vehicles in a network. Distance calculation is facilitated through link quality induction, employing the centroid method to compute localization error. Mobile anchor localization, particularly when employing a C-shaped trajectory commonly adopted by various topologies, consistently yields optimal positioning outcomes. However, this approach can be susceptible to the impact of noisy measurements, potentially reducing overall localization performance. To overcome this problem, we proposed a framework based on extended Kalman filtering (EKF), which is used to refine the coordinates of the vehicles. To compute the lower bounding of the vehicular node, an analytical framework is also proposed to enhance the localization error accuracy. Simulation results show that the EKF framework provides better positioning accuracy compared to the existing C-shaped solution, irrespective of noise statistics, topology selection, and anchor node density. With the help of the Extended Kalman Filter (EKF) framework, we achieved a comprehensive localization error of 0.99 m, accompanied by a standard deviation of 0.47.