Vehicle Localization in V2X-Enabled Environments: A Survey on Challenges, Enabling Technologies, and New Perspectives

Abstract

This paper presents a comprehensive review of state-of-the-art localisation techniques for autonomous vehicles, with a focus on technologies leveraging active and passive sensors, cellular vehicle-to-everything (C-V2X) communication, and data fusion methods. It highlights the challenges faced by global navigation satellite system (GNSS) and other localisation systems, particularly in urban environments where signal attenuation and multipath effects compromise accuracy. To address these challenges, the paper emphasises the critical role of data fusion techniques, which integrate inputs from sensors such as light detection and ranging (LiDAR), radar, inertial measurement unit (IMU), and cameras, thereby enhancing overall system reliability and precision. A dedicated treatment is given to intelligent reflecting surface (IRS)-aided localisation, which is positioned as an enabling layer that improves existing localisation solutions by reshaping the propagation environment. Furthermore, the necessity of (ISO 26262) functional safety (FuSa) compliance for safety-critical applications is examined, with a focus on the importance of real-time error detection. The paper categorises localisation technologies into four broad groups: memory-dependent, adaptive learning-based, stateless and hybrid localisation technologies. It highlights the strengths and weaknesses of each approach, emphasising the need for data fusion methods to balance cost, accuracy, and robustness in autonomous vehicle localisation. Finally, the paper explores the potential of machine learning, deep learning, and reinforcement learning to improve the robustness, adaptability, and efficiency of localisation systems in dynamic and complex environments.