Ego Vehicle Localization with GPS-IMU Fusion: A Key to Effective Scenario Generation

• Published in: 2025 3rd International Conference on Communication, Security, and Artificial Intelligence (ICCSAI) Vol. 3; pp. 1678 - 1680
• Main Authors: Verma, Mukesh Kumar, Panchal, Puneet, Hariharan, Udhayakumar, Rehan, Sobit, Singh, Dinesh, Yadav, Manohar
• Format: Conference Proceeding
• Language: English
• Published: IEEE 04.04.2025
• Subjects: Accuracy; Autonomous vehicles; Global Positioning System; Global Positioning System (GPS); Inertial Measurement Unit (IMU); Inertial navigation; Location awareness; Measurement units; Multisensor Data Fusion; Reliability; Scenario generation; Testing; Trajectory
• DOI: 10.1109/ICCSAI64074.2025.11063982

In order to provide realistic virtual driving scenarios for autonomous vehicle testing and simulation, accurate ego vehicle localisation is essential. In order to improve ego vehicle localisation, this paper combines data from the Inertial Measurement Unit (IMU) with the Global Positioning System (GPS). Although GPS provide position data, it is not a good tool for determining a vehicle's orientation because it is frequently subject to bias and noise. On the other hand, IMUs, which are made up of gyroscopes and accelerometers, provide relative information about angular velocity and acceleration but do not provide absolute position data. The suggested fusion approach overcomes the drawbacks of each sensor separately by integrating their strengths to provide an accurate estimation of the ego vehicle's position and orientation. The high-precision ego trajectory that is produced can subsequently be used to create dependable and captivating virtual driving situations. These scenarios facilitate the development and testing of autonomous vehicles by allowing for in-depth analysis and visualisation of recorded real-world circumstances. The study also discusses possible drift problems in trajectory prediction and offers lane recognition and high-definition map-based corrective techniques. This method offers improved simulation accuracy for the development of autonomous driving systems while providing a strong basis for scenario synthesis.