Résumé : Vehicle localization is essential for autonomous vehicle (AV) navigation and Advanced Driver Assistance Systems (ADAS). Accurate vehicle localization is often achieved via expensive inertial navigation systems or by employing compute-intensive vision processing (LiDAR/camera) to augment the low-cost and noisy inertial sensors. Here we have developed a framework for fusing the information obtained from a smart infrastructure node (ix-node) with the autonomous vehicles on-board localization engine to estimate the robust and accurate pose of the ego-vehicle even with cheap inertial sensors. A smart ix-node is typically used to augment the perception capability of an autonomous vehicle, especially when the onboard perception sensors of AVs are blocked by the dynamic and static objects in the environment thereby making them ineffectual. In this work, we utilize this perception output from an ix-node to increase the localization accuracy of the AV. The fusion of ix-node perception output with the vehicle's low-cost inertial sensors allows us to perform reliable vehicle localization without the need for relying on expensive inertial navigation systems or compute-intensive vision processing onboard the AVs. The proposed approach has been tested on real-world datasets collected from a test track in Ann Arbor, Michigan. Detailed analysis of the experimental results shows that incorporating ix-node data improves localization performance.