The objective of this paper is to estimate smartphones’ location which support services that demand lane-level precision like high-occupancy vehicle (HOV), lane Estimated Time of Arrival (ETA) estimation. We focus on developing a model based on raw location measurements collected in an open sky and light urban roads using datasets collected by hosts from Android smartphones. The application of mobile devices for most software products built for services such as cadastral surveying, mapping surveying applications, and navigation has been increasing due to the cost-effectiveness of GNSS smartphones. This paper aims to bridge the link between the geospatial information of detailed human behavior and the smartphone internet with improved granularity. It fixes the issue with the GNSS/INS integrated navigation system’s degrading data accu- racy during an GNSS signal outage. We aim to improve the currently used GNSS/INS integration algorithm built on the AI approach. The position of a vehicle during a GNSS loss can be predicted utilizing a GNSS/INS integration methodology for land vehicle navigation based on position update architecture (PUA) employing LightGBM regression. It models the connection between INS data and changes in vehicle location using LightGBM.