Improving Positioning Accuracy for Connected Vehicles: Integration of Low-Cost INS with GNSS

Nowadays, accurate positioning systems are becoming increasingly indispensable, mainly for applications such as Connected and Autonomous Vehicles (CAV), Vehicle-to-Vehicle (V2V) and Vehicle-to-Everything (V2X) communications, Precision Agriculture (PA), among others. The main technologies used for navigation purposes in those applications are of the Dead-Reckoning (DR)- and Position Fixing (PF)-types. Inertial Navigation Systems (INS) are examples of DR technology that provide solutions at high sampling rates. However, they accumulate errors over extended periods of time. Conversely, the Global Navigation Satellite Systems (GNSS) are examples of PF technology, which provide a bounded position solution, with global coverage at a lower sampling rate, but are susceptible to interference and signal losses. One recommended technique to maintain the advantages of both INS and GNSS is sensor fusion, i.e., INS/GNSS integration. This work aims to present practical results obtained with low-cost INS and GNSS receiver integration for CAV applications. The primary contribution is to showcase results of the so-called Loosely Coupled (LC) INS/GNSS integration, which exhibits accuracy improvements (w.r.t. standalone GNSS navigation solution), as validated through dynamic experiments. Tests reveal enhancements of 31.29%, 23.29%, and 21.79% in terms of horizontal, vertical, and total position accuracy, respectively, when compared to standalone GNSS. Moreover, the LC-INS/GNSS strategy can deliver 1.5 meter horizontal and 3.0 meter vertical position solutions at 22,5% and 52,49% of probability, respectively. Standalone GNSS can only deliver 0,42% and 16,32% of probability under the same conditions. These results confirm the relevance of INS and GNSS fusion to achieve CAV application requirements.