Enhancing GNSS and INS Integration for High-Precision and Continuous Positioning Using Odometer Trained With TSFNN

Abstract

Autonomous driving has become a hot topic in both the academic and industrial sectors in recent years, given the growing market demand. Among the various components of autonomous driving systems (ADS), the positioning and navigation module is of vital importance as it provides accurate information regarding a vehicle’s position in terms of integrity, continuity, reliability, and accuracy. Although the Global Navigation Satellite System (GNSS) and Inertial Navigation System (INS) have been widely used, their accuracy and reliability cannot be guaranteed during GNSS outages, such as in tunnels or underground parking areas. To address this limitation, an odometer sensor has been deemed as an ideal solution to aid INS during these outages. However, integrating INS and the odometer can lead to increasing navigation errors due to inaccurate odometer velocity measurement. To tackle this issue, this study proposes a T-S fuzzy neural network (TSFNN) to correct the odometer velocity measurement. Furthermore, a real-time GNSS/INS/Odometer integrated navigation system employing the global optimal filtering (GOF) algorithm is presented to provide continuous and accurate navigation solutions, even in the absence of GNSS signals. Road tests have been conducted to validate the effectiveness of the proposed system, and the results indicate satisfactory performance. Notably, improvements of 69.4%, 88.5% and 73.6% were achieved in test segments #2, #4 and #5, respectively.

Keywords

• Autonomous driving
• GNSS/INS/odometer integrated navigation system
• GOF
• GNSS outages
• MEMS inertial sensors
• TSFNN