Adaptive Filtering on GPS-Aided MEMS-IMU for Optimal Estimation of Ground Vehicle Trajectory

Area of Science:

• Navigation Systems Engineering
• Robotics and Autonomous Systems
• Signal Processing

Background:

• Global Positioning System (GPS) and Inertial Navigation System (INS) fusion is crucial for ground vehicle navigation in military and civilian sectors.
• Low-cost Micro-Electro-Mechanical Systems Inertial Measurement Units (MEMS-IMU) offer accessibility but suffer from lower accuracy compared to high-end systems.
• The rise of autonomous vehicle technology necessitates improved accuracy and reliability in navigation estimation while maintaining cost-effectiveness.

Purpose of the Study:

• To propose and evaluate a novel framework for fusing low-cost MEMS-IMU and GPS data for enhanced ground vehicle navigation.
• To investigate the efficacy of adaptive Kalman filter variants, specifically Sage-Husa and strong tracking algorithms, for sensor fusion.
• To implement and validate inertial sensor error models for achieving reliable and accurate navigation estimations.

Main Methods:

• Development of a fusion framework utilizing adaptive Kalman filters (Sage-Husa and strong tracking) for integrating MEMS-IMU and GPS data.
• Implementation of inertial sensor error models to enhance the accuracy of navigation state estimation.
• Simulation of the proposed fusion approach using real-world navigation data collected from a test vehicle equipped with commercial GPS and MEMS-IMU.

Main Results:

• The proposed adaptive Kalman filter fusion framework demonstrated superior navigation accuracy compared to conventional Kalman filter methods.
• Integration of MEMS-IMU and GPS data using the advanced filtering techniques provided more reliable estimations.
• The implemented error models contributed to the overall accuracy and robustness of the navigation solution.

Conclusions:

• The fusion of commercially available, low-cost MEMS-IMU and GPS using adaptive Kalman filtering presents a viable solution for enhancing ground vehicle navigation accuracy and reliability.
• The Sage-Husa and strong tracking algorithms, coupled with inertial sensor error modeling, offer significant improvements over traditional integration methods.
• This research contributes to the development of cost-effective and dependable navigation systems for autonomous and assisted driving applications.