Position, Velocity and Time Estimation in GNSS using MATLAB
implementing Maximum Likelihood Estimation (MLE) within the Extended Kalman Filter (EKF)
Project Overview
Leveraged Adaptive Kalman Filter along with Maximum Likelihood Estimation that mitigated 30% of noise. Designed and implemented MATLAB Simulation of Moving Window Averages increasing accuracy to 87%
Project Preview
Confidentiality
The Project source code and some other details are classified due to association with Space Application Center, ISRO due to some internal policies. However, I can share some concepts that were used in creation of the project.
Abstract
For navigation, surveying, and other uses that call for precise location data, the "Global Navigation Satellite System" (GNSS) plays the crucial tool. The Satellites constellations known as the Global Navigation Satellite System, (GNSS) send Signals to GNSS receivers on the ground or in orbit. GNSS receivers to calculate their position, speed and time use these signals. GNSS finds many applications in the position's determination of satellites in LEO and GEO orbits also. In our study focusing on enhancing Global Navigation Satellite Systems (GNSS) precision, we forged a theoretical foundation for implementing Maximum Likelihood Estimation (MLE) within the Extended Kalman Filter (EKF). Our groundwork involved a meticulous exploration of MLE's pivotal role in refining positioning accuracy by maximizing the likelihood function, specifically concerning GNSS residuals. While direct implementation awaited, our efforts concentrated on laying robust groundwork, including the practical implementation of a complex moving window average in MATLAB. This moving window average function serves as a critical component in our envisioned MLE equations for updating CNN (Covariance of the Clock Noise) and CEE (Covariance of the Code and Carrier Phase) matrices within the Extended Kalman Filter framework, thereby advancing the prospects of achieving superior GNSS-based positioning accuracy.