Multiple Estimation Filters Integration for Sensor Fusion Localization Module

When talking about Autonomous Vehicles (AVs), certain stages must be discussed, communication, perception, localization, navigation, and control. In this paper, localization is the main focus. Localization suffers from challenges in accuracy and reliability when simple and cheap sensors are used. Accuracy is lost and reliability is inconsistent. However, when estimation filters are used, the accuracy and reliability are significantly improved, even with the usage of simple and cheap sensors. In this paper, altitude estimation filters, such as Mahony Filter and Madgwick Filter, are used to improve the orientation angle estimation and correct sensors uncertainties. The Extended Kalman Filter (EKF) is used to estimate the position using odometry and GPS data to have corrected position states. The Usage of Mahony Filter for Yaw angle correction and estimation, Madgwick Filter for Roll and Pitch angles correction and estimation. Finally, the EKF was used for position states estimation. This concludes that the usage of multiple filters and their integration in a system results in desirable outcomes.