Integrating Machine Learning-based GNSS Signal Processing and 3DMA for Positioning in Mobile Multipath Urban Environments

Abstract

There has been an increasing demand for accurate positioning in urban environments from many applications. Location based services (LBS) and autonomous vehicles are at the forefront for such applications. Although there are many options for positioning sensors, incorporating GNSS is desirable for reasons like accuracy, availability, coverage, ease of use, and low cost. However, GNSS suffers from degraded positioning accuracy in urban environments because of multipath signals. There have been many efforts directed at mitigating the multipath problem. Multipath mitigation can work on the measurement level or the signal processing level. Approaches aided by 3D mapping (3DMA) have emerged as a promising solution to the multipath problem. Conventional 3DMA approaches work on the measurement level. This seminar will present an alternative approach that integrates 3DMA into a signal processing level algorithm. The algorithm is a machine learning-based that utilizes the tracked GNSS received signal parameters to provide an enhanced positioning accuracy in multipath urban environments. In mobile urban environments, when a receiver is moving, a signal status continues to change. A signal status can be either LOS, multipath, NLOS, or invisible. For example, a LOS signal can become NLOS signal when the surrounding structures block the direct signal and reflect it. A tracking module can follow the change in the parameters of the received signal and continue to track the NLOS signal, without knowing the change in the signal status. This leads to errors in the tracked signal parameters and  consequently errors in the positioning solution. This seminar will cover the following points: (1) how to detect a change in the signal status from the received signal parameters using machine learning; (2) how to utilize the detected signal status in the position computation; and (3) how the detected signal status is integrated into 3DMA. Experimental testing and results that verify the performance of the presented work will conclude the seminar. 

Speaker

Prof. Nesreen I. Ziedan holds the position of Professor and Department Chair at the Computer and Systems Engineering Department within the Faculty of Engineering at Zagazig University, Egypt. She earned her PhD in Electrical and Computer Engineering from Purdue University, USA. With several U.S. patents, a published book, and numerous papers on GNSS receiver design and processing, her expertise in the field is widely recognized. Additionally, Prof. Ziedan serves as a valued member of the editorial board and as a reviewer for various journals focusing on electrical engineering and navigation.