8:30 a.m. - 10:15 a.m.	Vision-Based Navigation: Techniques and Applications	Dr. Laura Ruotsalainen
10:30 a.m. - 12:15 p.m.	Estimation of Noise Parameters in State Space Models	Dr. Jindrich Dunik and Dr. Ondrej Straka
12:15 p.m. - 1:15 p.m.: Lunch provided for registered tutorial attendees
1:15 p.m. - 3:00 p.m.	GNSS in the Age of Machine Learning	Dr. Pau Closas
3:15 p.m. - 5:00 p.m.	LEO PNT: From Theory to Implementation	Dr. Zak Kassas

Vision-Based Navigation: Techniques and Applications

Time: Monday, April 28, 8:30 a.m. - 10:15 a.m.

This tutorial provides an introduction to vision-based navigation, a rapidly advancing field at the intersection of computer vision, robotics, and artificial intelligence. Vision-based navigation enables autonomous systems, such as drones and self-driving vehicles, to interpret and navigate their environment using visual data. Additionally, it offers potential benefits for pedestrians in infrastructure-free indoor navigation. We will cover key topics, including feature extraction, object recognition, visual odometry, simultaneous localization and mapping (SLAM), and path planning. Participants will learn how to process and utilize visual information for accurate positioning and navigation in dynamic environments. We will also discuss recent machine learning advancements in this field and examine why they have yet to be widely implemented in real-world applications.

Dr. Laura Ruotsalainen is a professor in computer science at the University of Helsinki and a steering group member for the Finnish Center for Artificial Intelligence (FCAI).

---

Estimation of Noise Parameters in State Space Models

Time: Monday, April 28, 10:30 a.m. - 12:15 p.m.

The tutorial introduces a more than six decades-long history as well as recent advances and state-of-the-art methods for estimating the properties of the stochastic part of the state-space model. In particular, estimation of the state and measurement noise means, covariance matrices, and other parameters are treated. The tutorial covers all major groups of noise statistics estimation methods, including correlation, maximum likelihood, covariance matching, machine learning, and Bayesian methods. The methods are introduced in the unified framework highlighting their basic ideas, key properties, and assumptions. Algorithms of individual methods are described and analyzed to provide a basic understanding of their nature and similarities. The performance of the methods is compared using a numerical illustration and available MATLAB implementations.

Dr. Jindrich Duník is an associate professor of cybernetics at the University of West Bohemia and a senior scientist at Honeywell Aerospace.

Dr. Ondrej Straka is an associate professor of cybernetics at the University of West Bohemia working in the areas of state estimation, fault detection, and GNSS navigation.

---

12:15 p.m. - 1:15 p.m.: Lunch provided for registered tutorial attendees

---

GNSS in the Age of Machine Learning

Time: Monday, April 28, 1:15 p.m. - 3:00 p.m.

The use of machine learning - based solutions is one of the emerging trend in GNSS to to enhance the accuracy, reliability, and robustness of PNT systems. The tutorial discusses the role and opportunities given by such techniques in the GNSS field, discussing its use with respect to analytical approaches. In particular, deep learning algorithms are data-driven models that, instead of using complex-to-derive physics-based models, use large datasets to learn the correlations in the data. This has been applied to address several typical issues of GNSS, such as multipath identification, robust acquisition, spoofing and jamming mitigation, scintillation detection and classification.

Dr. Pau Closas, associate professor, at Northeastern University, specializes in signal processing for GNSS and wireless communications. He has received several honors, including the ION Early Achievement Award.

---

LEO PNT: From Theory to Implementation

Time: Monday, April 28, 3:15 p.m. - 5:00 p.m.

This tutorial provides an introduction to positioning, navigation, and timing (PNT) with low Earth orbit (LEO) satellites, highlighting basic theory and implementation considerations. We will overview the challenges associated with exploiting megaconstellation LEO satellites for PNT purposes, namely their unknown signals, poorly known ephemerides, loose synchronization and oscillator instability, and propagation effects. We will present a cognitive software-defined receiver (SDR) to deal with unknown LEO signals. We will formulate and compare several LEO PNT frameworks to deal with their poorly known ephemerides and timing: open-loop, adaptive, machine learning, differential, and simultaneous tracking and navigation (STAN). We will present experimental demonstrations of LEO PNT on ground and aerial vehicles, navigating with multi-constellation LEO satellite signals (Starlink, OneWeb, Orbcomm, Iridium, and NOAA) to an unprecedented level of accuracy.

Dr. Zak Kassas is the TRC Endowed Chair in Intelligent Transportation Systems and professor at The Ohio State University. He is also the director of the U.S. Department of Transportation Center for Automated Vehicles Research with Multimodal AssurEd Navigation (CARMEN) .

---