| Abstract |
A multitude of technologies, like the Internet of Things, unmanned aerial vehicles, self-driving cars, as well as smart cities, are increasingly reliant on navigation systems. However, it is a complex task to achieve a dependable, seamless, and accurate solution using a single-location navigation system. For example, global navigation satellite system has limitations in indoor environments. As an alternative, multi-sensor integrated systems are gaining attention. They utilize the combined strengths of various sensors to address the limitations of individual technologies. This study conducts an in-depth examination of the multi-sensor data used in integrated navigation systems over the past decade. The chapter organizes and discusses various navigation systems from three viewpoints: (1) data sources, (2) system architecture algorithms, and (3) use cases. These are divided into further subcategories such as learning and analytic-based fusion. In our discussion on learning-based fusion versus analytic-based fusion, we showcase numerous deep learning, reinforcement learning, unsupervised, and supervised learning algorithms used in multi-sensor integrated navigation systems. We delve into the design considerations of these integrated systems from various perspectives and categorize their real-world applications. Finally, we explore potential future directions for their application and research. © 2025 Sanjeevi Ramakrishnan, Anuradha Jayaraman, Sandeep Tripathi, Prashantkumar B. Sathvara. All rights reserved. |
