Low-Cost Integrated Navigation Systems Using Nonlinear Observer Theory

New algorithms for strapdown inertial navigation systems have been developed and tested using unmanned aerial vehicles (UAVs). The new methods reduce the computational footprint significantly compared to state-of-the-art methods such as the ?Kalman filter?. The main focus on the research has been to develop effective algorithms or so-called nonlinear observers for loosely and tightly navigation systems, which can replace linear design techniques (Kalman filters) without performance degradation. Hence, small embedded computers with limited power demands can be developed and integrated into unmanned vehicle control systems. The strapdown inertial navigation systems have been developed and tested onboard UAVs using global navigation satellite systems (GNSS) and optical cameras for aiding. This add extra robustness to the system since an UAV can continue to fly and navigate after sensor failure. Robust navigation systems are necessary in order to operate autonomous vehicles in harsh environments such as the Artic. The strapdown inertial navigation systems are currently being benchmarked against commercial solutions. This is of great interest since it is important to compare the performance of the new algorithms with state-of-the-art algorithms. The commercial systems are much more expensive than our set-up, which uses low-cost MEMS-based inertial measurements units (IMUs). MEMS-based IMUs cost a fraction of fiber optic gyros and ring laser gyros and the overall goal is to demonstrate accurate low-cost navigation systems using MEMS technology. The results (2014-2017) have been been presented at 10 international conferences. The main theoretical results have been published in 8 high-quality journals and 4 book chapter. The core strapdown navigation algorithm was been published in Automatica in 2016, which is considered to be the best journal in automatic control. In addition, papers have been published in high-quality journals as IEEE Transactions on Control Systems Technology, IEEE Control Systems Magazine, IEEE Transactions on Aerospace and Electronic Systems, Control Engineering Practice and Journal of Dynamic Systems, Measurement and Control and the International Journal of Control.

The main motivation for the proposal is summarized below: - Develop low-cost integrated navigation using MEMS components with lower computational footprint than the Kalman filter and develop explicit stability guarantees for the nonlinear observers. - De monstrate how nonlinear observers can be implemented on light-weight embedded computers with limited processing capacity. - Nonlinear observer theory, MEMS technology and embedded systems are changing the rules of the game and opens for the development of low-cost compact position, velocity and attitude (PVA) estimation. The project will focus on methods which can be used in as low-cost consumer electronics, cars, navigation systems for autonomous underwater vehicles (AUVs), ships, unmanned aerial vehicle s (UAVs) etc. - Demonstrate that it is possible to develop effective computer algorithms that can be implementation using only a fraction of the source code when compared to a standard EKF implementation. This will simplify implementation, maintenance and software documentation. - Compare and verify the the performance to commercial EKF-based systems in cooperation with Maritime Robotics and FFI. - The project targets education of three PhD and eight MSc candidates (Working Packages 1-3) dealing with: WP1: Integrated observer design with a North-seeking strapdown MEMS-based gyrocompass WP2: Nonlinear tightly integrated filters for attitude, velocity and position WP3: Model-based nonlinear integration filters for INS and position measurements - T he results will be published at conferences and in the top ranked journals in the field.