Fluid Field Estimation and Source Localization by Dynamic Positioning of Autonomous Underwater Sensor Nodes

Motivation

The world's oceans have been a crucial part for life on earth in the past and present, but will also determine our future. They include the least explored areas on earth due to the high cost and the risks involved in offshore and deep-sea activities. Recent technological advances in underwater robotics promise to open up groundbreaking possibilities in offshore areas, such as oceanographic research, offshore wind and deep sea mining. Autonomous underwater vehicles (AUVs) are becoming increasingly important and will play a prominent role in the future. Today mostly single, highly specialized and very expensive vehicles are deployed. They usually have a low degree of autonomy and need a large vessel to be deployed. In order to optimize the use of these scarce resources, the next generation of autonomous underwater vehicles has to be able to evaluate mission performance and coordinate itself with other vehicles working on the same or similar task.

We develop an AUV system with a computational fluid dynamics (CFD) simulation in the loop. This allows the modeling of realistic environmental dynamics by the simulation of the fluid field and an estimation of sinks, sources (of heat, pollution, etc) and boundary conditions. Also, the results from the fluid dynamics simulation are fed back into the path planner for the AUVs. Depending on the mission objectives, regions with a high likelihood for sources can be of higher or lower interest as regions with a high field variable variance. Task allocation and consensus algorithms decide how the team of vehicles navigates within the field to satisfy both, robustness and a fast mission execution.

Vehicle Hardware

The current underwater vehicle setup represents a very flexible and robust control platform. We use a Pixhawk board with a 168 MHz Cortex M4F CPU (256 KB RAM, 2 MB Flash) and onboard sensors (3D ACC / Gyro / MAG / Baro). PX4 is an independent, open-source, open-hardware project (BSD licensed) for mobile robotics applications. The PX4 platform runs NuttX, a small footprint real-time operating system (RTOS), which provides a POSIX-style environment. The PX4 middleware runs on top of the operating system and provides device drivers and a micro object request broker (uORB) for asynchronous communication. Our mumAUV control stack is a custom, BSD licensed underwater vehicle control stack, providing remote controlled and fully autonomous (in progress) operations for our underwater vehicle Hardware.

Publikationen

• IEEE/RSJ, Hrsg.: A Micro Underwater Vehicle for Swarm Applications, 2015.
• A. Hackbarth and J.K. Hedrick and E. Kreuzer and E. Solowjow: A Nonlinear Estimator for Mobile Source Localization Applications. In Proc. of the 8th European Nonlinear Dynamics Conference, Vienna, Austria, 2014.
• Hackbarth, Axel and Hedrick, J. Karl and Kreuzer, Edwin and Solowjow, Eugen: A Nonlinear Estimator for Mobile Source Localization Applications. In 8th European Nonlinear Dynamics Conference, Wien, 2014.
• Bessa W.; Kreuzer, E.; Krumm, L. ; Pick, M.-A. ; Solowjow, E.: Adaptive Fuzzy Sliding Mode Controller and Observer for a Dive Cell. PAMM Proceedings in Applied Mathematics and Mechanics, 15(1): S. 263-264, 2015.
• A. Hackbarth and J.K. Hedrick and E. Kreuzer and E. Solowjow: Localization of a Diffusive Source Using Rao-Blackwellized Particle Filtering. Robotics: Science and Systems Conference, Workshop on Autonomous Control, Adaptation, and Learning for Underwater Vehicles, Berkeley, CA, USA, 2014.
• Bessa, Wallace M. and Hackbarth, Axel and Kreuzer, Edwin and Radisch, Christian: State and Parameter Estimation of an Electro-Hydraulic Servo System. In 8th European Nonlinear Dynamics Conference, Wien, 2014.
• A. R. Geist, A. Hackbarth, E. Kreuzer, V. Rausch, M. Sankur and E. Solowjow: Towards a Hyperbolic Acoustic One-Way Localization System for Underwater Swarm Robotics. IEEE International Conference on Robotics and Automation (ICRA), 2016.
• E. Kreuzer and E. Solowjow: Learning Environmental Fields with Micro Underwater Vehicles: A Path Integral - Gaussian Markov Random Field Approach. Autonomous Robots, 2017.
• A.D. Buchan, E. Solowjow, D-A. Duecker, and E. Kreuzer: Low-Cost Monocular Localization with Active Markers for Micro Autonomous Underwater Vehicles. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2017.
• W. Bessa, J. Lange, E. Kreuzer, M.-A. Pick, and E. Solowjow: Design and Adaptive Depth Control of a Micro Diving Agent. IEEE Robotics and Automation Letters & IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2017.
• H.-C. Lin, E. Solowjow, E. Kreuzer, and M. Tomizuka: A Data-Driven Exploratory Approach for Level Curve Estimation with Autonomous Underwater Agents. ASME Dynamic Systems and Control Conference (DSCC), 2017.
• A. R. Geist, A. Hansen, E. Solowjow, S. Yang, E. Kreuzer, and J.K. Hedrick: Data Collection for Robust End-to-End Lateral Vehicle Control. ASME Dynamic Systems and Control Conference (DSCC), 2017.
• V. Rausch, A. Hansen, E. Solowjow, E. Kreuzer, J.K. Hedrick: Learning a Deep Neural Net Policy for End-to-End Control of Autonomous Vehicles. IEEE American Control Conference (ACC), 2017.
• D.-A. Duecker, A.R. Geist, M. Hengeler, E. Kreuzer, M.-A. Pick, V. Rausch, E. Solowjow: Embedded Spherical Localization for Micro Underwater Vehicles based on Attenuation of Electro-Magnetic Carrier Signals. Sensors, 2017.
• T. Johannink, E. Kreuzer, and E. Solowjow: Sealing of Machine Parts and Modules Manufactured with Desktop 3D Printers. Konstruktion , 2017.