Fluid field estimation and source localization by dynamic positioning of autonomous underwater sensor nodes

• Applicant: Professor Dr.-Ing. Edwin Kreuzer
• Subject Area: Measurement Systems ; Mechanics ; Mechanical Properties of Metallic Materials and their Microstructural Origins ; Fluid Mechanics
• Term: from 2014 to 2018
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 250508151

In the scope of this research proposal fluid fields with uncertain sources, boundary and initial conditions shall be identified by mobile sensors. A simulation of the fluid field will therefore be assimilated with measurements from the mobile sensor nodes and predicted. The prediction of the uncertain fluid field will then be used to plan trajectories for the mobile sensors which return optimal measurements to improve the accuracy of the fluid field estimation.A system like that can be used to localize heat and pollution sources in primarily unknown fluid domains, but also to verify fluid dynamic calculations. Usually, currents and pollutants in rivers, harbors, estuaries, and oceans are measured by multiple stationary sensors, supported by measurements from manned vessels. With a centrally controlled mobile autonomous sensor network we can achieve a much higher dynamic resolution through optimal positioning of the sensor nodes. Global data assimilation in real-time is indispensable to discover in which areas more measurements are needed and to continuously generate paths and measurement locations for the mobile sensors.An automated process of boundary condition estimation, source identification, and current measurement assures a high flexibility and a fast deployment of the self-organizing underwater vehicles. Completely unknown fluid fields can be analyzed as long as all positions in the domain of interest can be reached by the vessels due to current conditions.At the Institute of Mechanics and Ocean Engineering several preliminary studies have been conducted about the theory of source identification and localization which were implemented successfully in simulations. Additionally, a control system has been developed which calculates the optimal trajectories for each mobile sensor and ensures an efficient coordination of the network. Furthermore, we will collaborate with partner institutions in the USA and Brazil in theoretical and practical parts of this project. Further research is especially needed in source localization and boundary condition estimation, but also the dynamic of the complete system. Because the complete system contains many uncertainties and parts, which cannot be modeled, a practical implementation is part of this research project.The verification of this concept will be done by an experimental setup, in which mobile underwater vehicles equipped with temperature sensors shall identify and localize a heat source in a pump driven flow field in a water tank and automatically detect boundary conditions. The estimated flow fields and its uncertainties are used to generate optimal trajectories for the vehicles. The theory and simulations shall be verified through this experiment in a controlled environment and technical issues and possibilities have to be analyzed before sea trials are conducted

• DFG Programme: Research Grants