The Advanced Marine Systems Laboratory is pursuing a wide range of research projects, all with the objective of advancing the leading edge of marine systems technology. In addition, the laboratory serves to bring together the various engineering specialties in the Ocean Engineering Department together with outside collaboration to develop multi-disciplinary solutions to challenging engineering problems. A primary thrust of the AMS Lab’s efforts is to use embedded computation and control systems to make systems more automated or "intelligent". Examples of these research efforts include the development of Autonomous Underwater Vehicles, Autonomous Oceanographic Sampling Network, Smart Sensors, Naval Shipboard Automation, and Intelligent Control Systems. The Advanced Marine Systems Lab has received support for its research from the Office of Naval Research, NSF, NOAA, navy laboratories, and the State of Florida. The Lab has 3 faculty members with collaborative ties to several others both at FAU and other institutions. The Lab supports 10 staff engineers along with several graduate and undergraduate research assistants. Graduate degrees are offered for both M.S. and Ph.D. degrees in Ocean Engineering with a specialization in intelligent systems. Graduate classes offered by associated faculty members include Underwater Vehicle Systems, Intelligent Systems, & Fuzzy Control Systems.

RESEARCH AREAS

Autonomous Underwater Vehicles

The following funded research projects are currently being pursued in the area of Autonomous Underwater Vehicles:

• Autonomous Oceanographic Sampling Network Technology Development. A network of AUVs, mooring, and ships are to be connected with acoustic and RF modems to perform synoptic, cooperative adaptive sampling of the 3D coastal ocean environment. The network is connected through RF and acoustic links to the Internet to allow remote monitoring of the system.

• Mine reconnaissance using multiple AUVs. The simultaneous operation of multiple AUVs with appropriate acoustic and optical sensors is used to perform rapid environmental assessment and detect mine like objects.

• Adverse weather measurements using AUVs. The goal of this project is to deploy an AUV prior to the arrival of stormy weather. The AUV sleeps on the sea floor docked to a power recharge and data upload station. The AUV wakes up once the storm has commenced to perform a sampling mission and then re-docks for eventual recovery. This approach may be one of the only ways to collect shallow water oceanographic data during a storm.

• Sensor system support: The AUVs developed by the AMS lab are being used as platforms for a variety of new underwater sensor systems and payloads. The modular design of the current generation of AUV is catalyzing the parallel development by several institutions of sensor systems and payloads. This effort is revolutionizing the sampling methods of coastal oceanography.

• Tactical Environmental Database and Synthetic Ocean Environment: This projects seeks to combine and coordinate the information generated by the AOSN collection efforts to provide playback and nowcasting capabilities for AUV base ocean measurements.

• AUV platform and technology development. In this project the next generation mini-modular AUV is being designed and built. The goal is to reduce the size and operations cost by an order of magnitude. Advanced molded injected plastics are being used to manufacture the AUVs interchangeable pressure vessels and wet sections.

Intelligent Control Systems

The following funded research projects are currently being pursued in the area of Intelligent Control Systems:

• AUV flight control using nonlinear, fuzzy, or neural network techniques. Different types of controllers have been developed and tested for the autopilot of an AUV. The controllers have been developed and tested at sea on working vehicles and also using a 6 DOF model of the AUVs running in a 3D graphical simulation.

• Development of Computational Methods for Optimal Nonlinear Control Systems Using Cell Mapping. This project involved developing new ways to automatically generate nonlinear control systems that optimize a given cost function. The controllers can be fuzzy or of some other nonlinear type. Cell mapping is a computationally efficient means of dynamic programming.

• Unified Nonlinear Control Systems for AUV. This project seeks to develop a unified architecture including both low level and high level control systems. The approach uses a hybrid hierarchical and layered control architectures based on elastic constraint propagation. This project also includes the development of a generic AUV simulation system using OpenGL that provides hardware in the loop simulation of AUV control systems.

• Multi-sensor fusion for AUV Navigation using either Fuzzy Heuristic and Asynchronous Kalman Filter Techniques. In this project a simplified position estimator was quickly developed that fuses the different types of velocity and position data available to the AUV using a heuristic rule based filter. The heuristic filter allowed good AUV navigation to be quickly achievable while a more complex Asynchronous Kalman filter is being developed to fuss the data more accurately but at more computational cost AUV planning systems using fuzzy cased based reasoning. The goal of this project is to develop efficient on line mission planning systems for AUVs using fuzzy case based reasoning. This combines the learning by example techniques of case base reasoning with the interpolative matching and reasoning methods of fuzzy logic.

Naval Shipboard Automation

The following funded research projects are currently being pursued in the area of Intelligent Control Systems:

• The general approach this research is taking is the development of techniques for component level intelligent distributed control systems (CLIDCS) for the automation of auxiliary ship systems. These systems consist of a network of low cost embedded processors communication over a serial network using the LONTalk protocol. Some research projects include:

• Underwater imaging techniques are also being developed to generate high resolution, three dimensional images of the ocean bottom terrain and the sub-bottom in real time. These images are for use by autonomous unmanned vehicles for navigation and surveying.

• Dependable Topologies with Network Fragment Healing in Component Level Intelligent Distributed Control Systems (CLIDCS) for Naval Shipboard Automation. In this project a network of multiply connected rings is used to provide a dependable damage resistant topology. The network can reconfigure itself to repair or "heal" fragments in the network due to loss, failure, or damage.

• An Automated Chill Water Demonstration System Using CLIDCS. The CLIDCS concept was applied to a benchtop working model of ship’s chill water system to demonstrate the technology.