Project Proposed: This project, acquiring two REMUS 100 autonomous underwater vehicles (AUVs), addresses the research challenge of multiple cooperating vehicles systems (MCVS), which involves a fleet of AUVs and autonomous safety vehicles (ASVs). Since successful operation of any MCVD often rests upon efficient communication among the vehicles, the dynamic mission planning problem (including search and classify and mapping capabilities) is studied utilizing a fleet of heterogeneous AUVs and ASVs. This challenging problem, considering underwater acoustic communication, navigation, and sensing constraints, differs from those usually studied and reported in the underwater literature. The work is contrasted to unmanned aerial or ground vehicles in which communication and navigation capabilities are generally NOT considered significant bottlenecks. The project capitalizes on two previous NSF achievements: - An underwater vehicle network simulatorsand - A location aware source routing (LASR) protocol developed for multiple communicating AUVs subject to realistic underwater communication constraints. Nonlinear and stochastic variations with the environmental conditions and sensor characteristics are expected. Because a closed form solution for an optimal controller design is not anticipated, given that control performance tends to be cost-prohibitive, when completely via experimentation. Hence, given the wide range of mission and environmental scenarios and controller objectives, the problem will be studied combining modeling, simulation, and experimentation. Some of the simulated results will be validated by carrying out targeted at-sea field experiments using the instrumentation. Local cost definitions will be explored as a means to constrain individual vehicle?s maneuvering and cooperation. Attention is paid to maximizing the overall mission efficiency while minimizing the impact of uncertainty. Expectations include development of: - An advanced event-base planning and control algorithms to improve robustness of communication and navigation uncertainty, throughput, and bandwidth efficiency, - An advanced multiple cooperating vehicle modeling software to support mission performance analysis, and - A science base for multiple AUVs and undersea acoustic networks.
Broader Impacts: The research problems constitute ideal topics for theses and dissertations. This multidisciplinary field will be integrated into the existing course curriculum, providing valuable theoretical and simulation knowledge to the students, as well as hands-on experiences. Using underwater robotics as a primary domain application, the existing effort, sustaining the teachers and students with interest in math and science, will be continued giving special consideration to female and underrepresented groups. Workshops, competitions, and summer classes are also planned to expose students the logistics of marine vehicles. The understanding gain is likely to provide insights to the U.S. Navy for missions using a fleet of autonomous underwater vehicles to better search and classify in homeland security missions.
The main goal of this project is to investigate the challenges of having multiple unmanned underwater and surface robotic platforms to work together autonomously and achieve functionality that is greater than the sum of the individual parts. Key challenges of this research are underwater communication between the platforms, information sharing and control of these platforms. The MRI instrumentation consists of two REMUS 100 autonomous underwater vehicles, a gateway buoy for command and communication for these vehicles. Our achievements consist of four separate efforts: 1) We have developed an in-house underwater acoustic tracking system that can be used to track the position of these vehicles; 2) we have developed a collaborative mission software layer (CMSL) so that important information can be exchanged between the vehicles. To minimize the user’s effort in developing new missions, a graphical user interface (GUI) was developed with pre-programmed vehicle behaviors that can be selected to suit a particular mission type; 3) Two high-level control algorithms were developed as part of this research: Rendezvous Docking (RD) and Waypoint-Heading (WPH) algorithms. The RD algorithm was developed for an USV to track a stern-approaching underwater vehicle. The concept is based on "quarterback-receiver" pattern where the unmanned surface vehicle has some prior knowledge about where the underwater vehicle is to traverse during the recovery phase of the mission, and has the means to acoustically track the UUV in terms of range and bearing information. The surface vehicle will adjust its track so that it is inline with the expected line of travel of the UUV. The WPH algorithm was developed to guide the USV to a specified waypoint with a specified heading at the waypoint. Twelve graduate students were benefited from working on their research using the MRI assets. Five undergraduate students conducted a Directed Independent Study using these vehicles. In addition, two outreach week-long workshops were organized so that 20 high school students had the opportunity to learn about the strate-of-the-art underwater vehicle technology and how to run simple missions at a nearby lake. These vehicles have been integrated into a senior data analysis course so undergraduate students learn how these vehicles work and how to analyze the data collected using these vehicles.
