Developing Fleets of Autonomous Underwater Vehicles

Fleets of AUVs can have a significant impact on the ability to search for mines in very shallow water, surf zone, and beach regions.

Autonomous underwater vehicles (AUVs) have a demonstrated capability to collect valuable data for scientific and military purposes. Historically, individual vehicles have been used. To reduce the overall time and cost of acquiring data over large areas, multiple vehicles must be used. A fleet of five AUVs, capable of underwater commendation, was fabricated. These AUVs include small submarines, referred to as “swimmers,” and small, two-tracked vehicles, referred to as “crawlers.” The control and communication algorithms developed in this work will enable AUVs to use formations to search for mines and to communicate with each other in order to implement cooperative behavior. Languages and logics were developed to enable collaborative operations among the vehicles.

The AUVs operated autonomously, in that they enabled their operations on their own, initiated and constrained by underwater acoustic communication and navigation against a general behavioral background provided by programmed logics. The operations were not choreographed in advance and programmed into the machines, nor were they the result of intervention by an operator on the surface. The vehicles performed deployment, formation-flying, vehicle replacement, divert-to-point of interest, and leader replacement behaviors.

The submarines were used to test the cooperative control algorithms and navigation procedures. Each vehicle was 1 meter long and 10 cm in diameter, and was equipped with a Woods Hole Acoustic micro-modem for communication and navigation. A distributed approach for control on each vehicle was used. The five functional units communicated using an Ethernet hub. The instrumentation unit collected and configured information from all of the sensors on the AUV except those requiring underwater communications.

The internal sensors included a battery monitor, a water detector, and a thermometer to monitor the internal temperature of the AUV. The external sensors consisted of a GPS unit, an electronic compass, a pressure sensor to determine the depth, and an accelerometer to determine pitch and roll. All sensors except the GPS were polled every 0.25 seconds. The mission control unit performed all of the control calculations necessary for navigation of the AUV and recorded all incoming and outgoing Ethernet packets to a 128-Mb XDRAM card. The 128-Mb data capacity of the log provided for approximately 30 hours of data storage. The locomotion unit controlled the speed of the propulsion motor and set the position of the control surfaces.

A Woods Hole Modem was used for underwater acoustic communication and navigation. When on the surface, wireless communications were accomplished with a 900-MHz MaxStream 9Xstream radio modem. A Linksys model WCF12 wireless CompactFlash 802.1 lb Ethernet card was used to allow configuration of the AUV from the base station.

Languages, logics, and algorithms were developed to enable collaborative operations among AUVs. Automatic formation control algorithms enable multiple AUVs to search cooperatively for mines in close formation. Organized in a swimmer-trailer formation, with one swimmer designated the leader, and programmed to conduct coordinated searches in a lawnmower pattern, various autonomous behaviors associated with large-area MCM have been modeled and simulated. These behaviors include deployment, vehicle replacement, leader replacement, divert to point of interest, and map building. The behaviors are supported by a version of AUVish, an agent communication language designed for the vehicles. Vehicle replacement is controlled by the leader with a 32-byte message broadcast.