Biometrics Makes Waves

University and military researchers are developing technologies based on biology that would allow unmanned undersea vehicles (UUVs) to be agile enough to dock in submarine tubes.

It takes a lot of brains to develop new technologies, and one U.S. Navy project is capitalizing on another type of brainpower. Navy researchers are examining work conducted jointly by the New York University Medical School and Russia’s Nizhny Novgorod State University and Institute for Applied Sciences that uses brain activity as the model for controlling movement in unmanned undersea vehicles. The advances culled from this research could support better designs for autonomous underwater vehicles that could hunt mines, deliver and retrieve sensors, track ship movement or gather plume samples.

The work is taking place at the Naval Undersea Warfare Center (NUWC), Newport, Rhode Island, with help from the Office of Naval Research (ONR), Arlington, Virginia. Techniques being explored would be used on unmanned undersea vehicles (UUVs) to solve two problems that limit their effectiveness. While current designs allow the Navy to conduct operations that might otherwise be impossible, the service is interested in improving the maneuverability and decreasing the noise of UUVs so the craft can conduct new types of missions in the future.

Dr. Thomas McKenna, program officer, division of cognitive, neural and social science and technology, Department of Human Systems Science and Technology, ONR, explains that the objective of the current work is to enable precise and quiet maneuvering as well as long endurance. With this capability, future UUVs could conduct surveillance missions without being detected or hover in one area for reconnaissance missions then be retrieved and possibly dock in a submarine tube.

To accomplish these tasks, however, requires basic research in a number of areas, and researchers are turning to biology for at least some of the answers. McKenna notes that in comparing the movements of fish and those of manmade vehicles, scientists have observed that neither has the advantage in speed when moving in one direction. However, when it comes to maneuverability and turning radius, fish are more agile than manmade vehicles by a large margin. To bring the two a little closer together, scientists have to determine both the structures that would facilitate movement and the connections to make these structures respond quickly.

McKenna explains that one facet the engineers are examining in the structures area is the high-lift principle. The principle emerged from the study of the wings of flies. The combination of the pitching and heaving motion results in enhanced lift when compared to fixed wings—up to five times higher. To improve propulsion, the researchers are exploring how to build foils in combination with motors that can mimic this superior lift. “This gives a higher thrust or can operate at a lower power, therefore emitting lower sound energy and increasing duration. If a number of these foils work in concert, there is the capability for much greater maneuverability,” McKenna states.

The NUWC is working on the high-lift actuators project. The challenge is that these actuators would require a number of degrees of freedom so that the pitching and heaving could be varied among the thrusters situated on various parts of the UUV. “It’s a nonlinear control problem, and we’re taking a look at a new, very rapid and adaptive nonlinear controller based on brain circuitry,” he offers. The prototype is currently the size of a circuit board; however, McKenna allows that the plan is to move to very large-scale integrated circuits in the future.

This nonlinear controller, called the universal control system, is based on how the cerebellum and a related structure known as the inferior olive work in the human body. By examining the physiology of this biological system, researchers at New York University Medical School and Nizhny Novgorod State University and Institute for Applied Sciences have built a model of chaotic oscillators.

The universal control system, which is an array of oscillators, can be divided into smaller elements, each controlling a different actuator. In addition, each oscillator can instantly shift phase in relation to other oscillators. The phase can then be used to encode the movement and send the control signals to the actuators. “It turns out you can control a very large number of actuators and adapt the control very rapidly. We’re building the electronic hardware prototype of the system, and we’re evaluating it now to determine how good that control is compared to the conventional way of controlling actuators,” McKenna relates.

This evaluation is taking place in the form of a “little competition,” he reveals. During the next year, two scale models of UUVs that feature the high-lift foils will be built. The NUWC will work with one of the models using a series of conventional and artificial neural network controllers. Researchers at New York University will employ their brain-control circuit system to assess the effectiveness of their technology. “Once we have those results, we’ll have some idea about whether the Navy should proceed with just the high-lift foils or the high-lift foils in conjunction with the brain-based controller,” McKenna explains.

The biomimetics aspect of the controller project is based on brain research that has been underway for about 20 years, he allows. It began with recording arrays in the cerebellum to examine the synchronized activity of neurons and different parts of the cerebellum as an animal’s brain performed motor activities. As a result of this work, researchers gained two key pieces of information that can help mimic brain activity in hardware today. First, it became apparent that transient synchronous structures were emerging in conjunction with a particular movement, McKenna notes. “The second part of the information was this brain nucleus called the inferior olive that the cerebellum projects to. In the inferior olive, those neurons are a big cluster that are electronically coupled and that synchronously oscillate at 10 hertz. Human movements key off that 10-hertz rhythm. The output of the inferior olive then influences the motor program—the sequence that activates your muscles,” he relates.

Transferring this biologically automatic process to the computer system environment poses some challenges. For example, McKenna notes that the neural-based controller is conducting analog computations, so one of the challenges is carrying this out in hardware where traditional circuitry is digital.

From the mechanical standpoint, the high-lift foils in the near term must be adapted to work with rotating motors. This could limit their range of motion and does not completely solve the issue of noise.

But researchers are exploring solutions to this problem as well. “In the future, we’re looking at linear actuators that are like actual muscles. There are a number of efforts, about six that ONR is supporting, to build muscle-like actuators out of various compounds—in particular electro-active polymers. But they are not quite ready to be used at this scale yet. All of them have different issues in terms of how much electricity they require and whether they operate fast enough. So we haven’t hit on just the right formula yet,” McKenna states. Once this issue is resolved, rotating motors would not be required and the noise would decrease tremendously, he adds.

Development of this material could affect the design of future UUVs. “Now you can apply the forces in lots of different directions and can start to act like a real fin. For example, you might actively control the actual shape of the fin cycle to cycle. There are projects examining how a fish’s pectoral fins are designed at all levels—molecular, attachment and muscles. And people are building artificial pectoral fins because we are just starting to understand from the hydrodynamics point of view what those fins are actually doing,” he notes.

A mobile autonomous research vehicle is being used to examine the advantages of biomimetic approaches to increase the agility and reduce the noise of UUVs.

As envisioned, these UUVs would be semi-autonomous. Missions would be preprogrammed into the vehicle; however, the information gathered by onboard sensors could be shared very quickly with the propulsion system, so they could adapt quickly. Coupled with a design that closely resembles the agility of fish, these UUVs could respond automatically to unforeseen changes in water currents. In addition, they would be able to hover without drifting off course. Precise navigation, including a tight turning radius, also would allow maneuvers required for hull inspections and docking, McKenna shares.

Work on this project has been going on intermittently for the past eight years. ONR has been collaborating with the Naval Sea Systems Command, which is responsible for UUVs, and McKenna anticipates that by late next year the researchers will have determined how well these approaches will work.

Web Resources
Office of Naval Research: www.onr.navy.mil/default.asp
Naval Undersea Warfare Center: www.nuwc.navy.mil
Naval Sea Systems Command: www.navsea.navy.mil
Chief of Naval Operations, Submarine Warfare Division: www.chinfo.navy.mil/navpalib/cno/n87/future/uuvs.html