Smart Structures:

Engineering Materials that can Think and Adapt to their Environments

Since the advent of large scale integrated circuits and personal computers based on them, the tendency has been for things to become smarter and smaller. Today, nearly every consumer electronic device has not only been reduced in size, but also caters more and more to our needs. The VCR can be set to tape things when we aren't home. Personal computers can answer and route our calls when we aren't there.

Despite these advances in technology, most of these devices still need to be programmed by a person to work - they need some supervision. But what if they could think for themselves, assess their state and make decisions based upon the data they collect? What if they could adjust to our needs without us having to tell them what to do?

"Smart structures" have the potential to do some of the above, and they are the focus of a large research project at Maryland. In the Department of Electrical Engineering and the Institute for Systems Research (ISR), two faculty members are conducting research on smart structures, as part of a Multidisciplinary Research Initiative (MURI) grant from the Army Research Office. The grant, worth $3 million over three years, establishes The Center for Dynamics and Control of Smart Structures (CDCSS), a three-university center conducting research on the design and control of smart structures. Maryland project participants include center Co-Director Prof. P. S. Krishnaprasad (EE, ISR), Prof. John Baras (EE, ISR), and Prof. Stuart Antman (Mathematics). Researchers from Harvard University (Prof. Roger Brockett, Prof. Howard Stone) and Boston University (Prof. John Baillieul, Prof. Tom Bifano) are partners in this project.

Smart structures are composed of materials that can determine their present state, decide what is the most desirable state, and carry out an appropriate response in a controlled manner. They are made up of: (1) sensors, which describe the physical state of their environment; (2) actuators, which adapt to their environment; and (3) a system or network which handles the transfer of information and real-time computation.

A hybrid motor prototype, which combines piezoelectric and magnetorestrictive actuators to generate motion. This motor was developed in the Intelligent Servosystems Lab hear at Maryland.

"The problem is that once you get a material into a structure, it will do the same thing forever-it has no smarts," said Krishnaprasad. "One cannot "dial-up" a prescribed behavior in ordinary materials. People want their needs to be met in different ways, in different variables, and at different times. For this type of functionality, a smart structure has to have a changeable quality-in a controllable way."

At the consumer level, the proliferation of smart structures could lead to devices such as seats which automatically adjust to the weight and shape of a person that sits on them, or "smart" insulin pumps for diabetics, Krishnaprasad explained. On a larger scale, smart aircraft wings could be designed to adjust their shape according to airflow, or robotic arms could be designed which accurately sense, feel, and respond to their environments in a human-safe manner.

"Smart Structures will affect the way people can interact with their environment," said Krishnaprasad, "especially to minimize noise, adjust indoor climate, or interact with various electromechanical gadgets. Once you have the basic principles and technology, especially sensors and actuators," Krishnaprasad added, "you can do all kinds of interesting things."

For instance, actuators embedded in devices may be able to respond to temperature or electromagnetic fields by changing their position, shape, stiffness, and/or other mechanical characteristics-depending on the stimuli. Four types of actuators are commonly used in smart structures: shape-memory alloys, piezoelectric devices, magnetostrictive materials, and magneto-rheological fluids. The last three are currently part of the research conducted at Maryland.

Piezoelectric actuators respond to applied voltages by expanding and contracting. They are efficient for dampening vibrations and controlling stress in materials. Lead zirconate titanate (PZT) is widely used. Magnetostrictive materials react to magnetic fields, and are also used to dampen vibrations. Terfenol-D (an alloy of Terbium, Iron and Dysprosium), is commonly used as this type of material. Magneto-rheological fluids respond to magnetic fields, and have been used as tuneable dampers in moving parts-such as clutches, brakes and robotic arms.

The modular ducted flow experiment at Boston University, used to study flow control using microactuators. (Courtesy: Prof. John Ballieul, Mechanical and Aerospace Engineering, BY).

Sensors gather information in smart structure systems. They are typically either optical fibers or piezoelectric materials. Piezoelectric polymers, such as polyvinylidene, can be applied in thin films and bonded to a variety of surfaces. Such polymers have proven to be sensitive enough to be useful in applications to read Braille.

• Smart Structure research within the CDCSS collaboration focuses on six areas:
• Research on Very Small-Scale Servo Systems
• Modeling of Materials for Sensing and Control
• Systems of Embedded Micro-actuators for the Control of Flow
• Over Airfoils and in Arrays of Microvalves
• Issues in the Control of Fluids on Small Length Scales
• The Communications Theory of Very Large-Scale Device Networks
• Numerical Methods and CAD

"We want to be able to build things that move - things that open and close, manipulate, control vibrations or the flow of a gas," said Krishnaprasad. "We want actuators to be able to do these things while they receive information from distributed sensors. Making them on a small scale, and making lots of them is a challenge."

"When you put many actuators and sensors into a host material and want it to do something, it leads to a 'wiring problem,'" Krishnaprasad explained, "because any one part of the system has to know what every other part of the system is doing. We are trying to find out what the scientific and mathematical principles are that will help us overcome these wiring problems."

Krishnaprasad's team will look at smart structures at all scales-using mathematical modeling, control, numerical methods, and CAD.

One example of smart structure research Krishnaprasad discussed was the manipulation of the shape of an airfoil. Practically, he said, it would be better for a wing to have a different shape at different times-one shape for cruising and drag minimization, and one for extra lift when the plane is taking off. With smart structures, designers could use a lighter but more responsive material.

"The idea is that you can program the behavior of a material either on the fly, or ahead of time in a controlled manner," said Krishnaprasad. "We are working with controls and systems, and developing mathematical models to use smart materials in dependable ways."

As smart structure technology becomes more mature, it could spur a revolution in the way engineers design things. Furthermore, as this technology begins to meet other growing technologies such as artificial intelligence, virtually every material structure and electronic device we come in contact with may someday be "smart."