A new AI material can learn behaviors and adapt to different circumstances

The artificial intelligence material could be used in aircraft wings or building structures in the future due to its flexibility.
Brittney Grimes
Photo of a mechanical neural network (MNN).
Photo of a mechanical neural network (MNN).

UCLA/Flexible Research Group 

Mechanical engineers at the University of California, Los Angeles (UCLA) have created a new type of material that uses artificial intelligence to learn behaviors over extended periods of time.

The study was published today in the journal Science Robotics.

The benefits of the material in different industries

The material consists of a structural system made up of tunable beams that can change shape and behaviors over time. This alteration is a response to dynamic conditions, the study said. The research team discovered this AI could be applied to constructing buildings, airplanes and imaging technology.

“This research introduces and demonstrates an artificial intelligent material that can learn to exhibit the desired behaviors and properties upon increased exposure to ambient conditions,” said Jonathan Hopkins, lead researcher of the study and mechanical and aerospace engineering professor at the UCLA Samueli School of Engineering.

He stated that the principles used in this research are the same as those used within machine learning, which gives the material its ability to adapt.

The example given in the study mentions the possibility of using the substance in aircraft wings. The AI material would have the ability to learn and change into the shape of the wings. This would occur based on the wind patterns during a flight to create more flexibility and efficiency.

The research team also explained the benefits of infusing buildings with the material, improving stability during earthquakes, hurricanes or other disasters.

Creating the material

The research team used concepts from already existing artificial neural networks (ANNs) to create the material. ANNs are the algorithms that drive machine learning. They call the material mechanical neural network (MNN). The MNN consists of independently tunable beams positioned in a triangular lattice pattern. Each separate beam contains a “voice coil, strain gauges and flexures that enable the beam to change its length, adapt to its changing environment in real time and interact with other beams in the system.” This allows for the material to maintain its adaptability within the environment.

Most Popular

Purpose of the voice coil, strain gauges and flexures

The study explains the purpose of the voice coil, strain gauges and flexures. “The voice coil, which gets its name from its original use in speakers to convert magnetic fields into mechanical motion, initiates the fine-tuned compression or expansion in response to new forces placed on the beam,” the study states.

“The strain gauge is responsible for collecting data from the beam’s motion used in the algorithm to control the learning behavior. The flexures essentially act as flexible joints among the moveable beams to connect the system.” The three work together to allow for elasticity and flexibility.

The finalization includes an optimization algorithm that regulates the whole system by taking data from the strain gauges and creating values of rigidity to control how the network should adapt. It controls the amount of force that should be applied. There are also cameras attached on the outer nodes of the system to check the validity of the strain gauge system.

Future goals

The current size of the MNN system is similar to a microwave oven. However, the research team wants to simplify the concept so that thousands of the systems can be created on a much smaller scale to be applied to various duties.