Elaborazione grafica a tema del progetto
23/09/2025
Research & Innovation

PoliTO researchers develop an “artificial ear”

Observing nature can lead to the development of new technologies: inspired by the cochlea, the inner ear organ that not only transmits sounds but also analyzes them to allow the brain to interpret them, the Nonlinear Elasticity and Metamaterials (Nemo) research group at the Department of Applied Science and Technology (DISAT) at PoliTO has developed a device that can recognize speech efficiently and with low energy consumption.

The results were recently published in the international journal Advanced Intelligent Systems, in an article co-authored by Politecnico researchers and professors Paolo H. Beoletto, Carlo Ricciardi, Federico Bosia and Antonio S. Gliozzi, together with Gianluca Milano of the National Institute of Metrological Research (INRiM).

The device, based on a 3D-printed spiral metamaterial, functions as a truly “artificial ear”: it does not merely record sounds, but processes them directly, transforming them into maps called “tonograms.” In this way, it can distinguish different words with an accuracy comparable to traditional digital methods, but with much lower energy consumption.

This innovation opens up new possibilities for portable devices, voice assistants, and edge computing, which need to run for long periods on small batteries, but also to increase the autonomy of cochlear implants, which need to be constantly recharged.

“In our research group, we believe that drawing inspiration from nature is fundamental to develop new technologies and smart materials. Biological systems have been optimized by billions of years of evolution, which is why they perform so well” explains Paolo Han Beoletto, lead author of the study. "The cochlea, for example, is not just a simple sensor, but a natural sound processor, capable of extracting relevant information before it even reaches the brain. Reproducing this function with artificial materials means reducing the digital calculations required and therefore the energy consumption of voice recognition technologies. This is an important step if we want a future in which billions of connected devices can interact with us in an environmentally sustainable way."