Integration of triboelectric sensing and machine learning

Robotic or automated manufacturing helps streamline workflow and advance manufacturing processes, allowing companies to stay globally competitive. One of the key factors in these processes is quality control which often requires validating or verifying the materials used in the processes.

Image Credit: Shutterstock.com/ Willyam Bradberry

Now researchers from the Institute of Nanoenergy and Nanosystems in Beijing have developed a ‘smart finger’ that can identify materials using ‘triboelectric’ sensors that test its ability to gain or losing electrons, as well as determining other characteristics such as its roughness, without risk of causing damage. Published in the journal Scientists progressthe team describes how they developed the triboelectric smart finger.

In principle, as each material has different abilities to gain or lose electrons, a unique triboelectric fingerprint output will be generated when the triboelectric sensor is in contact with the measured object.

Dan Luo, CAS Center for Excellence in Nanoscience, Beijing Institute of Nanoenergy and Nanosystems

Machine learning and quantification of material parameters

Humans rely on haptic feedback as an essential sensory function when in direct contact or communication with the surrounding environment.

Tactile perception comes from the response of subcutaneous tactile corpuscles to different environmental stimuli and from the brain’s recognition of afferent signals via nerve fibers.

Dan Luo, CAS Center for Excellence in Nanoscience, Beijing Institute of Nanoenergy and Nanosystems

In general, quantifying material parameters precisely at the psychological level of tactile perception can be a challenge when it comes to identifying the texture and roughness of a material. The smart finger developed by the researchers also uses machine learning to help improve triboelectric tactile perception in the mechanism and help human users of such systems.

Additionally, the team said they developed a smart finger that surpassed human tactile perception, enabling accurate identification of material type and roughness through the integration of triboelectric sensing and machine learning.

They claim the smart finger has at least 90% accuracy when sensing the material surface, suggesting the technology has potential use in automating robotic manufacturing tasks, including sorting materials. materials and quality control assessments.

Develop a smart finger

In recent years, various efforts have been made to design sensors or devices capable of identifying materials based on various strategies, such as computer vision, thermal conductivity, ultrasound, etc. As a result, computer systems and robots are becoming increasingly good at interacting with the world around them, but they will also need a sense of touch before they can reach their full potential.

When tested on a variety of samples, such as plastic, wood, silicon and glass, the smart finger demonstrated an average accuracy of 96.8% and an accuracy of at least 90% for all materials.

The system incorporates machine learning-based data analysis with four small square sensors, each made of a different plastic polymer that has been specifically chosen for its electrically conductive properties. The sensors are housed in a housing that looks like a finger, hence the name “smart finger”.

When the sensors come into contact with an object’s surface, the electrons in each square begin to interact with the surface in a different way, which the team was then able to measure.

Each of the sensors is connected to a processor and an organic light-emitting diode (OLED) display, which highlights the type of material being assessed. Indeed, the researchers were able to quantify tactile psychological parameters using the triboelectric effect, which could define a new paradigm in modeling human tactile perception.

Actual and future scenarios

In a real scenario, the processor could be directly integrated into a manufacturing control mechanism. The smart fingers could then perform quality checks and determine if the products are up to manufacturing standards.

Beyond the industrial/manufacturing setting, smart fingers could also be used in prosthetics as robotic limbs with a sense of touch to improve manipulation techniques and manipulation of objects.

The team also aims to introduce other sensors into the system, including pressure, temperature and humidity sensors, to help improve touch simulation.

In the future, artificial intelligence chips will be integrated into smart fingers to make them “smarter” and give them the ability to process data independently of the computer.

Dan Luo, CAS Center for Excellence in Nanoscience, Beijing Institute of Nanoenergy and Nanosystems

References and further reading

Qu, X. and Liu, Z., et al., (2022) Smart finger with artificial tactile perception for material identification based on triboelectric sensing. Scientists progress, [online] 8(31). Available at: https://www.science.org/doi/10.1126/sciadv.abq2521

Disclaimer: The views expressed herein are those of the author expressed privately and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork, the owner and operator of this website. This disclaimer forms part of the terms of use of this website.

Comments are closed.