Maker Pro

Sensing Module Developed by Purdue University Contributes to Machine Learning Precision for Manufacturing, EVs, and Smart Homes

December 22, 2019 by Luke James

Purdue University engineers have developed a new sensing technology that works with machine learning for precision data interpretation, and could improve machine learning precision for EVs, smart homes, and more.

Engineers from Purdue University have recently developed a sensing module that uses machine learning to monitor electric currents and gather information, including energy usage, system errors, and the best approaches to manufacturing. The Purdue team has stated that their sensor module is the first of its kind that is non-invasive, safe, and offers levels of precision that, until now, have never been achieved. 

Kaushik Roy, Purdue University’s Edward G. Tiedemann Jr. Distinguished Professor of Electrical and Computer Engineering, who partly leads the research team, said, “We have created the first-of-its-kind current sensor that is non-invasive, safe and much more precise than other options,”


Purdue’s Non-invasive Sensing Module

Current sensing modules for use in applications such as electric vehicles are somewhat limited and include the use of a resistor as a current sensor, or the use of a non-invasive Hall sensor, which cannot measure small currents.

The Purdue team’s sensor is different. It uses a machine learning algorithm that helps the sensor interpret and gather data such as energy usage, current problems and errors, and the best approaches to manufacturing. 

The new sensing technology is based on a new kind of sensor, named a magnet-resistive sensor, and a unique noise shaping technique that suppressed noise and increases signal sensitivity. It allows current to be measured in a non-invasive fashion from a few milliamps to hundreds of amps. In many cases, Purdue’s technology is more accurate than current invasive ones. 


A sensing module developed by Purdue University researchers.

The sensing module developed by Purdue University researchers. Image Credit:


Byunghoo Jung, a professor of electrical and computer engineering at Purdue’s College of Engineering, said that some of the potential applications include, “…machine learning to train manufacturing robots, provide precise tips for homeowners on cutting down their energy usage or help diagnose issues with electric vehicles and scooters.”

Other advantages of the Purdue sensor include easy installation and maintenance. It can also be used to transmit measured current information to any computer or computing system using Bluetooth, USB, and other methods, and it can also be trained through machine learning to detect precise information. 

The Purdue research engineers have already worked with the Purdue Research Foundation Office of Technology Commercialization to patent their design. Presently, the team is on the hunt for partners to license the technology. 

Related Content


You May Also Like