Memristor chips could see the patterns of pixels

May 29, 2017, 5:01 p.m. By: Pranjal Kumar


Researchers at the University of Michigan has developed ‘memristor chip’ that could handle the complex data in very less time. It could process complex data such as images and video faster than any available technology.

This technology is based on how the mammals see their surroundings.

This chip could have the large-scale implementation in the sectors like autonomous system such as self-driving cars.

Professor of electrical engineering, Wei Lu said,” The complexity of the system is increasing day by day. The amount of data that the machine has to handle have also increased in recent times. To compensate this increase, we need to improve the design of circuits.”

This chip works on the principal of ‘sparse coding’ to coax it’s 32 by 32 array of memristor to analyze several photos.

Memristor is an electric resistor that has storage capability. They could regulate the current on the basis of the history of voltage applied to them. They could also store and process the data simultaneously.

In the conventional system, logic and memory operation were located at different positions. This used to make system more power driven, expensive and large.

But the networks of memristors could solve this problem. They could perform many operations at the same time without having to move the data around. As a result, they could large data in parallel. This makes them capable of advanced machine learning.

They could also be used in deep neural networks.

Image Source: Electric component news