Dr. Chinese as a little sister

Editor’s note: This article is from the micro-channel public number “qubits” (ID: QbitAI), Author: Xiao Xiao.

Your actions are all under the control of these socks:

Whether you stand on one foot, twist your waist, or walk, these socks can predict your actions through changes in the pressure on your feet:

The most important thing is that the material is simple to manufacture.

There is no need to integrate a large number of high-precision sensors, and the method of knitting clothes can make socks with excellent detection results.

This pair of socks with “super powers”, developed by MIT, is now in the Nature journal:

How did such an accurate forecasting effect be achieved?

Combined with AI, accurately predict various postures

The principle of motion detection of socks is very simple in itself:

Using machine learning to associate human postures with tactile footprints to predict human motion.

This is because when a person changes from one posture to another, due to changes in the pressure on the feet (for example, the center of gravity shifts from the right foot to the left foot), the tactile footprints will also change accordingly.

By collecting a large amount of data, using self-supervised learning, and training the model, we can predict the movement of the human body:

However, how can this fabric be mass-produced while ensuring fine testing?

You must know that in mass production, it is impossible to accurately design the sensor positions one by one.

For this reason, this pair of socks is made of a mixture of two kinds of fabrics.

Includes ordinary textile fiber acrylic yarn, and functional fiber that can sense force.

As shown in the figure, the functional fiber contains force sensing materials.

Therefore, this socks not only retains the softness of the socks itself, but also has the ability to detect movements.

There will be overlaps between the two weaving threads.

Therefore, when the load pressure is sensed, the two lines will squeeze each other, and the sensing material inside can pass through the intersection to perceive the strength of the force.

Previously, many wearable devices required a large number of high-precision sensors to ensure the accuracy of measurement.

But different from the existing wearable electronic devices, the design of this socks can be used in the production of large-scale clothing, which greatly reduces the cost.

However, this type of socks may also have some problems during production.

If manufactured in batches, the sensors of some socks may not work properly.

At this time, it is necessary to combine machine learning algorithms to identify and adjust some sensors in the socks to find out the parts of the sensors that cannot work in time.

The specific test results are indeed notwrong.

For example, twisting the waist. The action predicted by the system is almost the same as the action being performed:

Squat down. The posture predicted by the machine is almost exactly the same as the real posture:

Try walking? Even the gait and stride are basically the same:

Let’s take a look at going down the stairs again. Except for the stride, everything else is basically the same.

Moreover, this smart fabric can not only be used on socks.

There are also vests, knees, gloves series

In addition to smart socks, researchers from MIT have also come up with various series of smart fabrics such as vests, knees, and gloves:

This magical suit is basically made of similar fabrics in the same principle as socks.

After sensing the pressure, using the pressure generated by the overlap of the two weaving threads, we can predict the actions of the person.

This suit can be used for sports and rehabilitation training of patients.

If the family allows it, it can be used to monitor the diseaseThe health status of the person in the hospital and the patient so that they do not fall or lose consciousness when they are alone.

‍

‍

Of course, it can also be worn by robots.

By learning the data in this suit, the machine may be able to understand how humans perform actions, and thus learn to make a variety of different actions.

With this suit, the robot may be able to “have” the sense of touch.

The first author of this study is Yiyue Luo, a Chinese PhD sister from MIT.

Yiyue Luo Middle School graduated from Guangzhou Zhixin Middle School. He once studied at the University of Illinois at Urbana-Champaign and is currently a third-year PhD student at MIT.

Her research direction is manufacturing, smart fabrics and tactile sensing.

The second paper is Yunzhu Li, a fourth-year doctoral student at MIT. He graduated from the Computer Science Department of Peking University. His research interests are computer vision, machine learning and robotics.

In addition, Pratyusha Sharma, Wan Shou, Kui Wu, Michael Foshey, Beichen Li, Tomás Palacios, Antonio Torralba, Wojciech Matusik also participated in the research of the paper.

Netizen: Socks also have WiFi

A netizen thought hard and gave this pair of AI socks a name: Neuro socks…

Some netizens joked: My socks will have WiFi soon!

However, some netizens expressed concern.

He believes that this type of socks may monitor our activities in the future: this is not good.

Do you want to have a pair of socks like this?

Thesis address:

https://www.nature.com/articles/s41928-021-00558-0

Reference link:

[1]https://www.csail.mit.edu/news/smart-clothes-can-measure-your-movements [2]https://twitter.com/MIT_CSAIL/status/1375505517917786119 [3] https://yyueluo.com/