WSJ: Making Senses

There was a special insert in the Friday July 8, 2021 issue of the WSJ titled the “Future of Everything.” In the center spread was an article titled “Making Senses” by Angus Loten & Kevin Hand. The article widened my mind scope on the “smart” and “eco-smart” project themes that are a part of the YCISL ITW-DTI workshops being offered this summer. Advances in sensory devices and dataset building with the alignment to human preferences should find amazing applications in sustainability design thinking.

Transparency. We are very familiar with computer vision and image recognition, but most of this is with regards to shapes and little else in terms of physical characteristics. The advancement described in the article about transparency will lead to better depth perception and recognition of vessel contents.

Taste. The need for sensory systems in food storage (eg, refrigerators that can detect bad food) is taking a long time. Let’s hope the advancements in “electronic tongue” technology means we are near the point where we can greatly reduce wasted food. A premium market for this may be in wine since wine can improve or decay depending on various factors.

Touch. I have seen many fascinating videos of product assembly lines and food production lines to know that there is a lot of mechanical tools involved. Upgrades to these and other applications could help healthcare as well as more common needs.

Smell. The AI sensory application to robotic noses and identifying vapors can be used in enhancing or neutralizing odors. There are so many places where pleasing odors could enhance productivity and removal of objectionable odors could reduce distractions.

Hearing. Noise cancellation hearing devices are quite popular presently. Ever tried using one during air travel? Isolating voices using AI could process sounds so that no idea is lost among the “chatter” and everyone is heard. Ever been to a call center?

The common thread in these advancing areas is the collection of reference data. Even so, there is so much variability that it is challenging to consider anything as a reference. Things change very quickly and conditions can tweak channels. For example, in the food waste application I mentioned above: how much reference data would be needed to tell whether it’s time to discard your jar of sauerkraut?


Comments are closed.