Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have developed a device that can detect our inner emotional states using wireless signals, called the “EQ-radio”. The EQ-Radio sends wireless signals that bounces off a person’s body and back to the device. This way it is able to capture subtle changes in breathing and heart rhythms, without the use on-body sensors. The CSAIL researchers have tested the EQ-Radio and the system is 87% accurate when assessing if a person is excited, happy, sad or angry.
The EQ-radio uses the same carrier frequency as WiFI, only with a 1000 times smaller amount of power. Using the signals that are send back to the device, the device can measure heartbeats with the same accuracy as an ECG monitor (with a margin of error of 0.3 percent). The device studies the heartbeats and breathing patterns and matches this with a certain behaviour. This is done by a machine-learning algorithm. Of course, people respond differently to certain emotions, so the researchers taught the machine to identify unique responses to emotions by examining 30 participants.
As already mentioned, the technology doesn’t require on-body sensors, nor does it use facial recognition software. Both these technologies have their pitfalls and the researches think that the EQ-Radio could be a great solution for companies who want to integrate emotional intelligence with their products.
Dina Katabi, the MIT professor who led the development of this technology, suggest that the device could be used in varying ways across the entertainment, consumer advertising and healthcare industry. Imagine integrating it with a smart TV, this way advertisers can come to know more about the responses of consumers. It could also be used by film makers or game designers to see how people respond to a movie or game.
What do you think about the potential uses of this technology? Would you like to use it yourself?
Sources:
http://news.mit.edu/2016/detecting-emotions-with-wireless-signals-0920
http://gizmodo.com/this-device-can-wirelessly-detect-your-emotions-1786802527
Dear Anneke, thank you for your blog! Very interesting research. I think this can be used mainly to analyze customer’s emotions with regards to specific advertisements and entertainment videos, and will provide more accurate responses than for example a survey customers have to fill in. Thus, this can greatly improve marketing efforts of firms. However, I feel that that are more emotions that could be added than just the four they have right now, which would also make the EQ-radio more accurate. Of course, I would suggest the EQ-radio is only used when people give permission, and not just randomly. I definitely think the technology has a lot of potential, but I doubt I would use it myself.
I think this is a fascinating piece of technology that requires futher research to ride the learning curve, as it has massive potential, especially in the entertainment industry. Cable and digital television providers can for example use this technology to assess the customer’s emotions and better target the customers by creating personalized offers. However, questions may be raised as to how privacy concerns may limit full development and deployment of this technology as it may be conceived as too intrusive. Nevertheless, I see great potential as conventional satisfaction measurement tools like surveys might be redundant and customer feedback can be processed real-time, even when the customer is unaware of this to improve end user experience and lower (marketing) costs for companies. I can also see how this technology might be used in a smart home setting, where a smart home application analyzes emotions and translates this into observable action, like dimming of lights or opening curtains for example whenever the user is feeling depressed.
Hey Anneke, thanks for sharing this blog!
I am really enthusiastic of the idea of emotional recognition by robots/electronic since I see a huge potential in various industries. However, I am not so sure if only recognizing an emotion by heartbeat is the correct way.
For example stress and excitement can both cause a high heart rate but are two totally different emotions. What I would see as a next step in emotional recognition, is that it should be comined with facial emotion recognition, a subject in which we are quite far with discovering already (http://nordicapis.com/20-emotion-recognition-apis-that-will-leave-you-impressed-and-concerned/). By combining these two both ‘outside’ and ‘inside’ characteristics of emotion can be measured and we will thus create a better picture of the emotion.
I am curious where this research will take us in the future!