Transcript: Researchers Use Smartphone Tech To Detect Drunk Walking
David Erickson: Do you like Tom Waits, BL?
BL Ochman: I do.
David Erickson: One of my favorite Tom Waits songs–one of the best songs ever written–is The Piano Has Been Drinking (Not Me). The piano has been drinking, not me, not me. I love that song. It’s just brilliant.
The Smartphone Has Been Drinking
David Erickson: So, I say that because, “the smartphone has been drinking, not me, not me.” This story is by Wired‘s Matt Simon, who reports that your smartphone should be able to figure out that you’ve been drinking now.
David Erickson: Apparently, there are subtle and not-so-subtle changes in your gait, that can betray your alcohol level. So if you’re carrying your smartphone around, the onboard accelerometer can pick up those changes.
Smartphones Can Detect Drunk Walking
David Erickson: There’s some scientists at the University of Pittsburgh with published research showing that they can use a smartphone’s motion data to detect if a subject is intoxicated with an average accuracy of 93%.
Drunk Gait Methodology
David Erickson: So what they did is they strapped an ordinary smartphone to the lower backs of the subjects in the study and got their baseline gaits. They had people walk 10 steps forward, turn around, walk 10 steps back, and then they repeated that when they were drunk. And they got them drunk really quickly because there’s only so much time to conduct a study, right?
David Erickson: Fun study, yeah,
Drunk Walking Experiment
David Erickson: It’s not like you can leisurely get drunk. Get drunk right away. So they did that and over the course of an hour–once an hour, over the course the seven hours, they had them doing that as their blood alcohol level piqued and waned because different people have different responses.
If you have a higher tolerance for alcohol you might not walk differently and blah, blah, blah. So there’s a bunch of different factors that came into that. But the smartphone was logging the motion of all the subjects during this study. They use mathematical models to compare the gaits and arrived at the accuracy.
David Erickson: There’s another Worcester Polytechnic Institute scientist named Emmanuel Agu who, he and his co-workers have used smartphones to measure “intoxication sway”–because that’s the technical term for it–and they even built a drunken-detecting app called AlcoGait. Is anybody gonna voluntarily use this app?, is my question.
Aid In Addiction Recovery
David Erickson: But by monitoring the gaits–so there’s some positive reasons for this–health experts could better predict when a person’s problematic drinking might occur. Addictions are not–
BL Ochman: And take the car keys away.
David Erickson: And take your car keys away, exactly. But also addictions are notoriously difficult to treat. It takes continuous treatment and case management and all that stuff over the course of a year, basically; where people go to their appointments, they don’t drink, they do all these things they’re supposed to do in order to get sober.
And so this could play a role in helping people overcome their addictions, I imagine it probably would be applicable to other types of addictions. I imagine you probably have a different gait if you’re high on, you know, heroin or something, you know. So–
BL Ochman: Yeah, you wouldn’t be moving.
David Erickson: There you go. But I don’t want my phone knowing I’m drunk!