Alec Moore on The Effects of Olfaction on VR Training of Assembly Tasks
Become a Patron! Support The Voices of VR Podcast Patreon
Theme music: “Fatality” by Tigoolio
Subscribe to the Voices of VR podcast.
Rough Transcript
[00:00:05.452] Kent Bye: The Voices of VR Podcast.
[00:00:12.037] Alec Moore: I'm Alec Moore. I'm at the University of Texas at Dallas. I'm working in the Future Immersive Virtual Environments Lab. And what my poster here is about is we had a training task which involved first subjecting users to trying to build a LEGO model with the instruction booklet in front of them. And then they would return 24 hours later and try to reconstruct the model from memory. Now what we were testing was the effects of olfaction on procedural memory, which is surprisingly not very tested. Most olfaction studies are dealing with memory, deal with facts and kind of static factual information and not really on like procedural tasks like assembly and building. So the way that we structured this study was we had on the first day the users were either subjected to no smell, a lavender scent, or a gasoline scent. And on the second day, all of the users were subjected to the gasoline scent because that was kind of like our realistic study, like the testing format. And what we found was actually that having the gasoline smell the first day, you might think that having the same smell on both days would provide you with a better recollection of the memory, but all of the users with the gasoline smell performed statistically worse than the other two smells. Now, our reasoning for that is that we think that the gasoline scent was so bad that it served as a distractor to prevent users from being able to encode the memories properly in order to be able to recall it better at a later date. So, what we think this means for virtual reality and training transfer is that virtual reality provides the unique opportunity to let you train somebody for an environment that might not smell good, but without that negative impact of having the smell in the first place.
[00:01:51.603] Kent Bye: I see. So if you have to work in an oil rig or something where it just smells terrible, that if you're on site getting trained, then that wouldn't be perhaps as effective as doing it in virtual reality is what it sounds like.
[00:02:03.011] Alec Moore: Exactly. Exactly. We think that a person being trained like on site at an oil rig might be distracted by the bad smell at the oil rig. And so an ideal scenario might be to train them in a virtual environment without that smell or even with a pleasant smell.
[00:02:16.540] Kent Bye: Now, in terms of having a pleasant smell, while learning, is it sort of an indiscriminate, like, smell something good and you'll do better than no smell at all?
[00:02:24.755] Alec Moore: Well, it's sort of indiscriminate. We didn't really test for positive smells. What we were testing for originally was really more of the fidelity of the smell. So our results were actually pretty surprising. The reason why we're saying bad smell versus good smell is that we had no smell in lavender and gasoline. And no one really commented on the lavender, but more than 10 users complained directly about the gasoline smell. They said it smelled very bad. And as the experimenter, I can tell you that putting up with that smell for more than a month Awful.
[00:02:57.115] Kent Bye: What is the process of actually dispersing these gasoline smells?
[00:03:01.199] Alec Moore: So in a previous paper from our lab we developed a scent emitting device which is composed of an Arduino, a PC fan, and a container of essential oil. And so this rig is basically placed in front of the user and they're told to keep their face above that so the smell disperses up directly into their nose.
[00:03:19.814] Kent Bye: And so you had to create an essential oil of gasoline and then blow it in someone's face?
[00:03:26.018] Alec Moore: Well, we actually didn't have to create it. There's a company out there that makes an essential oil that's designed to smell like gasoline.
[00:03:35.204] Kent Bye: Wow. OK. You learn something every day. So what were some of the references and research that you're kind of building on here?
[00:03:43.300] Alec Moore: Well, we don't have a lot of references or prior work based on assembly tasks or procedural memory with Olfaction. Most of the prior work is based on static factual memory when it comes to Olfaction. So, all of our reference work is based on that, so we kind of... I mean, if you're trying to treat it like the same kind of memory, it kind of juxtaposes that to a degree. But we think that with this different classification of memory, where encoding is very important for the proceduralness of the memory, that it's basically a fundamentally different type of learning that's required.
[00:04:19.157] Kent Bye: I see. So talk a bit about more of the psychological research that's been done in terms of differences of those learnings between doing procedural versus just recalling facts.
[00:04:28.326] Alec Moore: So, let's see, I can't think off the top of my head of references involving the difference in the two types of memory, but I do know that what we were initially expecting to see was kind of like the Proust effect, where you kind of have one smell that immediately brings you back to a prior experience, and we really didn't see that at all. I see.
[00:04:50.340] Kent Bye: And so what was the task of them building? And then what were the numbers in terms of being able to statistically measure that in terms of error or the time that it took to complete the task?
[00:04:59.788] Alec Moore: So we measured the amount of time that it took to complete each task and how long they spent on each step. As you probably know, in a LEGO manual, each step is numbered. And we organized all of the LEGO pieces by which step they were so users didn't have to spend time trying to search for that. So we were able to remove that possible effect. And then users were allowed to make up to three mistakes just in the placement, and we would mark those down as errors. And then if they made that third error, then we counted that as a critical error, and the experimenter would perform the step for them and give it back to the user.
[00:05:34.056] Kent Bye: I see. So just to kind of help them move along and to complete the task. Yes. OK. Great. And so when they were exposed to gasoline and they were making more errors when you were measuring them again, it sounds like.
[00:05:45.965] Alec Moore: Yeah, if they had to smell the Octane smell on that first day, they pretty consistently didn't do well on the second day, even though it was the same smell. But learning it in a, I guess, more enjoyable environment really helped them out for trying to perform the task of building the LEGO from memory.
[00:06:02.262] Kent Bye: Yeah, and at GDC there was something called feel real which was kind of this dystopian type of gas mask thing that Injected smells directly into your face and a lot of people didn't really like that But it makes me think that perhaps there's been a bad taste with a consumer VR at least when using smells and cooperation with VR but From your perspective, what do you see as the potential for combining that smell haptic feedback into another sense within the virtual reality experience?
[00:06:32.317] Alec Moore: Well, I think that you can definitely increase the immersiveness. But what we're showing is that even though you may increase how immersed the user is, it might not necessarily increase the training transfer for an industrial application.
[00:06:44.584] Kent Bye: OK, great. And finally, what's next for you moving forward with this type of research?
[00:06:48.531] Alec Moore: Well, next for me, I'm more or less passing this research on to the next guy who can put up with the smells. I'm focusing more of my work now on some other aspects of virtual reality, but I think that this is some important work and I think that we're going to see a lot more work about olfaction and procedural learning tasks and training transfer come out soon. And so what are you doing now then? Well, we just finished the 3D user interfaces competition. That's what I've been working on. Once I get back to the United States, I'm going to talk with my professor and we're going to discuss some new options for me to go in.
[00:07:23.095] Kent Bye: What do you see as kind of the ultimate potential for virtual reality and what it might be able to enable and how you want to be a part of that?
[00:07:30.063] Alec Moore: Well, a lot of my work is in training transfer, and I think that there's definitely a huge amount of potential for users to learn how to do tasks in environments that are necessarily dangerous or expensive to try to train in. And you can train that in a virtual environment for considerably less cost, especially with the decreasing cost of virtual environments these days. So I'm excited for that. And I think it'll be a very bright future for industrial applications. And I'm also looking forward to the more artistic video game type of implementations too. OK, great. Well, thank you.
[00:08:09.078] Kent Bye: Thank you very much. And thank you for listening. If you'd like to support the Voices of VR podcast, then please consider becoming a patron at patreon.com slash voices of VR.
