Tonight’s post hails from a different journal-Behaviour and Information Technology. The article-A reference level for assessing the acceptable visual demand of in-vehicle information systems-takes an in-depth look at something very common in this day and age: the GPS. The authors have taken note and some observation of a very obvious factoid; with the introduction of mobile devices made specifically for usage on the road, one has to wonder-just how safe is it to watch the GPS and not the road? In recent years, the number of accidents related to lack of focus to the road (specifically regarding a GPS) has risen. This analysis investigates just what in-vehicle information systems (IVIS) are doing to demand too much attention of the driver, and ultimately, become stumbling blocks down the road (no pun intended). The paper conducts a study using 60 participants across four vehicle, each with individual in-vehicle tasks that require the user to remove their attention from the road via visual occlusion. These results are then compared to expert usability analysis, multiple reference levels, social acceptability survey data, and alcohol impairment. This experiment thus hopes to define what the authors have called a demand reference level (DRL) for future work involving IVISs.
Their primary concerns stem from the user’s need to check the GPS constantly, not just because they could be potentially lost, but rather several other factors: poor imaging, distraction voice messages, fatigue, secondary tasks (i.e. driver distraction), general negligence, or quick glances that don’t provide the user with any relevant information. The goal is to create a metric such that IVIS manufacturers must abide by it such that the driver cannot be distracted by it for an extensive period. The researchers used liquid crystal display (LCD) glasses in each test, which were programmed to shut close and re-open every few seconds (thus providing a distraction). From there, an analysis of general driver performance and visual behaviour related criteria became apparent, as these data were used to compare the likelihood of an individual getting into an accident while trying to engage one, two, or everything listed. Alcohol testing was highly criticized, as the argument was made that not all drug induced accident involves specifically alcohol, and some of the side effects people could be experiencing could derive straight from age or genetics. The participants were given sets of tasks, that allowed them to execute each action with or without distraction. The primary tasks assessed were:
- Input: how did the driver enter information?
- Task: what needs to be done?
- Display: what was presented?
- Output: what results are displayed by the system?
The data shows that when comparing these factors to a social acceptability survey, most had confidence that the first two tasks should be done easily, while the latter two tasks (display and output) took the longest amount of time overall. However, when asked if drivers should be doing these tasks on the road at all, an overwhelming majority responded that they should not. After several tries, almost every participants could control the GPS without having to look. This experience was swiftly written off as conditioning. Lastly, the data pointed out that driving under heavy distractions as presented is roughly as lethal as driving drunk.
In conclusion, this article strongly feels that the DRL they propose should be introduced, with some limitations to how much regulation it can hold. The criterion would be based on a simple metrics, and would target populations that would be most likely to need them. It would be likened to a blood-alcohol concentration-that is, something the population would want under control, otherwise product would not hit the market. As a result, they are hoping to make the roads safer for all travelers.
I found this to be unique in that it deal with HCI, but in a non-conventional means. This study is dedicated to understanding why a person is relying heavily on something such as a GPS, and are aiming towards reducing the amount of usage it receives. Ultimately, these authors are not aiming to completely overhaul how we use our GPS, but rather a smarter way to use it. It would also avoid predicaments like the one I’m in, where my cell phone is my GPS, and I just have to hope that no one calls me while I’m driving with it on. I also find this to be rich, as our main focus of usability in class is how something is not working for a user, while in this case it can be argued it’s working too well; alerting the driver constantly, changing routes frequently without prompt, and potential lag are issues I’ve dealt with, and only the last one has no control. I think that this takes HCI to a different level, but still relates to everyday lifestyle. It could even start investigations of built-in systems that aren’t vehicle related (i.e. having a GPS in your cell phone) and make the same comparisons.
Alternatively, they could just make the passenger read the GPS while they focus on the road. Just a thought.
Reference: Stevens, Burnett, and Horberry. A reference level for assessing the acceptable visual demand of in-vehicle information systems. Behaviour & Information Technology, 29: 527-540. 22 February 2010.