By James Dobson (Dobsonj) on Thursday, March 18, 1999 - 05:40 pm:
Hello, my name is James Dobson. I'm a 17-year-old novice driver in Ontario (see Ontario GDL Report for more information about the licensing system in Ontario). I'm also interested in computer programming. Naturally, I was attracted to this paper. After reading it over, I had a general feeling of agreement. However, I do have several comments and concerns, which I will address below.
My primary concern is that simulation technology is seen, in many cases, as something which can be used to reduce the involvement of human instructors in the driver training process. In fact, I believe that proper use of computer technologies should be coupled with an increase in human involvement, augmenting the instruction techniques that are already in place. CAI, for example, could be improperly used to decrease the number of hours needed by students in a classroom setting. From my experience as a peer tutor, I have found that teaching a subject requires much more than a knowledge of that subject. It requires true understanding of that subject. This understanding must usually encompass a number of different angles, as an explanation that works for one student won't necessarily work for another. While computers can store a lot of information (a CD-ROM can store the text of about 600 books, and there are technologies foreseeable in the distant future, which could allow a device the size of a single book to store the equivalent information of the entire library of congress), computers are horrible at understanding. In fact, they don't understand a thing. Without this understanding, I find it very hard to believe that a CAI program could be used to cut down the involvement of the in-class instructor. Rather, it could be used in place of a text book & testing mechanism, and the instructor would be free to discuss "higher" concepts.
Simulation software, in my view, also should increase the involvement of the instructor. Why? Simulation software is very good at one thing: providing a realistic, controlled environment. The key word here is controlled. Controlled by whom? The instructor. Consider a classroom where the students are seated three to a simulator, with one instructor per simulator. The instructor would tell the student what to do, where to turn, etc, but the instructor would also control the situations that the simulator presents to the student. They could present a specific situation, and, based on the student's preformance, the next few minutes of simulation could be adjusted accordingly. Secondly, the instructor could help the student explore "what ifs" by tailoring the simulation to the specifications of the "what if" situation. During the simulation, each computer could record the situations that cropped up, and the students' responses. After the students had left, the instructors could search through the recordings, and pull up the "interesting" occurances. These highlights could be carefully selected and presented to the class at the beginning of the next session. The entire class could join in the analysis of the situations, thus learning from the mistakes of others. In fact, if the appropriate variables were recorded, the other students in the class could be given the opportunity to run through the same situation. They could be given multiple entry points into the situation, so they could study how to recover from it as well as how to avoid it.
The technologies used should not remove the instructor from the scene, rather the instructor should take on a central role that is empowered by the technology.
The paper also mentioned simulation technologies that are mildly interactive, where a video is presented and the "correct" choices must be made, at the "correct" times. This technology achives a very high sense of visual realism, but is very poor for most other types of realism. I have tried a simulator like this and found it extremely frustrating. There is no room to say "well, the best option is slam on my breaks, because I'd already be going slowly because I'd anticipated that this would..." There are only "black or white" choices, which leads to a decision process that is hardly applicable in the real world. This type of simulation software can be used for basic training, but after a very short time it becomes too simplistic.
To end off, I have a question about flat screens and depth perception in general. I know that the different images presented to each eye can be composed by the visual centre into a three dimensional image. But beyond some distance, this ability must break down, to be replaced with other perceptual clues (parallax, perspective, etc.) At what point does this breakdown occur? Secondly is this distance of any consequence in simulators that use flat screens which surround the participant, and are thus unable to present the stereoscopic view?
Thanks.
By peter christianson (Ydc) on Friday, March 26, 1999 - 12:57 am:
to James Dobson,
James I am absolutely impressed by your understanding of the issues raised by simulation which ultimately will play a large role in driver training and testing. The issue is not one of replacing instructors as much one of replacing examiners by a completely unbiased driving test measured by accurate laser tracking of the drivers eye movement without the examiner distraction of potential collisions. Simulation of all causes of collision can quickly teach proper driver response as a habit. You seem to have an understanding of the technology far beyond your years. Have you looked at the Mercedes simulator project that has over 70 million invested? Where do you see your interest in this subject taking you or is it just another interest for you? I am surprised that there is not more interest in this subject and I am still looking for an answer to the "sickness" issue that appears to be a major impediment to simulation. Did the aircraft simulators not have the same problem? Peter
By Mark William Simpson (Marksimpson) on Saturday, April 10, 1999 - 01:28 am:
Hi. My name is Mark Simpson. I am a driving instructor in a town called Mittagong in the Southern Highlands of New South Wales, Australia. I would like to mention that I have driven a simulator here in Australia, and experienced this "motion sickness". Apparently our bodies know that we are not actually moving, but our brains are tricked into believing it. It can be a very nasty experience if you are on the machine for any long period of time as I was, for over half an hour. The answer to the problem could be any of two suggestions or a combination of both.
1. A wrap around screen,
2. Convince the body that it is actually in motion.
I await any information others may have on this subject. Mark
