In-car Interaction Methods Affect on Driver

Info: 3948 words (16 pages) Example Literature Review
Published: 25th Oct 2021

Share this: Facebook Twitter Reddit LinkedIn WhatsApp

This paper presented below is the literature review of mainly a previous research paper, named, "You can touch, but you can't look". It further explains the studies and finding of the paper along with some other supporting facts and figures to analyse the effects of various factor, and study them for three main methods of interation.

INTRODUCTION

Automobile drivers are at present are provided with a huge variety of sense setups which require human sense, such as, education sense setups, weather control sense setups, song sense setups, and more. These sense setups mostly do require the driver's attentiveness, and, thus, this attentiveness of eyes, proves to have quite less time for eyes to be on the road. Here, we basically show and analyze three methods for interacting with automobile touchscreen system. They are, sense by gesture, touch, tactile. The aim of the paper is to analyze how these affect the driver during the drive. We evaluate these interacting strategies. The outcomes confirmed that the sense with gestures, has a tremendous impact on a huge variety of looks that the driver gives, specially, adjusting the eye back on the road. However, the sense by gestures, still needs quick looks, for good coordination of hand and eye. Secondly, the sense by touch, leads to speedy and efficiency of the job is improved. Tactile however, looked to have performed the least out of all the methods.

Planning for versatility is challenging, as versatile users are characterized by incredible differences, hence, it should be planned for diverse reasons and needs. Being architects, we should make users to be able to adjust with advances amid their ongoing life usage of the automobile. A good area in the sense of human and computer, for motion related work is, developing the interactive system in the automobile. Accompanying the critical propels of technological advancements, the automobile driver these days, advertised with a huge variety of automobile frameworks such as, course direction frameworks, weather displays, media sense setups. The developing automobiles constantly offer extra and more progressed automobile frameworks, and thus, the impacts of automobile frameworks have gotten to be more significant. Several inquire about considers outline that association with such automobile frameworks whereas driving profoundly challenged the attentiveness of the driver while driving the vehicle.

Being attentive, with the vision, seemed to be the most important characteristic [3]. As we imagined, analyzing the ocular data, shows, a very less time of "eye-on-the-road" as the drivers had to be interacting with the setups with great ocular demand, and it has direct effect on the drive ability [3].

Many automobiles are today shipped with a built-in touch screens which highly depend on Ocular attentiveness. Such touch-based screens are typically being employed to enable the driver to interact with novel functions as well as established functions [3]. Consequently, simple control operations (e.g. using a fan speed dial) have become more complicated requiring a number of discrete steps of senses (e.g. mode selection, option choice, adjustment setting) [3]. Taking the immediate safety critical aspects of driving into account, Green recommends that we need to identify and investigate new sense methods that are more suited, for automobile sense setups [3].

The paper will ultimately make us clear about the Effects of the Size and Location of touchscreen & touch buttons on the Usability and Accessibility of the Touchscreen and its overall effects on safety of the passenger while driving.

PREVIOUS RELATED WORK

This section highlights some findings and learnings from the studies of previous experiments, conducted by different researchers on this topic. As it shows the work that the reference paper was inspired from, the examples and facts have been taken from the reference paper itself.

Most automobile interacting systems, tend to find out as to how the work of the driver can be lessened and how the Ocular attentiveness need, should be lessened while interacting. The touch senses are getting popular in the automobiles (such as, people interacting with weather displays etc.), we tend to assume that try to look over a abilities of tactile interface, look for new methods that do not need much ocular attentiveness. Usually, the automobile does not require much attentiveness to get good at activities like, changing the gear, throttle, or the wheels control. [3]

The two methods, touch and tactile, are analyzed on the potential to enhance driver attentiveness while driving and interacting at the same time. Many of the previous works on this, find out the positives and negatives of these systems. "Bellotti et al." did such an experiment between haptic and touch systems. It showed that tactile was much better, on the things which needed less ocular attentiveness, as it needed less "eye-on-the-road" time. "Tsimhoni and Green" found something similar, in which they found that the touch system needed greater Ocular attentiveness, and also lessened the performance of the driver. It came out that they got lost in their own lanes a lot and got confused when they had to change it. The driver even had a lot of short stares and had gaps between two stares. When the condition was tough, and needed higher ocular attentiveness, the jobs were completed in a longer time.

Also, "Noy et al." discovered, the job that needed some manipulating work, with a screen, it took a lot more time, as it needed more mental application, than the simpler tasks like, increasing the volume as it needed greater ocular attentiveness because of less tactile input [3].

The "touch" and "tactile" look to have a lot of negatives, as to reduce the ocular attentiveness, the other two, tend to require less attentiveness and less interacting, such as voice, and gesture sense. Voice sense seems to be the best of the methods as it requires minimal use of hands and eye attentiveness for interacting.

"Barón and Green state, however, that Voice recognition is highly cognitive demanding, and it has been characterized as both impractical and flawed in regard to automobile sense". [3]

Gesture sense can be a good method as a replacement to voice sense, as the speech could have some dissimilarities that the system could not identify due to certain way that it is programmed.

"Alpen and Minardo" experimented the gesture sense in the real life setting of an automobile system. They used simulator for driving, where participants did entertaining jobs, like finding and playing music, searching a preset, adjusting the sound level), doing the same job, with, a regular radio, and a gesture sense system. The result showed that the people made lesser mistakes in the gesture sense system than the regular system, but the no. of mistakes were not a lot lesser. They also liked the gesture sense more as they did not have to look at the system much, to perform the job, and they could do it while concentrating on driving. They did not have to grab a knob, reach out for it, and the gesture sense made it possible for them to make inaccurate gesture and getting good outcomes. [3]

"Pirhonen et al." experimented between a gesture and touch sense system, for a music player. They wanted the driver and music player user, to pay attention to the things happening around, instead of having to look at the device, as the driver is in a pretty equal situation. It showed that the load was quite less, and the time to complete the job was lowered as well for the gesture sense, over the touch one. Also, it showed that the spped to the people taking the test remained equal to what it should have, in the gesture sense, rather than in the touch sense, but the number of errors made was not much different in both. [3]

METHOD OF EXPERIMENT

This part of the paper explains the experiment that was conducted by the authors in the reference paper. The experiment basically outlines the experiment conducted by the authors of the reference paper and concludes and analyses the results of the experiment.

The aim here, is, comparing the main three methods shown above.

Participants

16 people were surveyed, and they participated. All the 16 people had a valid license. Also, their experiences varied from a 100km in a year, to 20,000km in a year. The people also mentioned that they were fit to drive, and were having a regular vision. 7 from the 8 people mentioned that they have experienced a similar setting when they were practicing for their license. 6 of the 8 people also mentioned that they have used pedal and wheel in for a similar situation during playing a simulated game. So these were some differences kept in mind while considering the experience of people in the experiment.

Design

A mix of "within-subject" and "between-subject format" was used. The interfaced type, that is, tactile, touch, and gesture, was used like the "within-subject" variable and the settings for the driver, like "controlled and simulated" like the "betweensubject" thing. People were arbitrarily given, any setup out of the two.

They took part in each of the three session, one relating to each of the senses mentioned above.

The order in which they faced each setting was random. So, collectively, a total of forty eight (3x2x8) settings and situations were tried out (3 for sensing methods, 2 for the setting, 8 for the people). [3]

Driving Settings

The setting for drivers that was talked about earlier, are "controlled" and "simulated". This was needed as it told us, about different situations, and we got to find out about similar and dissimilar things in these two cases, and be able to figure out the issue related to both the condition. We here, talk about the three methods of sense and having effects of different condition on each of them.

"Controlled Driving"

This experiment for "controlled driving" happened at a place for driving in "Aalborg, Denmark" which was usually taken up to train people, as it can be seen in the figure. People had to drive on a 3.3 km road, used for testing, so there was no one else on the road during the experiment. As seen below, the road had various things on the road, to imitate the drive condition for a village setting, which included things like, "simple curve exercises and simple slalom areas". The track also had a camera, for speed, and it had traffic light as well. "Opel Astra" was used for this purpose.

Figure 1.

The sound system that was already present in the car was swapped with a different one, as well as the screen used was an 18" screen, which was flat, and only 8" of it was usable. This was placed at the center of the vehicle at the front, and the touch part of 8" was placed at the top area of the same setup.

"Simulated Driving"

For the simulated test, the authors went to "the HCI laboratory at Aalborg University" which is a "mediumfidelity driving simulator".

The setup had a set of pedals and a set of wheels that sent the feedback of force when steered, which was basically a setting for controlling a PS4 game called "Gran Turismo 3".

Figure 2.

The track chosen for this was a same kind of track. For making the experience as similar as possible, the screen was projected in a small setting just ahead of the driver at a similar distance to the actual one. The seats, one for driving and one for the guest, were put together so as to feel like a real situation, and the placement of the screen and its orientation was similar to the actual test case.

Tasks

There were around 30 different jobs given in every of the 48 session. The jobs were described by the guy who was assigned for reading out loud, and it was the same person in each case, to maintain similarity in hearing conditions.

The jobs were easy to perform using easy steps in each of the sense methods.

Procedure

So, the setting, steps and the environment for the two types of cases were tried to be kept as similar as it could be, and all the factors were made for the three senses just before the beginning of the test.

The procedure had an introduction to perform basic functions like, motion for regular gestures, and the structure of each gesture. Also, people were asked to similar small tasks so that the people got acclimatized to the way the steps are performed.

The steps and methods were given and explained to the participant. The people were told to drive around the road in the designated path, for a round or two, to get familiar with the conditions of the path and also the automobile. For this, and during it, the manager would check, familiarize and explain the track, and the process, and give regular instruction. They were told to drive in the path at the speed of 50 approximately. They were also told to stick to the regular rules and drive as they normally world, along with, doing the small jobs that they were told to, along the path, such as, near the light and the camera, when they had to speed up to 60 , as much as possible.

The jobs were created and demonstrated in a way to create a similarity to real life.

The automobile had 2 cameras, to get all the eye motions and the movement of humans to study their interaction. The vehicle also had a gps machine to get the data for the speed of the car. The actual speed stats were taken so as to mix and match with the computer stats which would show the importance of the speed for all the different senses on the drive path.

For the simulated test as well, an equivalent procedure was taken up. The people had been given time and practice sessions to get used to the sensitiveness for the steering and the sensitivities of the setup for simulation. To evaluate, subjects were said to stay within 50 and 70, and try to be in the center of the path.

There were cameras to record the eye motion and the activity of the motion of people.

BASIS OF ANALYSIS

Various factors were chosen to assess the sense methods. As being a review of a text on automobiles, various things were found and demonstrated to figure out the quality of the three sense methods, and their outcomes on the driving and safety of the driver.

So, the measures that were considered for the process were:

1. "Primary driving task performance"

2. "Secondary driving task performance"

3. "Eye glance behavior"

The attitude of the participants for the sense methods was also measured.

1) The main performance was gauged from the "lateral and longitudinal control errors".

"Lateral control incidents were defined as attentiveness related loss of automobile lateral control (lane excursions, steering wheel input, etc.) while longitudinal control errors denote incidents where participants had problems controlling automobile velocity (speed maintenance, acceleration, etc.)".

There errors were looking and analyzing the video logs and the clips recorded. The authors of the reference paper reviewed all the 48 clips individually, by finding, categorizing, and documenting each incident. The list, was matched and analyzed, and was combined into a final list. "An inter- rate reliability test of this analysis (weighted Cohen's Kappa) gave =0.92". The result above shows great adherence to the theory.

2) After the main job characteristics, the second ones had "task effectiveness (task completion and sense errors) and task efficiency (total task completion time)". The effectiveness of a job is a regular way to gauge the result while studying about an automobile.

Again, the video clips were used to find the Sense errors were identified, according to various factors, according to the type of each setting showing every sense method.

As an example, for tactile sense, the errors were if the person was unable to press a particular button, or choosing an incorrect option for a particular job, like changing the volume when told to change a song.

Similarly, some errors were made for tough and gesture controls, but they were recorded as an automatic response if it was a wrong tough or a wrong gesture.

The times that it took to finish the jobs for gesture and touch, were gauged by the log entries of each session.

The tactile sense didn't give out any similar data however, so it needed to manually gauge the times by looking at the videos. Similarly, the authors gauged times for each of the 48 sessions and finding out the time for each of the 1440 tasks. As to the type of this data, the other tests were not performed, however, the average from both the lists was taken out. Also, getting the total time to complete each task needed a constant to be put to the value of the log. This was found out from a stopwatch for a 20 task sample for both the sense setups and was found out to be 660 ms.

3) The movement of eye has to be something used quite often to gauge the attentiveness of the person driving and is considered as a good method of connecting Ocular attentiveness and performance while the drive. So, the looks of the eye were divided in 3 ways: < 0.5 sec, 0.5-2.0 sec and > 2.0 sec.

The looks that are above 0.5 secs are called fixations. These can usually tell what the driver is looking at and paying attention to.

This also, in a way, sets a limit for the other two options, that is < and > 0.5 secs. A study tells, that the person driving, are pretty inertial towards continuing for around 2 secs without any proper info, this is called the "2-second rule". The rule sets limit for < or > 2 secs.

So, according to the classification of the eye movement times as, <0.5, 0.5-2, >2, the authors again gauged the findings of each of the 48 session which were then matched and analyzed with the "inter-rate reliability test, with a constant of 0.71, to show a great adherence to text.

If there was ever a disagreement, it was decided that a 3^rd author would look at all the topics where these did not agree, and then put them accordingly in the category.

DISCUSSION and RESULTS

This paper does not go in the technical results of the experiment. The results can be studied in [3].

Discussing the major takes from the experiment, we can talk each of the type of interaction below, and learn more about the major outcomes for each sense type.

The tactile sense came out to be having less efficiency than what was assumed. It had the largest of the average time for completing each task, the greatest number of looks by the eye, so it is the least preferred out of the three.

Touch was quickest and simplest method. It showed a little better result than the tactile sense, if we see the average time and average number of looks by the eye, even though it was assumed to have more since it did not have any tactile feedback.

Gesture sense came to be the best alternative as it needed minimal time for eye movement, and thus, it was also observed that it did not record even a single case of a look of more than 2 seconds of the duration. Thus, it proved to be the best method out of the three.

Limitations

This experiment had a lot of limitations, which could be important for future studies on a similar topic ahead.

Firstly, this experiment, was different in the method of providing inputs as well as getting outputs from the participants, which would have been a huge factor to affect our results.

Then, practically, the touch and gesture panels had to be placed at a little different position, and hence it affected the outcomes. This also proves that the placement of these systems is very necessary, because a different position would have resulted in different touch and look time for eye.

Lastly, tactile system had more functions and thus increased the number of functions to use and decreasing the complexness of the system, which can result in longer times for eye movement, as it tends to distract and confuse the user off the right feature. Hence it would have resulted bad for the tactile system.

CONCLUSION

This study shows that there is a requirement for new systems to be put in the automobiles as the more general and old ones show a greater need to ocular movement and attentiveness, resulting in lesser efficiency and output of the driver. Hence, the newer methods, with less eye movement time could result in more time for eye "on the road" while driving. Here we mainly looked and analyzed three main interacting methods and tried to figure out their effects on the performance.

The outcomes indicate that gesture sense minimizes the time of eye to look at the screen for an easy secondary job, specifically, reducing long times over 2 secs. Also, no such difference was seen on errors like "lateral and longitudinal control". Gesture sense was not completely attentiveness less, as it needed the hand and eye to coordinate still. Touch was however, the quickest, but it also required the longest duration for the eye to look off the road, which means, the greatest number of looks above 2 secs.

Also, this test had some good limitations which would've negatively impacted the outcomes and thus would've altered the findings, thus, mainly, the newer researches should focus on providing the same input and output conditions and getting the results on them. Also, they should focus on how to get the best from every method and integrate them to develop a tool which requires minimal effort and does not hinder the performance, rather, improve the experience.