Mobile Phone Use In Cars example essay topic

The influence of the use of mobile phones on driver situation awarenessByAndrew Parkes, Transport Research Laboratory, Crowthorne, England. &Victor Hooijmeijer, Verkeersadviesburo Diepens en Okkema, Eindhoven, The bstractThe driving performance of 15 subjects in a simulated road environment has been studied both with andwithout a hands-free telephone conversation. The performance indicators used were choice reaction time,braking profile, lateral position, speed, and situation awareness. The driving task was relatively easy, and theyoung drivers studies were able to have a hands-free telephone conversation and perform well with respect tolateral position, the variation in lateral position of the car, and speed maintenance.

However, significantdifferences were found in choice reaction time, especially in the beginning stages of the telephone conversation,and in situation awareness. The subjects reacted significantly slower to an unexpected event in the first twominutes of the telephone conversation and were, for a large part of the telephone conversation, unaware oftraffic movements around them. IntroductionA survey in the United States has revealed that the vast majority (84%) of mobile phone users believethat using a phone is a distraction and increases the likelihood of an accident (IRC, 1999). The samerespondents report however that 61% of them use their mobile phone while driving and around 30% use theirphone frequently or fairly often. Since mobile phone use in cars is a relatively new phenomenon, and since theeffects of mobile phone use on traffic safety are still unclear, laws regarding this subject vary between differentcountries. Some countries use a mixture of legislation and recommendation, but are not consistent about thedifference in hands-free and hand-held phone use.

For example, in Italy only hands-free phones are allowed bylaw during driving. At the same time, however, the use of equipment that restricts the hearing senses (whichpresumably includes all types of mobile phones) is prohibited. The same situation exists in Spain, whereas inPortugal, Denmark, and Hungary only hand-held use of mobile phones is prohibited by law (Oei, 1998; UnitedNations, 1998). Outside Europe, a hand-held prohibition exists in Israel, Malaysia and some states of the U.S.A. (Oei, 1998). Germany, France, and Sweden are examples of countries in which no rules or jurisprudence areused to limit the usage of mobile phones during driving (Becker et al., 1995; Oei, 1998; Petica, 1993). Nevertheless, it is recommended in Finland and the UK to use hands-free phones only (Oei, 1998).

Thesituation is confused and changing continually. Only recently, The Netherlands (June 2000)) have jurisprudenceon using handheld mobiles during driving. A driver has been found guilty causing an accident because she washaving a phone conversation. It is likely that many other countries will develop case law in this way even iflegislation does not exist. At one point, it looked as though the problem for legislators would become easier. Research hadhighlighted the potential safety problems with driving and using handheld devices, and it seemed that themarket was leading to the point at which car manufacturers would integrate well-designed handsfree telephonesinto their vehicles.

Many interested experts claimed that driving and holding a carphone conversation was nomore difficult than talking to passengers, and so, if the handset were removed, the problem would go as well. Unfortunately the market has gone in a different direction. Personal mobile phones are ubiquitous due toaggressive and cheap pricing regimes, but handsfree adaption kits for use in vehicles, as yet, are not popular. Sothe use of handheld devices is actually increasing. Added to this is growing concern that the act of holding acarphone conversation is fundamentally different to other in-vehicle conversations with passengers (Parkes1991a and 1991b). So, we are not reducing the number of handsets, and even if we did, it is possible theproblem will remain.

This paper reports experimental work focused on an aspect of driving performance thathas not been looked at in-depth in previous studies. In addition to measures of performance such as driverreaction time and steering ability, we consider higher-order functions that identify not just the ability to controlthe vehicle, but also to maintain a clear picture of the traffic situation around the driver during a carphoneconversation. MethodIn the experiment, 15 volunteer subjects were used. The subjects were all (postgraduate) students at a UKuniversity, aged 22 to 31 (average age = 24.0 years, SD = 2.27 years), with more than 3 years of drivingexperience, and little or no experience with using a mobile phone while driving.

A static driving simulator was used. A medium sized saloon car stood in front of a purpose-built,cylindrical projection screen. Three projectors produced a horizontal forward field of view of 120^0. The verticalfield of view is 40^0. An additional projector, aimed towards the rear of the car gave a 50^0 horizontal x 40^0vertical image, and provided normal view through the vehicle's interior and driver-side wing mirror. The imageswere produced with a frame rate of 20Hz and a screen resolution of 960 x 620 pixels per channel.

Thecomputer monitoring the car outputs also controlled various elements of feedback to the driver, such as thedashboard lights, engine, road and wind noise as well as the sound of other vehicles in the scene. The route used in the simulation was designed to keep the attention of the driver on the road. A singlecarriageway in a countryside environment was used, with smooth horizontal and vertical curves equally spreadalong the track length. Other traffic on the simulated road did not disrupt the movement of the subject car. Inthis way, the subject could always drive at the desired speed. However, to make the driving environment as natural as possible a fairly high level of oncoming traffic was simulated.

Cars also appeared in the rear-viewmirror and far ahead of the subject car. The route had a total length of 15.5 miles. Measurement of theperformance indicators for each subject took place between the 4th and the 11th mile. Subjects were instructed to keep the vehicle in the middle of the lane and to keep closely to themandatory speed limit, indicated by regular roadside speed limit signs, at all times. The subjects were also toldthat other normal road environment conditions applied, that other traffic would be present and that some severeweather conditions (like wind gusts) were likely to occur. Subjects were asked a series of questions via the hands-free carphone, and were required to make averbal response when they felt able to do so.

A hybrid test was developed (Hooijmeijer 1999, based on Fox &Parkes 1991) that incorporated numerical and verbal memory, arithmetic and verbal reasoning. The test wasextensively piloted and was aimed to be demanding for the subject group. The subjects were not under timepressure, and they were told that their scores would not be used in any part of the analysis of the trial. Two different kinds of 'unexpected events' requiring choice reactions were simulated. The first one wasa green square (presented on two occasions) that appeared on the road in front of the car for approximately 2seconds. Subjects had to flash the lights of the car as quickly as possible in response to the green square.

Thesecond event was a red square that represented a danger on the road, and the subjects were expected to make anemergency stop immediately. A further indication of the driver performance was given by the braking distanceof the car after the appearance of the red square. This is of course directly related to the reaction time of thedriver and the speed of the car, but gives extra information about the performances of the driver. Choicereaction time measurements were taken at 5 miles (1st green square appearance), 6 miles (red squareappearance) and 8 miles (2nd green square appearance).

Wind gusts were simulated at 5.5 miles and at 10 miles,from the left side of the road at an angle of 90^0. Both gusts had a speed that gradually increased from zero to15mph and then decreased again gradually to 0mph. In total, each wind gust lasted over a distance of around500m. Lateral position and variability were measured, both on straight sections and when there were simulatedwind gusts from the side of the vehicle.

Speed maintenance was recorded during sections not involving othermeasures. Another indicator of performance with respect to speed is the adjustment to a change in themandatory speed limit. Along the test track the mandatory speed limit changed once from 80 to 50km/h (at 4.5miles) and once from 50 to 80km/h (at 7 miles) Situation awareness (SA) has been defined as 'a person's perception of the elements of the environmentwithin a volume of time and space, the comprehension of their meaning and the projection of their status in thenear future' (Endsley, 1994). The relation between poor situation awareness and poor performance has beenfound in several studies (Endsley, 1990; Venturino et al., 1989).

There are three different levels of SituationAwareness (Endsley, 1993): Level 1 SA: Perception of the elements in the environment. Level 2 SA: Comprehension of the current situation.  Level 3 SA: Projection of future status. All three levels of SA were measured in this research by questions directed to the subject at two fixedlocations at 6.5 miles and 9 miles after the start of the test. The screen went blank and the simulation wasstopped. The subject was then asked SA probe questions. At the moment the screen went blank, one car was inthe rear-view mirror and three or four cars were approaching on the opposite lane.

After asking the questions,the simulation was resumed at the same position as when the simulation was stopped. ResultsA one-sample t-test, was used to determine if there was a difference between the reaction time of thesubjects in both situations. The test showed for the first green square a t-valu e of 2.576 (the critical t-value is1.761, df=14, =0.05). The mean reaction time to the various stimuli in the phone and no-phone situationMean reaction time (sec.) Stimulus Without phone conversation With phone conversation1st green square 1.008 1.1312nd green square 1.115 1.187Red square 1.370 1.421The second green square showed however a non-significant t-value of 1.169 (df=14 and =0.05 andtcrit=1.761).

The reaction time to the red square gave the same non-significant result. The mean lateral position was calculated for each subject from the lateral position of the car between the4th and 5.5th mile and the 7th and 8.5th mile of track. A one-sample t-test showed no significant difference in themean lateral position between the phone and no-phone situation (df=14, =0.05, tcrit=1.761, t1=-1.390,t2=0.879). A further indication of the performances of the subjects is given by the variability in lateral position,estimated by the standard deviation of the lateral position.

The more the subject varies the lateral position, themore implications this might have on traffic safety. The standard deviation for each subject was also calculatedfor the same two sections as above. A one-sample t-test showed no significant difference between the phone andno-phone situation in each sections. The variance in lateral position was also used to measure the influences ofmobile phone usage on the unexpected event of a wind gust. The standard deviation of the lateral position ofeach subject was calculated from the lateral position of the car from the triggering of the wind gust until 500 metres after. A single-sample t-test gave non-significant values of -0.48 and 0. for wind gust 1 and 2respectively (with df=14, =0.05, tcrit=1.761).

Speed was also used to measure driver performance. The speed of the car in each trial was measuredalong part of the track. To avoid influences from the emergency stop (red squares) and the wind gusts, the speedof the car was measured in sections without these features. There was no significant difference between trials. Itwas also of interest to see if there was a difference in speed adjustment between the two situations, once themandatory speed limit changes. During the (possible) phone conversation the mandatory speed limit changedonce from 80mph to 50km/h and once from 50 to 80km/h.

For each subject the speed was recorded every 100metres from 500 metres before the mandatory speed limit sign until 500 metres after the speed limit sign. Analysis showed no difference in mean speeds around the 50-80 km / h change (df=14, =0.05, tcrit=1.761,t=1.13), but a significant difference was found when the speed changed from 80 to 50km/h (df=14, =0.05,tcrit=1.761, t=3.42). However, although the mean speeds of the subjects around the speed limit change differedsignificantly, this does not say anything about how quickly the subjects reacted to the speed change. Therefore,the mean speeds of the subjects were calculated at 100 metre intervals in both the phone and the no-phonesituation. These mean speeds are plotted for the 80-50 km / h and the 50-80 km / h speed changes respectively. This shows that there is no observable difference between the mean reaction to the change in speed limit from50 to 80 km / h in the phone and no-phone situation: both lines show approximately the same pattern.

The changefrom 80 to 50 km / h however seems to be slower in the phone situation: the mean speed of the subjects in the no-phonesituation was below the speed limit after almost 100 metres. Mean change in speed as a result of a mandatory speed limit change from 80-50 km/h01020304050607080-500 -400-300 -200-1000 100 200 300 400 500Distance from speed limit sign (in n metres) no conversationconversation Mean change in speed because of a mandatory speed limit change from 50-80 km / hSituation awareness of the subjects was measured by asking the subjects three questions at two fixedlocations during the simulation. The questions asked at each location were:1. Can you tell me what other traffic was surrounding you just beforeI stopped the simulation?2. Can you tell me the colour of the car that was in your rear-view mirror?3.

Was the car in your rear-view mirror driving faster than you or not?Performances on the situation awareness task with and without phone conversation (location 1) No. of correctanswers;Location 1WithoutphoneconversationWith phoneconversation 2 Critical value p-valueQuestion 1 14 4 13.89 12.12 <0.0005Question 2 14 6 9.60 9.14 <0.0025Question 3 13 6 7.03 6.63 <0.0100Performances on the situation awareness task with and without phone conversation (location 2) No. of correctanswers;WithoutphoneWith phoneconversation 2 Critical value p-value01020304050607080-500 -400-300 -200-1000 100 200 300 400 500Distance from speed lim it sign (in nmetres) Speed h) no conversationconversation Location 2 conversationQuestion 1 12 5 6.65 6.63 <0.01Question 2 12 4 8.57 7.88 <0.005Question 3 12 5 6.65 6.63 <0.01There were significantly more correct answers to the situation awareness questions in the no-phone situation atboth locations. DiscussionThe reaction time to the first choice reaction event, the responsiveness to a lower speed limit and thesituation awareness of the drivers showed a significant difference in favour of the no-phone situation. However,before drawing any conclusions it is important to interpret the outcome of the analysis within the context of theparticular limitations of this study. A main limitation is that the measurements took place in a safe, simulatedenvironment.

In addition to this, the simulated road was a countryside environment, with a reasonable amountof surrounding traffic, but no direct conflicts with other cars. The road driven did not have sharp curves or largejunctions and the subjects were asked to keep to the maximum speed limit, which was either 80 or 50km/h. As aresult, the driving task was rather easy. The phone task on the other hand consisted of a selection of verbal andnumerical questions and was rather difficult to perform.

The results may therefore only applicable directly to asituation of a relatively easy driving task combined with a relatively difficult telephone task. The reaction time to both the second and third choice-reaction events was not significantly longer in thephone situation, which questions the difference in reaction time to the first event. A reason for this differencemight be that the first square appeared rather soon after the start of the conversation (after approximately 2minutes), whilst the other squares appeared later (at approximately 5-7 minutes after the start of the phoneconversation). This would imply that drivers are slower in their reactions when a phone conversation just hasstarted, but the effect reduces over time. This result finds some support in real road studies (Parkes et al 1993) that showed a pronounced dip in speed at the onset of a carphone call. The carphone task seems to occupy thesubject more in the beginning stages, resulting in a reaction time which is on average 0.13 seconds longer,which means that (at a speed of 80 km / h) on average an extra 3 metres is covered before a reaction to an eventon the road is triggered.

The second significant difference in performance was found in the responsiveness to achange in speed limit from 80 to 50km/h. It took drivers in the phone situation on average 200 metres more toadapt to the new speed limit than in the no-phone situation. The third difference is more important. Significant deterioration was found in situation awarenessbetween the phone and the no-phone situation. Many subjects in the phone situation had very little idea aboutwhat was going on around them at the two points the simulation was stopped and were not able to report on thepresence or actions of traffic around them.

The results have shown that a young well-educated group of drivers were able to engage in a difficultcarphone conversation and cope with basic control elements of driving reasonably well. However even thisgroup showed a dramatic fall off in situation awareness due to the level of concentration demanded by thecarphone conversation. It is clear that more research is required into the nature and duration of typical carphoneconversations, and into the behavioural and performance consequences for a wider group of drivers, in bothsimulated and real world environments. Until then it is likely that legislation will continue to focus solely on theuse of hand held devices. Whilst this is important, it does not address the full extent of the problem. AcknowledgementThis study was carried out using the driving simulator at Leeds University UK, and could not have beencompleted without the enthusiastic support of Hamish Jamson, Technical Coordinator, and Andrew Bailey,Simulation analyst.

Bibliography

Becker, S. et al. (1995), Telefonieren am Steur.
Bundesanstalt f"ur Strassenwesen. Heft M 45, Bergisch Gladbach. Endsley, M.R. (1994).
Measurement of situation awareness in dynamic systems. Department of IndustrialEngineering, Texas Technical University, Lubbock, Texas. Endsley, M.R. (1993).
Towards a theory of Situation Awareness in Dynamic Systems. Departmetn of IndustrialEngineering, Texas Technical University, Lubbock, Texas. ECMT (1996).
New Information Technologies in the Road Transport Sector; Policy Issues, Ergonomics and Safety. Annex 6: Ergonomics and Safety of In-vehicle Information Systems: The example of cellular phoneFox, S. and Parkes, A.M., (1989), The effects of driving and handsfree telephone use on conversation styleand decision making ability.
HUSAT Memo. No. 434R. HUSAT Research Centre, Loughborough Universityof Technology.. Haigney, D. (1999).
Mobile phone use and driving: a critical review of the research. RoSPA, Birmingham. Hooijmeijer, V. (1999) The influence of the use of mobile phones ondriver behaviour.
Unpublished M. Sc. thesis. Institute for Transport Studies, University of Leeds, UK. Insurance Research Council, IRC (1999).
Cell phone owners prefer to ignore risks. http:// . ircweb. org / news/1297cell. htmlNCSL Transportation Reviews (1998), Cellular Telephones in Motor Vehicles,http:// .