Attitudes Towards Intelligent Speed Adaptation

Attitudes towards Intelligent Speed Adaptation 1

Attitudes towards Intelligent Speed Adaptation

Dr Charles Musselwhite

Senior Lecturer

Institute of Health and Community Studies

BournemouthUniversity

R111, Royal London House, Christchurch Road, Bournemouth, BH1 3LT

Telephone 01202 962787

Fax: (01202) 962194

ABSTRACT

This study investigated attitudes towards a variety of different types of Intelligent Speed Adaptation (ISA) in an attempt to explain the motivation behind behaviour exhibited when using such devices, in order to gain in-depth knowledge on acceptance, adherence and behavioural adaptation. A grounded theory approach was used, where emergent interviews were carried out with 47 participants. Salient issues arising from the interviews were then addressed in a questionnaire which was completed by 410 respondents. Drivers were categorised by their self-reported risk taking in ordinary driving situations. It was found that the most consistent risk takers are those least likely to use a voluntary ISA and perceive themselves as most likely to show behavioural adaptation through increased stress and lack of concentration, manifesting itself as shorter headway and increased acceleration and deceleration. Those who show most acceptance and least perceived behavioural adaptation are those who are in the least risky category. Further on-road trials are suggested with ISA devices that are devised and developed with these findings in mind.

KEYWORDS

Intelligent Speed Adaptation, speed, attitudes, behavioural adaptation.

INTRODUCTION

There is a growing interest in how technology can be used to aid transport, known as Intelligent Transport Systems (ITS), and how they can aid driver behaviour, more specifically known as Advanced Driver Support (ADS). Recent technological advances have allowed the possibility of the vehicle ‘knowing’ the speed limit through beacons, transmitters or tags conveying information from the infrastructure and environment to the vehicle (Hoedemaeker, 1999). Technology set up this way is known as Intelligent Speed Adaptation (ISA) or External Vehicle Speed Control (EVSC). Table 1 shows three differing versions of the ISA. The information on the speed limit could be passively displayed on the dashboard of the vehicle to help the driver see the current speed limit. This passive system could be complemented with a warning advising the driver when they were reaching the maximum speed through visual, auditory or haptic (in which the accelerator pedal becomes harder to press) feedback. Furthermore, information could be used to limit the vehicle from travelling over a speed limit through automatic application of the brakes. This could be a voluntary system, with an option to switch the device off, or be mandatory. Further advances in technology have led to the development of a more flexible system that could be spatially or contextually maintained, for example slower permitted speeds at sharp bends, or be time differentiated so only slow speeds were permitted outside schools at critical times or be dynamic enough to respond to weather or network conditions (Carsten and Fowkes, 2000a; Carsten and Fowkes, 2000b; Varhelyi and Makinen, 2001).

Insert table 1

Simulator studies have shown that people usingany ISA drove with lower speeds and have fewer trafficlight violations (Comte, 1996; Comte and Carsten, 1998; MASTER, 1999). On road studies have found reductions in excess speed and less speed variance for drivers using ISA, on a test-route (Carsten and Tait, 2000; Paatalo, Peltola and Kallio, 2001; Persson, Towliat, Almqvist, Risser and Magdeburg, 1993) or on open roads over a period of between 2 weeks and 2 months (Almqvist and Nygard, 1997; Duynstee, Katteler and Martens, 2001; Varhelyi and Makinen, 2001). Because of the link between speed and accidents, where increased speed increases the risk of an accident and the increase of accident severity (Fildes, Rumbold and Lenning, 1991; Finch, Kompfer, Lockwood and Maycock, 1994; Garber and Gadirau, 1988; Sabey, 1985; Staughton and Storie, 1977; Treat, 1980), use of an ISA should improve road safety. Indeed, there are observable improvements in safety with drivers using ISA, where less dangerous incidents, critical events and conflicts were observed by drivers using ISA (Almqvist and Nygard, 1997; Comte, 1999; Persson et al., 1993), although it must be noted that the number of events and conflicts involved were extremely small both before using ISA and during use of ISA.It has been estimated that there will be a reduction in injury accident of around 10%, if an advisory ISA is used (Carsten and Tait, 2000),to 20-40% for a system that allows for speed to be flexible to contextual, situational or network conditions (Carsten and Tait, 2000; Comte and Carsten, 1998; Verhelyi, 1996).

Field studies reported by Varhelyi and Makinen (2001) and Comte (2000) have shown that although driving speeds were slower overall when using a variety of different types of ISA,drivers adapted by not slowing at corners and intersections, driving faster than usual during poor weatherand driving with less headway due to drivers feeling increasingly stressed and frustrated while using the system.

Comte (1999), Comte, Wardman and Whelan (2000) and Varhelyi and Makinen (2001) studied the issue of acceptance following use on either simulator or field trials. They found drivers preferred the system in built-up areas and when it altered their driving least. Comte et al. (2000) found mandatory speed limiters very effective at reducing speeds but had a very low acceptance rate. Thus, it is concluded that the most effective systems are probably the least desired and least accepted (Dahlstedt, 1994; Kulmala, 1996).. A major methodological limitation to the field trials and simulator studies outlined above is that the sample size in all cases is relatively low, particularly for extended field trials which relies heavily on volunteers. This questions whether the sample is truly representative. It could be argued, for example, that drivers who intentionally drive with high risk or violation are less likely to want to trial a device that would limit their speed. Furthermore, the use a small sample of volunteers does not allow for comparative between-subjects research addressing individual differences such as gender, age, experience and attitudes towards driving and risk.

When driving a motorised vehicle, individuals are engaging in an activity with an accepted degree of risk which varies between drivers. For example, it has been shown that male drivers are significantly more likely than female drivers to drive with increased level of risk, shown, for example, by displaying faster speeds (Baxter, Manstead, Stradling, Campbell, Reason and Parker, 1990; Evans and Wasielewiski, 1982, 1983; Smeed, 1972; Wasielewiski, 1984) It is also known that younger drivers are significantly more likely than older drivers to drive with an increased level of risk, as shown, for example, by displaying faster driving speeds (Baxter et al., 1990; Fancher, Ervin, Sayer, Hagan, Bogard, Bareket, Mofford and Haugen, 1998; Fildes, et al., 1991; Quimby, Maycock, Palmer and Buttress, 1999; Quimby, Maycock, Palmer and Grayson, 1999; Smeed, 1972) However, in order to design interventions which attempt to reduce high levels of risky driving behaviour, it needs to be understood not only what group engage in risky driving behaviour but why such groups engage in risky driving behaviour. Therefore, understanding attitudes towards risk taking behaviour is of vital importance. Matthews and Moran (1986) found that younger drivers are less likely than older drivers to cite speed as a major cause of road traffic accidents and believe they were less likely to be involved in a road traffic accident. In addition Groeger and Brown (1989), McKenna Stanier and Lewis(1991) and Svensson (1981) have all found that younger drivers tend to overestimate their own skill and underestimate the skill of other drivers and this is more pronounced the younger the driver. In addition Rolls and Ingham (1992) have found that young male drivers view driving in terms of a social activity. Ellaway, Macintyre, Hiscock and Kearns (2003) found that driving is related to self-esteem for males, whereas driving is less symbolic and more practical for female drivers. These attitudinal differences should also be considered in the context of using ISA. This research will therefore investigate whether people showing different attitudes towards risk in an ordinary driving situation (i.e. without technology) will also show differing attitudes towards ISA and focus onwhether this has a perceived affect on driving behaviour.

Another important limitation of current research in the field is that the previous trials have run for a very short period of time, particularly those involving simulators. On the whole, field trails have usually run for less than 2 monthsand it could be argued that habituation and behavioural adaptation has not yet fully been realised. A further issue arises in that only a specific type of ISA is being addressed, meaning the results of simulator and field trials are likely to be an evaluation of that specific technology, rather than an investigation into possibilities of attuning different functions into an integrated device. A thorough investigation into attitudes and self-reported behaviour with regard to potential use of an ISA would help set a context for which these limitations could be overcome and challenged. Thus, this study investigates attitudes of different types of ISA and their perceived effect on driving behaviourand could be used as a basis to designing theoperation and function of ISA and aid continual evaluation and adaptation of such technology.

QUALITATIVE METHODOLOGY

Philosophical Framework

Using a grounded theory approach (Glaser, 2001; Strauss and Corbin, 1998), this study initially gathered in-depth knowledge through open-ended emergent interviews in a qualitative phase which was then analysed and contextualised before being tested using statistical analysis.

Qualitative Participants

Participants were approached whose name appeared on research volunteer databases at three Higher Education Institutions (HEI) in the South of England and in each case came from a variety of backgrounds and did not necessarily have a close connection with the HEI. It was essential volunteers were approached since the interviews, being emergent in nature and therefore largely unstructured, took a long time to complete. A total 47 participants (27 male; 20 female) were interviewed. Their ages ranged from 17 to 64 years-old and they had held a UK driving licence between 3 months and 47 years.

Qualitative Procedure

Interviews lasted around 1 hour. Each took place at a convenient time for the interviewee and were taped and transcribed. During the interview, interviewees were shown four different versions of the Intelligent Speed Adapter: A passive advisory system that displayed the current speed limit, a warning system that announced to the driver when it was going over the speed limit by a light, a buzzer or through increased resistance on the accelerator pedal (a haptic throttle) and two types of system that took over the car and limited the speed mechanically either a voluntary or a mandatory system. Interviewees were asked to comment on each ISA in general and specifically how it would affect their own driving behaviour. In line with grounded theory and emergent interviewing, salient issues raised by a previous interviewee were discussed with current and future interviewees. Transcriptions were analysed using Constant Comparative Analysis.

QUALITATIVE FINDINGS

The effect of Intelligent Speed Adaptation (ISA) on driving behaviour.

Most interviewees felt that any type of ISA could reduce accidents because attention could be directed to the road rather than on the speedometer. Although most cited that driving speeds would be reduced, the reduction in accidents linked to a reduction in speed was mentioned infrequently. Some individuals thought that accidents would increase when using any ISA system due to individuals feeling less in control and with over reliance on technology. This was particularly true for those using mandatory take over ISA.

It was mentioned that frustration and irritation may occur possibly leading to more erratic driving, particularly if using a mandatory take over system.Some individuals mentioned they could see themselves driving closer to the car in front with the mandatory take over system because they could become frustrated or irritated.

Acceptance and use of an ISA system.

The most popular system was an audible followed by a visual warning but many stressed that it should be intermittent with minimal annoyance or distraction. People tended to think using a haptic throttle and a take over system would be far too controlling and felt an advisory system would be of little benefit. Thus, the emphasis is on an aid to help driving rather than to punish fast driving through harsh and abrasive sounds. People would like around 10%-15% tolerance over the speed limit with the system. There was also a call for a more dynamic system which could take into account traffic conditions, road types, location and weather.

All drivers said that they would use a voluntary warning ISA on local residential roads, particularly those that are wide and easy to accidentally speed on. A number of drivers also mentioned the benefit of such a system on rural roads where the speed limits could be a little more ambiguous and changeable. Most cited that they would use a voluntary warning ISA in light to moderate flow of traffic with only a few mentioning it would be worth using in heavy traffic flow. A few individuals mentioned that if they were in a particular hurry the system may well be ignored.

QUANTITATIVE METHODOLOGY

A questionnaire was developed in order to analyse whether the issues deriving from the qualitative research could be rigorously tested through statistical treatment of a larger and more representative sample. In line with grounded theory, the questions reflected the answers given by the drivers in the interview section.

Quantitative Design

A questionnaire was developed consisting of 19 questions in three sections. The first section, covering questions 1 to 6, addressed background details of the respondents and addressed age, gender, length of time held UK driving licence, typical number of miles driven per week, number of accidents in previous 5 years and most common purpose for journey. The second section, consisting of question 7 with 11 sub-sections (a through to p), allowed for collection of data addressing frequency of performing certain driving behaviour in ordinary driving situations (i.e. not using ISA)
on a seven-point scale from 1 representing always to 7 representingnever (see table 2). This formed part of a wider research project in which categories of driving behaviour were determined from self-reported answers to these questions (see Musselwhite, submitted). By answering questions 7a through to 7p and analysing the results using a hierarchical cluster analysis utilising Ward’s method and Squared Euclidean distance, 386 of therespondents (it was impossible to categorise 24 of the individuals)could be assigned into one of four categories, unintentional risk takers (drivers who rarely, if ever take a risk), reactive risk takers (drivers who take a risk when reacting to external or internal stimuli, such as being late or being in a hurry), calculated risk takers (drivers who take a risk when it feels safe to do so) and continuous risk takers (drivers who take risks for most of the time while driving).

Insert Table 2

The third section on the questionnaire directly addressed attitudes towards voluntary and mandatory ISA. This included questions addressing the speed choice setting of a voluntary ISA and in addition, how the individual would like to be warned and how often they were likely to use it. With regard to mandatory take-over ISA how it hypothetically affected salient aspects of the driving experience were investigated. These were aspects that were mentioned in the initial interviews as being important, including speed, concentration, headway, stress, acceleration and braking levels and overall safety. In all cases respondents had to indicate on a seven-point scale how they thought they would be affected.

Quantitative Participants

Potential participants were handed questionnaires at motorway service stations and local garagesin the United Kingdom. In total around half the people entering refused to fill in the questionnaire and a half took but did not return the questionnaire.Overall data was collected for 410 individuals of which a total of 209 (54.3%) were male and 176 (45.7%) were female. The average age of the respondents who completed the question was 37.72 years of age, the distribution of ages in ranged from 17 to 75 years. On average across the sample, people had held their full UK driving license for 18.13 years. On average individuals stated they drove 117.11 miles per week, with a distribution which ranged from 25 miles per week to 525 miles per week. This compares favourably to the UK driving population at the time of the research (see DTLR, 2002 and Musselwhite, 2004 for an in-depth comparison).

Categorical Grouping

Using the hierarchical cluster analysis drivers were ascribed a grouping based on their self-reported driving behaviour. As would be expected the four groups of driver have distinct background characteristics (see table 3) and different self-reported driving behaviour (see table 4).