Passing Parked Police Cars: Speed as a Function of Emergency Lighting, Police Car Orientation, and Driver Experience

Transcription

1 PROCEEDINGS of the HUMAN FACTORS and ERGONOMICS SOCIETY 54th ANNUAL MEETING Passing Parked Police Cars: Speed as a Function of Emergency Lighting, Police Car Orientation, and Driver Experience Andrew K. Mayer, Jeff K. Caird, Shaunna L. Milloy, Nicole B. Percival, and Amanda D. Ohlhauser Cognitive Ergonomics Research Laboratory, University of Calgary Police vehicles and police officers working on the roadway shoulder are at risk for being struck by passing vehicles. The conspicuity of police vehicles may affect detection and speed regulation. Fifteen novice and fifteen experienced drivers participated in six experimental sessions over six months with a moderatefidelity driving simulator. Police cars were parked on the shoulder and oriented forward or backward with their emergency lights on or off. When the emergency lights were on, drivers slowed down more than when the lights were off. The orientation of the police vehicle had minimal effects on speed changes and novice and experienced drivers did not appreciably differ in their speed regulation. When compared to the criteria of reducing speed to 60 km/h when passing an emergency vehicle, which is the law in Alberta, only 16% of all drivers reduced their speed below this limit. Results suggest that drivers do not sufficiently reduce their speed in the presence of police cars and emergency vehicles should always use their emergency lights whenever they are parked on the side of the road. Copyright 2010 by Human Factors and Ergonomics Society, Inc. All rights reserved / X INTRODUCTION January 20th, 2010, a Rochester police cruiser was struck twice while responding to a traffic accident (Grossmith, 2010). January 31st 2010, a Chicago officer was injured when her squad car was rear ended while parked on the side of the road (Chicago Breaking News Center, 2010). In 2008, the Governor of Connecticut reported that 38 on-duty police vehicles were struck by passing vehicles while parked on the side of the road. These crash examples demonstrate that police vehicles and their occupants, when parked on the side of the road, can be struck by other drivers. Overall 18% of police fatalities occur when officers are struck either on the road or on the side of the road (Clarke & Zak, 1999). Of these, about two-thirds of fatalities occur within urban settings within the U.S. From 2003 to 2008, an average of 11 police workers were struck and killed each year (BLS, ). In addition, firefighters also are killed while working along roadways while directing traffic and performing roadside emergency rescues (Clarke & Zak, 1999). Police and firefighters being struck, killed and injured along roadways is a common crash type, but few studies have systematically examined why drivers make mistakes that lead to these crashes. Any vehicle parked on the roadside poses a hazard to passing drivers, but an emergency vehicle parked on the roadside posses an additional hazard because it is usually correlated with the presence of people, the emergency operator or others, working on or near the roadway within the vicinity of the vehicle. Passing drivers need to be vigilant when passing roadside emergency vehicles so that they can appropriately respond to unexpected events if they occur. Getting drivers to slow down is an important way to protect emergency operators. Governments and law enforcement agencies have worked to protect emergency workers engaged in roadside activities. Some police forces have made changes to their vehicles to improve visual conspicuity, such as adding retro-reflective decals (Fitzgerald, 2002). Drivers in the province of Alberta are required to decrease their speed to 60km/h or less if the speed limit is lower than 60 km/h or change lanes when passing road-side emergency vehicles (Driver Safety, 2005). The purpose of the law is...to make it safer for police officers, firefighters, ambulance workers...on Alberta roads (Driver Safety, 2005). Despite the law however, drivers often pass emergency vehicles at high speeds. It is unclear whether this is due to a failure to notice roadside emergency vehicles or a failure to appreciate the hazards associated with emergency vehicles. It seems reasonable to assume that drivers do not want to collide with emergency vehicles or operators, so the problem may be one of late detection, or detection errors, by passing drivers. Cognitive and perceptual errors are important reasons for the occurrence of late detection (Rumar, 1990). One type of cognitive error occurs when a driver has a partially correct expectation; they look in the correct area but do not properly interpret what they see (Rumar, 1990). For example, a driver may correctly perceive a police car parked on the shoulder but incorrectly interpret the vehicle to be moving with traffic (Langham, Hole, Edwards, & O Neil, 2002). A perceptual error can occur when an object in the environment fails to reach sufficient visual thresholds, either due to poor contrast or reduced ambient illumination (Rumar, 1990). A driver s ability to detect an object is reliant on the object s conspicuity; the probability of seeing a target object during a short viewing time. Perceptual conspicuity is determined by the physical characteristics of the object relative to its background (Hughes & Cole, 1986). Attentional conspicuity refers to how well an object is detected when the observer is not actively searching for the object (Hughes & Cole, 1986). Langham and colleagues (2002) investigated whether accidents involving roadside emergency vehicles were the result of cognitive rather than perceptual failures. They manipulated the orientation of a police car such that it was oriented in-line with traffic or at an angle to traffic. Participants were asked to respond to police cars they detected while watching videos depicting a drive along a roadway filmed from first-person view. The authors found that experienced drivers were faster to identify the police car when the vehicle was parked at an angle. The authors argued that drivers correctly detected emergency vehicles (i.e., they were

2 PROCEEDINGS of the HUMAN FACTORS and ERGONOMICS SOCIETY 54th ANNUAL MEETING seen) but misinterpreted what they saw (cognitive error). That is, a police vehicle oriented straight ahead may have been perceived as a moving vehicle because it matched drivers expectations. The police car parked at an angle violated this expectation (was more attentionally conspicuous) and was thus detected earlier as a hazard (Langham et al.). The study by Langham and colleagues (2002) is interesting but suffers a logical weakness. The authors discuss the importance of emergency lighting, but they do not actually manipulate the presence of police emergency lights. This is of particular importance because emergency lighting serves two important functions: (1) to help identify the vehicle as an emergency vehicle and (2) to help drivers detect an emergency vehicle from a greater distance (Dunn & Tunnicliff, 2009). In the case of emergency vehicles, errors may be attributable to both cognitive and perceptual errors. Improving characteristics of the vehicle that improve attentional conspicuity and perceptual conspicuity may improve safety. The objective for the present study was to determine if increasing the attentional conspicuity (by violating driver expectations and thus supporting cognition) in conjunction with increasing the perceptual conspicuity (improving emergency vehicle visibility and thus supporting perception) of roadside police cars would improve safety. Attentional conspicuity was manipulated by orienting the police car forward, facing the direction of traffic (lower attentional conspicuity) or by orienting the police car backward, facing against the direction of traffic (higher attentional conspicuity). Perceptual conspicuity was manipulated by turning the emergency lights on (higher perceptual conspicuity) or off (lower perceptual conspicuity). In addition, the effect of driver experience was evaluated (novice versus experienced drivers). It was expected that the combination of higher attentional and perceptual conspicuity would result in slower passing speeds by drivers and the greatest decrease in passing speeds. It was also expected that experienced drivers would exhibit greater decreases in speed when passing the more attentionally conspicuous police car since experienced drivers have a stronger mental model compared to novice drivers and the backward oriented police car would more greatly violate their expectations. Apparatus METHODS A moderate-fidelity driving simulator was used to examine driver behaviour when passing police cars. A Saturn car was interfaced to a network of I/O, graphics, and data collection computers. Hyperdrive (version ) was used to create the driving environment, events, and collect driving data. Participants Fifteen novice (eight male, seven female) and fifteen experienced (eight male, seven female) drivers participated. Participants were recruited using posters placed in the communities surrounding the University and community newsletters. Participants were paid $165 total for volunteering for seven 90-minute sessions over six months. Stringent inclusion criteria were used as the primary goal of the overall study was to investigate driving behavior during the first six months of independent driving. Participants were screened over the phone for age, simulator sickness, yearly driving history and driver s license status. Novice drivers were 16 or 17 years old and in their first month holding a graduated class 5 driver s license (M = 16.2 years of age, SD = 0.41). Experienced drivers held a class 5 driver s license, had at least 10 years of driving experience, and reported driving at least 10,000 kilometers per year (M = 32.90, SD = 5.06 years). To enter into the study, all participants had to meet the following visual standards: a minimum corrected visual acuity of 20/40 for driving in Alberta (Casson & Recette, 2000), score within the normal range of their age group on contrast sensitivity, and have normal color vision. Procedure The results presented in this paper are a subset of data from a large-scale longitudinal study of novice driver perceptual learning of hazards during the first six months of driving. Participants attended one screening and six experimental sessions over a six-month period (i.e., one experimental session every month). Each experimental session was comprised of a seven-minute practice drive and two experimental drives, each lasting about 12 to 15 minutes, for a total of 12 experimental drives. During the course of the study, participants drove through residential, urban, industrial, and freeway environments and encountered a number of hazards including, pedestrians, cyclists, unexpected lead vehicles braking, and late yellow lights. Participants were asked to drive as they normally would in their own vehicles and obey all of the rules of the road including traffic lights, signs and posted speed limits. Police Vehicles In eight of the twelve experimental drives, a police car was parked on the hard shoulder at the right side of the road. In four of the experimental drives, the police car was parked along a single lane road and in the other four drives the police car was parked along a double lane road. The results presented here are from the single lane roadway. The police car was black and white with blue and red emergency lights on top that could be turned on or off in the simulator (See Figure 1 for an example of the police cars used in the study). In two cases (drives 4 and 9) the car appeared after a curve in the roadway; in two cases (drives 3 and 10) the car appeared on a straight stretch of roadway. In drive 3, the police car was oriented backwards with the emergency lights on; in drive 4, the police car was oriented backwards with the emergency lights off; in drive 9, the police car was oriented forward with the emergency lights on; and in drive 10, the police car was oriented forward with the emergency lights off (See Table 1).

3 PROCEEDINGS of the HUMAN FACTORS and ERGONOMICS SOCIETY 54th ANNUAL MEETING ), F(1, 28) = 11.26, p =.002. Participants also decreased their speed more when the police emergency lights were on (M = 15.9, SD = 14.1) compared to when they were off (M = 10.6, SD = 12.1), F(1, 28) = 6.56, p =.016. There was no significant main effect of experience (F(1, 28) < 1) and no statistically significant interactions. Figure 1. Example of a police car oriented backwards (lights off) on the roadside. Table 1. Emergency vehicle condition (orientation and lights) during four experimental drives. Drive 3 Drive 4 Drive 9 Drive 10 Emergency Lights On Off On Off Orientation Backwards Backwards Forward Forward Road Geometry Straight Curve Straight Curve RESULTS Mixed-model ANOVAs were computed with police car orientation (forward, backward) and emergency lights (on, off) as the within-subjects factors and experience (novice, experienced) as a between-subjects factor. Speed as a percentage of the posted speed limit and change in speed were the dependent measures. Follow up analyses were calculated using a Bonferonni correction. Change in Speed Change in speed was calculated by subtracting drivers passing speed (i.e., the speed they were traveling when they were directly next to the parked police car) from the initial speed (i.e., the speed they were traveling before beginning to decelerate). Thus, positive numbers represent decreases in velocity. Overall, participants significantly reduced their speed when passing the police car, from an average of 84.2 km/h (SD = 4.1) to an average of 72.9 km/h (SD = 8.37), t(29) = 8.16, p <.001. Additionally, the initial speeds of novice (M = 84.3, SD = 4.37) and experienced drivers (M = 84.2, SD = 4.0) were statistically equivalent (t(28) =.06, p >.05) as were the passing speeds of novice (M = 72.2, SD = 9.3) and experienced drivers (M = 73.6, SD = 7.6) (t(28) =.46, p >.05). Figure 2 illustrates the change in speed data. When the police car was oriented forward (i.e., facing the same direction as traffic flow), participants slowed down more (M = 15.8, SD = 14.5) than when the police car was oriented backwards (i.e., facing the opposite direction as traffic flow) (M = 10.7, SD = Figure 2. Drivers change in speed when the parked police car was oriented either forward or backward and the emergency lights were either on or off. Passing Speed as a Percentage of Posted Speed Limit To calculate passing speed, the driver s speed when aligned directly next to the police car was taken as a percentage of the posted speed limit. Percentage values below 100 represent speeds that fall below the posted speed limit whereas percentage values above 100 represent speeds that exceed the posted speed limit. On average, novice drivers (M = 86.8, SD = 14.0) and experienced drivers (M = 86.8, SD = 12.81) drove below the posted speed limit when passing the police car on the side of the road. Participants drove slower (relative to the posted speed limit) when the police car was oriented forward (M = 84.5, SD = 15.4) compared to when the police car was oriented backwards (M = 88.4, SD = 11.7), F(1, 28) = 6.81, p =.002. When the police emergency lights were on participants drove slower (M = 83.8, SD = 15.2) compared to when the police lights were off (M = 89.0, SD = 11.9), F(1, 28) = 6.96, p =.013 (See Figure 3).

4 PROCEEDINGS of the HUMAN FACTORS and ERGONOMICS SOCIETY 54th ANNUAL MEETING Backward Lights Off 2 13 Experienced Forward Lights On 6 9 Forward Lights Off 0 15 Backward Lights On 0 15 Backward Lights Off 2 13 DISCUSSION Figure 3. Drivers speed as a percentage of the posted speed limit when the parked police car was oriented either forward or backward and the emergency lights were either on or off. There was a statistically significant orientation by emergency light interaction (F(1, 28) = 9.49, p =.005). When the police car had its emergency lights on and was oriented forward, participants drove 85.4% (SD = 18.2) of the posted speed compared to a statistically comparable 83.5% (SD = 12.5) of the posted speed when the police car was oriented forward with its emergency lights off, t(29) =.51, p > There was no statistical difference in speed as a percentage of the posted limit between the police car oriented forward with its emergency lights on (M = 85.4, SD = 18.2) and the police car oriented backward with its emergency lights on (M = 82.25, SD = 12.21), t(29) = 1.11, p > When the police car was oriented backwards with its emergency lights off participants drove much faster (M = 94.5, SD = 11.3) compared to when the police car was backwards with the lights on (t(29) = 5.15, p <.001) or when the police car was forward with its lights off (t(29) = 3.84, p =.001). Passing Speed Relative to the Legal Passing Speed in Alberta The legal speed limit for passing a roadside emergency vehicle in the province of Alberta is 60 km/h. Although participants, as a whole, reduced their speed when they passed the police cars, most participants failed to reduce their speed to the legal 60 km/h. More novice and experienced drivers slowed to the legal speed limit when the police car was oriented forward with the emergency lights on compared to the other conditions (Table 2). Table 2 Total number of participants, by condition, who drove above and below 60km/h when passing the police car. Condition # of Participants Below 60km/h # of Participants Above 60km/h Novice Forward Lights On 6 9 Forward Lights Off 1 14 Backward Lights On 2 13 Emergency vehicles and their occupants are often seen parked on the roadside for various reasons while traffic continues to flow around them. This can be a dangerous situation for emergency responders as evidenced by the frequency with which responders and their vehicles are struck by passing traffic. In the province of Alberta, legislators have attempted to improve safety for roadside emergency vehicles by creating laws that mandate slower passing speeds (Driver Safety, 2005). However, drivers continue to pass emergency vehicles at high speeds, putting emergency workers at risk. In the current study, the orientation of the police car and the illumination of the emergency lights were manipulated to determine if they would affect passing speeds. Most drivers did slow to some extent, but only 16% of all drivers passed the police cars at speeds at or below 60 km/h. The orientation of the police cars was manipulated to improve the attentional conspicuity of the vehicles. The backward police car was thought to violate expectations and result in slower passing speeds because it can be considered uncommon to have a vehicle facing the wrong direction on the side of the road. However, results show that drivers slowed down more when the police car was oriented forwards. It is possible that drivers interpreted the police car in terms of the action that it could take. A backward facing police is not likely to pull into traffic whereas a forward facing police car is more likely to do so. Participants may have slowed their passing speed when the police car was forward to compensate for this potential pullout action. In general, it was found that turning the emergency lights on resulted in a greater reduction of speed and slower passing speeds compared to when the lights were off. Emergency lights are designed to help identify emergency vehicles and to improve their perceptual conspicuity (Dunn & Tunnicliff, 2009). This appears to have been the case. Additionally, flashing emergency lights may improve the attentional conspicuity since they provide information to passing drivers, that is, that an emergency situation of some degree is occurring. This information is important to convey to passing drivers and encourages safer passing speeds. The safest driving behavior was observed when the presence of flashing emergency lights was combined with a forward orientation. This combination resulted in a 40% compliance with Alberta laws (i.e., slowing to 60 km/h). In contrast, orienting the police car backwards had a relatively small effect on passing speeds over and above effects related to turning the emergency lights on. Emergency vehicle operators are at risk every time they work around moving traffic. Eighteen percent of police

5 PROCEEDINGS of the HUMAN FACTORS and ERGONOMICS SOCIETY 54th ANNUAL MEETING fatalities in the US occur when police officers are on the road or on the side of the road (Clarke & Zak, 1999). Additionally, firefighters, EMS, and tow-truck drivers are at risk when working on the roadside. The greater risk for emergency operators is not when they are in their vehicle on the roadside but when they are outside their vehicles. For this reason, it is imperative that passing drivers significantly reduce their speeds so they can safely respond to unexpected events that can occur around emergency vehicles. It is clear from the current research that drivers do not sufficiently and safely reduce their speed when passing police cars. Further research is required to develop ways of improving safety for emergency operators and to alert drivers of hidden risks posed by roadside emergency situations. At the very least, emergency flashing lights should be used any time emergency vehicles are parked on the roadside. One limitation of the present study was that the luminance of the police car flashing lights was much less than what would be experienced on the road. Thus, these results probably underestimate the true effects of police car flashing lights. Additionally, the police cars were parked along different road geometries, however, this likely improves the overall generalizability of the findings. The findings from this study are somewhat at odds with the Langham et al. (2002) results. Langham et al. concluded that cognitive processing is of greater importance when detecting parked police cars. However, their manipulation (i.e., angled police cars versus forward facing police cars) may have actually changed the perceptual properties of the police car. The angled police car (where the rear and side of the car is visible) may project a larger image on the retina and markings (paint, decals, etc.) may differ between the angled and forward views. The present study held such differences constant. Additionally, the use of flashing emergency lights may provide added information that contributes to the overall conspicuity of the vehicle. CONCLUSION This study investigated the effects of emergency vehicle orientation and emergency lights on detecting and reacting to roadside police vehicles. Overall, the use of emergency lights had a greater influence on reducing drivers passing speeds than the orientation of the police car. However, orienting forward on the roadway does appear to be safer than orienting it backwards. It is reasonable to assume that these results would extend to other emergency vehicles as well. The findings have implications for emergency vehicles that have to park on the side of the road. It is recommended, based on the current study, that emergency vehicles should always turn on their emergency lights and orient their vehicles forward when parked on the side of the roadway to improve their safety. REFERENCES Bureau of Labor Statistics ( ). Fatal occupational injuries resulting from transportation incidents and homicides, All United States. [Retrieved June 12, 2010 from: Casson, E.J., & Recette,L. (2000). Vision standards for driving in Canada and the United States. A review for the Opthalmological Society. Canadian Journal of Opthalmology, 35, Chicago Breaking News Center (2010). Chicago cop injured after squad car rear-ended. Retrieved from the world wide web on February 20, 2010 from: go-cop-injured-after-squad-car-rear-ended.html Clarke, C. & Zak, M.J. (1999, Summer). Fatalities to law enforcement officers and firefighters, Compensation and Working Conditions, 3 7. Driver Safety (2005). Traffic Safety in Alberta. Retrieved from the world wide web on February 20, 2010 from: aq.html#q3 Dunn, K., & Tunnicliff, D. (2005). The safety and effectiveness of emergency vehicle lighting. Proceedings Road Safety Research, Policing and Education, Grossmith, P. (2010). Rochester police cruiser hit twice. NewHampshire.com. Retrieved from the world wide web on February 20, 2010 from: ochester+police+cruiser+hit+twice&articleid=3fd2f9 28-1dd6-4be5-b651-c9864e995eb0 Hughes, P.K., & Cole, B.L. (1986a). What attracts attention while driving? Ergonomics, 29(3), Hughes, P.K., & Cole, B.L. (1986). Can the conspicuity of objects be predicted from laboratory experiments? Ergonomics, 29(9), Langham, M., Hole, G., Edwards, J., & O Neil, C. (2002). An analysis of looked but failed to see accidents involving parked police vehicles. Ergonomics, 45(3), Rumar, K. (1990). The basic driver error: Late detection. Ergonomics, 33(10-11), ACKNOWLEDGEMENTS The authors are grateful to a number of students who helped to recruit and run participants including Aimee Pearson, Greg Halihan, and Sebastian Siwiec. This research was funded by AUTO21 Network of Centres of Excellence.