| 1. |
- Bianchi Piccinini, Giulio, 1982, et al.
(författare)
-
How Do Drivers Respond to Silent Automation Failures? Driving Simulator Study and Comparison of Computational Driver Braking Models
- 2020
-
Ingår i: Human Factors. - Chalmers University of Technology, Gothenburg, Sweden.; Volvo Group Trucks Technology, Gothenburg, Sweden.; Virginia Tech Transportation Institute, Blacksburg, USA.; University of Leeds, UK.; VTI, Gothenburg, Sweden. : SAGE Publications. - 1547-8181 .- 0018-7208. ; 62:7, s. 1212-1229
-
Tidskriftsartikel (refereegranskat)abstract
- Objective: This paper aims to describe and test novel computational driver models, predicting drivers’ brake reaction times (BRTs) to different levels of lead vehicle braking, during driving with cruise control (CC) and during silent failures of adaptive cruise control (ACC). Background: Validated computational models predicting BRTs to silent failures of automation are lacking but are important for assessing the safety benefits of automated driving. Method: Two alternative models of driver response to silent ACC failures are proposed: a looming prediction model, assuming that drivers embody a generative model of ACC, and a lower gain model, assuming that drivers’ arousal decreases due to monitoring of the automated system. Predictions of BRTs issued by the models were tested using a driving simulator study. Results: The driving simulator study confirmed the predictions of the models: (a) BRTs were significantly shorter with an increase in kinematic criticality, both during driving with CC and during driving with ACC; (b) BRTs were significantly delayed when driving with ACC compared with driving with CC. However, the predicted BRTs were longer than the ones observed, entailing a fitting of the models to the data from the study. Conclusion: Both the looming prediction model and the lower gain model predict well the BRTs for the ACC driving condition. However, the looming prediction model has the advantage of being able to predict average BRTs using the exact same parameters as the model fitted to the CC driving data. Application: Knowledge resulting from this research can be helpful for assessing the safety benefits of automated driving.
|
|
| 2. |
- Engström, Johan A Skifs, 1973, et al.
(författare)
-
Adaptive behavior in the simulator: Implications for active safety system evaluation
- 2011
-
Ingår i: Handbook of Driving Simulation for Engineering, Medicine, and Psychology. - 9781420061017 ; , s. 41-1-41-16-
-
Bokkapitel (övrigt vetenskapligt/konstnärligt)abstract
- The Problem. Driving is, most of the time, a self-paced task where drivers proactively control the driving situation, based on their expectations of how things will develop in the near future. Crashes are typically associated with unexpected events where this type of proactive adaptation failed in one way or another. These types of scenarios are the main targets for active safety systems. In evaluation studies, drivers’ responses to expected events may be qualitatively different from responses to similar, but unexpected, events. Hence, creating artificial active safety evaluation scenarios that truly represent the targeted real-world scenarios is a difficult challenge. Role of Driving Simulators. Driving simulators offer great possibilities to test active safety systems with real drivers in specific target scenarios under tight experimental control. However, in simulator studies, experimental control generally has to be traded against realism. The objective of this chapter is to address some key problems related to driver expectancy and associated adaptive behavior in the context of simulator-based active safety system evaluation. Key Results of Driving Simulator Studies. The chapter briefly reviews common types of adaptive driver strategies found in the literature and proposes a general conceptual framework for describing adaptive driver behavior. Based on this framework, some key challenges in dealing with these types of issues in simulator studies are identified and potential solutions discussed. Scenarios and Dependent Variables. Key variables representing adaptive driver behavior include the selection of speed, headway, and lane position as well as the allocation of attention and effort. It will never be possible to create artificial simulator scenarios for active safety evaluation that perfectly match their real-world counterparts, but there are several means that could be used to reduce the discrepancy. Problems with expectancy and resulting adaptive behavior may at least be partly overcome by various means to “trick” drivers into critical situations, several of which are addressed in the chapter. Platform Specificity and Equipment Limitations. The issues discussed in this chapter should apply across all types of driving simulator platforms. However, some of the proposed methods for tricking drivers into critical situations may require specific simulator features, such as a motion base.
|
|
| 3. |
- Engström, Johan A Skifs, 1973, et al.
(författare)
-
Effects of working memory load and repeated scenario exposure on emergency braking performance
- 2010
-
Ingår i: Human Factors. - : SAGE Publications. - 1547-8181 .- 0018-7208. ; 52:5, s. 551-559
-
Tidskriftsartikel (refereegranskat)abstract
- Objective: The objective of the present study was to examine the effect of working memory load on drivers' responses to a suddenly braking lead vehicle and whether this effect (if any) is moderated by repeated scenario exposure. Background: Several experimental studies have found delayed braking responses to lead vehicle braking events during concurrent performance of nonvisual, working memory-loading tasks, such as hands-free phone conversation. However, the common use of repeated, and hence somewhat expected, braking events may undermine the generalizability of these results to naturalistic, unexpected, emergency braking scenarios. Method: A critical lead vehicle braking scenario was implemented in a fixed-based simulator. The effects of working memory load and repeated scenario exposure on braking performance were examined. Results: Brake response time was decomposed into accelerator pedal release time and accelerator-to-brake pedal movement time. Accelerator pedal release times were strongly reduced with repeated scenario exposure and were delayed by working memory load with a small but significant amount (178 ms). The two factors did not interact. There were no effects on accelerator-to-brake pedal movement time. Conclusion:The results suggest that effects of working memory load on response performance obtained from repeated critical lead vehicle braking scenarios may be validly generalized to real world unexpected events. Application: The results have important implications for the interpretation of braking performance in experimental settings, in particular in the context of safety-related evaluation of in-vehicle information and communication technologies. © 2010, Human Factors and Ergonomics Society.
|
|
| 4. |
- Ljung Aust, Mikael, 1973, et al.
(författare)
-
A conceptual framework for requirement specification and evaluation of active safety functions
- 2011
-
Ingår i: Theoretical Issues in Ergonomics Science. - : Informa UK Limited. - 1464-536X .- 1463-922X. ; 12:1, s. 44-65
-
Tidskriftsartikel (refereegranskat)abstract
- Active safety functions intended to prevent vehicle crashes are becoming increasingly prominent in traffic safety. Successful evaluation of their effects needs to be based on a conceptual framework, i.e. agreed-upon concepts and principles for defining evaluation scenarios, performance metrics and pass/fail criteria. The aim of this paper is to suggest some initial ideas toward such a conceptual framework for active safety function evaluation, based on a central concept termed 'situational control'. Situational control represents the degree of control jointly exerted by a driver and a vehicle over the development of specific traffic situations. The proposed framework is intended to be applicable to the whole evaluation process, from 'translation' of accident data into evaluation scenarios and definition of evaluation hypotheses, to selection of performance metrics and criteria. It is also meant to be generic, i.e. applicable to driving simulator and test track experiments as well as field operational tests.
|
|
| 5. |
- Ljung Aust, Mikael, 1973, et al.
(författare)
-
Effects of forward collision warning and repeated event exposure on emergency braking
- 2013
-
Ingår i: Transportation Research Part F: Traffic Psychology and Behaviour. - : Elsevier BV. - 1369-8478. ; 18, s. 34-46
-
Tidskriftsartikel (refereegranskat)abstract
- Many experimental studies use repeated lead vehicle braking events to study the effects of forward collision warning (FCW) systems. It can, however, be argued that the use of repeated events induce expectancies and anticipatory behaviour that may undermine validity in terms of generalisability to real-world, naturalistic, emergency braking events. The main objective of the present study was to examine to what extent the effect of FCW on response performance is moderated by repeated exposure to a critical lead vehicle braking event. A further objective was to examine if these effects depended on event criticality, here defined as the available time headway when the lead vehicle starts to brake. A critical lead vehicle braking event was implemented in a moving-base simulator. The effects of FCW, repeated event exposure and initial time headway on driver response times and safety margins were examined. The results showed that the effect of FCW depended strongly on both repeated exposure and initial time headway. In particular, no effects of FCW were found for the first exposure, while strong effects occurred when the scenario was repeated. This was interpreted in terms of a switch from closed-loop responses triggered reactively by the situation, towards an open-loop strategy where subjects with FCW responded proactively directly to the warning. It was also found that initial time headway strongly determined response times in closed-loop conditions but not in open-loop conditions. These results raise a number of methodological issues pertaining to the design of experimental studies with the aim of evaluating the effects of active safety systems. In particular, the implementation of scenario exposure and criticality must be carefully considered.
|
|
| 6. |
- Ljung Aust, Mikael, 1973, et al.
(författare)
-
Manual for DREAM version 3.2
- 2012
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- The Driving Reliability and Error Analysis Method (DREAM) is based on the Cognitive Reliability and Error Analysis Method (CREAM; Hollnagel, 1998). CREAM was developed to analyse accidents within process control domains such as nuclear power plants and train operation, and DREAM is an adaptation of CREAM to suit the road traffic domain. The purpose of DREAM is to make it possible to systematically classify and store accident and incident causation information. This means that DREAM, like all other methods for accident/incident analysis, is not a provider but an organiser of explanations. For any of the contributing factor categories available in DREAM to be used, it must be supported by relevant empirical information. DREAM in itself cannot tell us why accidents happen (if it could, we would need neither on-scene investigations nor interviews).DREAM includes three main components: an accident model, a classification scheme and a detailed procedure description which step by step goes through what needs to be done in order to perform a DREAM analysis on an investigated accident/incident. Below, the accident model will be given more detailed descriptions. After this follows a description of the classification scheme, and then comes the analysis process, including example cases and recommendations for how to do the categorisation in certain typical scenarios.
|
|
| 7. |
- Nilsson, Emma, 1982, et al.
(författare)
-
Effects of cognitive load on response time in an unexpected lead vehicle braking scenario and the detection response task (DRT)
- 2018
-
Ingår i: Transportation Research Part F: Traffic Psychology and Behaviour. - : Elsevier BV. - 1369-8478. ; 59, s. 463-474
-
Tidskriftsartikel (refereegranskat)abstract
- The effects of cognitive distraction on traffic safety and driver performance are unclear and under debate. Based on increased response times to stimuli or events in controlled driving experiments, concerns, primarily about cell phone usage during driving, have been raised. But while cognitive load repeatedly have been shown to increase response times in artificial tasks such as the Detection Response Task (DRT), the generalizability of the results to response times in critical traffic situations is questionable. Method: Two experiments were conducted. In Experiment 1, response times in the DRT were measured during simulated driving with and without execution of a cognitively loading secondary task. In Experiment 2, brake response times in an unexpected lead vehicle braking scenario were measured with and without the same cognitively loading task. Results: In Experiment 1, DRT response times increased with increased level of cognitive load. In Experiment 2, brake response times were unaffected by cognitive load. Conclusion: The response time results from the artificial DRT did not generalize to the critical lead vehicle braking scenario. This finding can possibly be explained by the cognitive control hypothesis, which suggests that cognitive load selectively impairs driving subtasks that rely on cognitive control (i.e. novel or inconsistent tasks) but leaves automatic performance unaffected (Engström, Markkula, Victor, & Merat, 2017). While the DRT responses, because of the task novelty, can be assumed to require cognitive control, responses to visually expanding objects, such as a braking lead vehicle with short time headway, are triggered automatically. Common interpretations of the effect of cognitive load on traffic safety thus need to be re-examined. It seems inappropriate to generalize from effects of cognitive load on DRT, or other artificial laboratory tasks that rely on cognitive control, to unexpected real-world situations where responses are triggered primarily by looming cues.
|
|
| 8. |
- Bianchi Piccinini, Giulio, 1982, et al.
(författare)
-
Factors contributing to commercial vehicle rear-end conflicts in China: A study using on-board event data recorders
- 2017
-
Ingår i: Journal of Safety Research. - : Elsevier BV. - 0022-4375. ; 62, s. 143-153
-
Tidskriftsartikel (refereegranskat)abstract
- Introduction: In the last 30 years, China has undergone a dramatic increase in vehicle ownership and a resulting escalation in the number of road crashes. Although crash figures are decreasing today, they remain high; it is therefore important to investigate crash causation mechanisms to further improve road safety in China. Method: To shed more light on the topic, naturalistic driving data was collected in Shanghai as part of the evaluation of a behavior-based safety service. The data collection included instrumenting 47 vehicles belonging to a commercial fleet with data acquisition systems. From the overall sample, 91 rear-end crash or near-crash (CNC) events, triggered by 24 drivers, were used in the analysis. The CNC were annotated by three researchers, through an expert assessment methodology based on videos and kinematic variables. Results: The results show that the main factor behind the rear-end CNC was the adoption of very small safety margins. In contrast to results from previous studies in the US, the following vehicles' drivers typically had their eyes on the road and reacted quickly in response to the evolving conflict in most events. When delayed reactions occurred, they were mainly due to driving-related visual scanning mismatches (e.g., mirror checks) rather than visual distraction. Finally, the study identified four main conflict scenarios that represent the typical development of rear-end conflicts in this data. Conclusions: The findings of this study have several practical applications, such as informing the specifications of in-vehicle safety measures and automated driving and providing input into the design of coaching/training procedures to improve the driving habits of drivers.
|
|
| 9. |
- Bärgman, Jonas, 1972, et al.
(författare)
-
ANNEXT - Slutrapport
- 2013
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- Huvudmålet med ANNEXT har varit att utvärdera och illustrera styrkan i att använda redan insamlade externa naturalistiska kördata från DriveCams datainsamling (www.drivecam.com), för att öka förståelsen kring vad som orsakar trafikolyckor. Fram till nyligen har det inte funnits data som medger detaljerade studier av vad som sker i de viktiga sekunderna precis innan en olycka (eller nästan-olycka) i verklig trafik. Det vill säga, det har inte varit möjligt att studera t.ex. förares aktiviteter (t.ex. distraktioner såsom mobilanvändande) eller hur interaktionen mellan fordonen och dess förare faktiskt gått till under sekunderna före krock. Vi har använt oss av data from händelseinitierade inspelningsenheter i kommersiella fordon, där data inte samlats in som en del i något forskningsprojekt, utan som en del i en affärsverksamhet där DriveCam sålt en tjänst till företag som har flottor med fordon. Det vill säga, dessa händelseinspelare installeras inte i fordon för att specifikt forska på orsaker till olyckor, utan som en del i en tjänst för att förbättra trafiksäkerheten för just dessa företag. Ett tydligt mål är att få ner stilleståndskostnader för företagens fordon vid olyckor, men det har också effekten att antalet olyckor (och därmed skadade och dödade) minskar på vägarna där de körs. Fordonen körs som en del i den vanliga verksamheten hos företagen. När en olycka eller nästan-olycka identifieras med hjälp av kinematiska triggers (t.ex. tröskelvärden på acceleration), sparas data 8 sekunder föra och 4 sekunder efter triggern i inspelningsenheten. Data innefattar bl.a. GPS, video på föraren och framåt samt accelerometerdata. Denna skickas sedan tillbaks till DriveCam där en genomgång av alla händelser görs och varje event klassas för trafiksäkerhetsrelevans. I ANNEXT har vi fått tillgång till 100 påkörandehändelse (70 olyckor och 30 nästan-olyckor) samt 93 korsningshändelser (63 olyckor och 30 nästan-olyckor). Alla händelser har kodats av DriveCam-personal och vi på SAFER har sedan analyserat data. Följande beskriver processen genom projektet: Första steget var att ta fram en preliminär analysplan, baserad på projektansökan och ytterligare identifierade behov. Steg två var att kontakta DriveCam och få till ett kontrakt inom vilkets ramar projektet kunde genomföras. Scenarios identifierades och kriterier för hur händelser skulle väljas ut utvecklades och itererades mot DriveCam. Detaljerna är beskrivna i en publikation (Engström, Werneke, et al., 2013) från projektet (se separat avsnitt). Analysplanen har förfinats allteftersom i projektet. Som en del i analysplanen utvecklades en annoteringsbeskrivning, eller kodbok. Denna beskriver de variabler som vi identifierat som nödvändiga för analysen och som kräver manuell annotering av DriveCam-video. I ANNEXT utförde en dedikerad annoterare på DriveCam all primär videoannotering. För att förfina kodboken och verifiera annotering åkte SAFER-deltagare till DriveCam vid två tillfällen. När den huvudsakliga annotering var avslutad genomförde SAFER-partners 1) utveckling av metoder för extraktion av optiska parameterar baserat på annoterad video (Bärgman et al., 2013), och processade dessa data för att kvalitetssäkra inför analys, 2) iterativ vidareutveckling av kodningsschemat som kan kallas Kodbok för bidragandefaktorer (se resultat) och applicerade denna på tillgänglig data. Under våren och sommaren 2013 presenterades två vetenskapliga artiklar på konferenser. Ytterligare publikationer är under utveckling (se separat avsnitt)
|
|
| 10. |
- Bärgman, Jonas, 1972, et al.
(författare)
-
Using manual measurements on event recorder video and image processing algorithms to extract optical parameters.
- 2013
-
Ingår i: Proceedings of the Seventh International Driving Symposium on Human Factors in Driver Assessment, Training and Vehicle Design. ; , s. 177-183
-
Konferensbidrag (refereegranskat)abstract
- Vehicle kinematics and optical parameters such as optical angle, optical expansion rate, and tau are thought to underlie drivers’ ability to avoid and handle critical traffic situations. Analyses of these parameters in naturalistic driving data with video, such as commercial event recordings of near-crashes and crashes, can provide insight into driver behavior in critical traffic situations. This paper describes a pair of methods, one for the range to a lead vehicle and one for its optical angle, that are derived from image processing mathematics and that provide driver behavior researchers with a relatively simple way to extract optical parameters from video-based naturalistic data when automatic image processing is not possible. The methods begin with manual measurements of the size of other road users on a video on a screen. To develop the methods, 20 participants manually measured the width of a lead vehicle on 14 images where the lead vehicle was placed at different distances from the camera. An on-market DriveCam Event Recorder was used to capture these images. A linear model that corrects distortion and modeling optics was developed to transform the on-screen measurements distance (range) to and optical angle of the vehicle. The width of the confidence interval for predicted range is less than 0.1m when the actual distance is less than 10m and the lead-vehicle width estimate is correct. The methods enable driver behavior researchers to easily and accurately estimate useful kinematic and optical parameters from videos (e.g., of crashes and near-crashes) in event-based naturalistic driving data.
|
|
