Using Event Data Recorder (EDR) data to perform What-if simulations for safety benefit analysis by reconstructing real traffic kinematics and driver behaviors

Visa enkel post
dc.contributor.authorRao, Rakshith Mukunda
dc.contributor.departmentChalmers tekniska högskola / Institutionen för tillämpad mekaniksv
dc.contributor.departmentChalmers University of Technology / Department of Applied Mechanicsen
dc.date.accessioned2019-07-03T14:32:10Z
dc.date.available2019-07-03T14:32:10Z
dc.date.issued2017
dc.description.abstractAs the focus of traffic safety stakeholders shifts from passive safety to active safety, the need to predictively evaluate safety systems in addition to identifying driver behaviour in critical situations has come to the forefront. Availability of a wide variety of data has opened up new research possibilities; real world crash data is increasingly accessible through Event Data Recorder (EDR) data, although little information is available about the context of the crash. Naturalistic Driving Data (NDD) addresses this issue by monitoring the driver and vehicle with the help of cameras and sensors. However, there is a lack of real world crashes associated with NDD. Therefore, the need to combine data sources to identify driver behaviour and estimate safety benefit has never been higher. In this study Counterfactual or ‘What-if’ simulations are performed with EDR data of real rear-end crashes and driver glance behaviours inspired from NDD to assess the impact of different driver glance behaviour on possible outcomes. The data was extracted from the National Automotive Sampling System-Crashworthiness Data System (NASS-CDS) database for use in the What-if simulations. Artificial kinematics or Counter factual kinematics was created by removing the evasive (braking) manoeuver. A glance anchor point (AP) based on literature was chosen. Two distributions (Baseline glance distribution, Reaction time distribution) and a deceleration value were applied to the kinematics at the AP. The Baseline Glance distribution which represented normal everyday driving was inspired from NDD. The reaction time distribution and deceleration value was created and chosen based on literature respectively. The application of the combination of distributions resulted in counter factual outcomes with two possibilities: Crash or No Crash. The Impact speeds of all the counter factual events that resulted in a Crash were calculated. Another batch of simulations was performed replacing the Baseline Glance distribution with the Rockwell glance distribution. The Rockwell glance distribution represented the glance durations associated with a well-known secondary task of tuning a radio. Results were compared between outcomes from Baseline Glance Distribution and Rockwell Glance distribution with the Original crash data as reference. The results showed that the Baseline Glance distribution had lower percentage of crashes when compared to the Rockwell glance distribution. The impact speeds associated with the Rockwell glance distribution were much higher than the Baseline. However, the impact speeds resulting from both distributions were much lower compared to the impact speeds associated with the original crashes which clearly indicated the high severity of the original crashes. The methodology and results from this study provide the necessary framework to evaluate the benefit of rear end collision avoidance safety systems. Also, a basis for understanding driver behaviour prior to critical situations is provided.
dc.identifier.urihttps://hdl.handle.net/20.500.12380/250499
dc.language.isoeng
dc.relation.ispartofseriesDiploma work - Department of Applied Mechanics, Chalmers University of Technology, Göteborg, Sweden : 2017:31
dc.setspec.uppsokTechnology
dc.subjectTransport
dc.subjectHållbar utveckling
dc.subjectTeknisk mekanik
dc.subjectFarkostteknik
dc.subjectTransport
dc.subjectSustainable Development
dc.subjectApplied Mechanics
dc.subjectVehicle Engineering
dc.titleUsing Event Data Recorder (EDR) data to perform What-if simulations for safety benefit analysis by reconstructing real traffic kinematics and driver behaviors
dc.type.degreeExamensarbete för masterexamensv
dc.type.degreeMaster Thesisen
dc.type.uppsokH
Ladda ner
Original bundle
Visar 1 - 1 av 1
Bild (thumbnail)
Namn:
250499.pdf
Storlek:
1.86 MB
Format:
Adobe Portable Document Format
Beskrivning:
Fulltext
Ladda ner
Samling
Examensarbeten för masterexamen