Driver modelling for safety assessment of automated vehicle functionality in cut-in scenarios: Accumulation model development using SHRP2 data

Visa enkel post
dc.contributor.authorChau, Christoffer
dc.contributor.authorLiu, Qianyu
dc.contributor.departmentChalmers tekniska högskola / Institutionen för mekanik och maritima vetenskapersv
dc.contributor.examinerBärgman, Jonas
dc.contributor.supervisorBärgman, Jonas
dc.date.accessioned2021-09-07T09:13:37Z
dc.date.available2021-09-07T09:13:37Z
dc.date.issued2021sv
dc.date.submitted2020
dc.description.abstractThe development of active safety systems and automated vehicles has increased exponentially in recent years and can influence traffic safety in various ways. Prospective assessments of their safety impacts are required during the development of active safety systems and automated vehicle functionality. Virtual simulation is one of the most common approaches of safety prospective assessments; a method that uses models to represent drivers, vehicles and road environments, etc., and run simulations in computers to estimate the risk and benefit of safety systems. This thesis aims to extend and implement an existing rear-end accumulation driver model into a cut-in driver model in the virtual simulation tool Esmini. The model is for predicting the braking behaviour of the driver of a vehicle (going straight, denoted EGO vehicle) when another vehicle (denoted Principle Other Vehicle; POV) from an adjacent lane is cutting in in the front of the EGO vehicle. The model parameters were optimized to cut-in near-crashes from a set of Naturalistic Driving Data from The Second Strategic Highway Research Program (SHRP2) in the US. The stochastic machine learning method Particle Swarm Optimization was used to fit the braking onset timing and jerk, which represent the time when a driver starts to decelerate harshly and the ramp-up of the deceleration, respectively, as defined by a piecewise linear model of the behaviour observed in the SHRP2 data. Three different models were introduced and implemented in the virtual simulation framework. The performance of the models were good for some events, but could not capture the variability across all events sufficiently for direct use in safety assessment. Future work could include fitting the model to more data, as well as fixing some parameters to reduce the complexity of the current model and capturing more information, which could affect the driver response in a cut-in scenario. With further developments of the models, it may be used for safety assessment of active safety or other vehicle automation functions.sv
dc.identifier.coursecodeMMSX30sv
dc.identifier.urihttps://hdl.handle.net/20.500.12380/304058
dc.language.isoengsv
dc.relation.ispartofseries2021:55sv
dc.setspec.uppsokTechnology
dc.subjectDriver modelsv
dc.subjectCut-in scenariosv
dc.subjectSHRP2sv
dc.subjectNaturalistic driving datasv
dc.subjectAutomated vehiclesv
dc.subjectParticle swarm optimizationsv
dc.subjectVirtual simulationsv
dc.subjectActive safetysv
dc.titleDriver modelling for safety assessment of automated vehicle functionality in cut-in scenarios: Accumulation model development using SHRP2 datasv
dc.type.degreeExamensarbete för masterexamensv
dc.type.uppsokH
local.programmeAutomotive engineering (MPAUT), MSc
Ladda ner
Original bundle
Visar 1 - 1 av 1
Bild (thumbnail)
Namn:
2021-55 Christoffer Chau & Qianyu Liu.pdf
Storlek:
5.61 MB
Format:
Adobe Portable Document Format
Beskrivning:
Master thesis
Ladda ner
License bundle
Visar 1 - 1 av 1
Bild (thumbnail)
Namn:
license.txt
Storlek:
1.51 KB
Format:
Item-specific license agreed upon to submission
Beskrivning:
Ladda ner
Samling
Examensarbeten för masterexamen