Ifsttar PhD subject

French version
Detailed form :

Title : Cognitive and biomechanical modeling of drivers’ behavior to assess their safety in case of an accident

Main host Laboratory - Referent Advisor   -

Director of the main host Laboratory   -

PhD Speciality Informatique/Mécanique

Axis of the performance contract 1 - COP2017 - Efficient transport and safe travel

Main location Bron

Doctoral affiliation UNIVERSITE CLAUDE-BERNARD-LYON 1

PhD school MEGA (MECANIQUE, ENERGETIQUE, GENIE CIVIL, ACOUSTIQUE)

Planned PhD supervisor WANG Xuguang  -  Université Gustave Eiffel  -  TS2 - LBMC

Planned PhD co-supervisor BEILLAS Philippe  -  Université Gustave Eiffel  -  TS2 - LBMC

Planned financing Contrat doctoral  - Ifsttar

Abstract

Background
----------------
In the coming years, highly automated vehicles will be introduced gradually on the roads. Their occupants could have new leisure activities or work while driving and new interior layout may be designed to ensure a higher level of comfort. Moreover, automation is not the guarantee of the absence of accidents, at least during the phrase of transition. In this context, it is essential to be concerned about the safety issues that may arise in the future when using these vehicles, especially with regard to the new postures that could be adopted by occupants due to automation. The aim of this thesis will be to use numerical "multi-models" simulation to anticipate these new safety issues, in order to ultimately design safe highly automated vehicles. Different types of numerical human models are already used in the IFSTTAR to assist in the design of vehicle and its interior. They describe humans from different angles and aspects, all of which are important in the design process, especially for future automated vehicles. We can mention in particular:
- The so-called cognitive models, which attempt to simulate the driver's "awareness" of the situation, risk assessment, decision-making and, ultimately, the actions to cope with the driving situation (ex: Bellet et al, 2012). These models are important for predicting an "intention of action" that is about to engage the driver (to take over control of the automated vehicle, for example), but so far do not take into account the body movements and postures nor the biomechanical constraints.
- The so-called postural comfort models aim to predict the discomfort associated with a posture, and thus predict plausible postures that would be chosen by occupants (ex: Peng et al., 2017, 2018). These models are useful for the design of future vehicle interior.
- The so-called impact models are aimed at predicting the risk of injury associated with an accident (ex: Beillas and Berthet, 2017, 2018). These models are important for determining postures that will be acceptable in future interiors, for designing new systems of protection, and for taking into account the injury consequences in the decision algorithms of the future automated vehicle.

However, these different simulation models are currently relatively disconnected from each other. Thus, the cognitive models do not take into account the plausibility of the movement resulting from a decision from the biomechanical or in fine security point of view. Similarly, postural comfort models do not take into account (cognitive) occupant’s intention for motion prediction or the choice of kinematic strategies. Finally, impact models do not use the resulting intentions and motions to determine posture just before the impact.

The challenge now is to propose new methods and tools based on these different human models, to assess the risks and the changes induced by the driving automation using a combined "cognitive, behavioral and biomechanical" approach.

Objectives
-------------
The objectives of the thesis will be to investigate the possible connections between cognitive, comfort and impact human models. The reflection will focus on the nature and interest of possible connections, and the development of prototypes to evaluate different approaches. These could include the development of co-simulation approaches for cognitive and comfort models and their connection to injury simulation due to impact. The interests of proposed simulation approaches will be evaluated in autonomous vehicle driving scenarios.

Approach
-------------
The existing models available at LESCOT and LBMC will serve as a basis for this thesis research. After having familiarized with these models, simulation scenarios of "common" interest will be defined for the vehicles of the automation level 3 (demand for taking control of the driving at all times) and 4 (request for the manual control of vehicle programmed in advance).

Two types of scenarios of interest will be studied. The first is related to the new occupant’s posture in a pre-crash situation, partly determined by the activities of life on board adopted during the automated driving, and the second to that of taking over of vehicle control in situation of “emergency" (is it possible or not, cognitively and bio-mechanically speaking, for example), in case of incapacity of the automaton to manage alone the task of driving.

Without prejudging at this stage of the driving scenarios that will actually be investigated during the thesis, in a first step, the challenge is to collect the behavioral data in automated driving situations observed using a driving simulator. In a second step, it will then be necessary to rely on the various existing models to predict, model and "co-simulate" the activity of the occupant from a cognitive, behavioral and postural point of view, so as to jointly understand the perceptual, cognitive and biomechanical constraints associated with certain "intentions" of the occupant for example, or to evaluate the effects and consequences in case of accident.

Candidat
------------
Several disciplines will be implied during the thesis, including mechanics (with a strong biomechanical orientation), as well as the sciences of cognition and behavior. The aspects of numerical simulation will nevertheless be essential, whatever the discipline.
The ideal candidate will have a background in Computer Science, Robotics or Mechanics, with, if possible, a first experience in simulation in one of the aspects of the thesis. Insofar as this thesis addresses a subject involving several disciplines, curiosity and openness will be important qualities for the candidate.

Supervision team
----------------------
The doctoral student will be located at the LBMC (Laboratory of Biomechanics and Mechanics of Shocks, mixed unit Ifsttar and University Lyon 1) and LESCOT (Laboratory of Ergonomics and Cognitive Sciences). He/She will be co-supervised by Xuguang Wang, Thierry Bellet and Philippe Beillas, with complementary expertise on postural comfort, cognitive aspects, and impact.

References
----------------
Bellet, T., Mayenobe, P., Bornard, J.C., Gruyer, D., Claverie, B. (2012). A computational model of the car driver interfaced with a simulation platform for future Virtual Human Centred Design applications: COSMO-SIVIC, Engineering Applications of Artificial Intelligence, 25, pp. 1488-1504.

Beillas P, Berthet F. (2017) An investigation of human body model morphing for the assessment of abdomen responses to impact against a population of test subjects. Traffic Inj. Prev., 2017, 18(sup1):S142–S147.

Beillas P. and Berthet F. (2018) Development of simulation based liver and spleen injury risk curves for the GHBMC detailed models. IRCOBI Conference 2018, September 12-14, Athens, Greece

Peng J., Wang X. and Denninger L. (2017) Ranges of the least uncomfortable joint angles for assessing automotive driving posture. Applied Ergonomics, 61, 12-21, DOI.10.1016/j.apergo.2016.12.021

Peng J., Wang X. and Denninger L. (2018) Effects of anthropometric variables and seat height on automobile drivers’ preferred posture with the presence of the clutch. Human factors, Vol. 60, No. 2, March 2018, pp. 172–190, DOI: 10.1177/0018720817741040

Keywords : Automated vehicle, Human behavior, Pre-crash, Human model, Numerical simulation, Cognitive, Biomechanical, Safety

Main host Laboratory - Referent Advisor	-
Director of the main host Laboratory	-
PhD Speciality	Informatique/Mécanique
Axis of the performance contract	1 - COP2017 - Efficient transport and safe travel
Main location	Bron
Doctoral affiliation	UNIVERSITE CLAUDE-BERNARD-LYON 1
PhD school	MEGA (MECANIQUE, ENERGETIQUE, GENIE CIVIL, ACOUSTIQUE)
Planned PhD supervisor	WANG Xuguang - Université Gustave Eiffel - TS2 - LBMC
Planned PhD co-supervisor	BEILLAS Philippe - Université Gustave Eiffel - TS2 - LBMC
Planned financing	Contrat doctoral - Ifsttar
Abstract Background ---------------- In the coming years, highly automated vehicles will be introduced gradually on the roads. Their occupants could have new leisure activities or work while driving and new interior layout may be designed to ensure a higher level of comfort. Moreover, automation is not the guarantee of the absence of accidents, at least during the phrase of transition. In this context, it is essential to be concerned about the safety issues that may arise in the future when using these vehicles, especially with regard to the new postures that could be adopted by occupants due to automation. The aim of this thesis will be to use numerical "multi-models" simulation to anticipate these new safety issues, in order to ultimately design safe highly automated vehicles. Different types of numerical human models are already used in the IFSTTAR to assist in the design of vehicle and its interior. They describe humans from different angles and aspects, all of which are important in the design process, especially for future automated vehicles. We can mention in particular: - The so-called cognitive models, which attempt to simulate the driver's "awareness" of the situation, risk assessment, decision-making and, ultimately, the actions to cope with the driving situation (ex: Bellet et al, 2012). These models are important for predicting an "intention of action" that is about to engage the driver (to take over control of the automated vehicle, for example), but so far do not take into account the body movements and postures nor the biomechanical constraints. - The so-called postural comfort models aim to predict the discomfort associated with a posture, and thus predict plausible postures that would be chosen by occupants (ex: Peng et al., 2017, 2018). These models are useful for the design of future vehicle interior. - The so-called impact models are aimed at predicting the risk of injury associated with an accident (ex: Beillas and Berthet, 2017, 2018). These models are important for determining postures that will be acceptable in future interiors, for designing new systems of protection, and for taking into account the injury consequences in the decision algorithms of the future automated vehicle. However, these different simulation models are currently relatively disconnected from each other. Thus, the cognitive models do not take into account the plausibility of the movement resulting from a decision from the biomechanical or in fine security point of view. Similarly, postural comfort models do not take into account (cognitive) occupant’s intention for motion prediction or the choice of kinematic strategies. Finally, impact models do not use the resulting intentions and motions to determine posture just before the impact. The challenge now is to propose new methods and tools based on these different human models, to assess the risks and the changes induced by the driving automation using a combined "cognitive, behavioral and biomechanical" approach. Objectives ------------- The objectives of the thesis will be to investigate the possible connections between cognitive, comfort and impact human models. The reflection will focus on the nature and interest of possible connections, and the development of prototypes to evaluate different approaches. These could include the development of co-simulation approaches for cognitive and comfort models and their connection to injury simulation due to impact. The interests of proposed simulation approaches will be evaluated in autonomous vehicle driving scenarios. Approach ------------- The existing models available at LESCOT and LBMC will serve as a basis for this thesis research. After having familiarized with these models, simulation scenarios of "common" interest will be defined for the vehicles of the automation level 3 (demand for taking control of the driving at all times) and 4 (request for the manual control of vehicle programmed in advance). Two types of scenarios of interest will be studied. The first is related to the new occupant’s posture in a pre-crash situation, partly determined by the activities of life on board adopted during the automated driving, and the second to that of taking over of vehicle control in situation of “emergency" (is it possible or not, cognitively and bio-mechanically speaking, for example), in case of incapacity of the automaton to manage alone the task of driving. Without prejudging at this stage of the driving scenarios that will actually be investigated during the thesis, in a first step, the challenge is to collect the behavioral data in automated driving situations observed using a driving simulator. In a second step, it will then be necessary to rely on the various existing models to predict, model and "co-simulate" the activity of the occupant from a cognitive, behavioral and postural point of view, so as to jointly understand the perceptual, cognitive and biomechanical constraints associated with certain "intentions" of the occupant for example, or to evaluate the effects and consequences in case of accident. Candidat ------------ Several disciplines will be implied during the thesis, including mechanics (with a strong biomechanical orientation), as well as the sciences of cognition and behavior. The aspects of numerical simulation will nevertheless be essential, whatever the discipline. The ideal candidate will have a background in Computer Science, Robotics or Mechanics, with, if possible, a first experience in simulation in one of the aspects of the thesis. Insofar as this thesis addresses a subject involving several disciplines, curiosity and openness will be important qualities for the candidate. Supervision team ---------------------- The doctoral student will be located at the LBMC (Laboratory of Biomechanics and Mechanics of Shocks, mixed unit Ifsttar and University Lyon 1) and LESCOT (Laboratory of Ergonomics and Cognitive Sciences). He/She will be co-supervised by Xuguang Wang, Thierry Bellet and Philippe Beillas, with complementary expertise on postural comfort, cognitive aspects, and impact. References ---------------- Bellet, T., Mayenobe, P., Bornard, J.C., Gruyer, D., Claverie, B. (2012). A computational model of the car driver interfaced with a simulation platform for future Virtual Human Centred Design applications: COSMO-SIVIC, Engineering Applications of Artificial Intelligence, 25, pp. 1488-1504. Beillas P, Berthet F. (2017) An investigation of human body model morphing for the assessment of abdomen responses to impact against a population of test subjects. Traffic Inj. Prev., 2017, 18(sup1):S142–S147. Beillas P. and Berthet F. (2018) Development of simulation based liver and spleen injury risk curves for the GHBMC detailed models. IRCOBI Conference 2018, September 12-14, Athens, Greece Peng J., Wang X. and Denninger L. (2017) Ranges of the least uncomfortable joint angles for assessing automotive driving posture. Applied Ergonomics, 61, 12-21, DOI.10.1016/j.apergo.2016.12.021 Peng J., Wang X. and Denninger L. (2018) Effects of anthropometric variables and seat height on automobile drivers’ preferred posture with the presence of the clutch. Human factors, Vol. 60, No. 2, March 2018, pp. 172–190, DOI: 10.1177/0018720817741040
Keywords :	Automated vehicle, Human behavior, Pre-crash, Human model, Numerical simulation, Cognitive, Biomechanical, Safety

List of topics

Applications closed