Ifsttar PhD subject

French version
Detailed form :

Title : Multi-level risk and collective perception for high quality of service automated mobility in a highly dynamic V2X connected environment

Main host Laboratory - Referent Advisor   -

Director of the main host Laboratory   -

PhD Speciality « Sciences de l’information et de la communication » (discipline « Informatique, automatique »)

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

Main location Lille-Villeneuve d'Ascq

Doctoral affiliation UNIVERSITE DES SCIENCES ET TECHNOLOGIE DE LILLE 1

PhD school SCIENCES POUR L'INGENIEUR (SPI)

Planned PhD supervisor GRUYER Dominique  -  Université Gustave Eiffel  -  COSYS - PICS-L

Planned PhD co-supervisor TATKEU Charles  -  Université Gustave Eiffel  -  COSYS - LEOST

Planned financing Contrat doctoral  - Ifsttar

Abstract

The use of automated vehicle (AV) technologies such as self-driving cars is becoming more prevalent in daily life. These technologies aim to create fully- connected transportation systems, still there are concerns that remain unaddressed. Studies have shown that AVs can reduce collisions, ease traffic congestion, and provide transportation options for those who lack access. Yet, car manufacturers have already implemented certain automated features in their vehicles. One important aspect of AVs is improving communication between the vehicle and roadside. The objective of this study is to investigate the adaptability and suitability of the Chain branch leaf (CBL) communication model in cooperative systems to examine its impact on traffic responses. Additionally, the research aims to determine the role of Roadside Units and the effectiveness of multi-level perception in risk mitigation. The ultimate goal of this research is to improve communication and collaboration between autonomous vehicles leading to safer and more efficient traffic flow.
This thesis focuses on the estimation of obstacle attributes, the road, and the ego-vehicle to improve the quality of service on the road through communication, localization, and perception functions. We propose architectures and communication strategies that will take into account the information of surrounding vehicles to optimize coverage and estimate collision risk at different levels including local, extended local, extended branch, and global.
Subsequently, we use the most relevant metrics (Time to Collision (TTC), Time Headway (TH), Distance of Gruyer (DG), RISK (R), Risk Estimator with Uncertainties and Multidimensional Model (RIMUM)), to estimate the four (extended) collision risks. In optimal conditions first with perfect location and perception, and then the uncertainty scenario of perception with perfect location. Results show that the extended risks allow better anticipation of the collision than the local risk.
Furthermore, we have developed a new extended version of the Chain branch leaf-Gateway (CBL-G) model, which proves to be more efficient in terms of coverage. The hierarchical architecture of the model allows us to calculate collision risks with greater accuracy. The different levels of risk allow us to identify potentially dangerous situations earlier, which is considered to be very relevant for incident prevention.
In our future research projects, we plan to study other situations such as road intersections, highway exits, and entrances, as well as roundabouts. Additionally, we would also like to explore cases where we are unable to locate nodes through the chain (such as passing through tunnels). And elaborate risk indicators that explore all key components (ego vehicle, driver, obstacle, road, and environment).

Keywords : Cooperative Systems, Automated vehicles, Vehicle-to-Vehicle communications, IEEE 802.11p, Vehicle-to-Infrastructure communications, Clustering, Gateway, Extended Perception, Multi-level Collision Risk, extended collision risks, uncertainties

Main host Laboratory - Referent Advisor	-
Director of the main host Laboratory	-
PhD Speciality	« Sciences de l’information et de la communication » (discipline « Informatique, automatique »)
Axis of the performance contract	1 - COP2017 - Efficient transport and safe travel
Main location	Lille-Villeneuve d'Ascq
Doctoral affiliation	UNIVERSITE DES SCIENCES ET TECHNOLOGIE DE LILLE 1
PhD school	SCIENCES POUR L'INGENIEUR (SPI)
Planned PhD supervisor	GRUYER Dominique - Université Gustave Eiffel - COSYS - PICS-L
Planned PhD co-supervisor	TATKEU Charles - Université Gustave Eiffel - COSYS - LEOST
Planned financing	Contrat doctoral - Ifsttar
Abstract The use of automated vehicle (AV) technologies such as self-driving cars is becoming more prevalent in daily life. These technologies aim to create fully- connected transportation systems, still there are concerns that remain unaddressed. Studies have shown that AVs can reduce collisions, ease traffic congestion, and provide transportation options for those who lack access. Yet, car manufacturers have already implemented certain automated features in their vehicles. One important aspect of AVs is improving communication between the vehicle and roadside. The objective of this study is to investigate the adaptability and suitability of the Chain branch leaf (CBL) communication model in cooperative systems to examine its impact on traffic responses. Additionally, the research aims to determine the role of Roadside Units and the effectiveness of multi-level perception in risk mitigation. The ultimate goal of this research is to improve communication and collaboration between autonomous vehicles leading to safer and more efficient traffic flow. This thesis focuses on the estimation of obstacle attributes, the road, and the ego-vehicle to improve the quality of service on the road through communication, localization, and perception functions. We propose architectures and communication strategies that will take into account the information of surrounding vehicles to optimize coverage and estimate collision risk at different levels including local, extended local, extended branch, and global. Subsequently, we use the most relevant metrics (Time to Collision (TTC), Time Headway (TH), Distance of Gruyer (DG), RISK (R), Risk Estimator with Uncertainties and Multidimensional Model (RIMUM)), to estimate the four (extended) collision risks. In optimal conditions first with perfect location and perception, and then the uncertainty scenario of perception with perfect location. Results show that the extended risks allow better anticipation of the collision than the local risk. Furthermore, we have developed a new extended version of the Chain branch leaf-Gateway (CBL-G) model, which proves to be more efficient in terms of coverage. The hierarchical architecture of the model allows us to calculate collision risks with greater accuracy. The different levels of risk allow us to identify potentially dangerous situations earlier, which is considered to be very relevant for incident prevention. In our future research projects, we plan to study other situations such as road intersections, highway exits, and entrances, as well as roundabouts. Additionally, we would also like to explore cases where we are unable to locate nodes through the chain (such as passing through tunnels). And elaborate risk indicators that explore all key components (ego vehicle, driver, obstacle, road, and environment).
Keywords :	Cooperative Systems, Automated vehicles, Vehicle-to-Vehicle communications, IEEE 802.11p, Vehicle-to-Infrastructure communications, Clustering, Gateway, Extended Perception, Multi-level Collision Risk, extended collision risks, uncertainties

List of topics

Applications closed