Consideration of the characteristics of the infrastructure for the development of a driver assistance system for bicycles Université Gustave Eiffel

CONSIDERATION OF THE CHARACTERISTICS OF THE INFRASTRUCTURE FOR THE DEVELOPMENT OF A DRIVER ASSISTANCE SYSTEM FOR BICYCLES Réf ABG-102164 Sujet de Thèse 06/01/2022 Financement de l'Union européenne Université Gustave Eiffel - Site de Marne-la-Vallée Lieu de travail Champs sur marne - Ile-de-France - France Intitulé du sujet Consideration of the characteristics of the infrastructure for the development of a driver assistance system for bicycles Champs scientifiques * Sciences de l’ingénieur Mots clés Bicycle modeling, simulator, road safety, estimation, control DESCRIPTION DU SUJET To design and validate innovative solutions to improve the mobility of road users, it is important to take into account the behavior of these users when they travel. This involves various complementary approaches such as surveys, studies in natural driving, on tracks or on simulators. These last, have several advantages: control and reproducibility of travel conditions and situations, absence of risk, time savings and reduction in experimentation costs. They are therefore commonly used as training, development and research tools. Travel simulators are objects of interdisciplinary research, since they must restore the visual, sound, dynamic and haptic stimuli to which users are subjected in their movements. The importance of each type of restitution depends on the situations studied. They also make it possible to detect the behavior of drivers and other road users in various driving situations, while controlling dynamic variables and avoiding the risks associated with a real environment. In this thesis, we are mainly interested in the bicycle simulator. The latter is used for multiple purposes, such as training, sport and different research disciplines such as virtual reality, dynamic modeling and study of the behavior and performance of cyclists. Experimentation in a real environment is not always the most appropriate means because of its costs, uncontrolled variables and the risks faced by cyclists. Simulators therefore allow us to detect the behavior of cyclists and other road users in various driving situations, while controlling dynamic variables and avoiding the risks associated with a real environment. The objective of the proposed thesis is to study the behavior of the bicycle in interaction with the infrastructure and other road users and to propose methods to improve its safety. Indeed, we were able to demonstrate during Mr. Shoman's thesis that infrastructure plays a major role in the behavior of the cyclist, especially during bad weather conditions, such as snow. Taking into account infrastructure parameters, such as road adhesion, radius of curvature and road profile, is a major challenge in the development of tools and methods to estimate their influence on rider behavior. Depending on the parameters from the infrastructure, the dynamic variables of the bike, such as speed, acceleration and trajectory will be calculated and the risk of a possible accident (overturning, collision with other users, etc.) will be estimated. The method that will be developed as part of this thesis will also be able to offer recommendations to the infrastructure manager on the improvements to be made to the road in terms of profile, radius of curvature, maximum speed, etc., for a road. Safer and smoother. To achieve this, we will use accidents data to determine the boundary characteristics of the road. Indeed, the geometric characteristics identified as accident-causing by road safety studies (small radius, lack of adhesion, etc.) will be taken into account and considered as inputs to the bycicle model that we are going to improve. This will make it possible to propose a method for detecting the limits of driving the bicycle. Therefore, we propose in this thesis an assistance system. It cannot be achieved without in-depth modeling and analysis of the Cyclist / Infrastructure / bicycle system. This modeling is very important for science, engineering and human factors researchers working in the fields of transport, user mobility and vehicle dynamics. To achieve this, the proposed method consists in considering the driver as a control system making it possible to determine the necessary actions to ensure the trajectory of the vehicle. This will help to better understand the behavior of the cyclist and improve the quality of the bicycle modeling, in interaction with the infrastructure. Indeed, it is important to identify the type of cyclist and to be able to classify it according to the assistance you want to provide. Several parameters that can influence the cyclist will be taken into account, such as the age, gender, physical or mental capacities of the cyclist. The information we propose to study will be related to longitudinal or lateral stresses (braking and acceleration phase, speed levels, etc.). The study of these indicators will lead to a trajectory model specific to a category of cyclists defined according to a statistical classification. In addition, the use of the simulator will help to study the behavior of cyclists in different environmental determinants in order to * identify the elements in the environment that cyclists consider, to match their driving behavior to the road environment (speed, distance, etc.); * understand how cyclists adapt their driving practices according to the road users with whom they interact; * specify the driving situations considered to be the most risky in order to determine how cyclists anticipate these risks and what strategies (equipment or behavior) they employ to cope; * investigate cycling errors and behaviors inappropriate for the situation and the environmental and personal factors that contribute to such behavior. Like any road accident resulting from vehicle-driver-infrastructure interaction, we focus on the study of the interaction between the bicycle and the road by studying and analyzing the effects of the characteristics and geometries of the road surface on the behavior of the cyclist and thus determine whether these effects can lead to accident-prone situations. The full impact of this research will help improve the comfort and safety of cyclists and promote the proper sharing of urban space by cyclists and other road users. The behavioral aspect of the cyclist will be studied by comparing the results from the bicycle simulator with those from a real road situation, performed on an instrumented bicycle. This last step will be carried out in Italy, in collaboration with University of Bologna (Italy). PRISE DE FONCTION : 03/10/2022 NATURE DU FINANCEMENT Financement de l'Union européenne PRÉCISIONS SUR LE FINANCEMENT PRÉSENTATION ÉTABLISSEMENT ET LABO D'ACCUEIL Université Gustave Eiffel - Site de Marne-la-Vallée La thèse se déroulera au Laboratoire PICS-L, Université Gustave Eiffel - Site de Marne-la-Vallée PROFIL DU CANDIDAT Knowledge on: * Mechanics and modeling * Automation: Estimation, identification, non-linear control * Embedded system * A good level of english (written and oral) Date limite de candidature 21/03/2022