Search

International audience; In this paper, we propose a new approach which addresses the Positive Unlabeled learning challenge for image classification. Its functioning is based on GAN abilities in order to generate fake images samples whose distribution gets closer to negative samples distribution included in the unlabeled dataset available, while being different to the distribution of the unlabeled positive samples. Then we train a CNN classifier with the positive samples and the fake generated samples, as it would be done with a classic Positive Negative dataset. The tests performed on three different image classification datasets show that the system is stable up to an acceptable fraction of positive samples present in the unlabeled dataset. Although very different, this method outperforms the state of the art PU learning on the RGB dataset CIFAR-10.

International audience; Dans ce document, nous proposons une nouvelle approche répondant à la tâche de classification d'images à partir d'un apprentissage sur données positives et non-labélisées. Son bon fonctionnement repose sur certaines particularités des réseaux antagonistes génératifs (GANs). Ces derniers nous permettent de générer des fausses images dont la distribution se rapproche de la distribution des échantillons négatifs inclus dans le jeu de données non labélisé disponible, tout en restant différente de la distribution des échantillons positifs non labélisés. Ensuite, nous entraînons un classifieur convolutif avec les échantillons positifs et les faux échantillons générés, tel que cela aurait été fait avec un jeu de données classique de type Positif Négatif. Les tests réalisés sur trois jeux de données différents de classification d'images montrent que le système est stable dans son comportement jusqu'à une fraction conséquente d'échantillons positifs présents dans le jeu de données non labélisé. Bien que très différente, cette méthode surpasse l'état de l'art PU learning sur le jeu de données RVB CIFAR-10.

International audience; The paper investigates the effect of the bar number on the performances of a five-phase squirrel-cage induction machine with fractional-slot tooth concentrated winding. With a same stator, five different rotor bar numbers are chosen and the rotor magnetic circuits are designed using an analytical approach. Then, a finite-element analysis is done under two different supply conditions: fundamental currents and third harmonic currents. Finally regarding the used winding specificities the possibility of second harmonic current supply is evaluated. The results are presented in a comparative way in order to determine the impact of rotor bar number on torque quality for the different supply modes.

International audience; Spintronic devices currently rely on magnetization control by external magnetic fields or spin-polarized currents. Developing temperature-driven magnetization control has potential for achieving enhanced device functionalities. Recently, there has been much interest in thermally induced magnetisation switching (TIMS), where the temperature control of intrinsic material properties drives a deterministic switching without applying external fields. TIMS, mainly investigated in rare-earth–transition-metal ferrimagnets, has also been observed in epitaxial Fe/MnAs/GaAs(001), where it stems from a completely different physical mechanism. In Fe/MnAs temperature actually modifies the surface dipolar fields associated with the MnAs magnetic microstructure. This in turn determines the effective magnetic field acting on the Fe overlayer. In this way one can reverse the Fe magnetization direction by performing thermal cycles at ambient temperatures. Here we use element selective magnetization measurements to demonstrate that various magnetic configurations of the Fe/MnAs/GaAs(001) system are stabilized predictably by acting on the thermal cycle parameters and on the presence of a bias field. We show in particular that the maximum temperature reached during the cycle affects the final magnetic configuration. Our findings show that applications are possible for fast magnetization switching, where local temperature changes are induced by laser excitations.

International audience

International audience; Future thinking (FT) is the cognitive ability that enables humans to mentally imagine and pre-experience future events. Despite a sharp increase in basic research on future thinking in the last decade, we note the lack of transfer to Human Factors and Ergonomics (HF&E), although the field is increasingly concerned by future artefacts and although theory building on future-oriented HF&E has begun (e.g. prospective ergonomics). This article gives an overview of key findings regarding future thinking, namely i) underlying cognitive mechanisms, ii) functions and obstacles, iii) common descriptors of future thinking, iv) recommendations as to the possible improvement of future thinking during user research and, v) methods for the assessment of future thinking abilities. This synthesis can support HF&E practitioners in fine-tuning their future-oriented methods.

International audience

The ethics of autonomous vehicles (AV) has received a great amount of attention in recent years, specifically in regard to their decisional policies in accident situations in which human harm is a likely consequence. Starting from the assumption that human harm is unavoidable, many authors have developed differing accounts of what morality requires in these situations. In this article, a strategy for AV decision-making is proposed, the Ethical Valence Theory, which paints AV decision-making as a type of claim mitigation: different road users hold different moral claims on the vehicle’s behavior, and the vehicle must mitigate these claims as it makes decisions about its environment. Using the context of autonomous vehicles, the harm produced by an action and the uncertainties connected to it are quantified and accounted for through deliberation, resulting in an ethical implementation coherent with reality. The goal of this approach is not to define how moral theory requires vehicles to behave, but rather to provide a computational approach that is flexible enough to accommodate a number of ‘moral positions’ concerning what morality demands and what road users may expect, offering an evaluation tool for the social acceptability of an autonomous vehicle’s ethical decision making.

International audience; The emergence of visible light communication (VLC) as a subset of optical wireless communication (OWC) in the early 2000s has turned any light-emitting diode (LED) source into a potential data transmitter. The design process for any VLC or OWC system typically involves a link budget analysis performed by studying the signal-to-noise ratio (SNR) at the receiver. Since this SNR strongly depends on the radiant flux collected by the receiver, an accurate model for this parameter is required. The point-source model has been widely used since 1979 and generally provides a good approximation of the received radiant flux. However, it might be less accurate for typical extended lighting sources like LED panels or large-area organic LEDs. In this paper, the radiant flux distribution of flat Lambertian rectangular or circular sources is derived from a vector analysis of their irradiance. It is then validated through actual measurements in the case of a circular source. The resulting extended-source models thus better capture the light beam pattern of such transmitters to enable a more accurate link budget. It provides at the same time almost identical results to the point-source model for small sources and can, therefore, be seen as a natural extension of this already widely used model. INDEX TERMS LED pattern, irradiance pattern, light power pattern, non-imaging optics, link budget, optical wireless communication, visible light communication.

While the transportation sector is responsible for a growing share of greenhouse gas emissions, electric vehicles (EVs) offer solutions for greener mobility. The proportion of electric vehicles in transportation fleets is increasing, but wider adoption will not be possible without an appropriate charging infrastructure. The deployment of such infrastructure should follow a strategy that considers both the environment in which it is deployed and the behavior patterns of electric vehicle users. If these aspects are not taken into consideration, there is a risk of failing to meet users’ needs and generating additional costs. Here we review the literature on location problems for electric vehicle charging stations. We aim to draw up a comparative overview of approaches that have been used up to 2020 for optimizing the locations of charging infrastructure. We first briefly review the issues raised by the deployment of charging infrastructure, namely technical, economic and user acceptance concerns. We then look at the goals of the infrastructure location models in the literature. Schematically, those goals fall into two categories: minimizing the cost of charging infrastructure for a given level of service, or maximizing the service provided for a given cost. Finally, we focus on the approaches used to achieve these goals. Three categories of approaches are identified: node, path, and tour- or activity-based approaches. We then discuss these approaches in relation to technical, economic and user acceptance factors in order to provide a comprehensive analysis for stakeholders involved in EV charging infrastructure planning. Directions are given for future research to develop models that better reflect the real-world picture.

International audience; The autonomous platoon is today one of the key tools for better road utilization. In fact, by optimizing the distance between vehicles, the air drag is reduced, and researchers have shown that 20% of energy can be saved by using this concept. From a network point of view, reducing the distance between vehicles will allow a new point-to-point communication link between the vehicles in front and behind by using Vehicular Visible Light Communication (V-VLC), thus providing an opportunity to have a hybrid communication. Our new radio design based on the AS-DTMAC protocol guarantees a high Quality of Service for real-time applications. However, with a very high density, we can reach the bandwidth dedicated to V2X radio. In the case of a platoon, this scenario can cause dangerous platoon instability. Assisting the radio with another communication vector such as V-VLC can help to maintain the high level of reliability that is necessary for the control of a platoon. In this paper, we first carry out an analytical analysis to investigate the capacity of our new radio technology to support the platoon control use case in terms of the quality of service (QoS) required for this type of application. Secondly, we show through extensive simulations the current level of V-VLC technology, compared to radio technology, in terms of packet loss and delay.

International audience

International audience

International audience

The high penetration of distributed energy resources and electric vehicles is changing the way the electricity system is managed. In turn, the way utilities have been recovering their expenditures through tariffs needs reformulation. We investigate the impact of different retail tariff designs from a Californian scenario on private investment incentives and cost-shifting using solar PVs, stationary batteries, and electric vehicles. The commercial private facilities studied do not own the vehicles and the vehicle owners receive compensation for energy services provided, which strongly depends on the type of tariff applied. We found that energy-based tariffs with on-peak periods synchronized with solar PV production brought the highest private gains, but with high cost-shifting. On the other hand, the capacity-based tariffs reduced the economic benefits and cost-shifting concomitantly, mainly when on-peak periods defined by the rate matched the most constrained grid time window. Batteries are incentivized mostly to offset maximum demand charges rather than to arbitrage energy, but this will strongly depend on the spread between on-peak and off-peak periods. Coincident peak rates, coupled with EVs, can bring high remuneration for EV owners, second-highest net present value, and second-lowest cost-shifting among all rates. Finally, we derive policy implications from the results and earmark more sophisticated tariff designs for investigation.

From the beginning of the automotive industry, the car was long seen as simply an-other means of transport. Now, inside a vehicle, we can listen to music, regulate thetemperature, open the windows automatically, charge a smartphone, benefit fromacoustic insulation from the noisy outside with a silent electric motor. All this makesthe journey more pleasant, but safety applications are still not the main focus. Al-though vehicles already have a certain number of passive safety systems such as seatbelts, airbags, etc, and more recently, cars are being equipped with active safetysystems such as Advanced Driver-Assistance Systems (ADAS), these systems remainbasic in the sense that they are not connected. The automotive field is currentlyundergoing a massive transformation, with the car becoming a connected device andefforts are being made to develop automated driving vehicles with sophisticated activesafety systems. In this context, Vehicular Ad hoc Networks (VANETs) will play anessential role in the future of the automotive industry, allowing new usages and newservices to be offered, the main focus being on their support of safety applications.Nowadays, the number of road accidents is continuing to increase, and humanlosses are becoming more and more considerable. In a connected vehicle, receiving amessage within a few milliseconds can be enough to save a life, if the driver receives aDecentralized Environmental Notification Message (DENM) in time. Thus, latency isa very important metric to be taken into account in the design of the new generationof VANETs. Also, for critical applications, the system must be ultra-reliable sinceany failure may have fatal consequences. For example, augmented perception forautonomous vehicles will be revolutionized by the upcoming technologies for Vehicle-to-Everything (V2X): 5G (The fifth generation) NR (New Radio) & 802.11bd. Infact, the database for autonomous driving, which initially contains information obtainedfrom sensors (radar, camera, etc.), will be completed with information comingfrom Cooperative Awareness Messages (CAMs) and Collective Perception Messages(CPMs). Thus, the supervision system of the vehicle will be able to react more efficiently to any situation on the road, and maneuver quickly if any unexpected eventoccurs.As our highways become increasingly congested, for a better traveling experience,Intelligent Transportation Systems (ITSs) organizations plan to optimize road utilizationand traffic management. To address current mobility issues, conventional vehiclesystems for traffic flow like Adaptive Cruise Control (ACC) are not sufficient. ThisACC system adjusts the vehicle's speed automatically based on information from vehicles in front. A more dynamic regulation strategy requires a cooperative systemthat will provide more information on the positions, speeds and accelerations ofneighboring vehicles. Thanks to this cooperative system, new vehicular applicationscan be established (an example of such a system being Cooperative Adaptive CruiseControl (CACC)). Platooning is one of the best-known applications that can enhanceroad capacity. In fact, by optimizing the distance between vehicles, road infra-structure capacity can increase by over 30%. By using the same concept, the air drag isreduced: researchers have shown that up to 20% of energy can be saved. From anetwork point of view, reducing the distance between vehicles will make it possible toestablish new point-to-point communication links between the vehicles in front andbehind, thus providing an opportunity to have a hybrid communication.The subject of this thesis is vehicular wireless networks and, more precisely, the useof radio transmission and communication by light to improve vehicle safety. The thesisis motivated by the unreliability and scalability issues of the IEEE 802.11p standard.The idea is to move towards new techniques, particularly within future 803.11bdstandards, and associate radio transmission with Visible Light Communication (VLC)to allow for hybrid communication.The first part of the thesis concerns the development of low latency access techniques in vehicular networks within the framework of future standards. The aim ofthis first part is to propose a Medium Access Control (MAC) layer that is based onthe Time Division Multiple Access (TDMA) technique, enhanced with a mechanismthat allows nodes to benefit from reliable, low latency access on demand. This partalso introduces a special access scheme for high-priority emergency packets, while stillensuring reliable, low latency access. We complete this study by a mathematical analysis of the performance of this scheme, to show its high performances. In addition,we propose an original message forwarding strategy based on a Cross-Layer solutionfor disseminating Emergency Messages in a multi-hop manner (exploiting the relationbetween the MAC and network layers).The second part of the thesis deals with VLC to coordinate maneuvers in vehicular networks. The advantages of wireless communications for platoon control are,among other things, improved chain stability (which consequently improves safetyand allows a smaller gap between vehicles), vehicle grouping as well as maneuveringof separation. While the importance of wireless communications to the platoon is wellrecognized, certain aspects of communications have not been thoroughly investigatedin previous projects. One such aspect is the coexistence of VLC with radio vehicularcommunication technologies in a heterogeneous system to improve the overall performance (each vector of communication compensating for the drawbacks of the other).The idea is to develop an algorithm that selects the radio communication proposed inPart 1, and visible light communication for autonomous driving, based on the stateof the radio channel and platoon alignment.; Depuis le début de l'industrie automobile, la voiture a longtemps été considéréecomme un simple moyen de transport. Désormais, à l'intérieur d'un véhicule, onpeut écouter de la musique, réguler la température, ouvrir les vitres automatique-ment, recharger un smartphone, bénéficier d'une isolation acoustique de l'extérieurbruyant avec un moteur électrique silencieux.Tout cela rend le voyage plus agréable, mais les applications de sécurité ne sonttoujours pas au centre des préoccupations. Bien que les véhicules disposent déjà d'uncertain nombre de systèmes de sécurité passive tels que ceintures de sécurité, coussinsgonflables, etc., et plus récemment, les voitures sont équipées de systèmes de sécuritéactive tels que les systèmes avancés d'assistance à la conduite ADAS (Advanced DriverAssistance System), ces systèmes restent fondamentaux dans le sens qu'ils ne sontpas connectés. Le domaine de l'automobile subit actuellement une transformationmassive; la voiture devenant un appareil connecté et des efforts sont déployés pourdévelopper des véhicules à conduite automatisée dotés de systèmes de sécurité activesophistiqués. Dans ce contexte, les réseaux véhiculaires appelés VANETs (VehicularAd hoc Networks) joueront un rôle essentiel dans l'avenir de l'industrie automobile,permettant de proposer de nouveaux usages et de nouveaux services. L'accent étantmis principalement sur la prise en charge des applications de sécurité.De nos jours, le nombre d'accidents de la route continue d'augmenter et les perteshumaines sont de plus en plus importantes. Dans un véhicule connecté, recevoirun message en quelques millisecondes peut suffire à sauver une vie si le conducteurreçoit à temps un message DENM (Decentralized Environmental Notification Message). Ainsi, la latence est une métrique très importante à prendre en compte dansla conception de la nouvelle génération de VANET.Aussi, pour les applications critiques, le système doit être ultra-fiable car toutepanne peut avoir des conséquences fatales. Par exemple, la perception augmentéepour les véhicules autonomes sera révolutionnée par les technologies à venir pourVéhicule-à-Tout (V2X, Vehicle-to-Everything): 5G NR & 802.11bd. En effet, la basede données pour la conduite autonome, qui contient initialement des informationsobtenues a partir de capteurs (radar, caméra, etc.), sera complétée par des infor-mations provenant des messages CAMs (Cooperative Awareness Messages) et desmessages de perception collective CPMs (Collective Perception Messages). Ainsi, lesystème de supervision du véhicule pourra réagir plus efficacement à toute situationsur la route, et manœuvrer rapidement en cas d'évènement inattendu.Alors que nos autoroutes deviennent de plus en plus encombrées, pour une meilleureexpérience de voyage, les organisations des ITSs (Intelligent Transportation Systems)prévoient d'optimiser l'utilisation des routes et la gestion du traffic. Pour faire face auxproblèmes de mobilité actuels, les systèmes de véhicules conventionnels pour la circulation comme le régulateur de vitesse adaptatif ACC (Adaptive Cruise Control) nesont pas suffisants. Ce dernier système ajuste automatiquement la vitesse du véhiculeen fonction des informations des véhicules qui précèdent. Une stratégie de régulationplus dynamique nécessite un système coopératif qui fournira plus d'informations surles positions, les vitesses et les accélérations des véhicules voisins. Grâce à ce systèmecoopératif, de nouvelles applications véhiculaires peuvent être établies (un exempled'un tel système étant le régulateur de vitesse adaptatif coopératif CACC (Cooperative Adaptive Cruise Control)). Le peloton est l'une des applications les plus connuesqui peuvent améliorer la capacité routière. En effet, en optimisant la distance entre les véhicules, la capacité des infrastructures routières peut augmenter de plus de30%. En utilisant le même concept, la traînée de l'air est réduite tel que démontrépar les chercheurs qui ont affirmé que jusqu'à 20% d'énergie peut être économisée.D'un point de vue réseau, la réduction de la distance entre les véhicules permettrad'établir de nouvelles liaisons de communication point à point, entre les véhicules àl'avant et à l'arrière, offrant ainsi l'opportunité d'avoir une communication hybride.Le sujet de cette thèse porte sur les réseaux sans fil véhiculaires et, plus précisément,sur l'utilisation de la transmission radio et de la communication par la lumière pouraméliorer la sécurité des véhicules. La thèse est motivée par les problèmes de fiabilitéet d'évolutivité de la norme IEEE 802.11p. L'idée est d'évoluer vers de nouvellestechniques, notamment dans les futures normes 803.11bd, et d'associer la transmissionradio à la communication en lumière visible (VLC, visible light communication)pour permettre une communication hybride.La première partie de la thèse concerne le développement de techniques d'accès àfaible latence dans les réseaux véhiculaires dans le cadre des futures normes. L'objectifde cette première partie est de proposer une couche MAC (Medium Access Control) basée sur la technique TDMA (Time Division Multiple Access), enrichie d'unmécanisme permettant aux nœuds de bénéficier d'un accès fiable et à faible latenceà la demande. Cette partie introduit également un schéma d'accès spécial pour lespaquets d'urgence de haute priorité, tout en garantissant un accès fiable et à faible latence. Nous complétons cette étude par une analyse mathématique des performancesde ce schéma, pour montrer ses hautes performances. De plus, nous proposons unestratégie originale de transmission de messages basée sur une solution Cross-Layerpour diffuser les messages d'urgence de manière multi-hop (exploitant la relation entreles couches MAC et réseau).La deuxième partie de la thèse traite de VLC pour coordonner les manœuvres dansles réseaux de véhicules. Les avantages des communications sans fil pour le contrôledes pelotons sont, entre autres, une meilleure stabilité de la chaîne (qui améliore parconséquent la sécurité et permet un plus petit écart entre les véhicules), le regroupe-ment des véhicules ainsi que les manœuvres de séparation. Bien que l'importancedes communications sans fil pour le peloton soit bien reconnue, certains aspects descommunications n'ont pas été étudiés en profondeur dans les projets précédents. Unde ces aspects est la coexistence de VLC avec les technologies de communication radio véhiculaire dans un système hétérogène pour améliorer les performances globales(chaque vecteur de communication compensant les inconvénients de l'autre). L'idéeest de développer un algorithme qui sélectionne la communication radio proposéedans la partie 1, et la communication en lumière visible pour la conduite autonome,en fonction de l'état du canal radio et de l'alignement du peloton.

International audience; Named Data Networking (NDN), one of the most suitable candidates for the future Internet architecture, allows all network nodes to have a local cache that is used to serve incoming content requests. Content caching is an essential component in NDN: content is cached in routers and used for future requests in order to reduce bandwidth consumption and improve data delivery speed. Moreover, NDN introduces new self-certifying contents features that obviously improve data security and make NDN a secured-by-design architecture able to support an efficient and secure content distribution at a global scale. However, basic NDN security mechanisms, such as signatures and encryption, are not sufficient to ensure security in these networks. Indeed, the availability of the Data in several caches in the network allows malicious nodes to perform attacks that are relatively easy to implement and very effective. Such attacks include Cache Pollution Attacks (CPA), Cache Privacy Attacks, Content Poisoning Attacks and Interest Flooding Attacks. In this paper, we identify the different attack models that can disrupt the NDN operation. We conducted several simulations on NDNSim to assess the impact of the Cache Pollution Attack on the performance of a Named Data Network. More precisely, we implemented different attack scenarios and analyzed their impact in terms of cache hit ratio, data retrieval delay and hit damage ratio.

International audience; The evaluation of the road markings visibility has been defined according to human needs, but shall now be extended to the needs of vision-based Advanced Assistance Driving Systems (ADAS). Several publications propose minimum levels of daytime or night-time visibility of the road marking and their contrast with the surrounding pavement, to ensure optimum detectability of Machine Vision systems. However, the calculation methodology is rarely indicated (the heterogeneity of the markings is not always taken into account) and very little information about the Machine Vision system is provided. In this study, a daily experiment was conducted on a small dry road section using a mobile retroreflectometer and a vehicle equipped with a Machine Vision system. Several statistical analyses are performed on the collected data at different study scales (punctual or global scale). They show that despite very low levels of marking retroreflection values and visibility contrast ratios, the road marking lines are almost always very well detected by the camera's algorithm. That demonstrates that the current indicators characterizing the marking visibility according to standards are not enough to fully understand the behaviour of Autonomous Vehicle cameras.

ESREL2021, 31st European Safety and Reliability Conference, Angers, FRANCE, 19-/09/2021 - 23/09/2021; The quality and the ability of the road infrastructures play a very important role to ensure safe and convenient transportation. Current human needs oblige us to optimize the use of our travel resources. The autonomous driving is proposed as a solution to make the transport more efficient. The autonomous vehicle contains cameras and lidars that can communicate with the road infrastructure to understand the environment where they are traveling. Thus, the existing road infrastructure must therefore be developed and maintained to address the availability needs induced by the new uses of mobility. In consequence, the maintenance of the road infrastructure must be optimized. In previous work, a maintenance policy for lines and road cracks pavement of the road infrastructure was proposed by considering it as a 4 components system: median strip line, emergency line, broken center line and pavement. The proposed strategy is based on the individual optimal maintenance plan of each component. Then, over a finite planning horizon, the scheduled maintenance actions are grouped together to ensure both the proper functioning of the system and to minimize the cost of maintenance. To improve this previous work, a new long-term horizon (30 years) dynamic grouping maintenance strategy is proposed, dynamically considering new monitoring data. This new algorithm is applied to the maintenance optimization of the road infrastructure, using the Long-Term Pavement Performance database and French National Road 4 feedback data.

International audience

Understanding the behaviors and intentions of pedestrians is still one of the main challenges for vehicle autonomy, as accurate predictions of their intentions can guarantee their safety and driving comfort of vehicles. In this paper, we address pedestrian crossing prediction in urban traffic environments by linking the dynamics of a pedestrian's skeleton to a binary crossing intention. We introduce TrouSPI-Net: a context-free, lightweight, multi-branch predictor. TrouSPI-Net extracts spatio-temporal features for different time resolutions by encoding pseudo-images sequences of skeletal joints' positions and processes them with parallel attention modules and atrous convolutions. The proposed approach is then enhanced by processing features such as relative distances of skeletal joints, bounding box positions, or ego-vehicle speed with U-GRUs. Using the newly proposed evaluation procedures for two large public naturalistic data sets for studying pedestrian behavior in traffic: JAAD and PIE, we evaluate TrouSPI-Net and analyze its performance. Experimental results show that TrouSPI-Net achieved 0.76 F1 score on JAAD and 0.80 F1 score on PIE, therefore outperforming current state-of-the-art while being lightweight and context-free., Accepted to IEEE International Conference on Automatic Face & Gesture Recognition 2021 (December 15 - 18, 2021) 7 pages, 2 Figures

The emergence of demand responsive transport (DRT) services is restructuring mobility in the urban universe. They provide a high level of service and in many cases compete with public transportation modes. However, in less dense areas, where conventional public transit (CPT) services such as buses are inefficient and costly, DRT services could be an alternative that is both profitable and provides passenger satisfaction. This paper investigates the economic and socioeconomic potential of replacing CPT with DRT services. In particular, it focuses on the development and combination of two models. The first is an agent-based model, which describes movements of vehicles and assigns them to passengers according to a utility function. The assignment equilibrium problem is solved for demand that is elastic to generalized cost alone. The second is an economic optimization model, based on a simulated annealing algorithm, which aims to determine supply conditions that maximize the benefit for the operator, the user, the environment, and society. Four economic problems are discussed and formulated accordingly. In addition, supply optimization is carried out in particular with respect to fleet size, trip fare, and vehicle capacity. Finally, these models are applied to a real case in the Paris metropolitan area where a bus service has been replaced by a DRT system. The results show that though this shift is not beneficial from a societal point of view, bus demand is attracted by a DRT service consisting of 30 vehicles and charging a fare of €0.5. The operator would aim to propose a taxi-service, with small-sized vehicles and higher fares, in order to increase their profitability. User utility, on the other hand, would suggest that public authorities should regulate fares and vehicle capacity (more than 6 seats). Fare regulation, in particular, will depend on the fleet size, ranging linearly from €0 for 25 vehicles to €4 for 70 vehicles. Finally, we find through the sensitivity analysis that thresholds exist for demand and fixed costs (respectively 85% and 90% of reference values) beyond which the bus line is more beneficial to society than the DRT service.

Epique 2021, 11ème Colloque de Psychologie Ergonomique, Lille, France, 06-/07/2021 - 09/07/2021; Si l'automatisation des transports publics a pris un nouvel élan au cours des dernières décennies, le train autonome (TA) n'en est encore qu'à ses débuts. Pour guider sa conception, il est nécessaire d'étudier son acceptabilité par le public. L'objectif de cette étude est d'examiner pour la première fois la génération spontanée et provoquée d'idées relatives au TA. Inspirée de la méthode d'associations libres, cette étude est divisée en 4 parties respectivement consacrées au train, au TA en général, puis au TA sans conducteur mais avec du personnel à bord et enfin sans personnel à bord. Les résultats ont montré que les 30 participants interrogés pensent plutôt au voyage ou au monde du ferroviaire en évoquant le train alors qu'ils sont davantage focalisés sur les aspects liés à l'absence de personnel ou à la sécurité à bord pour le TA. Ces résultats pourront guider la conception des futurs TA, dans une approche d'ergonomie prospective.

International audience; Recently, the number of electric vehicles (EVs) is increasing due to the decline of oil resources and the rising of greenhouse gas emissions. However, EVs have not received full acceptance by consumers due to the limitations of the stored energy and charging problems. The dynamic or in-motion charging solution has become a suitable choice to solve the battery-related issues. Many researchers and vehicle manufacturers are working to develop an efficient charging system for EVs. In order to improve the efficiency of the dynamic wireless power transfer (DWPT), the electromagnetic coupling coefficient between the two parts of the coupler must be maximized. This paper was dedicated to find the optimal topology of a magnetic coupler with the best coupling factor while taking in consideration the displacement and the misalignment of the EV. The article is introduced by developing a methodology for characterizing the electrical parameters of couplers, followed by a comparative study of different forms of coils suitable for dynamic charging of electric vehicles. The particularity of the proposed study concerned the overall dimensions, or the areas occupied by the windings of the coils remaining the same for all the chosen shapes and corresponding to the surface that is actually available under the EV. Simulation and experimental tests were carried out to validate the proposed study.

International audience; Visible light communication could be an interesting complement to the IEEE 802.11p-based vehicle-to-vehicle communication systems that are sensitive to interferences and delays in dense traffic scenarios such as platooning. Visible light could also provide a redundant distance measure that is crucial for path control in this application. In this paper, a system called visible light communication rangefinder and performing simultaneously vehicle-to-vehicle communication and range-finding using the headlamps and taillights is proposed for the first time. By exchanging a clock signal contained in Manchester-encoded signals, both the following and leading vehicles can share information and estimate their inter-distance through phase-shift measurement. The system is first presented theoretically and it is shown in particular that the Doppler effect has no significant impact on both functions. Then, it is modeled and validated using Simulink. Finally, both the range-finding and communication function are validated experimentally. The range-finding function is functional up to 25 m, and has a resolution of around 24 cm at 10 m, whereas the communication function provides a 500 kbps link with a BER below 10 −6 up to 30 m.

International audience; At time of crisis, relief teams must have assured connectivity, not only just within the team but also across different service agencies in the area. Since emergency agencies and essential services always send service cars to affected zones, advanced technologies and computing resources aboard these vehicles can be pooled together to boost network capacity temporarily, just where it is crucially needed. These vehicles become automotive virtual edge communicators (AVECs). They are managed by a vehicular inter-agency service orchestration and resourcing (ViSOR) system that creates transient proximity-based “trust circles” to manage novel cooperative hosting, opportunistic virtualization, and “car sourcing” of crisis zone data. This study evaluates the feasibility for this challenging but highly rewarding concept and identifies gaps in emerging technologies.

21st Congress of the International Ergonomics Association (IEA 2021), VANCOUVER, CANADA, 14-/06/2021 - 18/06/2021; International audience; Using an integrative user-centered approach, the research project aimed at designing and assessing an adaptive HMI to improve the safety and the quality of the take-over in a level-3 automated vehicle. Future autonomous driving users confirmed the need for the monitoring systems to improve safety as long as they are simple and not intrusive. They request specific information to understand and initiate appropriated actions in case of a critical situation. However, to avoid overloading and detri-mental effects, it is important that the system limits the amount of infor-mation at the right moment.

International audience

International audience

International audience; The main objective of Vehicular ad hoc networks (VANETs) is to make road transportation systems more intelligent in order to anticipate and avoid dangerous, potentially life-threatening situations. Due to its promising safety applications, this type of network has attracted a lot of attention in the research community. The dissemination of warning messages, such as DENMs (Decentralized Environmental Notification Messages), requirse an efficient and robust routing protocol. In previous studies, the active signaling mechanism has shown its ability to prevent collisions between users trying to allocate the same resource. In this paper, we propose an original message forwarding strategy based on the active signaling mechanism. Our proposal disseminates warning messages from a source vehicle to the rest of the network while minimizing the access delay and the number of relay nodes. For this purpose, a special time slot is dedicated to forwarding emergency warning messages. To avoid access collisions on this slot, the active signaling scheme we propose favours the selection of the furthest node as the forwarder. We carry out a number of simulations and comparisons to evaluate the performances of the scheme.

International audience; We consider a shared transportation system in an urban environment where human drivers collect vehicles that are no longer being used. Each driver, also called a platoon leader, is in charge of driving collected vehicles as a platoon to bring them back to some given location (e.g. an airport, a railway station). Platoon allocation and route planning for picking up and returning automated vehicles is one of the major issues of shared transportation systems that need to be addressed. In this paper, we propose a coalition game approach to compute 1) the allocation of unused vehicles to a minimal number of platoons, 2) the optimized tour of each platoon and 3) the minimum energy consumed to collect all these vehicles. In this coalition game, the players are the parked vehicles, and the coalitions are the platoons that are formed. This game, where each player joins the coalition that maximizes its payoff, converges to a stable solution. The quality of the solution obtained is evaluated with regard to three optimization criteria and its complexity is measured by the computation time required. Simulation experiments are carried out in various configurations. They show that this approach is very efficient to solve the multi-objective optimization problem considered, since it provides the optimal number of platoons in less than a second for 300 vehicles to be collected, and considerably outperforms other well-known optimization approaches like MOPSO (Multi-Objective Particle Swarm Optimization) and NSGA-II (Non dominated Sorting Genetic Algorithm).

International audience; Autonomous vehicles (AVs) will profoundly modify our travel habits. The collective impact of AVs will differ according to the autonomous mode choice. In this paper, we apply a simultaneousequation model to a database from an original 2017 survey of French mobility users to analyse their acceptance of two forms of autonomous transport mode: autonomous shuttles and robotaxis (N = 3,297). Our results show that the intention to use autonomous shuttles is on average greater than robotaxis. Gender and age influence autonomous mode choice, as well as the current transport mode. In addition, location and urban representations play a central role.

International audience; Latency is a very important metric to be taken into account in the design of the Connected and Automated Driving (CAD) technolnogy. In fact, some studies have shown that a large number of accidents could be avoided if the drivers were warned at least 0.5s beforehand. This is why real time communication technologies are becoming a key factor to increase the safety of drivers and passengers. Today, connected vehicles have a dedicated technology named vehicle-to-everything (V2X). Many current research efforts and standardization activities aim to make the next generation of V2X technology able to offer new usages and services, with the main focus being on their support of safety applications. To satisfy the real-time constraints of safety applications, Time Division Multiple Access (TDMA) is a widely-used technique to control and share the channel. TDMA-based Medium Access Control (MAC) protocols have resolved many issues in VANETs and shown a good performance in literature by ensuring a bounded access delay to send an emergency message. However, latency can be further improved by eradicating some imperfections in the functionality of these protocols, such as the access collision problem which can occur when two or more vehicles in the same two-hop neighborhood set try to access the same slot at the same time. That is why we recently proposed an Active Signaling based TDMA MAC protocol, called AS-DTMAC, which operates above the existing DTMAC protocol. In this paper, we analytically study the performance of AS-DMTAC when we have an homogeneous arrival on each time slot with an error in the signaling process.

Nowadays, supervised deep learning techniques yield the best state-of-the-art prediction performances for a wide variety of computer vision tasks. However, such supervised techniques generally require a large amount of manually labeled training data. In the context of autonomous vehicles perception, this requirement is critical, as the distribution of sensor data can continuously change and include several unexpected variations. It turns out that a category of learning techniques, referred to as self-supervised learning (SSL), consists of replacing the manual labeling effort by an automatic labeling process. Thanks to their ability to learn on the application time and in varying environments, state-of-the-art SSL techniques provide a valid alternative to supervised learning for a variety of different tasks, including long-range traversable area segmentation, moving obstacle instance segmentation, long-term moving obstacle tracking, or depth map prediction. In this tutorial-style article, we present an overview and a general formalization of the concept of self-supervised learning (SSL) for autonomous vehicles perception. This formalization provides helpful guidelines for developing novel frameworks based on generic SSL principles. Moreover, it enables to point out significant challenges in the design of future SSL systems.

Robotic navigation requires complex algorithms to build an accurate and robust representation of a spatial environment. These systems must therefore constantly be able to detect useful information, to encode it and to merge it with the current representation. Despite the apparition of new efficient algorithms, they still struggle to offer a solution that is efficient and light enough to be deployed on embedded systems [2]. Nevertheless, biological systems, especially mammals, can solve these problems on a daily basis with very effective methods. Playing a key role both in the spatial cognition in mammals and in the episodic memory in humans, the hippocampal formation is a part of the brain where the interaction of several kinds of neurons results in a neural map of our spatial environment [3]. For instance, place cells [3] were discovered with the particularity to encode precise positions in an environment. Despite decades of research on this subject, a complete view of the information processing operated by these neurons is still lacking. Relying on these observations, a few neurorobotics navigation models have been proposed to better understand the neural mechanisms involved in animals navigation by mimicking the functioning of the hippocampal formation [6]. Among the different models proposed, the neural model LPMP (Log-Polar Max-Pi) allows to build place cells in an unsupervised way by merging current visual information and the absolute orientation of the robot [1]. It leans on several important principles such as one-shot learning or Hebb principle to build a neural representation of an environment. Despite its simplicity, the model has been shown to be competitive with the state of the art, and has given encouraging first results on an autonomous vehicle [1]. The tests carried out, however, also revealed computational cost limitations which prevent it from being deployed over long distances. To solve this problem, we propose the model HSD (Hierarchical Sparse Dictionaries), a new unsupervised model of visual information encoding intended for robot navigation. Inspired from the visual processing of primate, this model proposes to cascade "Topologic" Sparse Dictionaries (TSD) and pooling layers to build a lighter representation of a place. The integration of a topology in the sparse dictionary makes possible the application of an efficient pooling and allows

International audience; In VANET networks using TDMA access it is possible to approach the full capacity of the channel when the density of vehicles increases. When this occurs, some vehicles will be unable to communicate with the rest of the network. The use of Visible Light Communication (VLC) for certain vehicles can reduce the load on the radio channel and allow all vehicles to communicate on the network. The idea of our approach is to group vehicles with VLC capacity in platoons where only the lead vehicle would use radio transmission while the other vehicles in the platoon would communicate via the VLC system. We show that if the vehicles are randomly equipped with VLC transmission capabilities (with a probability p), the gain in bandwidth is significant even without rearranging the order of the vehicles. In this paper we estimate the performance gains obtained by using VLC in terms of bandwidth and probability of collision during an initial access.

This study deals with the reliability of a new power embedded PCB (printed circuit board) concept highly integrated into an electrical machine for aeronautical and automotive applications. The first part of this paper describes the assembly topology. The second part presents the main steps of module manufacturing using the silver sintering connection and the pre-impregnated composite fibers lamination. The third part shows the results of electrical and mechanical tests realized on this module. The purpose is to validate the design of a new assembly technology which offers an enhanced reliability based on the manufacturing time optimization, the reduction of the process steps, the module functionality and the feasibility.

PCB-Embedding technology is an established approach for the miniaturization of electronic systems and modules. Silver sintering, which is a reliable alternative to Lead-Free Bonding, is used to assemble the embedded power components in the PCB. Thermal cycles [−25 °C/+125 °C] have highlighted degradations in the die attach. The shear strength of dice decreases from 4 MPa before cycling to 2 MPa at 1040 cycles. The junction temperature has drifted during cycling especially at high temperatures. Simulations have shown that the maximal stresses are in the preg materials. This suggests cracks of the preg during the experiments. Simulations showed also a concentration of plastic stresses on the die attaches. The accumulation of stresses increases during aging. The presented methodology of thermal cycling simulation represents a good way to estimate the lifetime of embedded PCB.

VTC2020-Fall, IEEE 92nd Vehicular Technology Conference, virtuel, , 18-/11/2020 - 16/12/2020; Reinforcement learning is a commonly used technique in the field of traffic signal control. By iteratively testing actions given the current network state, a traffic signal agent gradually learns to optimally control traffic lights given the traffic situation at hand. While they usually outperform traditional control systems in the literature, these methods have to first go through an exploration phase where they test different actions in a trial-and-error fashion. The nature of reinforcement learning methods hence causes unstable performances and high computational costs. In order to limit the costs linked to this exploration phase, we propose a bootstrapping method that reduces the computation time needed for learning-based traffic control methods to reach acceptable performance levels. Our method models lanes of the road network as queues, and derives results on the average service time of vehicles at the intersection level in order to estimate the agent's policy. The performance of our bootstrapping method is then compared to a more traditional Q-Learning method using the SUMO simulator. Simulation results show that bootstrapping alleviates both of these issues by immediately reaching acceptable performance levels by quickly training the agent without any direct interaction with the simulation environment.

The main aim of this thesis is to extend navigation models built for indoor environments to autonomous vehicles. These models follow a neurorobotics approach: they reproduce neural substrates that sustain the mechanism responsible for spatial cognition in the mammalian brain. This thesis introduces and shows the potential of bio-inspired robotics and neurorobotics to the community of autonomous driving.This manuscrit begins by giving a state of the art of robotic navigation. It presents the subsidiary question of visual recognition.The notions of bio-inspired robotics and neurorobotics are then introduced. An overview of anatomical observations necessary to the comprehension of bio-inspired models is also given.The document describes a model for visual place recognition inspired by hippocampal place cells. This model is tested in outdoor on-road environments. The results show that the model is able to build a representation of the environment that can be used in navigation tasks.A reactive navigation model (PerAc) is then presented. This model uses the representation built by the visual place recognition model to learn sensorimotor units linking places and actions. A set of these associations allows the model to learn robust trajectories or goal positions. This model was initially built for robots evolving in indoor environments. It is embbedded and tested on an autonomous vehicle. The experiments show that the model is able to learn trajectory with a car in outdoor environments.My last contribution presented is about a visual contexts model. A second visual processing chain inspired by the primary visual cortex is proposed to build a coarse representation of the environment that modulates the activities of the visual place cells.The essay ends by giving a set of leads to improve different aspects of the proposed models, by modeling in more depth other brain structures or change the encoding of data for instance., L'objectif premier de cette thèse est de tester et d'étendre des modèles de navigation initialement développés pour de la navigation en intérieur et de les appliquer à la délégation de conduite des véhicules autonomes. La particularité de ces architectures de contrôle pour robot mobile est qu'elles sont issus d'une approche bio-inspirée et sont dite neurorobotiques : elles reproduisent les mécanismes neuronaux impliqués dans les comportements de navigation observés chez les animaux.Ces travaux concourent à introduire auprès de la communauté du véhicule autonome les approches bio-inspirées et leur potentiel au travers des modèles présentés.Cette thèse commence par aborder les technologies utilisées (GPS, lidar, ...) et les différentes stratégies de navigation (réactive, avec carte). Cette présentation se termine en circonscrivant le problème de la navigation des véhicules autonomes à la reconnaissance de lieux par la vision.Le manuscrit présente ensuite les approches bio-inspirée et neurorobotique avant de donner les notions d'anatomie facilitant la compréhension des modèles bio-inspirés décrits par la suite. Un modèle de reconnaissance de lieux par la vision est présenté en détail. Des expériences de simulations avec des données réelles permettent de tester ce modèle, initialement conçu pour la navigation intérieure, dans des environnements propres aux véhicules. Les résultats de ces tests montrent que le modèle construit une représentation de l'environnement capable de discriminer et de reconnaître des lieux. Une architecture neuronale de navigation réactive reposant sur le précédent système, le modèle PerAc est décrite ensuite. Ce modèle utilise les lieux construits par le modèle de reconnaissance et associe chaque lieu à une direction à suivre. Par un ensemble d'associations, il est possible d'apprendre à suivre de manière robuste des trajectoires ou une position à rejoindre. Ce modèle, conçu pour de la navigation en intérieur, est testé sur un véhicule dans des expériences servant à évaluer sa capacité à fonctionner sur véhicule. Les résultats prouvent que le modèle est capable d'apprendre des trajectoires avec une voiture en environnement extérieur.Ma dernière contribution porte sur un modèle de contextes visuels permettant d'accroître les performances computationnelles de la reconnaissance de lieux. Une chaîne de traitements visuelle inspirée du cortex visuel primaire effectue un découpage de l'environnement qui vient moduler la reconnaissance de lieux du modèle précédent. Ce manuscrit se termine en proposant un ensemble de pistes à même d'améliorer les différents aspects des modèles présentés, par exemple, en modélisant plus finement certaines zones du cerveau ou en utilisant un codage épars afin de réduire le coût computationnel.

This thesis is carried out in collaboration between the GeePs laboratory and the VEDECOM institute. The cost, volume and weight of electrochemical batteries still represent a major obstacle to the deployment of electric vehicles (EVs). One of the solutions being considered to extend the range of EVs without excessively increasing the capacity of the batteries, is to use contactless electrical energy transfer systems to power them while they are on the move. This thesis focuses on one of these techniques which is the resonant inductive energy transfer. The problems associated with this mode of energy transfer are mainly related to energy efficiency, the size constrained due to the need for integration into the vehicle and the road as well as compliance with electromagnetic emissionstandards. The efficiency is directly linked to the coupling of the two coils (magnetic coupler). A comparison of the coupling coefficients for different coupler geometries and different misalignment configurations is the subject of the first part of this work. In the second part, an approach based on equivalent sources is suggested for the analytical pre-dimensioning of shielding plate intended to limit the magnetic emissions in and outside the vehicle. In the last axis of the thesis, the study is dedicated to techniques for detecting the presence of the vehicle and the sequencing of the different ground coils. A conclusion giving an assessment of the work and perspectives that open up from this work, close this manuscript.; Les travaux de recherche de cette thèse sont menés dans le cadre d’une collaboration entre le laboratoire GeePs et l’institut VEDECOM.Le coût, le volume et le poids des batteries électrochimiques représentent encore un frein important au déploiement des véhicules électriques (VE). Une des solutions envisagées pour prolonger l’autonomie des VE sans augmenter démesurément la capacité des batteries, consiste à utiliser des systèmes de transfert d’énergie électrique sans contact pour les alimenter pendant leurs déplacements. Cette thèse porte sur une de ces techniques et plus particulièrement sur le transfert d’énergie inductif résonant. Les problématiques liées à ce mode de transfert d’énergie sont principalement liées au rendement énergétique, à l’encombrement contraint du fait de la nécessité d’intégration dans le véhicule et dans la route ainsi qu'au respect des normes d’émissions électromagnétiques.L’efficacité énergétique du transfert d’énergie est au premier ordre lié au couplage des deux bobines (coupleur magnétique). Une comparaison des coefficients de couplage pour différentes géométries de coupleurs et différentes configurations de désalignement fait l’objet d’une première partie du travail réalisé. Dans la seconde partie une approche à base de sources équivalentes est proposée pour le prédimensionnement analytique d’une plaque de blindage destinée à limiter le rayonnement du coupleur dans le véhicule et en dehors de celui-ci. Dans le dernier axe de la thèse, l’étude est dédiée aux techniques de détection de la présence du véhicule et au séquencement de l’alimentation des bobines au sol. Une solution originale, permettant de répondre à cette problématique est proposée. Le bilan des travaux ainsi que les perspectives envisagées, viennent clôturer ce manuscrit.

International audience; We address ethical dilemma situations that may arise during autonomous driving. To evaluate how this deliberation could work, we propose a decision-making algorithm based on a Markov Decision Process (MDP) which controls the vehicle in normal conditions. When a dilemma situation is detected, the collision severity is determined by an evaluation of the harm incurred by different types of road users. Then, to illustrate different moral approaches, three different policies are proposed: one based on ralwsian contractarianism, another on utilitarianism, and finally on egalitarianism. Each policy supports a different view of the concept of fairness, potentially producing different behaviors for the same dilemma situation.

International audience; Nowadays, many telecommunication systems (wifi, cable systems and 4G, 5G cellular networks) use Orthogonal Frequency Division Multiplexing (OFDM) as the physical layer standard. The design of efficient OFDM signal detection algorithms is very important to provide reliable systems, and this is particularly true for Vehicular Adhoc Networks (VANETs) involving autonomous vehicles, where missing a signal or detecting a fake one may cause a dangerous situation. The performance of these algorithms is generally evaluated in terms of their robustness against noise. In this paper, we evaluate the probability of error in signal detection in order to establish the minimum length of preamble needed for the active signaling process. This mechanism is used in AS-DTMAC (active signaling fully distributed TDMA-based MAC protocol) to reduce access collisions. Thus, by reducing the length of the preamble, greater time is given for the payload part of the packet, resulting in increased throughput.

This paper introduces an original maintenance decisionmaking for road marking. An ascending hierarchical classification segments the line into strategic zones and an EM algorithm estimates a lifetime model of markings. Two strategies are discussed: aged-based systematic andconditioned-based replacement policies., Cette communication introduit une approche originale de la maintenance des lignes de marquage. Une classification ascendante hiérarchique segmente la ligne en zones stratégiques et un algorithme EM estime un modèle de durée de vie des marquages. Deux stratégies sont discutées : systématique par rapport à l'âge et conditionnée par la dégradation courante.

International audience; Advanced vehicular applications such as autonomous driving are leading to a new evolution in vehicular radio access technologies. The Task Group BD on the Next-Generation V2X (NGV) have defined IEEE 802.11bd as the new standard for the Dedicated Short Range Communications (DSRC). Notwithstanding its the promising performances in terms of high reliability, low latency and high throughput, the design must also respect certain specifications, such as coexistence, interoperability, and compatibility with the previous standard. In this article, we study how the IEEE 802.11p protocol could coexist with a TDMA-based protocol named AS-DTMAC, which has recently been proposed to control access in Vehicular Ad Hoc Networks (VANETs). We carry out several analyses to show that the two protocols can coexist by several analysis when both are operating simultaneously on the same network. We also propose a modification of AS-DTMAC to handle the situation where both IEEE 802.11p and AS-DTMAC have to send urgent packets.

International audience; Recently, several cross-layer protocols have been designed for vehicular networks to optimize data dissemination by ensuring internal communications between routing and MAC layers. In this context, a cross-layer protocol, called TDMA-aware Routing Protocol for Multi-hop communications (TRPM), was proposed in order to efficiently select a relay node based on time slot scheduling information obtained from the MAC layer. However, due to the constant evolution of cyber-attacks on the routing and MAC layers, data dissemination in vehicular networks is vulnerable to several types of attack. In this paper, we identify the different attack models that can disrupt the cross-layer operation of the TRPM protocol and assess their impact on performance through simulation. Several new vulnerabilities related to the MAC slot scheduling process are identified. Exploiting of these vulnerabilities would lead to severe channel capacity wastage where up to half of the free slots could not be reserved.

International audience; The constantly growing number of vehicles on our roads has become an increasing major cause of serious injury and death. Efficient data dissemination in Vehicular Ad hoc Networks (VANETs) inevitably requires an efficient and robust routing protocol. In this context, several categories of routing protocols have been proposed in the literature to meet VANETs application requirements in terms of delay, packet loss and throughput. In this paper, we focus on cross layer routing protocols. We present a survey of state-of-the-art MAC aware routing protocols designed for VANETs. These solutions can broadly be divided into two categories: contention-free and contention-based MAC-aware routing. In this paper we carryout a comprehensive comparison of these approaches. Finally, we identify open research issues that should be addressed in order to improve MAC aware routing techniques in VANETs.