Cooperative Perception and Communication in vehicular technologies – CooPerCom

The WP2 is focused on theoretical work (dynamic distributed complex systems, uncertainties, reliability, and failure prediction). Fusion systems become more complex as more vehicles are involved in the information treatment chain steps, from the data sensor extraction to the decision. The main objective of this work package is to provide optimized answers to the following questions: How and which information to extract at a microscopic and macroscopic traffic level? How to represent and to aggregate information and the corresponding sources of uncertainties? How to decide that a global result is good leading to safe detection of vehicle behavior anomalies? How the vehicles and their drivers in a cooperative manner can adjust and use efficiently the output of these information processing chains?
The WP3 is dedicated to the perception and Risk assessment. In the perception stages, several level of information handle will be tackled (local and extended perception, positioning, fusion). The main objective of this task is to handle the theoretical work achieved in the last task in order to propose and to develop an efficient way to build dynamic and reactive maps. These maps give a local or an extended representation and modeling of the information resources. Each one of the map models is both a spatial and temporal representation of a specific situation.
The WP4 will tackle the communication Protocols and Dynamic Resource Allocations. This task will be carried out by Canadian and French researcher’s partnership. First, we will look at issues of efficient network protocols implementation for active safety and extended maps building. Second we will address issues of security and privacy related to exchange of information among vehicles.

From the previous tasks, several level of information will be available (local map and extended map). Moreover an efficient and robust way to communicate information from V2V and V2I will be provided. In this final task, we plan to apply the perception and communication results to the design and the development of 3 specifics platforms. The first platform consists to implement a modeling of the communication devices in a virtual environment dedicated to embedded sensors simulation (SiVIC). In the second platform, the perception and communication devices will be developed and implemented with scale model car in real time. The last platform will implement the same system (perception and communication) on LIVIC prototypes with real ADAS applications. The ADAS applications will be chosen among existing applications developed in Integrated European projects (CVIS and SafeSpot).

nothing yet

In progress. This information will be given soon

The inclusion of new embedded technologies in vehicular applications is precluded by several constraints and requirements among which optimizing solutions and finding trade-offs between 1) safety, 2) low-cost, 3) manufacturability, 4) environment friendliness 5) and standard policies and regulations. In the proposed project we intend to put emphasis on the safety aspects of the requirements while keeping in mind the other issues. One of the major concerns related to the introduction of new technologies in vehicular applications is safety. Apart from HMIs (Human Machine Interfaces) and driver’s distraction issues, safety is directly related to the level of reliability and robustness of the sensors and systems involved. In fact, one of the major challenges in industry is to achieve and guarantee a very high level of reliability and robustness of on-board equipments to insure sufficient safety at a cost low enough to enable large deployment and mass production in the automotive industry. The collaborative and distributed approach we propose for building an extended vehicular perception in clusters of vehicles address this problem.
One of the main objectives of this proposal is to determine how we can improve those reliability and robustness aspects of the embedded systems and sensors which are becoming ubiquitous, more numerous and more complex over time. As the number of sensors and systems increases, the challenge of insuring sufficient reliability and thus a high level of safety is becoming overwhelming. In addition, vehicular environment is highly dynamic and thus requires a high level of connectivity, which in turn requires reliability and robustness as well. Robustness is linked to the capability of the systems to adapt and adjust their performance/behaviour according to the situation and environment at hand. These adaptation and learning capabilities are directly related to safety as well.
Among the sub-objectives, we have in mind to explore and develop intelligent signal and information processing tools to take advantage of the presence of multiple vehicles and inter-vehicles communications capabilities to gather information from multiple sources, validate individual data pieces, assess their level of uncertainty and exploit potential redundancies in order to mitigate risk of unexpected failures and optimize reliability and robustness. In order to exploit various sources of information available from surrounding vehicles, we need to consider cooperative and distributed approaches that will rely on emerging information processing and communications tools. The proposed ANR-NSERC project is presented from the Canadian side by a team of researcher from the Université de Sherbrooke (Profs. Denis Gingras PI, Soumaya Cherkaoui) and the University of Toronto (Prof. Shahrokh Valaee) and supporting organization such as the Networked Vehicle Association Canada (NVA), Canadian Advanced Technology Alliance (CATA), the Montreal-based company Opal-RT and Transport Canada. From the French side, the team is composed of the mixed INRETS-LCPC laboratory LIVIC (Dr Dominique Gruyer PI, Dr Sébastien Glaser, Steve Pechberti), the Université d’Evry Val D’Essonne (Prof. Vincent Vigneron) and the Université de Paris-Sud (Dr. Alain Lambert).