Robust cooperative and communication-friendly localization in vehicular networks

Precise location services are seen as key enablers to future Intelligent Transport Systems (ITSs). Relying on Vehicle-to-Vehicle (V2V) communication links, one promising solution consists in performing distributed Cooperative Localization (CLoc). More specifically, Cooperative Awareness Message (CAM) broadcasts from neighboring vehicles (seen as "virtual anchors") are used to exchange positional information and to measure V2V radiolocation metrics such as the Received Signal Strength Indicator (RSSI). Then this information is fused with local onboard Global Navigation Satellite System (GNSS) data to enhance both estimated "ego" vehicle's position and that of neighboring vehicles, thus contributing to enrich Local Dynamic Maps (LDM). Conventional CLoc solutions may be unsuitable, or even non compliant, within the highly specific and challenging context of Vehicular Ad hoc NETworks (VANETs) (e.g., in terms of mobility patterns, protocol constraints, V2V channel load, network topology and connectivity...). This report thus describes and evaluates a new communication-friendly CLoc data fusion framework, along with adapted algorithms enabling robust location estimation, low-complexity cooperative link selection, low-overhead message approximation, and transmission/fusion rates adaptation.