Cooperative Position Prediction: Beyond Vehicle-to-Vehicle Relative Positioning

Abstract

Reliable and accurate relative position prediction techniques are vital to the integrity of cooperative intelligent transportation systems (C-ITS) because most C-ITS safety applications should constantly access a (relative-) localization mechanism that satisfies not only a certain (relative-) position accuracy range but also a minimal computational complexity. A global navigation satellite system (GNSS) coupled with dedicated short-range communications (DSRC) links is a candidate for the localization system required for vehicular relative positioning. The vehicle-to-vehicle (V2V) exchange of safety-related data such as raw GNSS measurements, as accurate and frequent as possible, via DSRC links allows the vehicular localization mechanism to maintain real-time relative positioning (RRP) vectors. However, V2V DSRC safety messages may include inaccurate positioning data or may fast become outdated, and/or DSRC links may occasionally fail due to various adverse scenarios including network overheads that drop the bandwidth of vehicular ad-hoc networks. Hence, each DSRC-enabled vehicle may not necessarily know the latest positioning data of neighboring vehicles. This calls upon reliable and accurate relative position prediction mechanisms to cover the DSRC-related deficiencies of relative positioning. This paper, at first, studies the benefits of a terrestrial communications system complementing DSRC-based V2V RRP and then considers position prediction techniques suitable for V2V RRP and examines a few existing solutions and makes a detailed comparison based on a set of required positioning performance parameters for vehicle safety application.