Distributed link selection and data fusion for cooperative positioning in GPS-aided IEEE 802.11p VANETs

In future Intelligent Transport Systems (ITS), advanced car safety applications may require that each vehicle determines not only its own absolute position (i.e., “ego” car localization on the road) but also the positions of its immediate neighbors in a continuous and accurate way. For this sake, intervehicle data transmissions via Dedicated Short Range Communications (DSRC) (e.g., compliant with the IEEE 802.11p standard) can be exploited to support both peer-to-peer ranging based on Received Signal Strength Indicators (RSSI) and Cooperative Positioning (CP) between GPS-equipped vehicles. Since the aggregation of heterogeneous (and possibly asynchronous) sources of information remains quite challenging in such Vehicular Ad hoc NETworks (VANETs), we herein describe and compare gradually cooperative solutions that perform distributed data fusion through Extended Kalman Filtering (EKF). One first stake consists in re-aligning in time the data received from cooperating cars following a generalized prediction approach. In addition, using a so-called validation gate based on innovation monitoring and/or a Cramer-Rao Lower Bound (CRLB) indicator accounting for conditional positioning performance, a low-complexity link selection mechanism is developed to identify the most relevant neighboring cars and/or the best RSSI candidates to feed the fusion engine. Preliminary simulation results, obtained under realistic IEEE 802.11p radio parameters and varying GPS accuracy conditions, illustrate benefits from selective cooperation, especially in terms of “ego” car navigation continuity.