Quality-Aware Broadcasting Strategies for Position Estimation in VANETs (1904.04062v1)

Abstract: The dissemination of vehicle position data all over the network is a fundamental task in Vehicular Ad Hoc Network (VANET) operations, as applications often need to know the position of other vehicles over a large area. In such cases, inter-vehicular communications should be exploited to satisfy application requirements, although congestion control mechanisms are required to minimize the packet collision probability. In this work, we face the issue of achieving accurate vehicle position estimation and prediction in a VANET scenario. State of the art solutions to the problem try to broadcast the positioning information periodically, so that vehicles can ensure that the information their neighbors have about them is never older than the inter-transmission period. However, the rate of decay of the information is not deterministic in complex urban scenarios: the movements and maneuvers of vehicles can often be erratic and unpredictable, making old positioning information inaccurate or downright misleading. To address this problem, we propose to use the Quality of Information (QoI) as the decision factor for broadcasting. We implement a threshold-based strategy to distribute position information whenever the positioning error passes a reference value, thereby shifting the objective of the network to limiting the actual positioning error and guaranteeing quality across the VANET. The threshold-based strategy can reduce the network load by avoiding the transmission of redundant messages, as well as improving the overall positioning accuracy by more than 20% in realistic urban scenarios.