Position based routing in crowd sensing vehicular networks

Using vehicles as sensors allows to collect high amount of information on large areas without the need to deploy extensive infrastructures. Although cellular technologies are presently the only solution to upload data from vehicles to control centers, in the next future short range wireless technologies could be used to offload part of this data traffic through vehicle to vehicle and vehicle to roadside communications. In such scenario, the greedy forwarding (GF) position based routing is an interesting algorithm to efficiently route packets from vehicles to the destination. However, GF suffers from the well known problem of local minima, which causes part of the packets to remain blocked in certain areas of the scenario. To deal with this issue, we propose two novel routing algorithms, specifically designed for crowd sensing vehicular networks (CSVNs): GF with available relays (GFAVR), fully distributed and independent of the scenario, and GF with virtual roadside units (GFVIR), exploiting a preliminary design phase where local minima are located. Through extensive simulations performed in different realistic urban scenarios, results demonstrate that both algorithms allow to improve data delivery by 10–40%, with negligible overhead and limited increase of complexity.