Vehicular Sensor Networks In Congested Traffic: Linking Stv Field Reconstruction And Communications Channel

It has been proposed that the use of speed and position information from a subset of vehicles in the traffic (probe vehicles) can provide for accurate traffic information. Furthermore, it can be seen as an economic and scalable alternative to the use of inductive loop detectors, cameras and radars. However, the impact of the communications channel performance in the estimation of the traffic states has been insufficiently studied.In this work we propose the use of the Wireless Sensor Network (WSN) paradigm to develop a Vehicular Sensor Network (VSN) in order to obtain accurate traffic information from a few probe vehicles. The problems that plague WSNs are 1) the deployment is to cope with the variations of the measured field, the Space-Time-Velocity (STV) field and 2) the communications channel influences data collection reliability. In order to assess these two issues, we perform 1) accurate simulation of discrete vehicular traffic in order to obtain spatiotemporal patterns that closely mimic traffic congestion and 2) accurate simulation of Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) wireless communications.We obtain experimental evidence that with a Fraction of Sensor Vehicles (FSV) as low as 1% we are able to obtain accurate measurements of traffic congestion accounting for Packet Loss from Rayleigh fading, Doppler spreading and multihop relaying. We present results for different FSVs and RoadSide wireless bridge (RSU) densities and assess that, for FSVs of 10% and RSU densities half the usual detector densities, the reconstructed STV fields are virtually indistinguishable from the ground truth.