Realistic Vehicle-to-Vehicle Link Simulation based on Measurement Traces

Vehicular Networks are establishing connectivity between different devices among vehicles and infrastructures, allowing for several applications such as, content distribution, vehicle data collection and safety. Measuring the performance of this kind of networks is becoming an important factor in order to understand the capability of vehicular networks to support these applications. Nevertheless, vehicular networks are not widely available and, many times, researchers have to make use of network simulators like NS-3. Network performance can be described through various metrics, including, Delay, Packet Loss and, one of the most important, Throughput, which are heavily affected by high mobility, vehicle and obstacle density. In order to enhance the estimation of performance in NS-3, a model was developed to take advantage of geostamped RSSI measurements and obtain a profile of the signal for a given scenario together with performance estimates. These estimates are then compared to real-world values extracted from several benchmarks to validate Throughput estimation. An approach is taken to correct possible errors from these results, by analysing the error pattern among different available scenarios. Results show that these models developed allow for more accurate experimentation on vehicular networks.Vehicular Networks are establishing connectivity between different devices among vehicles and infrastructures, allowing for several applications such as, content distribution, vehicle data collection and safety. Measuring the performance of this kind of networks is becoming an important factor in order to understand the capability of vehicular networks to support these applications. Nevertheless, vehicular networks are not widely available and, many times, researchers have to make use of network simulators like NS-3. Network performance can be described through various metrics, including, Delay, Packet Loss and, one of the most important, Throughput, which are heavily affected by high mobility, vehicle and obstacle density. In order to enhance the estimation of performance in NS-3, a model was developed to take advantage of geostamped RSSI measurements and obtain a profile of the signal for a given scenario together with performance estimates. These estimates are then compared to real-world values extracted from several benchmarks to validate Throughput estimation. An approach is taken to correct possible errors from these results, by analysing the error pattern among different available scenarios. Results show that these models developed allow for more accurate experimentation on vehicular networks