Sophisticated Spectrum Sharing Analysis: The Case of the 5.9 GHz Band
19 Pages Posted: 15 Mar 2018 Last revised: 18 Aug 2018
Date Written: March 14, 2018
Efficient spectrum sharing is becoming more critical as the demand for wireless services increases across many industries and available bandwidth remains scarce. To respond to this need, regulators must understand inter-system interactions in specific spectrum sharing scenarios in order to make informed policy judgments. These decisions are best made through risk-informed interference analysis, rather than worst-case advocacy, as noted by De Vries (2017) and many other scholars. Nowhere is this analytic need more apparent and urgent than the 5.9 GHz frequency band. 5.9 GHz is adjacent to existing Wi-Fi spectrum and a prime candidate to enable highly-demanded wireless broadband services, but it is today allocated exclusively to Dedicated Short Range Communications (DSRC), a technology in development for vehicle to vehicle (V2V) and vehicle to infrastructure (V2I) safety and commercial communications. A sophisticated and realistic technical understanding of coexistence between Wi-Fi and DSRC is therefore central to this archetypal spectrum sharing case study – an understanding that, to date, has been lacking, leading to a nearly five-year policy stasis since the FCC initially proposed sharing for this band in 2013. This paper uses radio frequency (RF) engineering and computer science techniques to provide the necessary technical foundation for spectrum sharing policy in 5.9 GHz. Specifically, we use lab-based RF measurements as the foundation for simulation of real-world system interaction between Wi-Fi and DSRC, with a particular focus on how the presence of adjacent channel Wi-Fi impacts the performance of the crash-avoidance safety application of DSRC according to the leading strategy for sharing 5.9 GHz, known as the "re-channelization" plan, which is described further in this paper. Lab-based RF measurements provide actual DSRC performance impacts at different inter-system configurations, which are then anchored to a network and road traffic simulation that captures the behavior of Wi-Fi and DSRC in an actual cityscape with observed vehicular volume and movement patterns. DSRC system architects define its key performance indicator as packet error rate (PER). In this paper we build on PER and define a risk-informed metric called safety alert failure rate (SAFR) which focuses on the failure rate for packets that have an impact on safety outcomes in real world traffic scenarios, consistent with risk-informed spectrum policy analysis. We use an open-source vehicular traffic simulator known as Veins (Vehicles in Network Simulation) to observe the DSRC SAFR for over 20,000 vehicles traveling during rush hour in Bologna, Italy, with the presence of Wi-Fi interference. This approach is data-intensive and enables calculation of SAFR in scenarios where vehicular traffic makes collisions probable – which is the inter-system interaction of greatest concern for spectrum policymakers. In so doing, we provide a path forward for policy in this important frequency band. We observe that DSRC SAFR is not impacted by the presence of Wi-Fi traffic in the adjacent channel. We also observe that the SAFR is above the levels deemed by DSRC architects to be harmful to system performance in the majority of locations simulated; however, since this result is not due to Wi-Fi interference, we conclude that it is most likely attributed to DSRC system instability in high vehicular traffic scenarios. We perform a number of checks to demonstrate the robustness of these conclusions. Our results therefore suggest that regulators can successfully enable Wi-Fi use of the 5.9 GHz band without impacting DSRC safety efficacy. This finding provides the necessary technical foundation for policy in an important frequency band, and provides a valuable case study for rigorous analysis in spectrum sharing scenarios more generally.
Keywords: TPRC, spectrum, spectrum sharing, Wi-Fi, DSRC, 5 GHz, FCC
Suggested Citation: Suggested Citation