Book Description

This dissertation considers the impact of blockage and mobility on collaborative sensing and millimeter wave (mmWave) based communication networks. We first study the character of interference and MAC performance in dense indoor mmWave wearable networks. Using simple stochastic geometric models for propagation in mmWave bands, we quantify the number of strong interferers as seen by a typical receiver and show that it is limited due to blockage. We propose a model to evaluate the performance of current MAC designs using clustering and hierarchical scheduling. Our results show that the MAC overheads are scalable, i.e., the performance optimal cluster size does not grow with user density in dense scenarios. Furthermore, we show that at high densities the per user throughput is eventually constant. Next we consider the impact of blockage mobility on MAC overheads and performance in such networks. We propose a stochastic geometric model to capture the temporal dynamics of strong interfering channels resulting from blocking in networks comprising both fixed and mobile blockages. Based on our analysis, we derive the rate of change in channels' states, i.e., Line-of-Sight (LOS) and Non-LOS (NLOS), and estimate the signaling overheads resulting from blockage mobility. We argue that while the overheads to track the interference environment may in fact be limited, MAC protocols will most likely be better off not coordinating with distant and/or mobile nodes. We then move on to another area where obstructions have a major impact, i.e., collaborative sensing for automated driving applications. Both the sensing and communication for collaborative sensing may be subject to obstructions (blockages) in such a collaborative setting. We introduce new models for vehicular collaborative sensing and networking under obstructions and evaluate how "performance" scales. In particular, we quantify the coverage and reliability gains obtained by collaborative sensing as a function of the penetration of collaborative vehicles. We further evaluate the associated communication loads in terms of vehicle-to-vehicle (V2V) and/or vehicle-to-infrastructure (V2I) capacity requirements and how these depend on penetration. Sensing by a single vehicle can be greatly limited by obstructing neighboring vehicles and objects, while collaborative sensing is shown to greatly improve sensing performance, e.g., improves coverage from 20% to 80% with a 20% penetration. Furthermore, the volume of sensor data a vehicle generates and needs to share for collaborative sensing does not necessarily increase with the density of objects. In scenarios with limited penetration and enhanced reliability requirements, infrastructure can be used to sense the environment and relay data. Once penetration is high enough, vehicular collaborative sensing provides good coverage and V2V connectivity. Data traffic can be effectively 'offloaded' to V2V network, making V2I resources available to support other services. Finally we present a more detailed evaluation of the performance of collaborative sensing assisted by sensing capable infrastructure, including Road Side Units (RSUs) and sensors on cellular infrastructure. We compare the performance of different infrastructure and deployment schemes in terms of collaborative sensing coverage. Unless deployed along roads, cellular based sensors off the roads are more obstructed and RSUs deployed at intersections and at even spacings appear more desirable. Simulation results show that RSUs see fewer environmental obstructions when placed higher than vehicles and can benefit from temporal diversity in sensing. Although RSUs have good sensing coverage, in order to communicate with the relevant vehicle, they will require relatively high communication range, rate and reliability. Even if RSUs provide complete coverage of the roads, to increase reliability of sensing, e.g., redundancy in sensing, collaboration amongst sensing capable vehicles may still be desirable.