Path Selection with Amplify and Forward Relays in mmWave Backhaul Networks

This paper focuses on the problem of path selection with amplify-and-forward (AF) relays for long-range ultra-high-speed millimeter wave (mmWave) backhaul networks in urban environments. Relays are selected between a pair of source and destination nodes to achieve the highest signal-to-noise ratio (SNR) at the destination. We first derive an equation for the end-to-end SNR of a relay path in a setting that approximates the urban mmWave backhaul environment. Based on the derived equation, we transform the maximum throughput relay selection problem to the shortest path problem in graphs. Dijkstra's algorithm can then be used to find maximum throughput relay paths, which however are shown to require a large number of relays. To address this, we propose a dynamic programming algorithm to find a highest throughput path with a given number of hops. Simulation results based on 3-D models of a section of downtown Atlanta show that these algorithms can be combined to find relay paths with a small number of hops and very high throughput.