Spectral Clustering Based Millimeter-Wave Beam Management in V2I Networks

The efficient beam management of millimeter-wave(mmWave) is a crucial component in Vehicle to Internet(V2I) networks and has attracted widespread attention in the automotive industry and academia. This paper uses the cellular automata theory to establish a detailed one-way two-lane vehicle mmWave communication network model. Then the beam resource allocation optimization problem is decomposed into two sub-problems: sub-network division and resource allocation in sub-networks. Inspired by the spectral clustering algorithm, this paper focuses on the characteristics of the V2I network and transforms the problem of resource allocation into a problem of graph division. Using a graph-based spectral clustering algorithm, the decomposition of sub-networks is relized. The simulation results show that the proposed algorithm can achieve a significant improvement in spectrum utilization and system throughput.