Millimeter wave (mmWave) communication is one viable solution to support Gbps sensor data sharing in vehicular networks. The use of large antenna arrays at mmWave and high mobility in vehicular communication make it challenging to design fast beam alignment solutions. In this paper, we propose a novel framework that learns the channel angle-of-departure (AoD) statistics at a base station (BS) and uses this information to efficiently acquire channel measurements. Our framework integrates online learning for compressive sensing (CS) codebook learning and the optimized codebook is used for CS-based beam alignment. We formulate a CS matrix optimization problem based on the AoD statistics available at the BS. Furthermore, based on the CS channel measurements, we develop techniques to update and learn such channel AoD statistics at the BS. We use the upper confidence bound (UCB) algorithm to learn the AoD statistics and the CS matrix. Numerical results show that the CS matrix in the proposed framework provides faster beam alignment than standard CS matrix designs. Simulation results indicate that the proposed beam training technique can reduce overhead by 80% compared to exhaustive beam search, and 70% compared to standard CS solutions that do not exploit any AoD statistics.

Millimeter wave channels exhibit structure that allows beam alignment with fewer channel measurements than exhaustive beam search. From a compressed sensing (CS) perspective, the received channel measurements are usually obtained by multiplying a CS matrix with a sparse representation of the channel matrix. Due to the constraints imposed by analog frontends, designing CS matrices that efficiently exploit the channel structure is challenging. In this paper, we propose an end-to-end deep learning technique to design a structured CS matrix that is well suited to the underlying channel distribution, leveraging both sparsity and the particular spatial structure that appears in vehicular channels. The channel measurements acquired with the designed CS matrix are then used to predict the best beam for link configuration. Simulation results for vehicular communication channels indicate that our deep learning-based approach achieves better beam alignment than standard CS techniques that use the random phase shift-based design.

Millimeter wave (mmWave) communication offers Gbps data transmission, which can support massive data sharing in vehicle-to-infrastructure (V2I) networks. In this paper, we analyze the blockage effects among different vehicles and coverage probability of a typical receiver, considering cross street BSs near urban intersections in a multi-lane mmWave vehicular network. First, a three-dimensional model of blockage among vehicles on different lanes is considered. Second, we compute the coverage probability considering the interference of cross street base stations. Incorporating the blockage effects, we derive an exact and semi closed-form expression of the cumulative distribution density (CDF) of the association link path gain. Then, a tight approximation of the coverage probability is computed. We provide numerical results to verify the accuracy of the analytic results. We demonstrate the effects of blockage and the cross street interference. Also, we compare coverage probability with different BSs intensities under various street settings. It is shown that in multi-lane V2I networks, blockage among vehicles is not significant. Also, deploying more BSs does not increase coverage probability efficiently in ultra-dense streets.

Millimeter wave (mmWave) networks are sensitive to blockages due to buildings in urban areas. This is critical for vehicle-to-infrastructure networks which are cellular networks designed to support emerging vehicular applications. Motivated by measurement and ray tracing results in urban microcells, instead of characterizing the pathloss by Euclidean distance, we calculate it by the weighted sum of segment length along the propagation path, i.e., Manhattan distance, and a certain corner loss at the intersections along the path. We analyze network performance by modeling the urban microcell network by a Manhattan Poisson line process. Our results show significant differences between Manhattan and Euclidean distance- based pathloss models. Assuming the receiver is associated with the base station (BS) with the smallest pathloss, we derive closed-form expression of the distribution of the associated link pathloss. We obtain the coverage probability and reveal the impacts of interference from the LOS and NLOS BSs. It is shown that in this scenario the interference from a NLOS parallel street is negligible.

This paper presents an analytical performance investigation of both beamforming (BF) and interference cancellation (IC) strategies for a device-to-device (D2D) communication system underlaying a cellular network with an M-antenna base station (BS). We first derive new closed-form expressions for the ergodic achievable rate for BF and IC precoding strategies with quantized channel state information (CSI), as well as, perfect CSI. Then, novel lower and upper bounds are derived which apply for an arbitrary number of antennas and are shown to be sufficiently tight to the Monte-Carlo results. Based on these results, we examine in detail three important special cases including: high signal-to-noise ratio (SNR), weak interference between cellular link and D2D link, and BS equipped with a large number of antennas. We also derive asymptotic expressions for the ergodic achievable rate for these scenarios. Based on these results, we obtain valuable insights into the impact of the system parameters, such as the number of antennas, SNR and the interference for each link. In particular, we show that an irreducible saturation point exists in the high SNR regime, while the ergodic rate under IC strategy is verified to be always better than that under BF strategy. We also reveal that the ergodic achievable rate under perfect CSI scales as log2M, whilst it reaches a ceiling with quantized CSI.

This paper investigates the antenna selection scheme choosing the antenna with the largest channel gain to the the cellular user for device-to-device (D2D) communication underlaying cellular networks. We derive an exact closed-form expression of the ergodic achievable rate and examine its asymptotic behavior in the high signal-to-noise ratio (SNR) regime. It is demonstrated that the high SNR approximation can be much improved by a higher transmit power ratio between the base station (BS) and the D2D transmitter in the small cell setting where all users are closely located. However, in the macro cell setting in which only the D2D terminals are fairly close, the influence of the transmit power ratio becomes insignificant. In addition, we present upper and lower bounds of the ergodic achievable rate. Based on these results, we illustrate that the D2D communication cannot help in the cellular network for elevating the rate when the transmit SNR at the BS grows high. If the BS SNR is lower, then the D2D communication can effectively increase the ergodic achievable rate. Numerical results are provided to justify the correctness of the expressions and the relevant performance analysis.