Mobile edge caching (MEC) and device to device (D2D) communications are two potential technologies to resolve traffic overload problems in internet of things (IoT). Previous works usually investigate them separately with MEC for traffic offloading and D2D for information transmission. In this paper, a joint framework consisting of MEC and cache-enabled D2D communications is proposed to minimize the energy cost of systematic traffic transmission, where file popularity and user preference are the critical criteria for small base stations (SBSs) and user devices, respectively. Under this framework, we propose a novel caching strategy where Markov decision process (MDP) is applied to model the requesting behaviours. A novel scheme based on reinforcement learning (RL) is proposed to reveal the popularity of files as well as users' preference. In particular, Q-learning (QL) algorithm and deep Q-network (DQN) algorithm are respectively applied to user devices and SBS due to different complexities of status. To save the energy cost of systematic traffic transmission, users acquire partial traffic through D2D communications based on the cached contents and user distribution. Taking the memory limits, D2D available files and status changing into consideration, the proposed RL algorithm enables user devices and SBS to prefetch the optimal files while learning, which can reduce the energy cost significantly. Simulation results demonstrate the superior energy saving performance of the proposed RL-based algorithm over other existing methods under various conditions.

No abstract

In this paper, we propose a compressive channel estimation techniques for IRS-assisted mmWave multi-input and multi-output (MIMO) system. To reduce the training overhead, the inherent sparsity in mmWave channels is exploited. By utilizing the properties of Kronecker products, IRS-assisted mmWave channel estimation are converted into a sparse signal recovery problem, which involves two competing cost function terms (measurement error and a sparsity term). Existing sparse recovery algorithms solve the combined contradictory objectives function using a regularization parameter, which leads to a suboptimal solution. To address this concern, a hybrid multiobjective evolutionary paradigm is developed to solve the sparse recovery problem, which can overcome the difficulty in the choice of regularization parameter value. Simulation results show that under a wide range of simulation settings, the proposed algorithm achieves competitive error performance compared to existing channel estimation algorithms.

The system capacity can be remarkably enhanced with the help of intelligent reflecting surface (IRS) which has been recognized as a advanced breaking point for the beyond fifth-generation (B5G) communications. However, the accuracy of IRS channel estimation restricts the potential of IRS-assisted multiple input multiple output (MIMO) systems. Especially, for the resource-limited indoor applications which typically contains lots of parameters estimation calculation and is limited by the rare pilots, the practical applications encountered severe obstacles. Former works takes the advantages of mathematical based statistical approaches to associate the optimization issue, but the increasing of scatterers number reduces the practicality of statistical approaches in more complex situations. To obtain the accurate estimation of indoor channels with appropriate piloting overhead, an offset learning (OL)-based neural network method is proposed. The proposed estimation method can trace the channel state information (CSI) dynamically with non-priori information, which get rid of the IRS-assisted channel structure as well as indoor statistics. Moreover, a convolution neural network (CNN)-based inversion is investigated. The CNN, which owns powerful information extraction capability, is deployed to estimate the offset, it works as a offset estimation operator. Numerical results show that the proposed OL-based estimator can achieve more accurate indoor CSI with a lower complexity as compared to the benchmark schemes.

The combination of unmanned aerial vehicles (UAVs) and millimeter wave (mmWave) multiple-input multipleout (MIMO) systems is considered as a key enabling technology for 5G networks, as it provides high data rate aerial links. However, establishing UAV-enabled mmWave MIMO communication is challenging duo to the high hardware cost in terms of radio frequency (RF) chains. As explored in this work, the joint optimization of the UAV's altitude and hybrid beamspace precoding is proposed for the UAV-enabled multiuser MIMO system, in which the DLA is exploited to decrease the amount of the RF chain. In the proposed scheme, the optimization problem is formulated as a minimum weighted mean squared error (MWMSE) method. Then an efficient algorithm with the penalty dual decomposition is proposed to optimize the altitude of UAV, selection of the beam and the method of digital precoding jointly. Simulation results confirm the superior performance of the explored scheme and perform close to full-digital beamforming in terms of achievable spectral efficiency.