Ishtiaq Ahmad, Ph.D.

We address the problem of a coordination among machine learning tools solving different problems of radio resource management. We focus on energy efficient device-to-device (D2D) communication in a scenario with many devices communicating adhoc directly with each other. In such scenario, deep neural network (DNN) is a convenient tool to predict the channel quality among devices and to control the transmission power. However, addressing both problems by a single DNN is not suitable due to a dependency of the power control on the predicted channel quality. Similarly, a simple concatenation of two DNNs leads to a high cumulative learning error and an inevitable performance degradation. Hence, we propose a mutual coordination of the DNNs for channel quality prediction and for power control via a feedback and a knowledge transfer to mitigate the accumulation of errors in individual learned models. The proposed coordination improves the energy efficiency by 10-69% compared to state-of-the-art works and reduces the training time of DNNs more than 3.5-times compared to DNNs without coordination.

User applications based on the deep neural networks (DNNs), such as object or anomaly detection, image recognition, or language processing, running on computation- and energy-constrained user equipment (UE) can be partially or fully processed in the edge computing servers to reduce a processing time and save an energy in the UE. To further reduce the processing time and the UE's energy consumption, DNN with multiple exit points can be incorporated. In this article, we address the problem of the decision on whether the computation should be offloaded from the UE to the edge computing server or processed locally by the UE and we solve this problem jointly and "on-the-fly" together with DNN exit selection. Since the formulated problem is very complex, we exploit the deep deterministic policy gradient for the exit selection and the offloading decisions (labeled DDPG-EOD) for the DNN-based applications. To this end, we first convert the problem into the Markov decision process, and then, we employ an end-to-end learning via DDPG with the actor-critic architecture. Second, we use a knowledge distillation-based technique to efficiently select the DNN's exit to minimize the delay and energy consumption. Simulation results show that the proposal is highly scalable, converges very quickly, and surpasses the best performing state-of-the-art approach by up to 120% and 100% in terms of the overall DNN processing delay and the energy consumption, respectively.

The co-channel interference for mobile users (MUs) of a public safety network (PSN) in the co-existence of heterogeneous networks such as unmanned aerial vehicles (UAVs) and LTE-based railway networks (LRNs) needs a thorough investigation, where UAVs are deployed as mobile base stations (BSs) for cell-edge coverage enhancement. Moreover, the LRN is employed for the train, and its control signal demands high reliability and low latency. It is necessary to provide higher priority to LRN users when allocating resources from shared radio access channels (RACs). By considering both sharing and non-sharing of RACs, co-channel interference was analyzed in the downlink network of the PSN, UAV, and LRN. By offloading more PSN MUs to the LRN or UAVs, the resource utilization of the LRN and UAV BSs was enhanced. In this paper, we aimed to adopt deep-learning (DL)-based enhanced inter-cell interference coordination (eICIC) and further enhanced ICIC (FeICIC) strategies to deal with the interference from the PSN to the LRN and UAVs. Moreover, a DL-based coordinated multipoint (CoMP) for coordinated scheduling technique was utilized along with FeICIC and eICIC to enhance the performance of PSN MUs. In the simulation results, the performance of DL-based interference management was compared with simple eICI, FeICIC, and coordinated scheduling CoMP. The DL-based FeICIC and CoMP for coordinated scheduling performed best with shared RACs.

Unmanned aerial vehicles (UAVs) equipped with reconfigurable intelligent surfaces (RISs) have emerged as a promising technology for numerous applications involving aerial networks. However, the UAV-RIS concept faces challenges related to the deployment of the UAV-RIS, especially in cases, where UAV-RIS is combined with emerging technologies, such as beamforming, sensitive to propagation channel variation. In this article, we first overview various use-cases of UAV-RIS beam-forming considering practical scenarios. Aiming to improve the performance of communication channels, we propose a machine learning-based beamforming policy for UAV-RIS by employing prioritized experience replay (PER) based deep Q-Network (DQN). Compared to traditional approaches, the proposed PER DQN-based beamforming for UAV-RIS communication provides significant enhancements in performance. Finally, we highlight some potential directions for future research.