Search

Facing the ever-increasing demand for high quality video streaming and the dramatic variations of wireless network environment, adaptive bitrate video streaming (ABS) emerges as a prominent video quality adaptation technique in improving users’ quality of experience (QoE). Reinforcement learning (RL)-based ABS and network-assisted ABS can improve users’ QoE through resource utilization efficiency enhancement. However, limited network capacity is still the bottleneck of QoE improvement. Considering the potential capacity gain of non-orthogonal multiple access (NOMA), we study network-assisted ABS in NOMA networks. Based on a synchronous video streaming model and the knowledge of channel state information for next segment, a joint optimization problem of resource allocation and bitrate adaptation is formulated to maximize users’ QoE and minimize energy consumption. The original problem is NP-hard and intractable due to the interference coordination problem introduced by NOMA and the multi-slot sum rate constraint. We decompose the original problem into the subproblem of joint subchannel assignment and power allocation (JSAPA), and the subproblem of bitrate adaptation (BA) with Lagrange relaxation method. A joint resource allocation and bitrate adaptation (JRABA) algorithm is further proposed to solve the JSAPA and BA subproblems iteratively. Specifically, the JSAPA subproblem is solved by an extended Gale-Shapley matching algorithm and differ of convex (DC) programming method iteratively. For the BA subproblem, appropriate video bitrate is explored with subgradient method. Numerical results validate the effectiveness of the proposed algorithm and reveal that network-assisted ABS in NOMA networks benefits users in achieving higher QoE with less energy consumption, especially when radio resources are insufficient or with large segment size. [ABSTRACT FROM AUTHOR]

Unmanned Aerial Vehicle (UAV)-enabled communication platform is recognized as a promising way for secure transmission in the air-to-air and air-to-ground (A2G) networks. However, the jitter feature of UAV platform due to the absence of fixed infrastructure, the inherent randomness of airflow, and the body vibration introduces non-negligible impact on establishing robust and energy-efficient secure communication links. In this paper, we investigate energy-saving and secure transmission in a downlink A2G wiretap system considering the impact of UAV jitter. Particularly, in order to minimize the total transmission power of UAV-mounted base station, a joint beamforming optimization of confidential signal and artificial noise signal is formulated with the constraints for secrecy performance of the worst scenario. Specifically, considering the impact of UAV jittering on the antenna array response, the worst scenario indicates the A2G wiretap system with the minimum legitimate data rate and the maximum eavesdropping data rate. The non-convex problem is then reformulated with linear approximation for the channel variations and linear matrix inequality alternation for the constraints. Then, semidefinite programming relaxation is adopted to obtain the optimal solutions for the reformulated problem. Furthermore, the problem and solution are extended to the multiuser downlink wiretap system. Finally, numerical results are provided to demonstrate the impact of key system parameters on power saving under the secure transmission constraints and the UAV jitter. [ABSTRACT FROM AUTHOR]

Presents correction to author biographical information. [ABSTRACT FROM AUTHOR]

Head motion occurring during brain positron emission tomography images acquisition leads to a decrease in image quality and induces quantification errors. We have previously introduced a Deep Learning Head Motion Correction (DL-HMC) method based on supervised learning of gold-standard Polaris Vicra motion tracking device and showed the potential of this method. In this study, we upgrade our network to a multi-task architecture in order to include image appearance prediction in the learning process. This multi-task Deep Learning Head Motion Correction (mtDL-HMC) model was trained on 21 subjects and showed enhanced motion prediction performance compared to our previous DL-HMC method on both quantitative and qualitative results for 5 testing subjects. We also evaluate the trustworthiness of network predictions by performing Monte Carlo Dropout at inference on testing subjects. We discard the data associated with a great motion prediction uncertainty and show that this does not harm the quality of reconstructed images, and can even improve it.