Search

Machine-type communication (MTC) is one of the main enabling technologies to support various applications with diverse quality of service (QoS) requirements. Finite blocklength transmission has great potential in meeting the strict delay requirements of delay-sensitive MTC devices (MTCDs), while also causing loss of network capacity due to the decoding error probability. Aiming at this problem, we introduce uplink non-orthogonal multiple access (NOMA) and buffer-aided relaying to assist the finite blocklength transmission with delay requirements for improving the achievable effective capacity (EC), which is defined as the maximum short-packet constant arrival rate under specific statistical QoS constraints. To solve the EC maximization problem, we derive the closed-form expression of time allocation coefficient. Then we establish a concave lower bound of EC using successive convex approximation (SCA) for power allocation of MTCDs, and formulate a non-cooperative game based distributed power allocation algorithm for relay. Furthermore, a joint time and power allocation (JTPA) algorithm is proposed to implement joint resource allocation. Simulation results show that under finite blocklength and statistical QoS constraints, adopting buffer-aided relaying can improve EC by 41.82% compared with no-buffer relaying. Moreover, the achievable EC of JTPA algorithm is only 3.12% lower than that of exhaustive search while reducing complexity. [ABSTRACT FROM AUTHOR]

This article investigates the displacement damage on AlGaN/GaN high-electron-mobility transistors (HEMTs) using the 30-MeV fluorine ions. The electrical degradation and the defects in HEMTs were examined before and after irradiation. Three defect states are detected in a deep-level transient analysis system (DLTS), including a broad surface state peak ($N_{1}$) and two bulk defect peaks ($H_{1}$ and $H_{2}$). The energy levels of these two defects are $E_{\mathrm {C}} 0.63$ eV ($H_{2}$) and $E_{\mathrm {C}} 0.91$ eV ($H_{1}$), respectively. There is a linear relationship between the concentration of $H_{1}$ and the fluorine ion fluence. The radiation damage and electrical performance degradation of AlGaN/GaN HEMTs are simulated by combining with simulation software of different methods. The calculated results are in agreement with the experimental data. The performance degradation caused by heavy ion of the device is mainly caused by the $H_{1}$ defect. The effect of defects with shallow energy level is shielded by the Fermi-level pinning in deeper energy level. Moreover, the defect distribution of GaN in AlGaN/GaN heterojunction plays a major role. It is found that the method of combining experimental data with multiscale simulation is an effective way to predict the radiation damage of devices. [ABSTRACT FROM AUTHOR]

Massive Machine-Type Communication (mMTC) is proposed by ITU to provide a large scale of connectivity services for billions of machine-type communications devices (MTCDs) via cellular networks. However, MTCDs are generally energy-constrained, security capability limited and short-packet transmission dominated. It has great challenges for massive MTCDs to perform short-packet transmission simultaneously while guaranteeing the confidentiality and energy efficiency of the information transmission. In this paper, we investigate the energy-efficient secure short-packet transmission of Non-Orthogonal Multiple Access (NOMA) assisted mMTC networks, in which MTCDs aim to transmit secrecy messages to the destination base station via trusted relays with a passive eavesdropper present. To realize secure, reliable and energy-efficient transmission, we maximize the secrecy energy efficiency (SEE) by implementing joint relay and power level selection. Furthermore, we formulate a decentralized stateless Q-learning-based multi-agent reinforcement learning (MARL) algorithm for the dynamic and complex mMTC systems. The simulation results show that adopting the proposed decentralized MARL algorithm and relaying with NOMA can improve the performance of SEE while reducing information exchange overhead. [ABSTRACT FROM AUTHOR]

In this article, we propose a novel framework on the state estimation for a class of positive real infinite-dimensional systems, which suffer from unknown actuator and sensor faults and uncertain exogenous disturbance. Four cases of fault occurrence are taken into account: single multiplicative actuator fault, single multiplicative sensor fault, simultaneous actuator and sensor faults, and single actuator fault with a dead-zone nonlinearity. We develop new adaptive laws for the estimation of fault parameters and embed the proposed laws into the infinite dimensional state observer design. Moreover, by considering the actuator fault with a dead-zone nonlinearity, an infinite dimensional observer associated with coupled parameter update laws is developed to help to complete the fault accommodation. This is achieved by reconfiguring the actuator input based on the estimated information of parameters in the actuator fault map with a dead-zone nonlinearity. [ABSTRACT FROM AUTHOR]