Wireless technologies are growing unprecedentedly with the advent and increasing popularity of wireless services worldwide. With the advancement in technology, profound techniques can potentially improve the performance of wireless networks. Besides, the advancement of artificial intelligence (AI) enables systems to make intelligent decisions, automation, data analysis, insights, predictive capabilities, learning, and adaptation. A sophisticated AI will be required for next-generation wireless networks to automate information delivery between smart applications simultaneously. AI technologies, such as machines and deep learning techniques, have attained tremendous success in many applications in recent years. Hances, researchers in academia and industry have turned their attention to the advanced development of AI-enabled wireless networks. This paper comprehensively surveys AI technologies for different wireless networks with various applications. Moreover, we present various AI-enabled applications that exploit the power of AI to enable the desired evolution of wireless networks. Besides, the challenges of unsolved research in this area, which represent the future research trends of AI-enabled wireless networks, are discussed in detail. We provide several suggestions and solutions that help wireless networks be more intelligent and sophisticated to handle complicated problems. In summary, this paper can help researchers deeply understand the up-to-the-minute wireless network designs based on AI technologies and identify interesting unsolved issues to be pursued in their research in a fast way.

This paper presents millimeter wave (mmWave) measurements in an indoor environment. The high demands for the future applications in the 5G system require more capacity. In the microwave band below 6 GHz, most of the available bands are occupied; hence, the microwave band above 6 GHz and mmWave band can be used for the 5G system to cover the bandwidth required for all 5G applications. In this paper, the propagation characteristics at three different bands above 6 GHz (19, 28, and 38 GHz) are investigated in an indoor corridor environment for line of sight (LOS) and non‐LOS (NLOS) scenarios. Five different path loss models are studied for this environment, namely, close‐in (CI) free space path loss, floating‐intercept (FI), frequency attenuation (FA) path loss, alpha‐beta‐gamma (ABG), and close‐in free space reference distance with frequency weighting (CIF) models. Important statistical properties, such as power delay profile (PDP), root mean square (RMS) delay spread, and azimuth angle spread, are obtained and compared for different bands. The results for the path loss model found that the path loss exponent (PLE) and line slope values for all models are less than the free space path loss exponent of 2. The RMS delay spread for all bands is low for the LOS scenario, and only the directed path is contributed in some spatial locations. For the NLOS scenario, the angle of arrival (AOA) is extensively investigated, and the results indicated that the channel propagation for 5G using high directional antenna should be used in the beamforming technique to receive the signal and collect all multipath components from different angles in a particular mobile location.

This paper considers the integration of channel decoding with fusion based decision, for coded collaborative spectrum sensing (CSS) employing local Neyman-Pearson (NP) testing at each sensor. We derive a belief-propagation (BP) algorithm for joint channel decoding and decision fusion (JCDDF), based on a factor graph model for coded CSS schemes. Using the Lloyd-Max method, we also propose a new methodology for the local sensor to quantize its observation. The design of the quantizer embeds the binary NP test outcome in the quantization bits. Using the JCDDF algorithm, we show that coded CSS paired even with a short (8,4) extended Hamming code outperforms not only uncoded CSS, but also schemes where channel decoding and decision fusion are executed separately. Then, we consider the design of good channel codes for such CSS schemes. We demonstrate that the JCDDF algorithm employing unequal error protection (UEP) coding improves performance and outperforms equal error protection coding. Furthermore, we present a simple code search algorithm for identifying short UEP codes. Using such UEP codes, we finally show that a performance improvement over uncoded CSS can be attained also without bandwidth expansion using higher order modulations.

We investigate the use of rate-compatible lowdensity parity-check (RC-LDPC) codes as part of a soft decodeand- forward (SDF) protocol over the half-duplex relay channel. We propose a new methodology to design the degree distribution of the RC-LDPC codes with a lower triangular parity-check matrix, enabling the additional parity bits to be linearly and systematically encoded at the relay. Our proposed methodology introduces the concept of a K-layer doping matrix to represent the structure of a lower triangular parity-check matrix. As a result of our methodology, the asymptotic performance of RC-LDPC codes can be analyzed and predicted using the multi-edge-type density evolution. Then, we derive the soft-re-encoding of the additional parity symbols at the relay using our designed RCLDPC codes. Moreover, we propose a novel method, which we refer to as soft fading, to compute the log-likelihood ratio (LLR) of the received signal at the destination for the SDF protocol. We demonstrate that our proposed soft fading method outperforms the best known method in the literature by up to 0.7 dB in terms of BER performance. Finally, we derive a new bound for the power multiplication factor at the relay, which limits the amount of soft-errors forwarded by the relay to the destination. The BER performance of our new RC-LDPC codes improves significantly once the power multiplication factor at the relay satisfies this bound.

In relay networks, we may not possess the ability to tune all system parameters in order to achieve the maximum achievable rate promised by theoretical analysis. This paper investigates the pragmatic issue of sub-optimal relay networks, where we can only optimize either on the time allocation, or on the power allocation. Our study concludes that optimizing on the power allocation, or on the time allocation, can achieve more than 95% of the relaying gain relative to the optimization of both time and power simultaneously. To produce the result, we derive the closed-form expression of the optimum power allocation for the source and relay that obtains the achievable rate of the half-duplex relay channel with fixed (equal) time allocation. We also derive the closed-form expression of the optimum time allocation between the source and relay transmission that obtains the achievable rate of the half-duplex relay channel with fixed power allocation. We demonstrate that for small SNR, where relaying is most advantageous, almost all relaying gain can be achieved by only optimizing the power allocation. Conversely, optimizing the time allocation alone is sufficient to achieve most of the relaying gain when the system's SNR is large. This result is important for pragmatic designs of emerging relay communication systems.

This paper studies the problem of determining the optimum degree distribution for distributed LDPC codes in two-way relay channels. Based on the framework of multi-edge type (MET) LDPC codes, we propose a methodology to asymptotically optimize the code's ensemble when different segments within the distributed codeword have been transmitted through different channels and experience different SNRs. An average noise threshold is formulated to compute the convergence threshold of the distributed LDPC codes under density evolution and acts as the performance gap between the optimized distributed codes and the theoretical limit. We demonstrate that the optimized distributed LDPC code using our proposed method performs asymptotically within a fraction of a dB away from the theoretical limit.

This paper proposes a new soft decode-and-forward (SDF) protocol using LDPC codes in the half-duplex relay channels. In order for the encoding of the additional parity-check symbols at the relay to be linear and systematic, we introduce a structured rate-compatible (RC) LDPC code. We then develop the soft-decoding and soft-re-encoding algorithms for the proposed RC-LDPC code, which allows the relay to forward soft messages to the destination when the relay fails to decode the source's message. Furthermore, we propose a new method, which we refer to as soft fading, to compute the log-likelihood ratio (LLR) of the received signal at the destination for the SDF protocol. We show that our proposed method performs better when compared to a previous reported method in literature.

We consider the design of bilayer-expurgated low density parity-check (BE-LDPC) codes as part of a decode and-forward protocol for use over the full-duplex relay channel. A new ensemble of codes, termed multi-edge-type bilayer expurgated LDPC (MET-BE-LDPC) codes, is introduced where the BE-LDPC code design problem is transformed into the problem of optimizing the multinomials of a multi-edge-type LDPC code. We propose two design strategies for optimizing MET-BE-LDPC codes; the bilayer approach is preferred when the difference in SNR between the source-to-relay and the source to-destination channels is small, while the bilayer approach with intermediate rates is preferred when this difference is large. In both proposed design strategies multi-edge-type density evolution is used for code optimization. The resulting MET-BE-LDPC codes exhibit improved threshold and bit-error-rate performance as compared to previously reported bilayer LDPC codes.

We investigate LDPC code design for a multi-source single-relay system, with uniform phase-fading Gaussian channels. We specifically consider the asymmetric channels for multiple sources, where the channel condition for each source in the system is different. We focus on LDPC code design when network coding (NC) at the relay is utilized. For the asymmetric sources, we firstly introduce a binary field rate splitting theorem which is used to discover an appropriate NC scheme at the relay. This NC scheme is then used to determine the achievable rates of each source and the whole system. These steps assist us in the development of the main contribution of our work, namely, network coded multi-edge type LDPC (NCMET-LDPC) code design. Extrinsic mutual information transfer (EXIT) chart analysis is utilized to optimize the code profiles. Our results demonstrate two key points. (1) From the whole system point of view, our NCMET-LDPC codes achieve better error performance than that of LDPC codes designed for the system without NC. (2) As a consequence of the binary field rate-splitting theorem, our NCMET-LDPC codes also guarantee better error performance of each asymmetric source. The improvement in error performance is typically about 0.3 dB relative to a system without NC.

This paper presents a new bilayer-expurgated LDPC (BE-LDPC) code design for decode-and-forward relay system. The code design is based on the multi-edge type LDPC (MET-LDPC) code structure. The lower and upper graphs' edges of the BE-LDPC codes are represented via different edge types. We call the codes as multi-edge type bilayer-expurgated LDPC (MET-BE-LDPC) codes. We derive the relationship between the lower and the overall graphs of the MET-BE-LDPC codes using the multi-edge type multinomials. This relationship acts as a constraint in optimizing the MET-BE-LDPC codes under the density evolution. We apply the differential evolution algorithm to search for the best MET-BE-LDPC codes. Applying the proposed methodology, two MET-BE-LDPC codes are designed. The codes perform asymptotically within 0.08474-0.5087 dB from the theoretical limits. Finally, we demonstrate that the proposed methodology can be used to design the BE-LDPC codes at a larger source-to-relay and source-to-destination SNR differences.

This paper presents an improved LDPC code design for decode-and-forward relay system. Our design is based on the recent work of bilayer LDPC code structures.We introduce an irregular check node degree distribution for the lower and hyper (overall) bilayer LDPC code Tanner graphs. We derive the exact relationship between the lower and hyper graphs in terms of degree distributions (node perspective). This relationship acts as constraint to perform the density evolution. We apply the differential evolution algorithm to search for the best degree distribution for both bilayer-lengthened and bilayer-expurgated LDPC codes. We show that the performance gap of the codes is within 0.07595dB and 0.28407dB from the theoretical limits.