Data availability is a crucial security feature in the perception layer of the Internet of Things (IoT) and Machine-to-Machine (M2M) communication devices. It enables devices through offering crucial data and services during communication failures. The IoT network is mostly equipped with cloned backups of data, power backups, and additional hardware for data and power to achieve the availability of crucial data and services, which is very costly. In this regard, we presented a novel anti-communication failure approach to limit crucial sensor data loss and enforced data loss attacks. The results show that the approach achieved the data availability features with no additional hardware cost.

The energy harvesting in a millimeter wave (MMW) region received great interest in wireless communication applications as it provides large bandwidth ranging from 30 GHz to 300 GHz. MMW communications are also an important technology for fifth-generation (5G) and beyond 5G cellular networks. The large antenna arrays implemented by the MMW system have significantly increased user performance. The performance of the MMW system depends on the gain of the antenna arrays. Therefore, a suitable antenna pattern model is crucial to predict the MMW system performance before the real base station (BS) implementation. This paper analyses the energy coverage probability of MMW wireless power transfer by leveraging the concept of stochastic geometry. The locations of BS and users are modeled following Poisson point processes (PPPs) and then the effect of the antenna model is considered in the analysis. This paper also proposes a new simplified antenna model, known as modified sectored (M-sectored) antenna. The results show that the energy coverage probability of the proposed antenna matches the actual antenna gain pattern. In addition, the effects of the BS density and blockage parameter are also observed.

The ability of millimeter-wave (mmWave) to deliver gigabit throughput has led to its widespread adoption in Fifth Generation (5G) networks. However, mmWave links between base stations and users can be easily blocked by obstacles. In vehicular networks with dynamic environments and mobile users, the mmWave link blockage issue is even more pronounced. In order to preserve the mmWave link in the vehicular network, it is necessary to predict blockages. For blockage prediction, sensor information from Lidar, Radar, and cameras has been considered. Nonetheless, these non-radio frequency (RF) methods necessitate the use of additional equipment and signal processing, which raises the implementation cost and complexity. Existing works also consider the use of base station and user radio frequency signatures to predict blockage. However, mobility of the user has not been taken into account. Intelligent Reflecting Surface (IRS), on the other hand, have been viewed as a promising method for providing an alternate path by reflecting the mmWave signal between the base station and user in order to prevent link blockage. For blockage prediction, the use of the IRS's RF signature has not been considered. Therefore, the research investigates the IRS-assisted blockages prediction in order to determine the future link status in the vehicular environment with respect to user mobility. The proposed solution employs a number of active elements that are randomly distributed on the IRS to obtain the RF signatures and employs machine learning techniques to learn the pre-blockage wireless signatures, which can predict future blockages. The results indicate that the proposed method can predict blockages between a single IRS and a moving user with a greater than 98 percent accuracy one second before they occur.

The 5G network is intended to accommodate a significant quantity of mobile data traffic as well as a great number of wireless connections. It improves cost and power consumption; it offers ultra-low latency and ultra-high dependability to enable new services in a variety of sectors. However, the general public is concerned about the possible health dangers linked with 5G equipment's Radio Frequency (RF) radiation, and numerous localities are actively lobbying to prevent 5G implementation. This research aims to measure the maximum exposure emitted by a 5G base station operating on 3.5 GHz. Considering the transmitted power changes over time with data traffic, the analysis is based on a code selective method. To investigate the impacts of distance and time on the level of RF-EMF radiation, measurements were conducted at two different distances and three different times. The maximum radiation from the base station is determined to be 11.69V/m, which is far less than the accepted limit by the ICNIRP standard.

Demand for Unmanned Aerial Vehicle (UAV) to fly Beyond Visual Line of Sight has been getting intensive for the past few years. To do so, the UAV must have a reliable long-range Command and Control (C2) link. The UAV is also expected to avoid any obstacle autonomously with a reliable Detect and Avoid (DAA) mechanism and the ability to detect nearby aircraft since the sky is occupied by manned aircraft. Lastly, the UAV should not endanger others' life or property. Therefore, this article proposes a hardware and software configuration that utilizes the cellular network, and DAA system with the deployment of a parachute system to achieve a safe BVLOS UAV operation. The quality of cellular networks at multiple altitudes, the C2 link performance, the ability to perform long-range flights, and the parachute system's reliability were investigated. The results indicate the effectiveness of the developed systems as the starting point to allow BVLOS drone operations

Collaborative beamforming (CB) in wireless sensor network (WSN) increases the transmission distance, which allow nodes to transmit data directly to the base station without relying on multi-hop transmissions. The choice of sensor node location directly affects the sidelobe levels of the radiation beampattern, which can affect the energy efficiency of the WSN. Therefore, an Artificial Intelligence algorithm, i.e. the Backtracking Search Algorithm (BSA) has been developed and implemented to optimize the radiation beampattern performance by finding the suitable location of the sensor nodes based on the linear array sensor nodes arrangement. The proposed Modified BSA-based Linear Array (LinearBSA) shows a superior improvement compared to the conventional linear array (LFA) in desired performance, i.e. minimize sidelobe level (SLL) and adaptive main beam angle. Performance on 8-, 12- and 16-node LinearBSA shows an improvement in radiation beampattern performance particularly in SLL minimization in any desired adaptive main beam angle.

Currently, the Internet of Medical Things (IoMT) and smart sensors are widely used. IoMT devices generate valuable and beneficial data for healthcare organizations. Chronic diseases are seriously threatening human health. In this situation, the IoMT provides essential monitoring of the status of patients who have chronic diseases. This paper proposes a new framework using telemedicine techniques for monitoring patients who have chronic diseases but are too far from a hospital. Furthermore, we present a triage algorithm using Random Forest machine learning. The results demonstrate accurate results of 82.2%. The proposed framework successfully predicts the severity of the status of the patients.

The growing demand for broadband access drives the evolution of access networks to fiber-based solutions. School Wi-Fi that is reliable, fast, and secure will empower teachers to maximize their use of the internet in the classroom. As part of its Malaysian Education Blueprint 2013-2025, the Ministry of Education has taken another initiative in order to provide an effective learning experience for Malaysian students. The program aims to give every Malaysian public-school access to high-speed Internet access and a virtual learning environment (VLE). Energy consumption is becoming more important in schools as their network infrastructure expands, not only from an environmental standpoint but also from a capital expenditure perspective. This paper proposes the cost and energy consumption model for fiber-to-the-school (FTTS) network. The developed model was used to assess the energy consumption of the FTTS network. For the cost assessment, there is not much different in the cost variation between different scenarios considered in the analysis. It was also found that subscription cost contribute the highest percentage of 69% in total FTTS cost compared to other network costs. Meanwhile in energy consumption assessment, it is been discovered that the FTTS through scenario with energy saving approach by considering the idle state of the network elements will save approximately 24.33%. The energy saving is improved when considering energy consumption model when the network elements are switched off during inactive hours which is approximately 66.67%.

The next-generation gigabit passive optical network (NG-PON) is a promising expansion and growth solution for access networks. NG-PON aims to connect more people as the fiber-to-the-home (FTTH) concept is implemented. However, immediate attention is required due to threats to its security infrastructure, such as degradation attacks in the presence of transmission protocol (TCP) traffic environments. Throughout the attack, TCP congestion management is activated to manage collisions. This research investigates the relationship between TCP response in the transport layer while enduring attack in the MAC layer. Two types of TCP were evaluated for resistance to the attack. New Reno TCP reduced link consumption by 54%, whereas high-speed TCP (HSTCP) only reduced it by 37%. The capability of TCP's congestion window to regain quickly in HSTCP demonstrates that TCP flavor contributes to sustaining network performance under attack.

For tropical areas, rainfall is the main factor in the attenuation of millimeter wave transmission. This paper considers KLIA Sepang region of Malaysia as an example to measure and collect rainfall rate data via rain sensor, establish a mathematical model for millimeter wave rainfall attenuation in tropical areas, simulate and calculate rainfall attenuation at different frequency ranges of 20-300 GHz through MATLAB software, with emphasis on the analysis of millimeter wave communication attenuation characteristics in the widely used 30-50GHz frequency band. The results show that the greater the rainfall rate, the higher the frequency, and the greater the rainfall attenuation. Moreover, the vertical polarization achieves the lowest attenuation compared circular polarization and horizontal polarization. The attenuation and attenuation rate increase with the increase of frequency and rainfall rate. Specifically, the attenuation value appears linearly in the range of 15dB to 90dB in the rainfall rate of 20-60mm/h for the frequency bands of 30-50GHz.

Voice pathology detection has gained a lot of attention in the last few decades. Furthermore, this field is considered an active topic in the healthcare area. However, most machine learning techniques are proposed to differentiate the healthy voice from the pathological voice only, where there is a lack of identification of a certain voice disease. Therefore, this work presents a method for detecting Dysphonia Disease (DD), which belongs to the pathology detection application. The proposed method uses the Naive Bayes (NB) algorithm as a classifier in order to identify the dysphonia (pathological) class from the healthy (normal) class. In addition, the Mel-Frequency Cepstral Coefficient (MFCC) is used for extracting the acoustic features. The acoustic signals used in this method were gained from the Saarbrucken Voice Database (SVD). Several evaluation measurements have been used to assess the proposed method. The experiment results indicate that the NB classifier obtained an accuracy of 81.48%, 65% sensitivity, a specificity of 91.17%, and a 76.98% G-mean. Further, the precision and F1-score are 81.25% and 72.22%, respectively.

Multi-agent Deep Reinforcement Learning (MADRL) has been applied to a plethora of state-of-the-art applications such as resource allocations and network routing in both centralized and distributed manners. This paper investigates the performance of centralized and distributed MADRL in Dynamic Spectrum Access (DSA). We consider a multichannel wireless network with a shared bandwidth divided into k channels. The objective of the MADRL is to develop a spectrum access strategy by learning in both a centralized and distributed manner. To evaluate the performance of centralized and distributed MADRL, we tackle the spectrum access problem by applying centralized MADRL and distributed MADRL. Experimental results show that distributed MADRL outperforms the centralized MADRL by 15% in collision avoidance and accumulated rewards in DSA.

With the proliferation of connected devices and the increase in Machine Learning, more confidential data is being generated. Traditional Machine Learning, whereby data is sent to a server for training and processing using models, is becoming less suitable due to privacy concerns. Thus, distributed approaches including Federated Learning(FL) are becoming more popular. In the latter approach, the model is sent to the clients, where it is trained using the client's data. The updated model is sent to a server to be aggregated. Federated Learning is expected to be used extensively in wireless networks. Therefore, researchers are interested in optimizing Federated Learning for wireless networks. This paper aims to study the performance of FL in terms of accuracy and amount of data exchanged in a wireless network considering the impact of delay using different datasets. The accuracy of the FL model was found to be reliable when benchmarked to the centralized approach(less than 0.1 difference in accuracy). The data transfer size in the FL system was significantly smaller than in the centralized approach.

Electricity consumption has become a forefront issue of global energy demand management, and one of the biggest contributors to Malaysia's high electricity demand is the residential sector. Hence, user monitoring of energy consumption is critical for global energy efficiency. The aim of this project is to develop a smart energy metering and appliance control using a microcontroller ESP32 and Arduino IDE, provide prediction for bills to allow decision-making based on energy conservation measures using artificial neural network (ANN) model on MATLAB, provide a dashboard to monitor energy consumption, display accumulated bill and control of appliances on Adafruit IO platform, and provide notifications through email when bill exceeds limit using If This Then That (IFTTT) software platform. At about 94% accuracy of bill prediction, the developed system is believed to be able to contribute significantly to an efficient household utility management system.

With CSPs making a lot of investment in 5G infrastructure, generating a return is bound to be a top priority. For 5G network slicing, the ability to service all types of connectivity across common network infrastructure, as well as enable operators to open their networks to B2B2X industries using analytics is key to tapping into the massive projected revenue opportunities. Optimizing Slice allocation/selection using machine learning, Deep Learning, and AI will be key to generate revenue for CSPs for various B2B2X uses cases on basis of type of device, mobility, Latency requirement, etc. For this purpose, we are introducing O-RAN DeepB2B2XSlice Engine. O-RAN DeepB2B2XSlice Engine will perform the following tasks using deep learning and machine learning algorithms.

• Service degradation parameter deduction and optimizing Slice to handle failures in future by deep learning algorithms, improve efficiency and meet SLA performance requirements
• Slice selection optimization based on SLA requirement: DeepB2B2XSlice Engine will recommend optimize Slice RAN resources parameters based on data analysis from historical data

The FBMC-OQAM transceiver is increasingly investigated for its use as alternative to OFDM transceiver. A general matrix description of the transceiver is able to describe the existing interferences among different transmitted symbols, and is often employed to design the prototype filter and the basic equalizer parameters. In the present paper the general model of FBMC-OQAM transceiver, already proposed by the authors and used for single-tap equalization, is considered and a method for its fast evaluation is proposed. Some applications of the proposed procedure to single-tap transceiver equalization are suggested; computer simulations are carried out in order to confirm the value of the proposed method.

In Multi-Carrier (MC) systems, all equalization and detection procedures are classically performed at the sub-carrier level after the received signal is projected into the frequency domain. Besides, in the presence of rapid channel variation, conventional receivers suffer from critical performance degradation caused by interference. To address this specific problem, we propose to add a low-complexity time domain preprocessing prior to the frequency domain equalization process. We specifically study Orthogonal Frequency Division Multiplexing (OFDM) and Filter-Bank Multi-Carrier (FBMC) over single-path fast-varying Rayleigh channels when using the proposed scheme. Two time domain preprocessing techniques are considered. Their impact on equalization performance is evaluated in perfect and imperfect Channel State Information (CSI) scenarios, showing robustness to reasonable channel estimation errors. For both systems, a reduction in Bit Error Rate (BER) is obtained thanks to the preprocessing. Furthermore, the proposed scheme allows for capturing time diversity leading to improvement in performance for faster channel variation, rather than inducing performance degradation, showing the relevance of our approach. For OFDM, this preprocessing allows reaching the best performance compared to the existing equalizers at significant lower complexity. For FBMC, it permits to obtain a performance gain of 10 dB at FdTS = 0.25, while avoiding the BER floor effect. This gain is observed at BER= 2 × 10−2 for scenarios accepting Inter Symbol Interference (ISI), and at BER= 10−3 when assuming perfect ISI cancellation. For the latter scenario, we obtain performance very close to the (computationally intractable) optimum Maximum-Likelihood equalizer.

Ionosphere behavior is characterized by geographical and temporal changes. It was possible to determine the mobility of the ionosphere by observing how the total electron content (TEC) changed over time. The study on TEC variation was conducted in 2008 and 2009 over Malaysia, where latitude ranged from 1° N to 8° E and longitude ranged from 95° to 120° E, with each year reflecting the same degree of solar activity. The hourly mean vertical TEC (VTEC) is used to monitor the difference of the TEC under various geomagnetic factors such as disturbance storm time (Dst), planetary 3-hour range index (Kp), and solar flux F10.7 variations. This paper considers four cases of geomagnetic disturbances were considered 9 March, 27 March, 4 September 2008, and 22 July 2009. A percentage deviation was calculated to further analyze the geomagnetic disturbances in the equatorial region on quiet days. Results showed the TEC during geomagnetic disturbances was lower than the quiet day, with the percentage deviations between -20% and 10%. An increment of TEC was shown after approximately 24 hours, which is in the storm recovery phase. Also found were that GPS-TEC reaches its highest around post noon and its lowest in the small hours of the morning.