Privacy concerns arise from sharing electric load data. For instance, this data can be hijacked and deductions can be made about building occupancy. Anonymized data does not fully solve the issue, as there have been several successful attempts of re-identifying individuals from anonymized data. Federated Learning (FL) involves training a global model among several clients at the edge without sharing data, but instead, sharing model weights. Besides training a global model to reduce the universal error, it is important to focus on reducing the local error for each client. Personalization helps improve local model convergence without affecting the global model learning process. Therefore, individuals can have a more accurate understanding of their energy usage behavior. It also reduces the number of FL rounds needed for training, which reduces load on the smart grid edge computing resources, communication network, and therefore reduces costs. Previously, research has only looked at stochastic gradient descent (SGD) and Adam for fine-tuning clients’ local models at the edge in the process of FL. This research simulates different FL environments to explore electric load forecasting using different FL aggregation algorithms at the server, as well as seven local optimizers for FL personalization (SGD, Adam, Adagrad, Adadelta, Adamax, Nadam, and RM-SProp). An analysis of the different local optimizer algorithms is also presented.

COVID-19 pandemic has brought major uncertainty in load forecasting. Enforcing and relaxing lockdown rules, infection numbers, and the changing habits of people are the main causes of this uncertainty. Electric load forecasting maintains the balance between electric supply and demand. It also assists electric utilities in pricing their services, planning, and managing their infrastructure. This paper proposes two pandemic-aware load forecasting models (i) a city-level model, applied on the cities of Ottawa and Toronto, predicting hourly load using weather and pandemic-related features including population mobility and the number of daily COVID-19 infections, and (ii) a second open-source model forecasting quarter-hourly residential-level loads using weather and population mobility features for the city of Pune in India. Both models utilize multitask learning to jointly learn and predict future electric loads. The quarter-hourly model uses Bi-directional Long Short-Term Memory (LSTM) to learn from COVID’s specific features, and a Convolutional Neural Network (CNN) to learn from the historical load data before the pandemic. The multitask nature of the model allows for incorporating multiple datasets with different numbers of features. The residential-level multitask model allowed for learning from long-term data before COVID-19 using weather features, short-term load data, and the mobility data. Multitask learning has also enabled the use of two datasets with different numbers of features due to the lack of mobility data pre-COVID.

Although the market valuation and adoption of IoT in various sectors is in an uptrend, the actual deployment is lagging when compared to the industrial forecasted data. The main reason behind this drawback is the lifespan of IoT devices due to their limited battery capacities. A solution to the problem is to deploy energy harvesting units (e.g., solar to replenish the batteries). However, due to the time varying availability of both irradiance and temperature and their effect on the power output, it is essential to predict both variables. To this end, we propose in this paper a prediction system that does not consume a lot of energy and that can be deployed on low computational nodes. This model consists of a Seasonal Auto Regressive Integrated Moving Average (SARIMA) with a Kalman filtering (KF) component. We build this model using an actual dataset for Ottawa, Ontario, Canada. We then demonstrate its effectiveness by presenting the results for randomly selected days in the Winter Season. In this context, we show that the SARIMA-KF outperforms the SARIMA in all scenarios with an average error reduction of 59.3%.

Diabetes Mellitus Type II (T2D) is a Chronic Disease and is the most common type of Diabetes in the world, responsible for 95% of all Diabetes patients. T2D is a very complex disease and requires a large amount of self-management from the patient in order to maintain a healthy and threat-free lifestyle. Therefore, we develop in this paper a data analytics solution to assist in the self-management of T2D patients through several methods consisting of a rule-based system, anomaly detection, and threat forecasting.

Smart Grid is a major element of the Smart City concept that enables two-way communication of energy data between electric utilities and their consumers. These communication technologies are going through sharp modernization to meet future demand growth and to achieve reliability, security, and efficiency of the electric grid. In this paper, we implement an IPv6 based two-way communication system between the transformer agent (TA), installed at local electric transformer and various customer agents (CAs), connected to customer's smart meter. Various homes share their energy usage with the TA which in turn sends the utility's recommendations to the CAs. Raspberry Pi is used as hardware for all the CAs and the TA. We implement a self-healing mesh network between all nodes using OpenLab IEEE 802.15.4 chips and Routing Protocol for Low-Power and Lossy Networks (RPL), and the data is secured by RSA/AES keys. Several tests have been conducted in real environments, inside and outside of Carleton University, to test the performance of this communication network in various obstacle settings. In this paper, we highlight the details behind the implementation of this IPv6-based smart grid communication system, the related challenges, and the proposed solutions.

Emerging technologies such as the Internet of Things (IoT) and their anticipated massive deployment stimulate the need for developing adaptive energy efficient modulation schemes to maximize network lifetime. IoT systems are typically comprised of limited energy heterogeneous devices in the sensing layer, imposing significant challenges in developing cross-layer schemes to solve the network lifetime problem. In this paper, we present a multi-objective adaptive modulation scheme for the physical layer of a heterogeneous IoT environment. We consider channel conditions to opportunistically maximize device prioritization, energy efficiency, and spectral efficiency. The problem is modeled as a Mixed Integer Linear Program (MILP) in GAMS and is solved by CPLEX under Rayleigh fading channel conditions. Performance evaluations show that considering device heterogeneity is crucial in order to exploit energy savings and spectral efficiency in IoT sensing nodes.

With the popularity of smart electrical appliances and home energy management systems, there has been a massive amount of data generated by the power consumption. This data can be beneficial for the utility as it provides the behavior patterns of customers, and thus useful decisions can be made to optimize the load on the grid. In this work, we establish a bidirectional communication system between some homes through the customer agents (CAs), which are installed at home, and the transformer agent (TA) which is installed at the local transformer. Once data is collected at the TA, it is sent to the cloud through LTE. We then use IBM Cloud services to filter and analyze this data to forecast energy consumption and make recommendations to different customers based on their real-time changing behaviors. To this end, we use six different machine learning models predicting the energy consumption: support vector regression (SVR) using linear kernel, SVR using Gaussian kernel, SVR using the polynomial kernel, linear regression, polynomial regression, and feed- forward neural networks (FFNN). To measure the accuracy of these models, we compute three different error metrics, the normalized mean absolute percentage error (NMAPE), the normalized root mean square error (NRMSE), and R2also known as the coefficient of determination. Based on these results, we observe that the performance of the forecasting model depends on the dataset properties including the size and variations. For example, while linear and polynomial regressions perform well for small-scale datasets, FFNN gives higher accuracy for large-scale datasets.

In this paper, we consider an overlay cognitive radio (CR) scenario where the primary transmitter (PT) and the primary receiver (PR) communicate via the help of a secondary users' (SUs) system. Under a worst-case scenario, we assume that the link between the primary users (PUs) is broken and the help of a selected secondary transmitter (ST) is required. Taking advantage of this opportunity, this ST will be able to transmit its own data. The communications of the PUs and the SUs take place over two phases. In the first phase, receive space shift keying (R-SSK) is employed at the PT in order to activate one ST for reception. This ST is scheduled to transmit its own data during the second phase using conventional SSK, which also allows the PR to decode the PT's message. The proposed scheduling scheme is initiated by the PT based on its incoming bits which provides fairness among STs. The proposed system comes with other advantages including the low receivers' complexity and the improved energy efficiency (EE) all gained by the use of SSK. We analyze the performance of the proposed scheme in terms of the average bit error probability (ABEP). We finally provide comparisons to existing schemes and we generate numerical results through which we confirm the derived analysis and we demonstrate the effectiveness of the proposed overlay cognitive scheduling scheme.

We propose in this paper a physical‐layer security (PLS) scheme for dual‐hop cooperative networks in an effort to enhance the communications secrecy. The underlying model comprises a transmitting node (Alice), a legitimate node (Bob), and an eavesdropper (Eve). It is assumed that there is no direct link between Alice and Bob, and the communication between them is done through trusted relays over two phases. In the first phase, precoding‐aided spatial modulation (PSM) is employed, owing to its low interception probability, while simultaneously transmitting a jamming signal from Bob. In the second phase, the selected relay detects and transmits the intended signal, whereas the remaining relays transmit the jamming signal received from Bob. We analyze the performance of the proposed scheme in terms of the ergodic secrecy capacity (ESC), the secrecy outage probability (SOP), and the bit error rate (BER) at Bob and Eve. We obtain closed‐form expressions for the ESC and SOP and we derive very tight upper‐bounds for the BER. We also optimize the performance with respect to the power allocation among the participating relays in the second phase. We provide examples with numerical and simulation results through which we demonstrate the effectiveness of the proposed scheme.

We propose in this paper a physical layer security (PLS) scheme for dual-hop cooperative networks in an effort to enhance the communications secrecy. The underlying model comprises a transmitting node (Alice), a legitimate node (Bob) and an eavesdropper (Eve). It is assumed that there is no direct link between Alice and Bob, and the communication between them is done through trusted relays over two phases. In the first phase, precoded spatial modulation (PSM) is employed, owing to its low interception probability, while simultaneously transmitting a jamming signal from Bob. In the second phase, the selected relay detects and transmits the intended signal, whereas the remaining relays transmit just the jamming signal received from Bob. We analyze the performance of the proposed scheme in terms of the ergodic secrecy capacity (ESC) and secrecy outage probability (SOP) where we obtain closed form expressions for those metrics. We also optimize the performance with respect to the power allocation among the participating relays in the second phase. We provide examples with numerical and simulations results through which we demonstrate the effectiveness of the proposed scheme. We also provide a comparison of the bit error rate (BER) at Bob and Eve by simulation.

Based on the concept of reconfigurable antennas (RAs), SSK-RA has been recently proposed as a novel transmission scheme to improve the performance of space shift keying (SSK). In this context, it has been shown that RAs' reconfigurable properties can be used as additional degrees of freedom to enhance the throughput, implementation complexity, and error performance of SSK. In this paper, we extend SSK-RA to cognitive radio (CR) systems in an effort to improve the secondary system's performance in a spectrum sharing scenario. Taking advantage of the interplay between RAs and the propagation channels for both the secondary and interference links, we propose a RA-based scheme with pointing-direction reconfiguration aiming at improving the secondary system's performance while verifying an outage interference constraint to the primary user (PU). In this paper, we analyse the performance of the proposed scheme in Rician fading channels and provide simulation examples confirming these analytical results. The proposed schemes are shown to offer better bit error rate (BER) performance and lower implementation complexity when compared to conventional antenna-based spectrum sharing systems.

Based on the concept of reconfigurable antennas (RAs), space-shift keying (SSK)-RA has been recently proposed as a novel transmission scheme to improve the performance of SSK. In this context, it has been shown that RAs' reconfigurable properties can be used as additional degrees of freedom to enhance the throughput, implementation complexity, and error performance of SSK. In this paper, we study the implementation of SSK-RA within underlay cognitive radio systems in an effort to improve the performance of the secondary user while verifying the constraints set by the primary user (PU). Taking advantage of the interplay between RAs and the propagation channels for both the secondary and interference links, we propose an RA-based scheme with beam-direction reconfiguration aiming at improving the secondary system's performance while verifying an outage interference constraint to the PU. In this paper, we analyze the performance of the proposed scheme in Rician fading channels and provide simulation examples confirming these analytical results. The proposed schemes are shown to offer enhanced bit error rate performance and lower implementation complexity when compared with conventional antenna-based spectrum sharing systems.

In this work, we use the concept of the new emerging technology of reconfigurable antennas (RAs) in an effort to improve the performance of space shift keying (SSK). Indeed, the reconfigurable properties of RAs can be used as additional degrees of freedom to enhance the performance of SSK in terms of throughput, system complexity, and error performance. In this context, considering correlated and nonidentically distributed Rician fading channels, we propose a number of SSK-RA schemes using SSK with antenna-state selection while taking advantage of the effect of RAs on the multipath propagation channel. More specifically, based on the variation of the Rician K-factor and the correlation coefficients with different antenna states, the proposed schemes jointly optimize the fading and correlation parameters to enhance SSK's performance. In this paper, we analyze the performance of the proposed SSK-RA schemes over Rician fading channels in terms of average spectral efficiency (ASE) and bit error rate (BER). We also provide several simulation examples through which we corroborate these analytical results. The proposed schemes are shown to enhance the performance of SSK both in terms of ASE and BER. For the same ASE, compared to that of SSK, not only do the proposed SSK-RA schemes provide much better error performance, they also reduce the implementation cost and the overall system complexity considerably.

In this work, we use the concept of reconfigurable antennas (RAs) in an effort to improve the performance of space shift keying (SSK). Indeed, the reconfigurable properties of RAs can be used as additional degrees of freedom in order to enhance the performance of SSK in terms of throughput, system complexity, and error performance. In this context, we propose a number of SSK-RA schemes using SSK with antenna-state selection while taking advantage of the interplay between RAs and the propagation channel. In this paper, we analyse the performance of the proposed SSK-RA schemes in Rician fading channels in terms of average spectral efficiency (ASE) and average bit error rate (ABER).We also provide several simulation examples through which we corroborate these analytical results. The proposed schemes are shown to enhance the performance of SSK both in terms of ASE and ABER. For the same ASE as SSK, SSK-RA schemes not only provide much better error performance but also considerably reduce the implementation cost and the overall system complexity.

We have recently proposed adaptive spatial modulation (ASM) for multiple antenna systems with the aim of improving the energy efficiency through spatial modulation (SM) and improving the average spectral efficiency (ASE) through adaptive modulation. In this paper, we extend ASM to spectrum sharing systems in an effort to improve the secondary system's performance in terms of energy efficiency and ASE. Based on a limited feedback from the primary user (PU), the secondary transmitter (ST) uses power adaptation in order to maximize its ASE while respecting a peak interference constraint to the PU. When compared to the recently-proposed fixed power scheme (FPS), the adaptive power scheme (APS) proposed in this paper offers lower transmission delays, higher effective throughput, and comparable error performance. In this work, we analyze the performance of APS in terms of ASE, effective throughput, and average bit error rate (ABER). We also provide several simulations through which we corroborate these analytical results.

We have recently introduced adaptive spatial modulation (ASM) for multiple antenna systems, with the aim of improving the energy efficiency through spatial modulation (SM) and improving the average spectral efficiency (ASE) through adaptive modulation (AM). In this paper, we extend ASM to cognitive radio (CR) systems in an effort to improve the secondary system's performance in terms of energy efficiency and ASE. To this end, we propose two ASM schemes, one referred to as fixed power scheme (FPS) and the other as adaptive power scheme (APS). In both schemes, the secondary transmitter (ST) has limited knowledge of the channel state information (CSI) of the interference link. The difference between the two schemes, however, lies in the way the limited CSI is used to adapt the transmit power and/or modulation. As a benchmark, we also consider the scenario where the ST has perfect knowledge of the CSI of the interference link. For all cases, we analyze the performance in terms of ASE, average delay, and bit error rate (BER). We show that the proposed schemes offer tradeoffs in terms of the previously mentioned performance metrics, thus offering different options for applying ASM to CR systems. We also provide several simulation examples through which we corroborate the analytical results.

We have recently introduced adaptive spatial modulation (ASM) [1], which comprises both adaptive modulation (AM) and spatial modulation (SM), with the aim of enhancing the average spectral efficiency (ASE) of multiple antenna systems. This technique was shown to offer high energy efficiency and low system complexity thanks to the use of SM while achieving high data rates thanks to the use of AM. Motivated by this technique and the need of such performance in a cognitive radio (CR) scenario, we extend in this paper the concept of ASM to a spectrum sharing system. In this context, we propose the ASM-CR scheme as an energy-efficient, spectrally-efficient, and low-complexity scheme for CR systems. Compared to selected existing spectrum sharing schemes, ASM-CR is shown to offer a considerable ASE improvement while experiencing a slight degradation in the average bit error rate (ABER) performance. In order to address this, a hybrid scheme compromising between ASE and ABER is proposed. The performance of both proposed schemes is analyzed in terms of ASE and ABER and confirmed with selected numerical results using Monte-Carlo simulations.

Using spatial modulation (SM) jointly with adaptive modulation (AM), we propose a low-complexity and spectrally-efficient transmission scheme in a multiple-input multiple-output (MIMO) system. While in the conventional SM technique a fixed data rate is achieved, the proposed adaptive spatial modulation (ASM) technique is throughput-optimized by taking advantage of the wireless channel variations in order to increase the spectral efficiency of SM. ASM has been previously studied in [1] in order to improve the average bit error rate (ABER) performance of SM while only providing a fixed data rate. On the other hand, the ASM technique is introduced in this paper in order to achieve high data rates while keeping the ABER below a certain threshold. We propose two variations of ASM compromising between the spectral efficiency and the error performance. The ABER and the average spectral performance results of both variations are presented via Monte-Carlo simulations and confirmed with analytical results including asymptotic performance bounds on the ABER. These results show that the proposed ASM techniques come with a considerable spectral efficiency gain compared to SM while only requiring a limited feedback from the receiver.

In this paper, we propose a number of adaptive transmission techniques in order to improve the performance of the secondary link in a spectrum sharing system. We first introduce the concept of minimum-selection maximum ratio transmission (MS-MRT) as an adaptive variation of the existing MRT (MRT) technique. While in MRT all available antennas are used for transmission, MS-MRT uses the minimum subset of antennas verifying both the interference constraint (IC) to the primary user and the bit error rate (BER) requirements. Similar to MRT, MS-MRT assumes that perfect channel state information (CSI) is available at the secondary transmitter (ST), which makes this scheme challenging from a practical point of view. To overcome this challenge, we propose another transmission technique based on orthogonal space-time block codes with transmit antenna selection (TAS). This technique uses the full-rate full-diversity Alamouti scheme in order to maximize the secondary's transmission rate. The performance of these techniques is analyzed in terms of the average spectral efficiency (ASE), average number of transmit antennas, average delay, average BER, and outage performance. In order to give the motivation behind these analytical results, the tradeoffs offered by the proposed schemes are summarized and then demonstrated through several numerical examples.

Under the scenario of an underlay cognitive radio network, we propose in this paper two adaptive schemes using switched transmit diversity and adaptive modulation in order to increase the spectral efficiency of the secondary link and maintain a desired performance for the primary link. The proposed switching efficient scheme (SES) and bandwidth efficient scheme (BES) use the scan and wait combining technique (SWC) where a transmission occurs only when a branch with an acceptable performance is found, otherwise data is buffered. In these schemes, the modulation constellation size and the used transmit branch are determined to minimize the average number of switched branches and to achieve the highest spectral efficiency given the fading channel conditions, the required error rate performance, and a peak interference constraint to the primary receiver (PR). For delay-sensitive applications, we also propose two variations of the SES and BES schemes using power control (SES-PC and BES-PC) where the secondary transmitter (ST) starts sending data using a nominal power level which is selected in order to minimize the average delay introduced by the SWC technique. We demonstrate through numerical examples that the BES scheme increases the capacity of the secondary link when compared to the SES scheme. This spectral efficiency improvement comes at the expense of an increased average number of switched branches and thus an increased average delay. We also show that the SES-PC and the BES-PC schemes minimize the average delay while satisfying the same spectral efficiency as the SES and BES schemes, respectively.

Under the scenario of an underlay cognitive radio network, we propose in this paper an adaptive scheme using switched transmit diversity and adaptive modulation in order to increase the spectral efficiency of the secondary link. The proposed bandwidth efficient scheme (BES) uses the scan and wait (SWC) combining technique where a transmission occurs only when a branch with an acceptable performance is found, otherwise data is buffered. In our scheme, the modulation constellation size and the used transmit branch are determined to achieve the highest spectral efficiency given the fading channel conditions, the required error rate performance, and a peak interference constraint to the primary receiver. Selected numerical examples show that the BES scheme increases the capacity of the secondary link when compared to an existing switching efficient scheme (SES). This spectral efficiency comes at the expense of an increased average number of switched branches and thus an increased average delay.

Adaptive modulation and diversity combining represent very important adaptive solutions for future generations of wireless communication systems. Indeed, in order to improve the performance and the efficiency of these systems, these two techniques have been recently used jointly in new schemes named joint adaptive modulation and diversity combining (JAMDC) schemes. Considering the problem of finding low hardware complexity, bandwidth-efficient, and processing-power efficient transmission schemes for a downlink scenario and capitalizing on some of these recently proposed JAMDC schemes, we propose and analyze in this paper three joint adaptive modulation, diversity combining, and power control (JAMDCPC) schemes where a constant-power variable-rate adaptive modulation technique is used with an adaptive diversity combining scheme and a common power control process. More specifically, the modulation constellation size, the number of combined diversity paths, and the needed power level are jointly determined to achieve the highest spectral efficiency with the lowest possible processing power consumption quantified in terms of the average number of combined paths, given the fading channel conditions and the required bit error rate (BER) performance. In this paper, the performance of these three JAMDCPC schemes is analyzed in terms of their spectral efficiency, processing power consumption, and error-rate performance. Selected numerical examples show that these schemes considerably increase the spectral efficiency of the existing JAMDC schemes with a slight increase in the average number of combined paths for the low signal-to-noise ratio range while maintaining compliance with the BER performance and a low radiated power which yields to a substantial decrease in interference to co-existing users and systems.

Adaptive modulation and diversity combining represent very important adaptive solutions for the future generations of communication systems. In order to improve the performance and the efficiency of wireless communication systems these two techniques have been recently used jointly in new schemes named joint adaptive modulation and diversity combining (JAMDC) schemes. Considering the problem of finding low-complexity, bandwidth-efficient, and processing-power efficient transmission schemes for a downlink scenario and capitalizing on one of these recently proposed JAMDC schemes, we propose and analyze in this paper two fully joint adaptive modulation, diversity combining, and power control (FJAMDC) schemes. More specifically, the modulation constellation size, the number of combined diversity paths, and the needed power level are jointly determined to achieve the highest spectral efficiency with the lowest possible combining complexity, given the fading channel conditions and the required error rate performance. Selected numerical examples show that the newly proposed schemes considerably increase the spectral efficiency with a slight increase in the average number of combined path for the low signal to noise ratio (SNR) range while maintaining compliance with the bit error rate (BER) performance and a low radiated power which yields to a substantial decrease in interference to co-existing systems/users.

We propose and analyze in this paper two new adaptive power controlled minimum selection generalized selection combining (MS-GSC) diversity schemes. The key idea behind these two schemes is to request the transmitter to vary its power level during very adverse channel conditions in order to communicate with the minimum required quality of service. For each scheme, four power control variants accounting for practical implementation constraints including discrete power levels and transmitter gain saturation are proposed and studied. Selected numerical examples, show that ideal continuous power control lowers the average bit error rate of MS-GSC and makes it always meet the specified target quality of service at the expense of an increase in the transmitter gain. Additional numerical examples, show that the power control variants that take into account practical implementation constraints conserve the main features of the ideal continuous power algorithm.