Internet of Things (IoT) can significantly enhance various aspects of today’s electric power grid infrastructures for making reliable, efficient, and safe next-generation Smart Grids (SGs). However, harsh and complex power grid infrastructures and environments reduce the accuracy of the information propagating through IoT platforms. In particularly, information is corrupted due to the measurement errors, quantization errors, and transmission errors. This leads to major system failures and instabilities in power grids. Redundant information measurements and retransmissions are traditionally used to eliminate the errors in noisy communication networks. However, these techniques consume excessive resources such as energy and channel capacity and increase network latency. Therefore, we propose a novel statistical information fusion method not only for structural chain and tree-based sensor networks, but also for unstructured bidirectional graph noisy wireless sensor networks in SG environments. We evaluate the accuracy, energy savings, fusion complexity, and latency of the proposed method by comparing the said parameters with several distributed estimation algorithms using extensive simulations proposing it for several SG applications. Results prove that the overall performance of the proposed method outperforms other fusion techniques for all considered networks. Under Smart Grid communication environments, the proposed method guarantees for best performance in all fusion accuracy, complexity and energy consumption. Analytical upper bounds for the variance of the final aggregated value at the sink node for structured networks are also derived by considering all major errors.

A Software Defined Vehicular Network (SDVN) is a new paradigm that enhances programmability and flexibility in Vehicular Adhoc Networks (VANETs). There exist different architectures for SDVNs based on the degree of control of the control plane. However, in vehicular communication literature, we find that there is no proper mechanism to collect data. Therefore, we propose a novel data collection methodology for the hybrid SDVN architecture by modeling it as an Integer Quadratic Programming (IQP) problem. The IQP model optimally selects broadcasting nodes and agent (unicasting) nodes from a given vehicular network instance with the objective of minimizing the number of agents, communication delay, communication cost, total payload, and total overhead. Due to the dynamic network topology, finding a new solution to the optimization is frequently required in order to avoid node isolation and redundant data transmission. Therefore, we propose a systematic way to collect data and make optimization decisions by inspecting the heterogeneous normalized network link entropy. The proposed optimization model for data collection for the hybrid SDVN architecture yields a 75.5% lower communication cost and 32.7% lower end-to-end latency in large vehicular networks compared to the data collection in the centralized SDVN architecture while collecting 99.9% of the data available in the vehicular network under optimized settings.

Aphasia is a type of speech disorder that can cause speech defects in a person. Identifying the severity level of the aphasia patient is critical for the rehabilitation process. In this research, we identify ten aphasia severity levels motivated by specific speech therapies based on the presence or absence of identified characteristics in aphasic speech in order to give more specific treatment to the patient. In the aphasia severity level classification process, we experiment on different speech feature extraction techniques, lengths of input audio samples, and machine learning classifiers toward classification performance. Aphasic speech is required to be sensed by an audio sensor and then recorded and divided into audio frames and passed through an audio feature extractor before feeding into the machine learning classifier. According to the results, the mel frequency cepstral coefficient (MFCC) is the most suitable audio feature extraction method for the aphasic speech level classification process, as it outperformed the classification performance of all mel-spectrogram, chroma, and zero crossing rates by a large margin. Furthermore, the classification performance is higher when 20 s audio samples are used compared with 10 s chunks, even though the performance gap is narrow. Finally, the deep neural network approach resulted in the best classification performance, which was slightly better than both K-nearest neighbor (KNN) and random forest classifiers, and it was significantly better than decision tree algorithms. Therefore, the study shows that aphasia level classification can be completed with accuracy, precision, recall, and F1-score values of 0.99 using MFCC for 20 s audio samples using the deep neural network approach in order to recommend corresponding speech therapy for the identified level. A web application was developed for English-speaking aphasia patients to self-diagnose the severity level and engage in speech therapies.

Traditional networking is hardware-based, having the control plane coupled with the data plane. Software-Defined Networking (SDN), which has a logically centralized control plane, has been introduced to increase the programmability and flexibility of networks. Knowledge-Defined Networking (KDN) is an advanced version of SDN that takes one step forward by decoupling the management plane from control logic and introducing a new plane, called a knowledge plane, decoupled from control logic for generating knowledge based on data collected from the network. KDN is the next-generation architecture for self-learning, self-organizing, and self-evolving networks with high automation and intelligence. Even though KDN was introduced about two decades ago, it had not gained much attention among researchers until recently. The reasons for delayed recognition could be due to the technology gap and difficulty in direct transformation from traditional networks to KDN. Communication networks around the globe have already begun to transform from SDNs into KDNs. Machine learning models are typically used to generate knowledge using the data collected from network devices and sensors, where the generated knowledge may be further composed to create knowledge ontologies that can be used in generating rules, where rules and/or knowledge can be provided to the control, management, and application planes for use in decision-making processes, for network monitoring and configuration, and for dynamic adjustment of network policies, respectively. Among the numerous advantages that KDN brings compared to SDN, enhanced automation and intelligence, higher flexibility, and improved security stand tall. However, KDN also has a set of challenges, such as reliance on large quantities of high-quality data, difficulty in integration with legacy networks, the high cost of upgrading to KDN, etc. In this survey, we first present an overview of the KDN architecture and then discuss each plane of the KDN in detail, such as sub-planes and interfaces, functions of each plane, existing standards and protocols, different models of the planes, etc., with respect to examples from the existing literature. Existing works are qualitatively reviewed and assessed by grouping them into categories and assessing the individual performance of the literature where possible. We further compare and contrast traditional networks and SDN against KDN. Finally, we discuss the benefits, challenges, design guidelines, and ongoing research of KDNs. Design guidelines and recommendations are provided so that identified challenges can be mitigated. Therefore, this survey is a comprehensive review of architecture, operation, applications, and existing works of knowledge-defined networks.

The functionality of Vehicular Ad Hoc Networks (VANETs) is improved by the Software-Defined Vehicular Network (SDVN) paradigm. Routing is challenging in vehicular networks due to the dynamic network topology resulting from the high mobility of nodes. Existing approaches for routing in SDVN do not exploit both link lifetimes and link delays in finding routes, nor do they exploit the heterogeneity that exists in links in the vehicular network. Furthermore, most of the existing approaches compute parameters at the controller entirely using heuristic approaches, which are computationally inefficient and can increase the latency of SDVN as the network size grows. In this paper, we propose a novel hybrid algorithm for routing in SDVNs with two modes: the highest stable least delay mode and the highest stable shortest path mode, in which the mode is selected by estimating the network contention. We distinctly identify two communication channels in the vehicular network as wired and wireless, where network link entropy is formulated accordingly and is used in combination with pending transmissions to estimate collision probability and average network contention. We use the prospect of machine learning to predict the wireless link lifetimes and one-hop channel delays, which yield very low Root Mean Square Errors (RMSEs), depicting their very high accuracy, and the wireless link lifetime prediction using deep learning yields a much lower average computational time compared to an optimization-based approach. The proposed novel algorithm selects only stable links by comparing them with a link lifetime threshold whose optimum value is decided experimentally. We propose this routing framework to be compatible with the OpenFlow protocol, where we modify the flow table architecture to incorporate a route valid time and send a packet_in message to the controller when the route’s lifetime expires, requesting new flow rules. We further propose a flow table update algorithm to map computed routes to flow table entries, where we propose to incorporate an adaptive approach for route finding and flow rule updating upon reception of a packet_in message in order to minimize the computational burden at the controller and minimize communication overhead associated with control plane communication. This research contributes a novel hybrid routing framework for the existing SDVN paradigm, scrutinizing machine learning to predict the lifetime and delay of heterogeneity links, which can be readily integrated with the OpenFlow protocol for better routing applications, improving the performance of the SDVN. We performed realistic vehicular network simulations using the network simulator 3 by obtaining vehicular mobility traces using the Simulation of Urban Mobility (SUMO) tool, where we collected data sets for training the machine learning models using the simulated environment in order to test models in terms of RMSE and computational complexity. The proposed routing framework was comparatively assessed against existing routing techniques by evaluating the communication cost, latency, channel utilization, and packet delivery ratio. According to the results, the proposed routing framework results in the lowest communication cost, the highest packet delivery ratio, the least latency, and moderate channel utilization, on average, compared to routing in VANET using Ad Hoc On-demand Distance Vector (AODV) and routing in SDVN using Dijkstra; thus, the proposed routing framework improves routing in SDVN. Furthermore, results show that the proposed routing framework is enhanced with increasing routing frequency and network size, as well as at low vehicular speeds.

Coronavirus disease 2019 (COVID-19) has been causing negative impacts on various sectors in Sri Lanka, as a result of the public health interventions that the government had to implement in order to reduce the spread of the disease. Equivalent work carried out in this context is outdated and close to ideal models. This paper presents a mathematical epidemiological model, called SEQIJRDS, having additional compartments for quarantine and infected people divided into two compartments as diagnosed and non diagnosed, compared to the SEIR model. We have presented the rate equations for the model and the basic reproduction number is derived. This model considers the effect of vaccination, the viral load of the variants, mask use, mobility, contact tracing and quarantine, natural immunity development of the infected people, and immunity waning of the recovered group as key developments of the model. The model has been validated for the COVID-19 pandemic in Sri Lanka by parameter derivation using mathematical formulations with the help of the existing data, the literature, and by model fitting for historical data. We present a comparison of the model projections for hospitalized infected people, the cumulative death count, and the daily death count against the ground truth values and projections of the SEIR and SIR models during the model validation. The validation results show that the proposed SEQIJRDS model’s 12-week projection performance is significantly better than both the SEIR and SIR models; the 2-, 6-, 8-, and 10-week projection performance is always better, and the 4-week projection performance is only slightly inferior to other models. Using the proposed SEQIJRDS model, we project mortality under different lockdown procedures, vaccination procedures, quarantine practices, and different mask-use cases. We further project hospital resource usage to understand the best intervention that does not exhaust hospital resources. At the end, based on an understanding of the effect of individual interventions, this work recommends combined public health interventions based on the projections of the proposed model. Specifically, three recommendations—called minimum, sub-optimum, and optimum recommendations—are provided for public health interventions.

COVID-19 has been causing negative impacts on various sectors in Sri Lanka as a result of the public health interventions that government had to implement in order to reduce the spreading of the disease. Equivalent work carried out in this context is outdated and close to ideal models. This research is carried out in a crucial time which the daily deaths are rapidly increasing which arise the requirement for an accurate and practical model to predict the mortality in order to take decisions regarding public health interventions. This paper presents a mathematical epidemiological model called SEQIJRDS to predict on COVID-19. The model has been validated for the COVID 19 pandemic in Sri Lanka. The results show that the model outstands many of the state-of-the-art SEIR epidemiological models such as Imperial, IHME once properly parameterized. At the end; this work recommends public health interventions at this crucial time to save people's lives based on the predictions of the proposed model. Specifically, 3 recommendations called minimal, sub-optimal and optimal recommendations are provided for public health interventions.

There is no comprehensive study on the mental health of Sri Lankan undergraduate in higher education as most existing studies have been done for medical students only. It is unknown how academic and environmental factors contribute for the prevalence of psychiatric illnesses. Further, there is no sufficient information on the student/university based remedies to reduce the psychological distress of students. This research is carried out to find the overall psychological distress, well-being, prevalence percentages of psychiatric illnesses, associated risk factors and student/university remedies to overcome them. All 13 psychiatric illnesses were found with a moderate correlation among diseases having a mean prevalence percentage of 25.92 and standard deviation of 12.93 despite the prevalence of well-being factors among students and only 8 % are clinically diagnosed. 89 % of the students were suffering from at least one psychiatric illness and 68 % were found to be psychologically distressed. Sets of overall and individual demographic, academic and environmental risk factors contributing for the prevalence of a psychiatric illness in general and in particularly were identified respectively after a Binary logistic regression analysis. 61 % of the students don’t receive psychiatric help from the university and are using their own remedies.