Using transfer learning-based plant disease classification and detection for sustainable agriculture.

Subsistence farmers and global food security depend on sufficient food production, which aligns with the UN's "Zero Hunger," "Climate Action," and "Responsible Consumption and Production" sustainable development goals. In addition to already available methods for early disease detection and classification facing overfitting and fine feature extraction complexities during the training process, how early signs of green attacks can be identified or classified remains uncertain. Most pests and disease symptoms are seen in plant leaves and fruits, yet their diagnosis by experts in the laboratory is expensive, tedious, labor-intensive, and time-consuming. Notably, how plant pests and diseases can be appropriately detected and timely prevented is a hotspot paradigm in smart, sustainable agriculture remains unknown. In recent years, deep transfer learning has demonstrated tremendous advances in the recognition accuracy of object detection and image classification systems since these frameworks utilize previously acquired knowledge to solve similar problems more effectively and quickly. Therefore, in this research, we introduce two plant disease detection (PDDNet) models of early fusion (AE) and the lead voting ensemble (LVE) integrated with nine pre-trained convolutional neural networks (CNNs) and fine-tuned by deep feature extraction for efficient plant disease identification and classification. The experiments were carried out on 15 classes of the popular PlantVillage dataset, which has 54,305 image samples of different plant disease species in 38 categories. Hyperparameter fine-tuning was done with popular pre-trained models, including DenseNet201, ResNet101, ResNet50, GoogleNet, AlexNet, ResNet18, EfficientNetB7, NASNetMobile, and ConvNeXtSmall. We test these CNNs on the stated plant disease detection and classification problem, both independently and as part of an ensemble. In the final phase, a logistic regression (LR) classifier is utilized to determine the performance of various CNN model combinations. A comparative analysis was also performed on classifiers, deep learning, the proposed model, and similar state-of-the-art studies. The experiments demonstrated that PDDNet-AE and PDDNet-LVE achieved 96.74% and 97.79%, respectively, compared to current CNNs when tested on several plant diseases, depicting its exceptional robustness and generalization capabilities and mitigating current concerns in plant disease detection and classification.

Security Framework for Network-Based Manufacturing Systems with Personalized Customization: An Industry 4.0 Approach.

Smart manufacturing is pivotal in the context of Industry 4.0, as it integrates advanced technologies like the Internet of Things (IoT) and automation to streamline production processes and improve product quality, paving the way for a competitive industrial landscape. Machines have become network-based through the IoT, where integrated and collaborated manufacturing system responds in real time to meet demand fluctuations for personalized customization. Within the network-based manufacturing system (NBMS), mobile industrial robots (MiRs) are vital in increasing operational efficiency, adaptability, and productivity. However, with the advent of IoT-enabled manufacturing systems, security has become a serious challenge because of the communication of various devices acting as mobile nodes. This paper proposes the framework for a newly personalized customization factory, considering all the advanced technologies and tools used throughout the production process. To encounter the security concern, an IoT-enabled NBMS is selected as the system model to tackle a black hole attack (BHA) using the NTRUEncrypt cryptography and the ad hoc on-demand distance-vector (AODV) routing protocol. NTRUEncrypt performs encryption and decryption while sending and receiving messages. The proposed technique is simulated by network simulator NS-2.35, and its performance is evaluated for different network environments, such as a healthy network, a malicious network, and an NTRUEncrypt-secured network based on different evaluation metrics, including throughput, goodput, end-to-end delay, and packet delivery ratio. The results show that the proposed scheme performs safely in the presence of a malicious node. The implications of this study are beneficial for manufacturing industries looking to embrace IoT-enabled subtractive and additive manufacturing facilitated by mobile industrial robots. Implementation of the proposed scheme ensures operational efficiency, enables personalized customization, and protects confidential data and communication in the manufacturing ecosystem.

A Comprehensive Study on Cyber Attacks in Communication Networks in Water Purification and Distribution Plants: Challenges, Vulnerabilities, and Future Prospects.

In recent years, the Internet of Things (IoT) has had a big impact on both industry and academia. Its profound impact is particularly felt in the industrial sector, where the Industrial Internet of Things (IIoT), also known as Industry 4.0, is revolutionizing manufacturing and production through the fusion of cutting-edge technologies and network-embedded sensing devices. The IIoT revolutionizes several industries, including crucial ones such as oil and gas, water purification and distribution, energy, and chemicals, by integrating information technology (IT) with industrial control and automation systems. Water, a vital resource for life, is a symbol of the advancement of technology, yet knowledge of potential cyberattacks and their catastrophic effects on water treatment facilities is still insufficient. Even seemingly insignificant errors can have serious consequences, such as aberrant pH values or fluctuations in the concentration of hydrochloric acid (HCI) in water, which can result in fatalities or serious diseases. The water purification and distribution industry has been the target of numerous hostile cyber security attacks, some of which have been identified, revealed, and documented in this paper. Our goal is to understand the range of security threats that are present in this industry. Through the lens of IIoT, the survey provides a technical investigation that covers attack models, actual cases of cyber intrusions in the water sector, a range of security difficulties encountered, and preventative security solutions. We also explore upcoming perspectives, illuminating the predicted advancements and orientations in this dynamic subject. For industrial practitioners and aspiring scholars alike, our work is a useful, enlightening, and current resource. We want to promote a thorough grasp of the cybersecurity landscape in the water industry by combining key insights and igniting group efforts toward a safe and dependable digital future.

High-Speed Network DDoS Attack Detection: A Survey.

Having a large number of device connections provides attackers with multiple ways to attack a network. This situation can lead to distributed denial-of-service (DDoS) attacks, which can cause fiscal harm and corrupt data. Thus, irregularity detection in traffic data is crucial in detecting malicious behavior in a network, which is essential for network security and the integrity of modern Cyber-Physical Systems (CPS). Nevertheless, studies have shown that current techniques are ineffective at detecting DDoS attacks on networks, especially in the case of high-speed networks (HSN), as detecting attacks on the latter is very complex due to their fast packet processing. This review aims to study and compare different approaches to detecting DDoS attacks, using machine learning (ML) techniques such as k-means, K-Nearest Neighbors (KNN), and Naive Bayes (NB) used in intrusion detection systems (IDSs) and flow-based IDSs, and expresses data paths for packet filtering for HSN performance. This review highlights the high-speed network accuracy evaluation factors, provides a detailed DDoS attack taxonomy, and classifies detection techniques. Moreover, the existing literature is inspected through a qualitative analysis, with respect to the factors extracted from the presented taxonomy of irregular traffic pattern detection. Different research directions are suggested to support researchers in identifying and designing the optimal solution by highlighting the issues and challenges of DDoS attacks on high-speed networks.

Using a novel convolutional neural network for plant pests detection and disease classification.

BACKGROUND: Early plant diseases and pests identification reduces social, economic, and environmental deficiencies entailing toxic chemical utilization on agricultural farms, thus posing a threat to global food security. METHODOLOGY: An enhanced convolutional neural network (CNN) along with long short-term memory (LSTM) using a majority voting ensemble classifier has been proposed to tackle plant pest and disease identification and classification. Within pre-trained models, deep feature extractions have been obtained from connected layers. Deep features have been extracted and are sent to the LSTM layer to build a robust, enhanced LSTM-CNN model for detecting plant pests and diseases. Experiments were carried out using a Turkey dataset, with 4447 apple pests and diseases categorized into 15 different classes. RESULTS: The study was evaluated in different CNNs using logistic regression (LR), LSTM, and extreme learning machine (ELM), focusing on plant disease detection problems. The ensemble majority voting classifier was used at the LSTM layer to detect and classify plant disease labels. Furthermore, an autonomous selection of the optimal LSTM layer network parameters was applied. Finally, the performance was validated based on sensitivity, F1 score, accuracy, and specificity using LSTM, ELM, and LR classifiers. CONCLUSION: The presented model attained 99.2% accuracy compared to the cutting-edge models on different classifiers such as LSTM, LR, and ELM, and performed better compared to transfer learning. Pre-trained models, such as VGG19, VGG18, and AlexNet, demonstrated better accuracy when the fc6 layer was compared with other layers. © 2023 Society of Chemical Industry.

A Systematic Literature Review on Service Composition for People with Disabilities: Taxonomies, Solutions, and Open Research Challenges.

Integrating smart heterogeneous objects, IoT devices, data sources, and software services to produce new business processes and functionalities continues to attract considerable attention from the research community due to its unraveled advantages, including reusability, adaptation, distribution, and pervasiveness. However, the exploitation of service-oriented computing technologies (e.g., SOC, SOA, and microservice architectures) by people with special needs is underexplored and often overlooked. Furthermore, the existing challenges in this area are yet to be identified clearly. This research study presents a rigorous literature survey of the recent advances in service-oriented composition approaches and solutions for disabled people, their domains of application, and the major challenges, covering studies published between January 2010 and October 2022. To this end, we applied the systematic literature review (SLR) methodology to retrieve and collate only the articles presenting and discussing service composition solutions tailored to produce digitally accessible services for consumption by people who suffer from an impairment or loss of some physical or mental functions. We searched six renowned bibliographic databases, particularly IEEE Xplore, Web of Science, Springer Link, ACM Library, ScienceDirect, and Google Scholar, to synthesize a final pool of 38 related articles. Our survey contributes a comprehensive taxonomy of service composition solutions, techniques, and practices that are utilized to create assistive technologies and services. The seven-facet taxonomy helps researchers and practitioners to quickly understand and analyze the fundamental conceptualizations and characteristics of accessible service composition for people with disabilities. Key findings showed that services are fused to assist disabled persons to carry out their daily activities, mainly in smart homes and ambient intelligent environments. Despite the emergence of immersive technologies (e.g., wearable computing), user-service interactions are enabled primarily through tactile and speech modalities. Service descriptions mainly incorporate functional features (e.g., performance, latency, and cost) of service quality, largely ignoring accessibility features. Moreover, the outstanding research problems revolve around (1) the unavailability of assistive services datasets, (2) the underspecification of accessibility aspects of disabilities, (3) the weak adoption of accessible and universal design practices, (4) the abstraction of service composition approaches, and (5) the rare experimental testing of composition approaches with disabled users. We conclude our survey with a set of guidelines to realize effective assistive service composition in IoT and cloud environments. Researchers and practitioners are advised to create assistive services that support the social relationships of disabled users and model their accessibility needs as part of the quality of service (QoS). Moreover, they should exploit AI/ML models to address the evolving requirements of disabled users in their unique environments. Furthermore, weaknesses of service composition solutions and research challenges are exposed as notable opportunities for future research.

A Two-Phase Machine Learning Framework for Context-Aware Service Selection to Empower People with Disabilities.

The use of software and IoT services is increasing significantly among people with special needs, who constitute 15% of the world's population. However, selecting appropriate services to create a composite assistive service based on the evolving needs and context of disabled user groups remains a challenging research endeavor. Our research applies a scenario-based design technique to contribute (1) an inclusive disability ontology for assistive service selection, (2) semi-synthetic generated disability service datasets, and (3) a machine learning (ML) framework to choose services adaptively to suit the dynamic requirements of people with special needs. The ML-based selection framework is applied in two complementary phases. In the first phase, all available atomic tasks are assessed to determine their appropriateness to the user goal and profiles, whereas in the subsequent phase, the list of service providers is narrowed by matching their quality-of-service factors against the context and characteristics of the disabled person. Our methodology is centered around a myriad of user characteristics, including their disability profile, preferences, environment, and available IT resources. To this end, we extended the widely used QWS V2.0 and WS-DREAM web services datasets with a fusion of selected accessibility features. To ascertain the validity of our approach, we compared its performance against common multi-criteria decision making (MCDM) models, namely AHP, SAW, PROMETHEE, and TOPSIS. The findings demonstrate superior service selection accuracy in contrast to the other methods while ensuring accessibility requirements are satisfied.

An Ensemble Learning Based Classification Approach for the Prediction of Household Solid Waste Generation.

With the increase in urbanization and smart cities initiatives, the management of waste generation has become a fundamental task. Recent studies have started applying machine learning techniques to prognosticate solid waste generation to assist authorities in the efficient planning of waste management processes, including collection, sorting, disposal, and recycling. However, identifying the best machine learning model to predict solid waste generation is a challenging endeavor, especially in view of the limited datasets and lack of important predictive features. In this research, we developed an ensemble learning technique that combines the advantages of (1) a hyperparameter optimization and (2) a meta regressor model to accurately predict the weekly waste generation of households within urban cities. The hyperparameter optimization of the models is achieved using the Optuna algorithm, while the outputs of the optimized single machine learning models are used to train the meta linear regressor. The ensemble model consists of an optimized mixture of machine learning models with different learning strategies. The proposed ensemble method achieved an R2 score of 0.8 and a mean percentage error of 0.26, outperforming the existing state-of-the-art approaches, including SARIMA, NARX, LightGBM, KNN, SVR, ETS, RF, XGBoosting, and ANN, in predicting future waste generation. Not only did our model outperform the optimized single machine learning models, but it also surpassed the average ensemble results of the machine learning models. Our findings suggest that using the proposed ensemble learning technique, even in the case of a feature-limited dataset, can significantly boost the model performance in predicting future household waste generation compared to individual learners. Moreover, the practical implications for the research community and respective city authorities are discussed.

Moisture Computing-Based Internet of Vehicles (IoV) Architecture for Smart Cities.

Recently, the concept of combining 'things' on the Internet to provide various services has gained tremendous momentum. Such a concept has also impacted the automotive industry, giving rise to the Internet of Vehicles (IoV). IoV enables Internet connectivity and communication between smart vehicles and other devices on the network. Shifting the computing towards the edge of the network reduces communication delays and provides various services instantly. However, both distributed (i.e., edge computing) and central computing (i.e., cloud computing) architectures suffer from several inherent issues, such as high latency, high infrastructure cost, and performance degradation. We propose a novel concept of computation, which we call moisture computing (MC) to be deployed slightly away from the edge of the network but below the cloud infrastructure. The MC-based IoV architecture can be used to assist smart vehicles in collaborating to solve traffic monitoring, road safety, and management issues. Moreover, the MC can be used to dispatch emergency and roadside assistance in case of incidents and accidents. In contrast to the cloud which covers a broader area, the MC provides smart vehicles with critical information with fewer delays. We argue that the MC can help reduce infrastructure costs efficiently since it requires a medium-scale data center with moderate resources to cover a wider area compared to small-scale data centers in edge computing and large-scale data centers in cloud computing. We performed mathematical analyses to demonstrate that the MC reduces network delays and enhances the response time in contrast to the edge and cloud infrastructure. Moreover, we present a simulation-based implementation to evaluate the computational performance of the MC. Our simulation results show that the total processing time (computation delay and communication delay) is optimized, and delays are minimized in the MC as apposed to the traditional approaches.

WEAMR-a weighted energy aware multipath reliable routing mechanism for hotline-based WSNs.

Reliable source to sink communication is the most important factor for an efficient routing protocol especially in domains of military, healthcare and disaster recovery applications. We present weighted energy aware multipath reliable routing (WEAMR), a novel energy aware multipath routing protocol which utilizes hotline-assisted routing to meet such requirements for mission critical applications. The protocol reduces the number of average hops from source to destination and provides unmatched reliability as compared to well known reactive ad hoc protocols i.e., AODV and AOMDV. Our protocol makes efficient use of network paths based on weighted cost calculation and intelligently selects the best possible paths for data transmissions. The path cost calculation considers end to end number of hops, latency and minimum energy node value in the path. In case of path failure path recalculation is done efficiently with minimum latency and control packets overhead. Our evaluation shows that our proposal provides better end-to-end delivery with less routing overhead and higher packet delivery success ratio compared to AODV and AOMDV. The use of multipath also increases overall life time of WSN network using optimum energy available paths between sender and receiver in WDNs.

An enhanced backbone-assisted reliable framework for wireless sensor networks.

An extremely reliable source to sink communication is required for most of the contemporary WSN applications especially pertaining to military, healthcare and disaster-recovery. However, due to their intrinsic energy, bandwidth and computational constraints, Wireless Sensor Networks (WSNs) encounter several challenges in reliable source to sink communication. In this paper, we present a novel reliable topology that uses reliable hotlines between sensor gateways to boost the reliability of end-to-end transmissions. This reliable and efficient routing alternative reduces the number of average hops from source to the sink. We prove, with the help of analytical evaluation, that communication using hotlines is considerably more reliable than traditional WSN routing. We use reliability theory to analyze the cost and benefit of adding gateway nodes to a backbone-assisted WSN. However, in hotline assisted routing some scenarios where source and the sink are just a couple of hops away might bring more latency, therefore, we present a Signature Based Routing (SBR) scheme. SBR enables the gateways to make intelligent routing decisions, based upon the derived signature, hence providing lesser end-to-end delay between source to the sink communication. Finally, we evaluate our proposed hotline based topology with the help of a simulation tool and show that the proposed topology provides manifold increase in end-to-end reliability.