Artificial Intelligence for Sustainable Development

The idea of “smart agriculture” is still relatively new, but it refers to the use of modern technology to provide information about agricultural areas and then take appropriate action based on feedback from consumers. It combines important information and communication technology with sensor technologies to provide effective and efficient agricultural services. A wide range of cutting-edge technologies, such as cloud computing, robotics, drones, artificial intelligence, and wireless sensor networks are used in smart agriculture. Utilizing such technologies in intelligent agriculture may enable all agricultural stakeholders to make better managerial choices that will boost productivity. The deep fusion of modern information technology and conventional farming has resulted in the era of agriculture 4.0, sometimes referred to as smart agriculture, which promotes automation and intelligence. In this article, a survey of smart agriculture is focusing on various processing techniques in smart farming. The article also provides an overview of different technologies that are integrated with farming to make agriculture smarter. Finally, some general security issues and solutions are also mentioned in this paper.

Maximizing total profit by determining the most suitable crop for each piece of land in a complex agricultural region poses a significant challenge in cultivation management. However, numerous factors such as cost, yield, and selling price often lack clarity, rendering precise programming approaches impractical. The objective of this study is to identify an optimal crop prediction model that can assist farmers in selecting the appropriate crop based on local climatic conditions and soil nutrient levels. This research presents a novel hybrid model called the Modified Lion Optimizer (MLO)-based Convolutional Neural Network (CNN). By incorporating MLO, an efficient algorithm inspired by nature, this paper addresses the aforementioned challenge. The MLO algorithm assesses each solution using three metrics: expected, optimistic, and pessimistic values. This combination empowers the algorithm to discover reliable solutions even in the presence of uncertainties. Experimental results demonstrate that the proposed technique surpasses traditional deep learning models in terms of performance.

The participation of artificial intelligence (AI), particularly in the medical imaging field has enhanced the domains of health innovation. As the lack of condition which is unknown to the affected ones, there is no effective or suitable approaches for preventing and treating the breast cancer. Early detection may increase the possibilities of a full recovery from the disease. A timely analysis of an effective method of identifying and controlling breast cancer. The best method for early breast cancer identification is mammography. This device also makes it possible to identify additional diseases and may reveal details about the type of cancer, such as whether it is normal, malignant, or benign. Basic definitions of concepts like “machine/deep learning” are given in this article, which also examines how AI has been incorporated into radiology. With the advancement of digital imaging technologies, analyzing medical images to diagnose diseases has become increasingly crucial. Clinical medicine can advance through the smart segmentation, identification, and size categorization of breast cancer images using digital image processing technology. This research introduces approaches of medical image identification technology for breast cancer. The investigation of smart segmentation and deep learning for breast cancer is discussed.

Radiology is an expansive field that relies on specialized medical knowledge and insights to effectively identify brain tumors. Recent advancements in biomedical image analysis and processing techniques have made it possible to utilize Magnetic Resonance Imaging (MRI) for improved tumor detection. In this study, MRI images were utilized as input to identify the location of brain tumors, employing a segmentation and detection approach. This approach is complicated by the wide variety of tumor tissues seen in people as well as the similarity between normal and tumor tissues. On the other hand, using automated computer-aided techniques can significantly enhance tumor diagnosis. This research presents a deep learning model that uses a faster R-CNN with transfer learning for brain MRI image classification. The proposed method achieves a remarkable 93% classification accuracy, outperforming the other algorithms analyzed in the research.

More applications for wearable technology are being investigated and developed due to the significant technological growth of medical sensors and nanoelectronic devices. A new area of study is now possible because of wearable biomedical technology that has been merged with AI and ML. As it is used to track human behaviors and diagnose, this subject offers exciting potential. Neurodegenerative diseases (NDs) are becoming more prevalent in an aging society. The occurrence of a neurodegenerative illness takes place when the body’s central nervous system gradually deteriorates. Although this is rare, millions of people will be impacted worldwide. Despite the clinical importance of keeping an eye on ND’s symptoms, current practice makes it difficult since it is difficult to recall and describe symptoms effectively and because clinical sessions are infrequent. There are many neurodegenerative disorders among older people, such as Alzheimer’s diseases, Parkinson’s disease, and so on. So far, resting-state functional connectivity analysis has been followed to detect Alzheimer diseases (AD). Nevertheless, the Resting-state practical connectivity approach fails to take into account the distinctive features of different frequency bands, which encompass the brain’s most crucial atrophies. Hence, this work proposes an automatic Alzheimer disease detection algorithm based on their applications for various bands. Initially, the proposed detection algorithm has been learned using SVM and KNN to deal with AD disease. By adjusting different settings, we also explored additional machine learning and deep learning techniques and achieved high levels of accuracy. With only three bands, our suggested model performs well without external feature selection. The findings demonstrate that our approach is accurate (93.71%)/AUC (0.9363) in separating AD subjects from healthy controls.

Medical practice is being gradually transformed by artificial intelligence (AI). Diabetes mellitus (DM) is a disease characterized by inadequate control of blood glucose levels, and has the potential to create a healthcare crisis worldwide. Pregnant women who have gestational diabetes mellitus (GDM) may put their unborn children at risk. This form of diabetes can result in larger-than-normal offspring, making vaginal birth more difficult. This chapter presents a case study that uses an analysis of different machine learning methods to suggest a machine learning model for the early detection of gestational diabetes mellitus and the probability that it would proceed to type 2 diabetes.

In medical applications, deep learning has shown to be a powerful tool, especially when it comes to identifying patterns in healthcare datasets. Radiologists’ evaluation of CT images is crucial to the prompt identification of cerebral bleeding. The dataset used in this investigation included 3000 patients’ full-body DICOM CT scans. After segmenting these scans to separate the brain pictures, clustering was used to put them in groups according to visual similarity. This method increases the possibility of reliably and effectively identifying cerebral hemorrhage, which may have an effect on patient outcomes. Further convolutional neural network (CNN) is applied to find patterns in Brain CT scans of patients to correctly detect internal bleeding and classify hemorrhage and nonhemorrhage images.

Mental health disorders are becoming very common, and depression and stress have grown into a significant issue in our culture. The majority of the population is currently affected by depression, a highly serious and severe mental ailment brought on by a variety of factors, including stress from jobs, studies, personal relationships, other illnesses, and other factors. It is also known as serious depression and is a key contributing factor in suicide, particularly among teenagers. Even though depression is a highly prevalent disorder, it is nevertheless frowned upon to discuss it publicly. People are hesitant to discuss this illness for fear that others would think they are crazy. This resistance can occasionally be highly detrimental to the patient, advancing his condition to the point where he cannot be restored to health. According to WHO data, depression remains the second most common factor contributing to the world’s disease burden. Social media platforms possess shown to be a fantastic medium for people to talk about themselves in the emergence of such problems. Social media accounts can therefore reveal a lot about a user’s emotional condition and mental health. In the proposed study, we employ deep learning approaches to use social media to detect depressed individuals and quantify their depression intensity. The objective-setting method divides the strategy into two parts: the first one depends on the content’s time and reviewing the posted content or tweet for analysis, and writing patterns. With the use of various word embedding approaches and suggested metadata features, the performance of deep learning algorithms that were trained using training data is assessed to have results. The predictive strategy is used to identify depression or additional mental illnesses early on. The goal is to assist individuals with this condition in recognizing depression’s early signs, which could be advantageous to both them and their families.

The transportation industry has become a significant contributor to the rising usage of fuel as well as greenhouse gas (GHG) emissions. In order to overcome the problems, we have introduced Electric vehicles (EV) which are an alluring answer to such issues. The significant penetration of electric cars may result in various issues with the distribution network and its dependability owing to the fluctuation in charging demands. Therefore, a variety of strategies are used to forecast the demand for charging EVs and minimize the associated difficulties. Artificial intelligence (AI) approaches are very interesting for the development of electric vehicles (EV) as well as their energy management systems (EMS). Because of EVs high potential for performing complicated parameterization jobs in an efficient manner, AI approaches can be a perfect option. The goal of this article is to offer a comprehensive understanding of smart energy management techniques by reviewing the literature in these domains. EVs should have charge schedules to communicate with power sources, and charging stations and manage charging schedules. Blockchain technology and federated learning (FL) are two new approaches to handling data privacy issues. The analysis of different machine learning approaches for current EV energy management and charging of vehicles, as well as energy trading and challenges of EVs analyzed through the literature is presented in this paper.

Artificial intelligence gives urban designers and planners the tools they need to develop flexible urban landscapes with the use of real-time data-supported methodologies. Artificial intelligence (AI) real-time data can help to fully integrate sustainable practices into the urban environment. Due to flawed computations and information research tools, this enormous volume of information is exhausted without differentiating possible lessons. Urban planners can use AI to help them choose routes that will improve traffic management, provide equitable public transportation, and have smarter utilities. An innovative and flexible learning framework for smart cities is explored with the aim of adding a new dimension to educational conversations. The multilevel nature of the massive amounts of data generated by smart cities is capitalized on by this design. With the application of AI, city planners may design more responsive and efficient “smart cities” for their constituents. This chapter covers the foundational AI architecture for urban computing applications, as well as current research on urban deep learning and machine learning.

To meet the demands of a future society and enable new applications, sixth-generation mobile networks (6G) are expected to achieve previously unheard-of communication capabilities. Artificial Intelligence (AI) and Machine Learning (ML) can greatly contribute to the intelligent computing aspects of these futuristic communication systems. In this chapter, we highlight the (i) directions of advanced communication that AI can support (ii) trending AI-based techniques, algorithms, and advanced models for 6G and (iii) a study on spectrum management using intelligent cognitive radio and reinforcement learning technique.

Intelligent transport is now a vital facilitator of the smart city concept, with navigation serving as a crucial component. Recently, software as well as hardware research have focused a lot of emphasis on automated vehicles (AV). The AV now provides more flexible and efficient industrial and transportation system solutions. The navigation strategy used by an AV is crucial to its functionality. Even if using AV navigation seems appropriate and sufficient, making the choice is not simple. Systems based on the Automated Vehicles (AV) arise to offer car users navigation services, and they are distinguished by using Roadside Units (RSUs) to gather data on the state of the roads from surrounding vehicles. However, the design of the vehicle navigation systems must adhere to strict performance standards for autonomous operations. Relevant performance metrics take into account the system’s accuracy as well as its capacity to identify sensor problems within a given Time-to-Alert (TTA) window and without producing a false alarm. However, they solely concentrate on collecting features from the traffic patterns of isolated or nearby intersections, despite recent research using deep reinforcement learning algorithms for traffic light control showing promising outcomes. In this paper, different navigation techniques and algorithms that are used for smart navigation are explained. In order to provide readers an idea of how Reinforcement Learning may be used in an autonomous vehicle for navigation is discussed in this article incorporation of artificial intelligence (AI).

The frequency of roadside accidents is rising quickly along with the number of automobiles. The driver’s irresponsibility is to blame for the bulk of these collisions. The current communication networks in the automotive ad hoc networks face enormous obstacles due to the constantly rising traffic, numerous delay-sensitive services, and energy-constrained needs. Researchers from all over the world are constantly creating new protocols and architecture for intelligent transportation systems. As a result, many nations are increasingly embracing and investing heavily in vehicular ad hoc networks (VANET) in order to assure the safety of drivers. On the other hand, before VANET technology is widely used, there are a number of problems in this area that need to be fixed. Numerous attacks may take place in the event of low or no security, which could impact the system’s dependability and effectiveness. Software-Defined Networking technology (SDN) was introduced to increase the effectiveness of VANET systems. In this article, a short study on Intelligent reflecting surface (IRS) and artificial intelligence (AI)-enabled energy-efficient communication solution for SDN vehicular Network is explained. SDN-based VANET was the quick term for this method. The data transmission in AI-based IRS and the three planes in this framework are also explained in this paper. The cluster head selection of a vehicle and communication between vehicle to vehicle are also provided in this article. To enhance vehicular communication, an IRS-aided data transfer is suggested.

The present attention of computer vision study is on AI emotion identification, which comprises the automatic acknowledgment of facial terminologies of feeling and the evaluation of sentiment in visual database. In order for artificially intelligent systems with visual capabilities to comprehend human interactions, the study of human–machine interaction is essential. Artificial emotional intelligence, sometimes referred to as affective computing and emotional AI, is a subfield of artificial intelligence that concentrates on the comprehension, examination, and replication of human emotions. Its goal is to advance the sincerity and organic nature of interactions between people and robots. Textual content, voice tone, facial expressions, and gestures are just a few of the cues that emotional AI uses to understand people’s emotions and alter its answers accordingly. Using computer vision technology, Visual Emotion AI analyzes facial expressions in photos and videos to determine a person’s emotional state. This study uses both audio and visual inputs to investigate the recognition of emotions using artificial intelligence.

The foremost aim of the recommended, automatic identification of handwritten Tamil text is offline identification of the manually written Tamil script in an editable display. The novel system used to obtain the best accuracy rate with less processing time by the transfer learning skill with the Inception-v3 deep learning model. This suggested research plan uses the modified RAdam optimizer, which optimizes the entire model in a better way to achieve stable training and model generalization. An attempt initiated to improve the recognition accuracy with the help of spelling error detection and correction of the recognized Tamil text in the post-processing stage. Additionally, the identified manually written Tamil content converted into the English language with an altered Google translation framework. Experimental results on the Tamil handwritten text identification verify the perception and prove the effectiveness and robustness of the modified RAdam optimizer. The trial outcomes achieved from this proposed research work reveal the efficiency of the automatic Tamil script identification system in an enhanced height.

Object detection or identification is the one of the fundamental problems in computer vision application. Even though, object detection is a successful research area, detection of small object from remote sensing images is complicated. Remote sensing image-based automatic ship detection is part of marine surveillance system. Marine safety is also one of the main security sectors for national security. So, to avoid the risk of pirates and extremists entering the harbor zones, early detection of ship is necessary. Similarly, when there are accidents of ships in maritime, identifying the ship is a challengeable task. So, when considering oceanic security and safety, automatic detection of ship is obligatory. The deep learning model, particularly MobileNet, without forgetting architecture, was considered for automatic and early ship detection. The experimental setup produced the 98.2% accuracy rate with Kaggle ship dataset. The experimental setup is evaluated with the performance analysis and finally compared with some other techniques.

Industries are changing in ways that are unimaginable due to the rapid progress of Artificial Intelligence (AI) technology. It has become the keystone that will drive several industries into previously unheard-of levels of growth in the future. Artificial Intelligence (AI) has a great deal of promise to transform military capabilities in various fields. AI’s importance is growing in a world where terrorist activity is a persistent danger to both international peace and national security. Its capacity to handle enormous volumes of data, decipher intricate patterns, and act quickly gives defense forces the advantage they need to remain ahead of the constantly changing security scenario. To protect public safety and international peace, AI-driven systems can help with threat identification and mitigation, improved surveillance and reconnaissance, and faster response times. With AI’s further advancements, the military industry stands to benefit greatly from its ability to combat the growing dangers posed by terrorism. It is clear that integrating AI into military operations gives a revolutionary advantage, supporting efforts to make the world a safer and more secure place overall. This study presents the significance of AI toward various domains of military sector, its supporting algorithms, applications, and future directions for rapid decision-making and resource support.

Natural disasters possess the capacity to cause substantial and extensive harm, resulting in noteworthy economic ramifications. Interestingly, there has been a noticeable increase in the amount of loss and damage brought on by these occurrences in recent years. As such, disaster management organizations have an even greater need to proactively protect communities through the development of efficient management plans. Artificial intelligence (AI) approaches have been used in a number of research projects to analyze catastrophe-related data, improving the caliber of decision-making related to disaster management. The volume and diversity of data from satellite photography make it difficult to comprehend, despite the large amount of data it offers for a variety of uses. Manual ground inspections are usually required for damage assessment, which is a time-consuming and ineffective procedure. To address these issues, this work presents a novel deep learning algorithm for classifying buildings in satellite photos as damaged or undamaged.

Humans are directly and indirectly impacted by the frequency and severity of natural disasters due to climate change Rather than relying solely on event response, proactive preparedness actions are needed to prepare they have given themselves to the worst. Steady measures must be taken immediately to protect lives and property from unforeseen natural disasters. Survivors frequently suffered from trauma leading to anxious thoughts, persistent anxiety, disturbed sleep, and depression before, during, and after such events. Geographic Intelligent System (GIS), Integrated Decision Support Systems (IDSS), and other sophisticated technologies such as cloud computing and artificial intelligence (AI) are proposed in this case as tools to enhance readiness They can provide more efficient ways of recovery. Based on the findings, GIS and IDSS are two important tools for disaster management. Subsequent studies have shown that AI and cloud-based collaborative systems are viable options for managing natural disasters and extreme weather events.

A group of algorithms known as Nature Inspired Optimizers (NIO) are inspired by how things behave in the natural world. Animal actions, biological processes, chemical reactions, etc., have all served as inspiration for NIOs. Natural methods are easily divided into numerous intricate subprocesses. Because of this, every algorithm is distinct and potent. By addressing problems with selection of algorithms, parameter tuning/modification, and updating problems, the learning of NIOs aims to expand the performance of nature-inspired algorithms. Natural occurrences are constantly changing, and eventually, an algorithm’s behavior must also change. As a result, we must continually discover and modify existing ones. The genetic algorithm, Artificial Bee Colony Algorithm (ABCA), Bat Algorithm (BA), Grey Wolf Optimizer, and others are samples of NIOs. An in-depth study of various NIOs and their use in different Artificial Intelligence (AI) scenarios are provided in this chapter. This chapter offers a familiarized assessment of the novel NIO while identifying and examining the main difficulties experienced in creating NIOA.

Artificial intelligence (AI) is a vital component that helps businesses run more efficiently by organizing corporate data and optimizing workflows. Lately, scholars from several disciplines have been investigating how artificial intelligence affects business results. Businesses that are able to create value over long periods of time do so by using AI technology to strategically integrate, adapt, and reinvent their business models. The findings of the study provide insight into how artificial intelligence is revolutionizing the commercial world. The main insights on AI support for business modeling using both cutting-edge and current AI tools and technology are presented in this chapter. The work ends with a presentation of the general AI framework for commercial applications and a case study of an organization that now offers products or services that incorporate AI technology.

Human Activity Detection (HAD) refers to the process of categorizing and identifying human motion. HAD studies have received significant attention in the estimation of frequency and duration. This has facilitated the use of sophisticated assistive technologies and significant manual examination. The artificial intelligence used by HAD enables experts to deduce human behavior based on data obtained from sources like wearables or physical objects. Deep tissue massage is becoming more often used for HAD (healthcare-associated infections), particularly in relation to everyday tasks and functions. HAD primarily focuses on discerning activities such as ambulation, locomotion, leaping, and engaging in recreational activities. As part of this research project, a new Convolutional Neural Network framework for HAD has been developed. The results of the experiments show that the proposed model performs better than traditional machine learning techniques in correctly recognizing and classifying human actions.

One of the natural resources in the world is forest, because the ecosystems are directly influenced by the forest area. So, to preserve the strong and blossoming ecosystem, forest maintenance is undoubtably necessary. In recent time, especially at summer season, forest fire is one of the serious problems. This forest may collapse the eco system. So early detection of fire in forest area is important. The computer vision with artificial intelligence is one of the successful research areas, which is still providing exact solutions for many problems. Accordingly, this chapter proposed Vision Transformer Model (ViTM) for early forest fire detection. The performance of the proposed system is analyzed with various performance metrics successfully. The planned early fire detection model produced 98.7% accuracy on the classification of fire and no-fire dataset. From the result, we can conclude that the proposed ViTM is an exclusive fire detection model.

The forest is a natural territory for wildlife species like animals, birds, trees, and shrubs. Precision monitoring of wild animals and birds is important to save the biodiversity. Wild life monitoring is important to maintain the richness of biodiversity; to save the rare animals, birds, and plants; and to monitor the population of wild species. The manual monitoring and surveying are difficult due to the vast area, which is also a time-consuming process. So automatic wildlife monitoring system introduced a modified YOLO model, Recurrent You only look once (ROLO) for surveillance. The experimental results present a hopeful result in identification of wild animals and birds. Particularly, the model setup was compared with other deep learning models, and proved that the proposed model has high precision rate than other models.