Keywords

1 Introduction

The network of devices that are linked to the Internet is referred to as the Internet of Things (IoT). IoT is an umbrella phrase that refers to any device that is able to transmit and receive data via a network. In most cases, these gadgets take the form of sensors that monitor various physical parameters, such as the amount of light, humidity, or temperature. IoT is continuing to proliferate across a variety of corporate sectors, which has resulted in the development of brand-new possibilities for cooperation and innovation. Devices are able to gather data, take actions depending on that data, and then report the results of those activities back to a centralized server. This technology has been around for quite some time, but only lately has there been an explosion of new inventions in this sector of the industry. Artificial intelligence (AI) and the Internet of Things (IoT) are the two primary motivating factors behind the recent uptick in interest about linked gadgets. When coupled, AI and IoT provide even more opportunities for analyzing data and discovering new insights into how we may enhance the quality of our lives. In this article, we will discuss AI and IoT, focusing on how these two technologies might collaborate to provide innovative answers to common challenges. Some of the core areas where IoT has been used at its peak are as follows:

  1. 1.

    5G: The next generation of wireless communication, 5G, will increase Internet access and network connection. The innovative design of 5G allows it to simultaneously connect more devices at greater speeds with less delay. This cellular IoT application was developed for the advantage of users at minimal cost and with increased speed. Using this innovative program, we can fully automate several sectors with the aid of smart grids. In the future, 5G will expand to support larger devices across wider regions, bridging the gap between smart cities and wireless vehicle communication. The public and commercial sectors may both profit greatly from this potential use of IoT. With 5G, it will be possible to remotely manage even more types of devices via software. The infrastructure for smartphones, tablets, and other mobile devices, as well as for sensors, medical equipment, and automobiles, is being established by IoT and 5G.

  2. 2.

    The Internet of Things and Augmented Reality: The connection between IoT and augmented reality is becoming stronger all the time. While IoT bridges the gap between physical assets and digital infrastructure, augmented reality (AR) brings digital components into the physical world. The future of AR and IoT seems bright in the medical field. Surgeons may, for instance, use software designed to reconstruct a body part in three dimensions, together with equipment monitoring the required statistics in real time. This software can be found on mobile devices. All these things could make difficult operations simpler for surgeons and more rewarding for patients who are patient.

  3. 3.

    Smart Cities on the Rise: The popularity of smart city technologies is at an all-time high right now, and analysts predict that interest in and investment in smart city technology will continue to soar in the not-too-distant future. City councils and other local authorities believe that smart city solutions are effective ways of involving the general public in the day-to-day operations of municipal administration and maintenance. Providing more comfort and convenience for the populace is just a minor portion of the whole bargain. Interactivity is being given a significant amount of consideration in the development of smart city initiatives all around the globe. There is a growing consensus that engaging individuals in a manner that requires greater hands-on participation is beneficial in all matters pertaining to the day-to-day maintenance of an urban space, which pays huge dividends for everyone concerned. Safe, habitable, and environmentally sustainable cities will not be possible without IoT-based smart city technologies. As city populations are expected to grow dramatically in the future, these strategies will become even more important.

  4. 4.

    Blockchain Technologies and Information Security: One of the most important developments in IoT technology is blockchain. Put together, IoT and blockchain technology are living up to their hype. As of today, the reliable exchange of money and data between IoT devices become possible once the blockchain technology provides them with a simple infrastructure for doing so. When these two phenomena are put together, they live up to their hype. The decentralized nature of blockchain is analogous to the dispersed nature of the Internet of Things. The latter provides anonymity and security to numerous networks and the owners of those networks, while the digital signatures and private keys that accompany each transaction assure that the environment of the Internet of Things will be secure. IoT devices are intended to simplify a person’s or an operational unit’s day-to-day activities in some way. This necessitates the ongoing creation of personal data and introduces significantly increased opportunities for cybercriminal activity. In addition, it is anticipated that the quantity of data will exponentially increase in the same way as the number of machine-to-machine interactions increase. As a consequence of this, the decentralized structure of blockchain will cause massive numbers of data to be accessible whenever they are required and with a minimal number of associated security issues.

All the aforementioned applications require the use of IoT in various domains because IoT acts as a bridge between hardware sensors and the digital infrastructure. This work will focus on all the applications that employ IoT device to obtain real-time data, which will be further fed into the AI-based prediction models. In this review analysis, seven domains have been identified where AI and IoT are being employed: healthcare, sustainability, information security, education, pollution monitoring, robotics, and autonomous vehicles.

Healthcare Applications

The IoT of the future will make it much simpler for doctors and nurses to keep track of their patients’ activities and vital signs. This will be made possible by advancements in 5G wireless technology, AI, and sensor technology. A smart glucose-monitoring system and smart insulin pens will also be of assistance since they will automatically transmit the patient’s important information to a monitoring system. This will serve to direct assistance, particularly in the situation involving insulin. The pen will be able to evaluate the quantity of insulin that has to be administered on the basis of the data that is derived from the patient. Security and patient safety are of the utmost importance in the clinical setting, and IoT may assist in enhancing the monitoring and transfer of patient data. The whole world has been forced to reconsider the significance of remote healthcare. It won’t be long until patients and physicians won’t even have to physically interact with one another, which might be quite helpful in times of lockdown. IoT developments will soon have stronger effects on healthcare, which will result in the proliferation of increasingly intelligent medical equipment.

Better Way to Store Data and Perform Data Analytics

We’ve just recently become used to storing information on the cloud before it became necessary to make a switch. Edge computing, in its simplest form, enables interconnected gadgets to carry out computations, store results, and locally view results. Edge computing is a hybrid method of data processing that is altering the trajectory of the Internet of Things. The actual value that IoT may bring about via data analysis is in managing and analyzing these data. Therefore, machine learning and artificial intelligence will play increasingly crucial roles. These developments will aid in making our lives simpler and more comfortable and will provide efficient methods for completing jobs.

2 Recent Works on IoT and AI in Various Domains

2.1 Healthcare

The healthcare industry has always relied on a small number of centralized agents freely disseminating raw data to the public. This system still faces substantial threats and weaknesses. With AI, however, the system would consist of several agents working together and effectively interacting with their preferred host. The most cutting-edge and fascinating innovations in the area of intelligent healthcare include federated learning (FL), AI, and explainable AI (XAI). FL operates in a decentralized way and keeps the communication based on a model in the favored system without transmitting raw data. Multiple healthcare constraints may be alleviated with the integration of FL, AI, and XAI methods. In [1], Rahman provides a comprehensive evaluation of FL as it relates to the use of AI in forward-thinking medical settings. They used FL-AI in several healthcare technologies and categorized the results.

Using multisensing, edge-based, and on-device AI components, T. Montanaro et al. [2] constructed a real-time IoT-aware healthcare system that comprises three layers: an edge computing layer, a data visualization layer, and an intelligent data-acquisition layer. (i) Three sublayers make up the intelligent data-acquisition layer. (1) Advanced sensors include motion, temperature, location, and electric charge buildup in skin sensors. These devices gather data for deductions. (2) Computation and data preprocessing devices control the selection and collection of sensor data. (3) AI modules contain microcontrollers linked with AI algorithms. These algorithms identify data irregularities, accurately differentiate between behaviors and between people, classify measured data, etc. (ii) The edge computing layer receives data from the preceding layer, provides a gateway to the top layer, handles multiple protocol communications, and performs additional analyses. This layer may accept data from the bottom layer, transmit changes to it, transfer data to the top layer, and receive notifications and updates from the upper layer. This layer allows communicates with employers, families, and intimate partners. (iii) The data visualization layer connects storage and user interactions. Web dashboards provide authorized users with local device warnings and historical occurrences. Healthcare providers may utilize this dashboard to give people ideas. Advanced data analysis is also possible. Finally, REST APIs connect with lower levels and contain a database to store data.

M. M. Kamruzzaman et al. [3] identified new difficulties, possibilities, case studies, and edge-AI applications for linking healthcare in smart cities. Relevant publications and journals were studied, analyzed, and appraised, and this review also included secondary data sources such as Google Scholar, Science Direct, etc. Only papers including AI, edge AI, IoT, and deep learning (DL) were reviewed. The study selection and data abstraction yielded 22 relevant articles/research papers, which were grouped into two subtopics: edge AI and healthcare in smart cities. They addressed how the machine-learning (ML), DL, and IoT models could be used in healthcare. The accuracy of the models implemented in various research papers was assessed. A. Alghamdi [4] developed a VGG-Net model for analyzing electrocardiogram (ECG) images. VGG-MI-1 showed sensitivity, specificity, and accuracy values of 98.76%, 99.17%, and 99.02%, respectively, and the VGG-MI2 model showed sensitivity, specificity, and accuracy values of 99.15%, 99.49%, and 99.22%, respectively, which were the best so far.

Zhao-xia Lu et al. [5] examined the technical aspects of IoT, cloud computing, big data analysis, and machine learning in clinical medicine. They highlighted the application of AI and IoT in various medical scenarios and discussed challenges and future prospects in this rapidly evolving field. COVID-19 exacerbated the global scarcity of nurses and physicians. Automated tools employing IoT, cloud computing, ML, etc. may be helpful. IoT devices monitor real-time data and send them to the cloud. 5G has increased medical staff’s accuracy and speed. Cloud computing services include processing power, machine learning, and storage. They provide data exchange, remote consulting, etc. Large data sets are analyzed by using big data. They can cluster, classify, and visualize data and can mine text. Clustering sorts data by proximity. Classification mining maps data to labels by using decision trees, neural networks, etc. Text mining methods shape unstructured medical data through preprocessing, segmentation, and semantic analysis. Visualizing data uses charts and graphs. This may prevent, diagnose, and cure illnesses. Machine learning creates data-driven models. High-dimensional, high-variance data are analyzed. Medical analysis uses supervised and unsupervised learning. Preprocessing techniques express data by using piecewise linear representation and Fourier transform. By using the sliding window, top-down, bottom-up, and other methods, time series data are segmented, but sliding window is the best method. Diminution involves feature selection and change, k-clustering, SVMs, etc. Decision trees, k-nearest neighbors, SVMs, naïve Bayes, etc. identify and classify data. Behavior detection monitors patient behavior by using CNNs or RNNs. Unsupervised learning detects abnormal values. Clustering is simpler. Patient data privacy is also vital. K anonymizes identifiable information but not attributes. L-diversity and T-closeness models reduce granularity. Many clinical uses of IoT and 5G are have been explored. IoT and 5G allow remote diagnostics. Thanks to using IoT patient data and ML graphs, medical practitioners can cooperate. Neonatology, cardiology, and skin cancer screening have been used with near-professional precision. ML may be used for supplementary diagnostics, triage, and alerts for patient vitals. Patients’ histories and conditions may be used to produce exercise and food regimens. Cloud-based medical photos may be examined by ML systems to diagnose patients. IoT data, cloud sharing, and ML and AI analysis may assist in diagnosis, spot warning signals, notify emergency services, and provide remote services. This study examines IoT-assisted wearable sensor systems, AI, blockchain difficulties, and other concerns that must be addressed to improve their use in Health management system (HMS).

The study by Junaid et al. [6] aimed to explain the need for and uses of new technologies (sensor-IoT-AI-blockchain) in the healthcare sector by analyzing these technologies as well as past approaches and methodology. In the healthcare administration system, Junaid et al. [6] utilizes papers and survey research on topics such as sensors, IoT, AI, and blockchain. The majority of the data used in their study came from a network of intelligent wearable IoT devices. These devices monitor a patient’s vital signs and other pertinent data and display them in real time. For instance, the LIFE Shirt is a multisensor extended HMS that collects and analyses a patient’s health data. This information is obtained from the patient by means of the device. In addition to that, an AI-based data synthesis was carried out in order to provide data for testing and validation purposes. There are no well-known data sets that are available for purchase in a prepackaged manner. A smart health ecosystem that limits access to a patient’s electronic health record (EHR) might be built with the use of smart contracts. As a result of this research, Junaid et al. [6] concluded that a single ledger that is maintained by healthcare stakeholders may record a patient’s entitlements, and smart contracts automate information gathering and distribution and calculate benefits in real time. The acquisition and analysis of real-time patient data from hospital and home devices is facilitated by using sensing technology. It’s possible that a real-time analysis may increase the accuracy of patient safety monitoring and incident prediction. When using ML, decisions are made on the basis of studying the data rather than on the basis of making intuitive assumptions. The proposed HMS is able to function by collecting data from its users via the use of smart wristbands and then feeding that information into an artificial neural network (ANN) for risk assessment. The findings of the studies demonstrate that the suggested HMS is capable of accurately assessing the health states of patients.

Ramasamy et al. [7] offered an AI-enabled combination of the Internet of Things with a cyber-physical system (IoT-CPS) for doctors to identify illnesses in patients. Human intelligence improves AI. Computers are better at arithmetic and numbers, whereas humans excel at logic and reasoning. AI might make things “thinkable.” Two algorithms make up the AI-powered IoT-CPS algorithm. The first component of the algorithm generates classification rules by categorizing the training data set of patients’ illnesses. The second part of the algorithm classifies patients according to their symptoms in order to make disease predictions for the disease-testing data set. The experimental findings show that the suggested AI-enabled IoT-CPS method outperforms state-of-the-art algorithms on accuracy, precision, recall, F-measure, and execution time when diagnosing patients’ illnesses. Figures 1 and 2 display the IoT-CPS process.

Fig. 1
Four illustrations explain that the medical data of the elderly person slash patient in the home is collected and transmitted via the cloud server to duty doctors in the hospital and the son of the elderly person working in the office.

IoT-CPS [7]

Full size image
Fig. 2
A flowchart is as follows. Start, disease training data, apply disease classification algorithm, yes, classification rules, apply D P A, sensed data, yes, predicted patient disease, end, apply disease classification algorithm, no, end.

Flowchart of an AI-enabled IoT-CPS [7]

Full size image

The importance of the Internet of Things and artificial intelligence in COVID-19 was the primary focus of the research carried out by Praveen Kumar et al. in 2022 [8]. The Internet of Things was studied on a new level: how it manages the COVID-19 epidemic. In this research, long short-term memory (LSTM) with a recurrent neural network (RNN) was employed for diagnostic purposes because this architecture is particularly helpful in assessing the acoustic aspects of coughing and breathing. Real-time data on a patient’s temperature and respiration levels are gathered via the use of a large number of IoT sensors.

The purpose of this chapter is to examine the use of various sensing devices to gather various types of information. The research focuses mostly on using AI systems on data sources. The major objective of this work is to gather data on the temperature and respiration rate of COVID-19 patients in real time by using a number of sensors.

Severe Acute Respiratory Syndrome Coronavirus (SARS-CoV), Middle East Respiratory Syndrome Coronavirus (MERS-CoV), and Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) constitute serious public health hazards. These viruses risk countless lives and inflict economic damage. Recent IT and networking advances have led to IoT and AI applications in numerous sectors. The healthcare and diagnostic sectors are impacted by IoT and AI. By interfacing with smart gadgets and biometric sensors, they have expanded into telemedicine, healthcare, and disease prevention. Even though IoT and AI may improve disease diagnoses, surveillance, and quarantines, their influence is limited insofar as they aren’t integrated or deployed quickly for a sudden outbreak. Conventional procedures fail to prevent large-scale illnesses and halt worldwide outbreaks via prediction, resulting in many deaths. Sungho Sim et al. [9] proposed a combined Internet of medical Things and AI (IoMT-AI) framework model to handle COVID-19 outbreaks. IoMT uses (1) remote monitoring, (2) prescriptions tracking, and (3) biometric sensors to provide health data to doctors. Image signal processing and virus infection detection using AI can carry out disease surveillance, risk prediction, medical diagnosis and screening, curative research, virus modeling and analysis, and the management of lockdown measures, as shown in Fig. 3 (Table 1).

Fig. 3
An organizational chart of artificial intelligence categorizes into the disease surveillance, risk prediction, diagnosis and screening, curative research, virus modelling and analysis, and enforcing lockdown measures.

Applications of AI in medical IoT [10]

Full size image

2.2 Sustainability

Internet of things (IoT) makes it possible to digitalize a wide variety of tasks and procedures, including the distribution of water, preventive maintenance, and smart manufacturing. Although IoT technologies and ideas such as edge computing hold a great deal of potential for the digital transition to sustainability, they are not yet making contributions to the sustainable development of the IoT industry. This sector has a significant carbon footprint because it makes use of a limited supply of raw materials and a significant amount of energy in its production, operation, and recycling processes. However, its sustainable vision collides with edge artificial intelligence (edge AI), which demands more energy, which is why the green Internet of Things (GIoT) paradigm has emerged as a study topic to reduce carbon footprints. In article [11, 12], the authors investigate the process of designing and developing edge-AI GIoT systems. The concepts discussed in the paper are highlighted by using a real-world example of an Industry 5.0 application case. To enhance operator safety and operation tracking, a smart Industry 5.0 workshop should be held. This application takes advantage of an IoT mist architecture that is equipped with AI. After the application situation has been explained, the energy consumption and carbon footprint of the application are analyzed. This article offers guidelines for aspiring developers who want to design edge-AI GIoT systems using the aforementioned technologies (Table 2).

Table 1 Recent works on IoT and AI in healthcare
Full size table
Table 2 Recent works on IoT and AI for sustainability
Full size table

2.3 Information Security

The Internet of Things (IoT) has altered how humans live and has permeated all facets of human life, but it has also given rise to worries over data security, which in turn have led to a variety of technical ethical and security difficulties. For the information security of a company, these problems often center on the demand to obtain user data and maintain privacy when the IoT is deployed. This requirement becomes more sensitive when discussing the security of corporate information. The development of AI-based security solutions for IoT sensor networks has been a recent trend.

In [27, 28], the authors built a platform for the administration of information security that was comprised of four parts: the management of IoT data mining, management of equipment, management of keys, and management of databases. Testing for concurrency, stress, high data volume, and security were improved, together with the original architecture for physical security, which was also reorganized.

The purposes of this chapter are to provide some fresh suggestions for applying information security technology to IoT-based corporate management and to advocate for the use of IoT in industrial and commercial management in the near future. The following features are included in the functionality of the platform: authorization and revocation, staff scheduling, data storage and backup, rank role administration, and data encryption and mining (Table 3).

Table 3 Recent works on IoT and AI for information security
Full size table

2.4 Education

In most cases, the services and administrations that rely on traditional libraries cannot function without them. A reader must follow a multistep procedure to borrow books, which includes entering a library to preserve books of interest, bringing books to a certain spot (i.e., the circulation desk), showing the librarian their identity for verification, and lastly confirming the books to borrow. The many occurrences of the targeted books being borrowed by others throughout the preceding method make it difficult to ignore them. If a reader doesn’t know in advance which branch has the books they’re looking for, they’ll have to make many trips to different libraries. However, in the era of the “smart library,” obtaining books requires only a few steps via a smart terminal device: confirm and make an appointment for the intended books and fetch the books under the smart guide and devices furnished by the library. This is the optimal schedule for the borrowing process. Thanks to the immense potential of AI and IoT, library circulation efficiency has greatly improved over more-conventional approaches.

One study [29] conducted an extensive literature review on the use of AI and IoT technologies in proposed future smart libraries in order to formally offer a systematic, organized, and comprehensive strategy for such a possible topic. When compared to librarianship that relies on only human effort, the smart library model, which incorporates AI and IoT, provides far superior service, as seen in the aforementioned smart circulation service. Smart service, smart sustainability, and smart security are the three primary foci of the author efforts. There is a wide variety of use cases for AI and IoT in a smart library. According to the findings of [29, 30], its authors have formally described the development of the smart library (Fig. 4).

Fig. 4
A flowchart is as follows. Data preparation with C R FID, data encoding with data structure transformation, R N N structure with deep learning structure design, and demand behavior recognition with the performance evaluation.

The workflow of the personalized activity-learning system [31]

Full size image

Machine-learning and deep-learning techniques, including natural language preprocessing, deep-learning models, recurrent neural networks, and deep-learning-based recommendation systems, have been used throughout this study.

Using IoT with the help of artificial intelligence may significantly lessen the likelihood that sensitive data or valuables will be compromised. A new authentication system is being suggested to ensure the anonymity of readers. IoT and AI can work together to create a sustainable timetable that takes into account real-world requirements. The article proposes many AI-assisted IoT methods for controlling the library’s lighting in order to maximize the efficiency of its use of natural light (Table 4).

2.5 Pollution Monitoring (Table 5)

Table 4 Recent works on IoT and AI for education
Full size table
Table 5 Recent works on IoT and AI for pollution control
Full size table

2.6 Robotics (Table 6)

Table 6 Recent works on IoT and AI for robotics
Full size table

2.7 Other Related Works

Apart from the above-mentioned domain, Several papers have been identified that are relevant to the theme. Gültekin et al. [32] presented a deep learning approach based on multisensory data fusion for fault diagnosis in an industrial autonomous transfer vehicle. They focused on real-time fault detection and condition monitoring using edge artificial intelligence for industrial autonomous transfer vehicles [33]. Malik et al. [34] discussed the latest trends in the design and application of smart antennas. Rogers and Malik [25] explored the opportunities and challenges of planar and printed antennas for IoT-enabled environments. Abdul Rahim and Praveen Kumar Malik [36] analyzed and designed a fractal antenna for efficient communication networks in vehicular models. Shaik and Malik [39] conducted a comprehensive survey on 5G wireless communication systems, addressing open issues, research challenges, channel estimation, multi-carrier modulation, and applications. Malik, Wadhwa, and Khinda [40] conducted a survey on device-to-device and cooperative communication for future cellular networks. Tiwari and Malik [41] focused on wideband microstrip antenna design for higher “X” band frequencies. Kaur and Malik [42] presented a study on multiband elliptical patch fractal and defected ground structures microstrip patch antennas for wireless applications. Shaik and Malik [43] retrospectively analyzed channel estimation techniques for 5G wireless communications. Finally, Malik and Singh [44] proposed a multiple bandwidth design of a microstrip antenna for future wireless communication. These papers collectively contribute to the understanding and application of IoT and AI in various domains.

3 Conclusion and Future Directions

IoT and AI have offered wonderful opportunities for organizations to enhance efficiency, visibility, and profitability. AI and IoT are altering our world. More companies than ever are making use of IoT. ML, AI, rapid feedback, and remote monitoring are here and not slowing down. Businesses that embrace the IoT revolution early avail themselves of several possibilities. The intention of this work is to present various new future directions in the field of AI and IoT by exploring their applications in various domains, such as healthcare, education, sustainability, robotics, pollution control, and much more. Every application of AI and IoT has unique advantages and disadvantages. But security and privacy concerns are the most common shortcomings emerging as open challenges. These will be stepping stones for future research. More real-time case studies are needed in order to measure the real-time effects of AI-assisted IoT on the environment.