Emerging Trends in Artificial Intelligence Based IoT: Techniques, Applications and Security

The growth of the Internet of Things (IoT)-embedded intelligent systems with both the industrial perspective and research enables communication and data transfer among heterogeneous devices in real-time without any physical human intervention. These devices are resource-constrained devices in terms of power and storage. This autonomous and constrained behavior of smart IoT devices leads to privacy and security vulnerabilities. The existing cryptographic techniques for security and privacy preservation are not suitable enough to deal with the issues of these systems. Blockchain technology provides a platform for decentralized computation and secure information sharing in a distributed manner among multiple authoritative participants. These participants do not trust each other but cooperate, coordinate, and collaborate in a rational decision-making process. Blockchain uses a cryptographic environment to secure the data and network. Different characteristics of blockchain technology, such as decentralization, immutability, transparency, security, and reliability, make it the ideal and effective candidate for integration with smart IoT systems. The main objective of this paper is to analyze the use of blockchain technology and its concept in the context of smart IoT to improve capabilities and enhance its security. This paper first highlights the introduction of smart IoT with security and privacy concerns for these systems. This is followed by a detailed systematic analysis of blockchain architecture, its characteristics, consensus algorithms, various platforms, and application areas in smart IoT with the aim of analyzing the concept and workings of blockchain technology. Various benefits and challenges in the integration of blockchain with IoT are then analyzed and illustrated.

Blockchain technology, with its globalized features like a decentralized consensus and tamper-proof immutability, expands the liberty of miners through smart contracts to participate in the Blockchain network. The method of decentralized consensus creation, which involves all relevant alterations in sequence allocation of the informational environment, is a key aspect we will analyze in this survey. Blockchain's redesign of the financial domain protects against the rivalry of several cryptocurrencies from various applications in highly demanding areas. It enhances the efficiency and security of today's high-demanding cryptocurrencies. But perhaps most importantly, blockchain plays a vital role in the financial industry, providing a sense of security and stability by maintaining the market's equilibrium and improving economic outcomes. Anti-trust policy inference embattles several blockchain applications, like the extrication of consensus records from a public perspective.

Nature is the most valuable aspect of humanity. Many civilizations and cultures have flourished in the lap of nature. The forest is the most beautiful treasure in nature. It always fulfills the basic needs of the lives of the earth’s dwellers. Today, the forests are depleting quickly. The major cause behind this is forest fires or wildfires. An uncontrollable fire either occurs naturally or due to human interruption or any other disturbance caused by nature that may or may not be suppressed by artificial control. A fire broke out in the Amazon rainforest in 2021 and had a significant impact on this planet. Researchers have measured the impacts of wildfire on the habitats of 14,000 species of plants and animals, finding that 93% to 95% suffered from some consequences of the fires. In 2020, Australia’s forest fires caught the world’s attention. Wildfires are a major challenge for humans in today’s time. Early identification of forest fires is the only way to mitigate the risk of damage.

Researchers are working on various techniques to detect the fire early. Several existing approaches, like wireless sensor networks, machine learning, and remote sensing, are used to identify wildfires. Some researchers are using UAVs to identify forest fires. In most cases, the researchers are focused only on prediction using some environmental parameters sensed by the sensors or images captured by the satellites. This paper highlights the various challenges in the prediction of forest fires using the WSN paradigm.

Fuels, gases, and other such substances are widely used in domestic and industrial settings daily. However, they frequently result in significant mishaps like fires and gas leaks. If prompt notification is received, such accidents can be avoided. Installing a gas leakage and fire incident detection system in strategic locations is one approach to achieve this. Here, we demonstrate the construction of a straightforward system that sends an SMS using a GSM module in the event of a fire or gas leak. Additionally, a temperature sensor simultaneously detects the temperature of that difficult circumstance and transmits information to a web server. This is achieved with the use of the Internet of Things (IoT), neural computing, and machine learning. We employ a system with multi-sensing and interaction with the current centralized M2M (Machine-to-Machine) home network and external networks in place of discrete units with basic functionality (such as the Internet). Then, using machine learning, we apply a data mining technique to the sensed data and find anomalous changes for early risk prediction. The system's goals are to increase security and safety and safeguard properties. 

Evaluation of answer scripts is an essential component of any test or educational institution to measure learner outcomes. In a digital script evaluation system, write-up scripts are scanned and uploaded to the central server and assigned to the evaluator for grading. A fair, consistent, unbiased, and error-free evaluation protects an examination system's integrity and is essential for all educational institutions. Most of the applications available in the market were concerned with certificate administration. For an examination system, security, transparency, accessibility, and trust are the major factors that can easily be audited when a blockchain infrastructure is followed. There are a few challenges found in adopting blockchain technology in this field, like scalability, cost of adaption, trust in technology, etc. It is intended to create a totally transparent and effective examination and reward system that may be utilized to replace the traditional system. This chapter presents a self-sustaining educational environment on top of a blockchain for fair assessment. The proposed method proved to be quite beneficial due to its decentralized nature. In our framework, all the steps of evaluation processes are traced and recorded through blockchain to improve the security, transparency, and trustfulness of the system. This makes it simpler to figure out how a candidate obtained the score that he or she did, providing credibility to the gained certificate. 

The unwanted and excessive use of pesticides and fertilizers creates health hazards and makes the soil unproductive. Only a small portion of the crop catches disease, but the farmers spray the pesticides evenly over the entire field, resulting in wastage and overuse of pesticides. They are not aware of the exact composition of the soil. Certain nutrients, pH levels, moisture, and fertility support a certain kind of crop. All of these problems have been observed in the nearby villages of District Sonepat (Haryana), such as in the villages of Bhigan and Murthal. Robots with sensors to match the levels of pH, temperature, humidity, and soil moisture to the levels required for growing crops are used in the present research to identify the best crop for a given soil. The sensors are connected to Raspberry Pi to send the data to ThingSpeak, which compares the sensed data with the standard data and gives the user the required information. A drone hovering over the field clicks the pictures of the field to be used by the intelligent image processing system for disease detection. The productivity can be increased without adversely affecting the soil. This will reduce the use of chemical fertilizers by 30% to 75% and create economic and environmentally sustainable crop production. The equipment manufacturing companies are potential buyers for the patent of the proposed technology. Marketing of this technology may be done in partnership with fertilizer suppliers. 

Establishing a secure and safe trust system for a dispersed cloud-based IoT environment is very tough. Nowadays, the differences between cloud and IoT are blurring, and they are increasingly being applied together in several applications. A robust trust establishment model is the need of the hour. In this paper, we are advocating a trust-dependent authorization model to be implemented in a clouddependant IoT environment that also functions as a two-way or dual-mode trust model, taking into account the requirements of service giver and taker. In the suggested paradigm, trustworthiness is assessed on both the user and supplier sides. A transaction is only permitted if both trust values exceed a predefined or set threshold. We cannot entirely rely on the recommender's comments to determine how trustworthy a node or entity is. Therefore, the primary presumption in this situation is that a small number of hostile entities may provide false information about other entities. In the actual world, we can anticipate several malevolent actors attempting to sabotage the efficient operation of the cloud system by providing misleading reports and feedback, and the IoT ecosystem is ultimately dependent on the cloud for data processing and allied services, which generally consists of an application layer, network layer, and sensor layer. A new phrase for gauging this inaccurate or biased reporting is included in the model. In addition, the ongoing transactions are made more trustworthy by removing skewed feedback from both the user and supplier end. We also put forward a rule that allows trust values to be dynamically modified and active nodes to be given more weightage when determining trust. 

Malware has emerged as a significant threat with growing infection rates and degrees of sophistication as the number of devices and technologies related to the Internet of Things (IoT) has increased and more are put into service. Without robust security procedures, many confidential records are left open to vulnerabilities. As a result, it is simple for cybercriminals to use this data to carry out various unlawful acts. Therefore, advanced network security mechanisms that are capable of executing a traffic analysis in real time and mitigating harmful traffic are required. These mechanisms must also be able to detect malicious traffic. We propose a revolutionary technique for IoT malware traffic analysis that uses deep learning and graphical demonstration to detect and categorize new malware more quickly. This will allow us to handle the difficulty that has been presented (zero-day malware). Due to the utilization of deep learning technology, the suggested method for detecting malicious network traffic operates at the package level, significantly reducing the time required for detection and producing promising outcomes. A dataset called “1000 pcap files of ordinary and malicious traffic that were collected from various network traffic sources” is created to evaluate our method's performance. This dataset is used to assess how well our method works. The experimental findings of the Residual Neural Network (ResNet) are highly encouraging, delivering a rate of accuracy for the identification of malicious traffic that is 95.09%.

IoT is a future-proof technology that is extremely capable of providing the required solution in real time for everyday life, healthcare, smart city, agriculture, automation industry, disaster management, etc. IoT is taking sensors, computing, and communication to the next level collectively. Due to the explosion of demand for services, the IoT platform provides customizable health services. Managing health records and patient’s personal information can improve healthcare quality and efficiency simultaneously. For smart and safe usage, a unique identification is provided in IoT; in this way, large amounts of information are accessible to all. The solutions provided through IoT are more accurate and accountable. Wearable devices can help in dealing with various diseases with the ability to constantly monitor the present situation. One can even predict the worst condition to arrive and prepare for it or can even delay it or completely eliminate it. There are various devices that are already available that can be worn on hands and tell about blood pressure, temperature, step count, calories burnt, heart rate, sleeping hours, walking speed, etc. Wearable devices provide continuous monitoring at every place, like at home or office or while swimming and playing games. However, the accuracy and the ability to transmit the information to medical service providers in real time are also attained with the help of innovative IoT technology. The Immediate Health Monitoring System uses IoT, which enables it to monitor the patient's temperature along with the oxygen level in an instant; this system transmits the same information to the doctor or medical service provider at a distant location. If the values of the parameters change from the standard values, then an alert message is given to the medical service provider or the doctor concerned with the patient. This instant health monitoring plan based on IoT helps doctors effortlessly collect real-time numbers at their location, which is far from the patient's location. The next level in the healthcare industry can be achieved with the help of wearable devices, which are part of IoT.

COVID-19 was a global disaster for humans in 2020. It actually causes breathing problems, sneezing, and coughing. Early detection is essential to minimize the viral effects. RT-PCR test is not suitable against the rapid spread of coronavirus since it takes time to detect positive cases. Moreover, COVID-19 is a transmittable disease; therefore, there is always a chance of spreading the disease by RT-PCR method. Nowadays, radiographic images have already proven to be popular and alternate diagnostic tools in the context of disease prediction. Several research papers have already proposed different techniques using deep learning or CNN methods or transfer learning methods with radiographic images for automatic screening of COVID-19 disease. Unfortunately, most of these existing screening methods have either worked on small datasets or have produced the results with low accuracy. In this paper, an IoT-enabled automated screening model has been presented to classify the COVID-19 cases with the CNN-LSTM+Capsule network. The LSTM-Capsule network model not only diagnose the COVID-19 cases quickly but also maintains low hardware costs. The proposed model has worked on 280 COVID-positive and 290 normal patients’ images taken from two benchmark datasets and tried to find out the spreading tendency of the virus by getting some realistic data from IoT sensor devices in the affected zone. The proposed model achieves an accuracy of 97%. Moreover, crossvalidation is applied in this model to avoid over-fitting. This paper concludes that the proposed model gives an insight into the CNN-LSTM+Capsule network working for COVID-19 detection that helps get richer feature mapping from the radiographic images and further classifies the COVID-19 cases from the normal ones effectively. Moreover, the LSTM-Capsule network model has outperformed relative to other related works in the early identification of COVID-19 cases.

The Internet of Things (IoT) is a booming trend in the current automation industry and is estimated to cover a considerable share of the international market in the near future. Sharing of IoT data, especially data with a potential risk of sensitive data exposure, poses challenges to organizations. The rapid growth of IoT data is found in diverse applications, and it is connected to different hardware and software platforms. This leads to a major issue at the time of executing the application. Data identification is made through anonymization procedures and de-identification procedures before sharing with a third party. The detection of data points that have the potential to expose sensitive information can be a tedious task, especially if done manually. Automating the task helps make identification much easier when dealing with a lot of data sets, both small and large. The current solution has been envisaged with the intention of helping to identify potential leakage of sensitive information using an easy-to-implement framework, as well as a solution for detecting potential quasiidentifiers in data.

Blockchain is a widely recognized technology that has grown in prominence over time. It first gained popularity due to its use in Bitcoin. Because of its reliable security features, blockchain has been extensively employed in a variety of domains. The agricultural supply chain is one field where blockchain can be especially useful, as it provides benefits such as data immutability, transparency, and the ability to disregard middlemen. As a consequence of this, academics are working hard to figure out how to build a smart infrastructure for supply chain management that combines the security of blockchain with the promise of the Internet of Things (IoT). The primary goal of this study is to examine the structure and relationships of all entities participating in the agricultural supply chain. Smart contracts are crucial in such an environment. They are used to facilitate secure and automated interactions between various parties, such as farmers and consumers, farmers and delivery workers, and farmers and quality control authorities. These smart contracts ensure that every participant works in a coordinated manner. We used the remix IDE and the solidity programming language's smart contracts to create this solution. We compared blockchain transaction submission messages and node communications during the implementation phase. Customer messages, in particular, tend to be larger due to the inclusion of additional information such as registration, verification, purchase placing, and order confirmation. The authority's messages, on the other hand, are the smallest, as they merely demand the product's initial rating. This research intends to develop a robust and efficient agricultural supply chain management system that benefits all stakeholders by using the potential of blockchain technology and integrating IoT.

The rise of IoT-connected gadgets has resulted in a deluge of information. Network bandwidth is heavily used due to the large amount of data being delivered to the cloud or fog. The IoT network could be vulnerable to network congestion. Furthermore, some IoT implementations need multimedia data transmission with various Quality of service (QoS) standards. Multinetwork information architecture (MINA) aims to satisfy various QoS criteria for different areas of IoT. While it can adapt to new network conditions, it cannot provide the quality of service required for critical IoT applications. This study proposes enhancing the software-defined networkbased IoT network with the application-aware QoS routing algorithm (AQRA) and MINA to guarantee a wide range of quality-of-service (QoS) requirements for critical IoT applications. The best possible outcome was achieved by using the particle swarm optimization (PSO) method.

Data security is crucial when interacting with internet-based networks, clouds, servers, etc. In today’s world, the most prominent percentage of communication is internet-based. Using sensors like temperature and soil moisture sensors, the collected data is encrypted. The electronic gadgets transform the normal text into unintelligible cipher text, which is then stored in the cloud at the transmitting end. The AES128 key and hash code used on the transmitting side are used to decode data on the receiving side. In this process, if any unauthorized person tries to hack the data, they fail to get the original data since the data is in the form of cipher text. The application of IoT and 5G communication technologies introduces smart farming ecosystems to a wide range of cyber security risks and vulnerabilities. This type of cyberattack will cause economic disruption in nations with significant agricultural dependence. Hence, there is a need for data security in smart farming. In this paper, an efficient and easy way to encrypt the data using Cipher text is shown.