Artificial Intelligence in IoT

In recent years we are beginning to see the convergence of three distinct research fields: augmented reality (AR), intelligent virtual agents (IVAs), and the Internet of things (IoT). Each of these has been classified as a disruptive technology for our society. Since their emergence, the advancement of knowledge and development of technologies and systems in these fields were traditionally performed with limited input from each other. However, over recent years, we have seen research prototypes and commercial products being developed that cross the boundaries between these distinct fields to leverage their collective strengths. In this paper, we review the body of literature published at the intersections between each two of these fields, and we discuss a vision for the nexus of all three technologies.

Ensuring the security of the Internet of Things (IoT) is deemed as one of the most critical challenges and needs that have to be addressed in order to guarantee the successful deployment of IoT in emerging technologies like 5G. In an effort to address this challenge, in this work, an improved and flexible physical layer security technique, referred to as orthogonal frequency-division multiplexing with subcarrier index selection and artificially interfering signals (OFDM-SIS-AIS), is developed for protecting the transmission of OFDM-based waveforms against eavesdropping in 5G and beyond wireless networks. In this technique, the frequency response of correlated subchannels is first converted into a completely randomized and independent response by means of adaptive interleaving. Then, the whole OFDM block is divided into small subblocks, each containing a set of subcarriers, from which a subset of these subcarriers, which are corresponding to high subchannel gains, are selected and used for data transmission, while the remaining ones, which are corresponding to low subchannel gains, are used for sending artificially interfering signals. The selected subcarriers are determined through an optimization problem that can effectively maximize the signal-to-noise ratio (SNR) at only the legitimate receiver. The obtained results demonstrate a significant improvement in the secrecy gap performance without considering the knowledge of the eavesdropper’s channel nor sharing any keys while maintaining low complexity and high reliability at the legitimate user. These numerous advantages have the potential to make the proposed scheme a consistent candidate technique for secure IoT-5G based services.

Emotional artificial neural network (EANN) is a cutting-edge artificial intelligence method that has been used by researchers in the engineering and medical sciences over the recent years. First introduced in the 1999s, EANN is the combination of physiological and neural sciences for investigation of complex processes. Rainfall-runoff is a complex hydrological process that may be modeled by EANN methods to attain information about the response of a catchment to a rainfall event. In practice, the response is surface runoff either in the form of streamflow or flood in the catchment of interest. Thus, a reliable rainfall-runoff model is an inevitable component of a watershed so that decision-makers may use it to reduce the relevant vulnerability against extreme rainfall events. Undoubtedly, one way to empower the capabilities of rainfall-runoff models is the integration of recent achievements in the Internet of Things (IoT) with robust modeling algorithms such as EANN. Relying on the huge amount of knowledge within IoT components, the hybrid IoT-EANN can yield in the high-resolution space-time estimations of runoff that is a practical way to mitigate potential hazards of flooding through real time or in advance actions. With this chapter, we provide a short overview of the state-of-the-art EANN and its application in rainfall-runoff modeling. In addition, a concise review of the applications of IoT in hydro-environmental issues is provided. The chapter reveals that integrations of IoT with hydro-environmental studies are in their infancy. Being a new class of investigation, there is no hybrid rainfall-runoff model within the literature coupling IoT technology with artificial intelligence.

Smart tourism destination (STD) concept has taken serious attention as a result of the smart city initiatives. Technology connects all organizations, entities, activities, and elements. Tourism is a multidimensional service system covering different actors and organizations. When a tourism destination gets smarter, the tourists’ needs and demands are expected to be fulfilled more efficiently to create a better tourist experience. This paper aims to examine the content of smart tourism destination and its link with smart city addressing a model for Antalya as a candidate for a smart tourism destination.

Spectral clustering is a class of graph theoretic procedure, which is popular for finding natural groupings. Over the last decade, it has become a widely adopted tool – utilized in solving image segmentation problems, via normalized cut (NCut) methodology. Few challenges faced by image segmentation based on spectral clustering include its inability of processing large images due to high computational cost and memory requirements and its sensitivity to irrelevant and noisy data. This chapter presents an unsupervised image segmentation technique using spectral clustering, aimed at salient object detection, followed by extraction. The presented technique addresses all of the aforementioned challenges by means of a weighted binary tree-based fast spectral clustering (WBTFSC). The algorithm integrates dimensionality reduction with spectral clustering by introducing an effective preprocessor, comprising two fundamental steps of color quantization and unique pixels selection. The experiments, performed on color images using the proposed algorithm, show improved performance in extracting objects of interest with high accuracy. We also test the algorithm on several noisy images; the obtained results reveal better performance in comparison to few existing techniques.

The Internet of Things (IoT) is going to be the next technological revolution. According to the Internet, the revenue generated from IoT products and services are going to be approximately 300 billion in 2020. Simultaneously, with the massive amount of data that the IoT will generate, its impact will be reflected across the entire Big data universe that will coerce the organizations to upgrade current tools and technology to evolve to accommodate this additional data volume and take advantage of the insights. IoT and Big data basically are two sides of the same coin according to some experts. It is a challenging task to manage and extract insights from IoT data. Therefore, a proper analytics platform/infrastructure to analyse the IoT data is a vital aspect for any organization when it is also true that not all IoT data is important.

Water is a critical and indispensable resource for the sustainability of life, economic development, and the environment. According to the United Nations (UN) estimates, 70% of the world’s population will live in cities by the year 2025, and the current centralized piped infrastructure relied upon by water utilities will be inadequate. Leveraging on the advancement in emerging blockchain, Internet of Things (IoT), and sensor technologies offers a means for efficient water management. In this era of Fourth Industrial Revolution (4IR), human creativity will be a critical requirement in this regard. This chapter explores the impact of blockchain and IoT on water management and examines the feasibility of its adoption in multiple case scenarios and instances such as stormwater management, water quality monitoring and reporting directly to consumers and other relevant stakeholders, and smart payment and contract, in order to sustainably deal with the challenges of global water crisis induced by climate change and rapid population growth. This chapter makes special and unique emphasis on the relevance of the research through an African perspective and view. Furthermore, the technical advantages, socioeconomic gains, and technological benefits of synergizing blockchain and IoT such as enhanced security and transparency, reduced operational cost, overall efficiency, and other merits are expatiated.

With the rapid development of wireless sensor network, the transmission and processing of multimedia data gradually become the main task of wireless sensors. To reduce the data bandwidth, many wireless sensors use frame rate up-conversion (FRUC) to recover the dropped frames at the receiver. FRUC is actually a temporal-domain tampering operation of video at the receiver, and FRUC forgery can be found by analyzing the statistical feature of the video. In this chapter, a forensics algorithm based on edge feature is proposed to discover forged traces of FRUC by detecting the edge variation of video frames. First, the Sobel operator is used to detect the edge of video frames. Then, the edge is quantified to obtain the edge complexity of each frame. Finally, the periodicity of the edge complexity along time axis is detected, and FRUC forgery is automatically identified by hard threshold decision. Experimental results show that the proposed algorithm has a good forensics performance for different FRUC forgery methods. Especially after the attacks of de-noising and compression, the proposed algorithm can still ensure high detection accuracy.

Underground mining is considered as one of the most hazard-prone industries, and serious work-related fatalities have arisen as a consequence of processes related to it; this chapter deals with occupational hazards and related risk factors. Artificial neural network-based risk assessment approach in underground copper and zinc mine case study is proposed. Occupational health and safety (OHS) history dates back to ancient human history ever. Mankind date was obliged to do business in order to sustain life. OHS studies aim to increase the safety standard with reducing risk level in an acceptable degree. Safe workplaces with respect to OHS increase health, safety, and welfare standards of whole workers. Throughout the world major hazards categorized as physical, chemical, biological, psychosocial, and ergonomic risks can be observed. Although technological developments provide rapid growth in almost all industries, it can be observed that there is a lack of attention being paid and advanced occupational safety practices in the mining industry. A case study is carried out in one of the largest underground mining companies using neuro-fuzzy approach. Neuro-fuzzy logic-based risk assessment study supplies opportunity to provide more adequate decision-making process and gives meaningful classifications of hazard. Neuro-fuzzy approach is a combination of advantages of artificial neural networks and fuzzy logic. It gives more appropriate and comprehensive risk assessment in OHS. After all the neuro-fuzzy approach is applied for classification of risk types in each department of the copper and zinc mine, the necessary control measures for each department and for a whole system are presented. In the study, adaptive neuro-fuzzy inference system (ANFIS)-focused model is applied to the copper and zinc mine risk analysis problem based on three-step neuro-fuzzy approach. Improvements are shown on the study to show the efficiency and flexibility of the method. The main target by integrating the neuro-fuzzy logic application into the risk analysis is to obtain a more effective risk assessment and getting better results than the conventional models used. In conclusion, besides its theoretical contribution, obtained results of this study contribute toward improving occupational safety levels of copper and zinc mine with more comprehensive risk assessment process.

“Internet of Things” (IoT) is expected to revolutionize the application of services by enabling the creation of smart spaces such as smart cities, smart houses, smart transportation, and smart outdoor monitoring (SOM) in the near future. Such smart spaces will require the deployment of a significantly large numbers of devices connected to the Internet such as sensors, actuators, and wearable computing devices and the likes. The growing population in urban areas will pose a significant challenge toward the utilization of public resources. Smart cities provide a promising solution by enabling a smart and efficient way to handle challenges such as waste management, traffic, security, and so on. Consequently, there is a large number of devices deployed in a smart city so as to enable collection and transition of data and data analysis. Therefore, communication is a key aspect in a smart city topology. In this work, we provide an overview about how the IoT provides an efficient communication platform and how 5G will be the main enabler by providing competitive bandwidth, high integrity, low latency, high spectral efficiency, and viable network capacity.