Expert system for identifying and analyzing the IoT devices using Augmented Reality Approach

Biometrics in conjunction with the new development of the Internet of Things (IoT), augmented reality (AR) expert systems are evolving to visualize 3D virtual models of the real world into an intelligent and interactive virtual reality environment that facilitates physical identification of objects and defines their specifications efficiently. The integration between AR and IoT in a complementary way helps identify network-related items' specifications and interact with the Internet of Things more efficiently. An identity is a dedicated, publicly known attribute or set of names for an individual device. Typically, identifiers operate within a specific area or network, making it difficult to identify things globally. This paper explores the use of Augmented Reality (AR) expert system for identifying devices and displaying relevant information about the device to the user. Based on the developed model network, the developed system of identification of IoT devices was tested. Also, the traffic generated by the AR device when generating requests to the organization server was investigated. According to the test results, the system is undemanding to the main network indicators. The system-generated traffic is self-similar. The test results show that the server software can solve the problems of identifying IoT devices through interaction with augmented reality devices. network packets. the proposed system tracks the network flow data and extracts specific features to establish a fingerprint for each device in the network. The researchers adopt a novel supervised machine learning technique for IoT identification tasks. The proposed approach can automatically recognize white-listed device types and unknown devices with abnormal behavior connecting to the network by constraining and enforcing privileges rules for IoT device


Introduction
The Internet of Things is currently a generally accepted concept for developing communication networks in the short and long term. It is also considered an advanced platform for creating digital intelligence in the idea of a "smart state" (Albahri et al., 2018). According to most consulting analytics firms, over the next five years, more than 30 billion devices will be present in each area of human activity. Thus, we can talk about the pervasive nature of the Internet of Things' penetration into our daily life (Albahri et al., 2019c).
The phrase "Internet of Things" was first heard by Kevin Ashton in 1999 at the presentation of innovative solutions from Procter & Gamble. Ashton suggested applying RFID tags to its products and thus ensuring its interaction with the radio receiver. Kevin Eton indicated that such data collection could be used to solve many problems in the real In the last decade, the Internet of Things has become one of the breakthrough technologies generally recognized by all world countries. IoT allows people and things to interact anywhere, anytime, and in any combination using the IoT infrastructure. The IoT ecosystem involves collecting data from sensors (or sending commands to actuators), transmitting them through a communication network to cloud platforms for subsequent analysis to provide intelligent services for people. Figure 1 shows the key components required to build an IoT system. According to the model, sensors and information retrieval devices collect various data types about a particular object. These data can then be further processed and analyzed to extract useful information to provide intelligent services (Al-Bahri et al.,2020).
The Internet of Things can be seen as a collection of four main elements:  Internet: to ensure communication anytime and anywhere between any participants in the internetworking exchange. Cloud computing, intelligent web services, etc.  Hardware: includes communication equipment and pickup terminals such as sensors, tags, actuators, and transmitters.  Middleware: used for storing data, calculating, and analyzing transmitted data;  Interface: used to visualize and interpret the collected results for different platforms and applications.
Various IoT applications are aimed at solving specific problems. Typical applications include data management, analytics, visualization, management of heterogeneous networks, research goals, etc. Nevertheless, IoT research is still in its infancy due to the existence of many unresolved problems. For example, issues related to battery life, the simplicity of the "lightness" of data transfer technologies, performing actions depending on the context of what is happening, identification and security issues, the cost of terminal devices, scalability, and heterogeneity (Standardisation, 2017). Despite all the advantages of the Internet of Things, there have recently been cases of disclosure of data collected by IoT devices, which raises concerns about the identity of devices and applications within the IoT concept framework.
Indeed, identifying things plays a vital role in the classification and recognizing system (Hasoon et al., 2011). For example, attackers can use portable RFID / NFC readers to steal personal information from bank cards in public transport using technology vulnerabilities such as PayPass. This is possible due to the lack of confirmation of the identity of the RFID reader's owner. Another example is the possibility of an attacker intercepting data from networks of IoT devices to obtain IMEI-identifiers of various terminal devices equipped with modems to subsequent broadcasting of intentionally distorted messages.
Current solutions worldwide are mainly aimed at binding an IoT device based on expert system application with an identifier similar to an IP address or a mobile phone number. One can understand who is using a particular device.
Research in this area was initiated due to discussing these issues in the BEREC (Body of European Regulators for Electronic Communications) (Standardisation, 2017). Simultaneously, identification has a much broader scope and is more appropriate for many applications and entities (subjects) in IoT. In addition to identification purposes in communications, ongoing research includes issues of identification of physical and virtual things, such as services for users using IoT, collected data, location. Today, various identification schemes have already been standardized and  The new generation of the Fourth Industrial Revolution has recently emerged, enabling connectivity and access to real-time insights across processes, products, and people. That give user unlimited opportunities for using various services and applications over the internet. Recently, scientists have become more interested in the Internet of Things (IoT) and Augmented Reality (AR) expert system implementation.
to make spaces more intelligent and interactive. The amazing development of the Internet of Things technology enables the possibility of connecting and exchanging data between various devices through the Internet and communication protocols (Hu et al., 2021). The Internet of Things supplies related data in the physical environment (sensory, auditory, visual) into augmented reality to provide a convenient and intuitive way for users to visualize and interact with objects in the real world.
One of the exciting areas of this concept is the provision of services based on augmented Reality Technology.
Augmented reality supplements the existing world with the necessary data. For example, looking at a shop window glasses, augmented reality can be seen and the mannequins the entire range of products, sizes, and prices, without going inside. There is already an extensive selection of augmented reality glasses, differing in functionality and communication network requirements. Some of them are almost indistinguishable from regular glasses (Hu et al., 2021). The availability and simplicity of equipment stimulate the creation of a variety of services. So, the use of augmented reality technology is difficult to overestimate; it is used both in medicine and education. Also, it uses for solving everyday problems, in industry and agriculture, in VANET networks, and flying sensor networks.

Literature Survey
The Phupattanasilp and Tong (Phupattanasilp and Tong, 2019) introduced the AR-IoT method based on augmented reality (AR) to support IoT data visualization. The proposed method uses IoT Multi-camera data over real-world objects to improves the identification of 3D images with a non-destructive and low-cost imaging platform of the IoT.
The results show a high integration of IoT data with AR resolution, effectively updating accuracy and precision with the Internet of Things for a better shopping experience. AR technologies and frameworks are scalable to deal with any IoT product due to its incorporation into the IoT platform. An intuitive augmented reality-based visualization and interaction allow the provided AR service to reduce latency significantly. The researchers focus on three major architectural components that are required for simplified, scalable AR services, and expertise for IoT-ready products have been identified: object-focused data processing and visualization, entry, control, and interaction with objects, and interoperability. Lund et al. discussed the concept of "Identity in the Internet of Things" and introduced the abbreviation IDoT. Besides, a parallel and analysis is carried out on the subject: why it is so unique compared to the concept of "identity of users" (IDoU) in traditional networks and communication systems. Using ideas of "Identity" of the user (IDoU) from conventional approaches and networks, a stack for "identity" in the Internet of Things is proposed as in Figure 3. The presented information stack has four categories: inheritance, association, knowledge, and context (Lund et al., 2014).  Figure 4 illustrates the abstract concept of traditional identifier (IDF) and locator separation. According to the recommendation, identity is not explicitly associated with the corresponding identifier (s) in all existing ILS schemas. The essence of a thing (or object) may perform many functions in future network architecture, allowing for identifier separation. Therefore, the authors proposed a new IIS schema and the ILS paradigm, which is promoted within a single framework with various potential value-added services (Halavachou & Fei, 2020). Al-Bahri (Al-Bahri et al.,2020) studied the ability of Digital Object Architecture (DOA) Technology to identify IoT devices. The authors mentioned that DOA technology allows for unambiguous persistent identification of objects in which these objects' copyright holders are interested. This makes it reasonable to develop DOA technology as a global identification system with equal rights for all members. The paper shows that there have been no works in which the DOA would be analyzed in detail as a method for identifying devices and applications of the Internet of Things. The proposed approach shows how people can shop in the future using digital and analogs like the real world. The proposed system indicates that shopping augmented reality is more likely to lead to increased usability and customer loyalty. Through pilot tests, significant support was obtained by the synergies and benefits of the trial. However, there remains a concern that the system should standardize the protocol for sharing data and displaying content.

Research Methodology
This work adopted an experimental investigation to explore the Augmented Reality expert system implementation in identifying the connected IoT devices in a system and display the relevant information about each of them. Also, a quantitative research method was deployed to analyze the collected data. As noted earlier, the response time to a change in the environment is one of the leading indicators of the quality of augmented reality services. A possible scenario for providing augmented reality services using SETCO cloud services is considered, and hands of the quality There are already several solutions based on augmented reality technologies for identifying and inventorying various objects. The identification process assumes that information about the objects searched for by the identifier is stored in a system for storing and processing large amounts of data. As such a system, it is proposed to use "SETCO: Enterprise" to automate the enterprise. This system is successfully used for the inventory of objects owned by the enterprise.

Experimental Method
This section proposes a system for identifying IoT devices. This system is based on augmented reality technologies and server software "SETCO: Enterprise" and a model network developed and tested. Figure  This model has an identification device, an augmented reality device (namely augmented reality glasses), to identify the Internet of Things devices. The identification device then sends a request to the identification server, a web version of the SETCO: Enterprise software. The identity server processes the request and accesses the database, which returns the sought data and then sends the identity device's data.

Model network for identification of IOT devices
Based on the above architecture, a model network was developed, as shown in Figure 2. Consisting of:  Identification devices in Augmented Reality glasses are where a program is run to recognize IoT devices' identifiers (using Bluetooth technology, BLE). Then requests are created to the SETCO server (using HTTP REST), traffic is intercepted and analyzed.  Identification object is an IOT device with its identifier, on which the software is responsible for interacting with the IoT functions (using Bluetooth technology, BLE).  Identity servers -a server that is a web version of the SETCO: Enterprise application and a Microsoft SQL 2012 database that stores object identifiers and information.  Interaction with the identity server takes place using the HTTP REST interface.  Network jamming devices -a NetDisturb software device allows to simulate a public communication network's operation.
Identification devices were implemented based on augmented reality glasses -Epson Moverio BT-300, operating based on the Android operating system. The Java Programming Language and Android SDK Toolkit (Android Bluetooth, Android HTTP Library) were used to develop programs responsible for interacting with OI and SI. To create the software responsible for intercepting and analyzing traffic, the C ++ was used to the programming language, Android NDK tools, and libations libraries. The identification object was implemented based on an Intel Edison microcomputer. The software responsible for interaction with the Identification Device was implemented using the C ++ programming language and the "libblepp" library.
The identification server is implemented based on SETCO software company "SETCO: Enterprise", SETCO programming language, Microsoft SQL 2012 database. The HTTP REST interface, developed based on the REST concept, includes the following commands:  GET request. It searches for an item in the database by an object identifier and returns information about the identified object.  DELETE request. It removes an element from the database by an object identifier.  POST request. It adds a new object by identifier and information about it.

Model network testing
The proposed system of identification of IoT devices was tested based on the developed model network. Also, was investigated the traffic generated by the AR device when generating requests to the SETCO server. The work was carried out for periodic GET and POST + DELETE requests for the following network parameters: latency, bandwidth, Hurst parameter. Some parameters were performed on the identification device: Interaction, traffic analysis, and calculation of network parameters. The delay between the arrivals of network packets was calculated using the UNIX Time Stamp system. Throughput was calculated as the sum of the sizes of all packets received in one second.
The Hurst parameter (H) parameter characterizes the system's self-similarity and is used in time series analysis (Coelho, 2011) The value of H can take on the following: where: 1≤k≤n 5. The standard deviation Sa is calculated for each interval Ia as determined in equation 5: where: i  (1,2,3... A);

Experimental Results
The Testing was carried out for the case in which the Augmented Reality expert system requests information by identifier using the HTTP GET method. The total testing time (traffic interception) is 180 seconds. The countdown starts after receiving the first HTTP packet, which is a request to the identification server.   According to the test results, we can determine that this scenario's system is undemanding to the leading network indicators. The traffic generated by the system is self-similar. The developed model network is resistant to low and average service quality indicators but shows severe deviations from the norm with increasing network restrictions. A particularly severe degradation in system performance can occur when bandwidth is limited; when the jitter indicator's value rises, the traffic changes its properties to antpersistent.  Limited Bandwidth (kbps) Number of packages delivered

Conclusion
This study presents the expert system architecture for identifying the IoT devices using augmented reality technologies and server software "SETCO: Enterprise". The use of augmented reality technology to inventory at enterprises and general identification of IoT devices seems interesting. The combined use of SETCO: Enterprise software and augmented reality allows creating several services that can speed up the execution of specific processes in enterprises and improve the quality of perception in the provision of augmented reality and the Internet of Things services. The model network was created and tested based on the proposed architecture. The test results show that the SETCO: Enterprise server software can solve the problems of identifying Internet of Things devices through interaction with augmented reality devices. During testing, it was found that the developed system is resistant to changes in network characteristics. This feature is essential since existing networks transmitting large volumes of different traffic types do not always guarantee the fulfillment of service quality indicators' established values.

Acknowledgment
The research leading to these results has no Funding. Insertion jitter… Latency (ms)