Cyber-physical systems, built from the integration of cyber and physical components, are being used in multiple domains ranging from manufacturing and healthcare to traffic con- trol and safety. Ensuring the security of cyber-physical systems is crucial because they provide the foundation of the critical infrastructure, and security incidents can result in catastrophic failures. Recent publications report that machine learning models are vul- nerable to adversarial examples, crafted by adding small perturbations to input data. For the past decade, machine learning security has become a growing interest area, with a significant number of systematic reviews and surveys that have been published. Secu- rity of artificial intelligence in cyber-physical systems is more challenging in comparison to machine learning security, because adversaries have a wider possible attack surface, in both cyber and physical domains. However, comprehensive systematic literature re- views in this research field are not available. Therefore, this work presents a systematic literature review of the adversarial attacks on artificial intelligence in cyber-physical sys- tems, examining 45 scientific papers, selected from 134 publications found in the Scopus database. It provides the classification of attack algorithms and defense methods, the sur- vey of evaluation metrics, an overview of the state of the art in methodologies and tools, and, as the main contribution, identifies open problems and research gaps and highlights future research challenges in this area of interest.

Innovations such as combustion engines, electricity and assembly lines have all had a significant role in manufacturing, where the past three industrial revolutions have changed the way manufacturing is performed. The technical progress within the manufacturing industry continues at a high rate and today's progress can be seen as a part of the fourth industrial revolution. The progress can be exemplified by ”Industrie 4.0”; the German government's vision of future manufacturing. Previous studies have been conducted with the aim of investigating the benefits, progress and relevance of Industry 4.0-technologies. Little emphasis in these studies has been put on differences in implementation and relevance of Industry 4.0-technologies across and within industries. This thesis aims to investigate the adoption of Industry 4.0-technologies among and within selected industries and what types of patterns that exists among them. Using a qualitative multiple case study consisting of firms from Aerospace, Heavy equipment, Automation, Electronics and Motor Vehicle Industry, we gain insight into how leading firms are implementing the technologies. In order to identify the factors determining how Industry 4.0-technologies are implemented and what common themes can be found, we introduce the concept production logic, which is built upon the connection between competitive priorities; quality, flexibility, delivery time, cost efficiency and ergonomics. This thesis has two contributions. In our first contribution, we have categorized technologies within Industry 4.0 into two bundles; the Human-Machine-Interface (HMI) and the connectivity bundle. The HMI bundle includes devices for assisting operators in manufacturing activities, such as touchscreens, augmented reality and collaborative robots. The connectivity-bundle includes systems for connecting devices, collecting and analyzing data from the digitalized factory. The result of this master thesis indicates that depending on a firm’s or industry’s logic of production, the adoption of elements from the technology bundles differ. Firms where flexibility is dominant tend to implement elements from the HMI-bundle to a larger degree. In the other end, firms with few product variations where quality and efficiency dominates the production logic tends to implement elements from the connectivity bundle in order to tightly monitor and improve quality in their assembly. Regardless of production logic, firms are implementing elements from both bundles, but with different composition and applications. The second contribution is within the literature of technological transitions. In this contribution, we have studied the rise and development of the HMI-bundle in the light of Geels (2002) Multi-Level Perspective (MLP). It can be concluded that an increased pressure on the landscape-level in the form of changes in the consumer-market and the attitudes within the labor force has created a gradual spread of the HMI-bundle within industries. The bundles have also been studied through Rogers (1995) five attributes of innovation, where the lack of testability and observability prevents increased application of M2M-interfaces. Concerning Big Data and analytics, the high complexity prevents the technology from being further applied. As the HMI-bundle involves a number of technologies with large differences in properties, it is hard draw any conclusion using the attributes of innovation about what limits their application.

The purpose of this thesis is to describe and implement how a specific form of IoT, sensors, can be integrated with BIM in order to improve the building performance, when the perspective taken is the end-users. It seeks to explore different perceived values of BIM and sensor integration for the occupants who directly use the building facilities. The thesis also describes the concept, frameworks and cases of how BIM and sensors integration can be setup. These are used for an implementation at a case facility. Three main methods are used – literature review, comparative case study, and a smallscale implementation, containing a survey and sensor implementation based on the respondents’ satisfaction with the office air quality. A basic literature review is used to gather the fundamental concepts used within the relevant areas, and to review the empirical research connected to these. The conceptual part of the thesis review frameworks for BIM and sensor integration, and points toward a more user-centric framework that is later developed in relation to the thesis’ empirical results. The theoretical framework integrates Information Systems Theories with Knowledge Management for a framework of understanding how knowledge about new kinds of Information Systems in developing areas function. The empirical part of the thesis is structured into two main phases, one descriptive comparative case study, and the other an implementation based in the first phase results. The first phase is descriptive, where two cases of sensor and BIM implementation processes for FM are described. The main case of Tyréns company (Tyréns), and a reference case of Mästerhuset is used for understanding how different organizational structures may lead to different perceived values and processes of BIM and sensor integration for the end-users. The second phase is an implementation at the main case, Tyréns’ headquarter building. Here the end-user perspective is employed with a survey that is constructed in accordance with some of the fundamental concepts and research reviewed, in order to measure the perceived satisfaction with the air quality of the end-users working environment. The answers show concerns with air quality in the meeting rooms, and this is used as the basis for a small-scale implementation of sensors, where CO2 and temperature sensors are set up. The results show how different organizational-specific conditions generate different perceived values of BIM and sensor integration depending on ownership relation to the end-users. The case study also illustrate the different processes of BIM and sensor integration may be setup to supplement building performance. This points to a needed add-on into frameworks that conceptualizes BIM and sensor integration without the inclusion of the end-users’ perspective. Based on this an end-user conceptual framework of BIM and sensors is proposed with the supplementary part of a knowledge layer, named analytic layer and data source from occupants.

The Intel Galileo, inspired by the well-known Arduino board, is a development board with many possibilities because of its strength. The Galileo is has an Intel processor capable of running GNU/Linux and can be connected to the internet, which opens up the possibility to be controlled remotely. The programming language that comes with the Intel Galileo is the same as for the Arduino development boards, and is therefore very limited and does not utilize the Galileo’s entire strength. Our aim with this project is to integrate a more suitable programming language; a language that can make better use of the relatively powerful processor to control the components of the board. The programming language of choice is Erlang, and the reason is obvious. Erlang can be described as a process-oriented programming language based on the functional programming paradigm and its power in concurrency. The result of the project was the successful integration of a complete version of GNU/Linux on the board and the cross-compilation of Erlang/OTP onto the board. Having Erlang running on the system opens up many possibilities for future work, amongst all: creating Erlang programs for the Intel Galileo, integrating an effective API, and measuring the pros and cons of using Erlang on an Intel Galileo. / Intel Galileo är ett utvecklingskort som bygger på Arduinos succe. Den kommer med en kraftigare processor jämfort med Arduino Uno, och den har möjlighet att kunna köra GNU/Linux. Den har också en port för att kunna kopplas till internet och på så sätt kommunicera med andra enheter. Programmeringsspråket som rekommenderas för Intel Galileo är densamma som används för Arduinos utvecklingskort. Det finns däremot en möjlighet att kunna kombinera utvecklingskortet med ett programmeringsspråk som kan erbjuda mer funktionalitet och fortfarande vara enkelt. Vårt val hamnade på Erlang för den är ett funktionellt språk och har möjlighet att hantera olika processer. Tanken är att kunna behandla olika komponenter kopplade till utvecklingskortet som processer, som kan kommunicera med andra komponenter och med internet. Projektarbetet bestod av att undersöka ifall det är möjligt att kunna kombinera Erlang/OTP med Intel Galileon samt skriva en guide för hur implementeringen gick till. Att kombinera de två var lyckat och det öppnar upp möjligheter för fortsätta arbeten och försök.

The purpose of this research is to develop best practices for deploying air quality sensors in a remote location such as Mongolia. I discussed the architecture and design constraints when collecting remote air quality sensors data, the challenges that emerge while implementing a sensor-based network in a remote location such as Mongolia. The tradeoffs of using different architectures are described. I observed the usage of electrical heaters in modified gers in remote locations and conclude how effective they are in reducing PM2.5 levels by analyzing air quality data and go through the process of cleaning up the data and removing humidity from low-cost sensors used to deploy in a remote location such as Mongolia so that the PM2.5 reading is more accurate. In order to help many humanitarian efforts dealing with better air quality in developing countries, an air quality sensor was designed to keep low cost as much as possible. The cost is about $200 to build, which is cheaper than other low-cost sensors, yet provides more functionality (e.g., CO2 sensing) and used cellular connectivity to upload data in real-time. This sensor has implications beyond Mongolia. The sensor can be used anywhere WiFi connectivity is not available, such as parks, bus stops, and along roadways, breaking the constraints that other low-cost sensors have. Removing the need for WiFi is a necessary step in allowing ubiquitous air quality sensing. The contributions in this thesis are: First, I presented the challenges one should consider while deploying air quality sensors in developing countries. Second, since Mongolia offers a unique environment and constraints, I shared experiences in deploying sensors in a remote location like Mongolia. This experience goes beyond air quality sensors and can inform anyone who is deploying sensors in remote areas. Third the analysis of the PM2.5 on the gers gives us better insights as to whether modifying gers with insulation and using electrical heaters as opposed to burning coal to heat up the gers makes a difference in regard to better air quality in the gers.

With the emergence of the Internet of things (IoT) networks, the replacement of batteries for IoT devices became challenging. In particular, the battery replacement is more expensive and cumbersome for scenarios where there are many IoT devices; or where the IoT devices are in unreachable locations; or when they have to be replaced often. Some IoT devices might be lost or forgotten, and there is a risk of hazardous chemicals leakage and e-waste in large scale in nature. Radio frequency (RF) wireless power transfer (WPT) is an alternative technology for powering those devices. It has been shown that only less than one millionth of the transmitted energy is absorbed by the receivers, the rest is absorbed by the objects in the environment. We can utilize the existing infrastructure for wireless communications such as base stations (BS) to charge IoT devices. The present work is devoted to analyze the feasibility and limitations of the battery-less operation of IoT devices with RF WPT technology and energy harvesting from existing infrastructure for wireless communications. We study the indoor and outdoor scenarios for powering of IoT devices. In the first scenario, we consider an outdoor environment where an IoT device periodically harvests energy from an existing BS and transmits a data packet related to the sensor measurement under shadow fading channel conditions. We analyze the limits (e.g., coverage range) of energy harvesting from a BS for powering IoT devices. We characterize the "epsilon-coverage range, where" is the probability of the coverage. Our analysis shows a tradeoff between the coverage range and the rate of sensor measurements, where the maximal "epsilon-coverage range is achieved as the sensor measurement rate approaches zero. We demonstrate that the summation of the sleep power consumption and the harvesting sensitivity power of an IoT device limits the maximal "epsilon-coverage range. Beyond that range, the IoT device cannot harvest enough energy to operate. The desired rate of the sensor measurements also significantly impacts the "epsilon-coverage range. We also compare the operational domain in terms of the range and measurement rate for the WPT and battery-powered technologies. In the second scenario, we consider the remote powering of IoT devices inside an aircraft. Sensors currently deployed on board have wired connectivity, which increases weight and maintenance costs for aircraft. Removing cables for wireless communications of sensors on board alleviates the cost, however, the powering of sensors becomes a challenge inside aircraft. We assume that the IoT devices have fixed and known locations inside an aircraft. The design problem is to minimize the number of WPT transmitters given constraints based on the cabin geometry and duty cycle of the IoT devices. We formulate a robust optimization problem to address the WPT system design under channel uncertainties. We also derive an equivalent integer linear programming and solve that for an optimal deployment to satisfy the duty cycle requirements of the cabin sensors. / <p>QC 20220223</p><p></p>

<p>International Data Corporation^[3] (IDC) data estimates that 152,200 Internet of things (IoT) devices will be connected to the Internet every minute by the year 2025. This rapid expansion in the utilization of IoT devices in everyday life leads to an increase in the attack surface for cybercriminals. IoT devices are frequently compromised and used for the creation of botnets. However, it is difficult to apply the traditional methods to counteract IoT botnets and thus calls for finding effective and efficient methods to mitigate such threats. In this work, the network snapshots of IoT traffic infected with two botnets, i.e., Mirai and Bashlite, are studied. Specifically, the collected datasets include network traffic from 9 different IoT devices such as baby monitor, doorbells, thermostat, web cameras, and security cameras. Each dataset consists of 115 stream aggregation feature statistics like weight, mean, covariance, correlation coefficient, standard deviation, radius, and magnitude with a timeframe decay factor, along with a class label defining the traffic as benign or anomalous.</p><p>The goal of the research is to identify a proper machine learning method that can detect IoT botnet traffic accurately and in real-time on IoT edge devices with low computation power, in order to form the first line of defense in an IoT network. The initial step is to identify the most important features that distinguish between benign and anomalous traffic for IoT devices. Specifically, the Input Perturbation Ranking algorithm^[12] with XGBoost^[26]is applied to find the 9 most important features among the 115 features. These 9 features can be collected in real time and be applied as inputs to any detection method. Next, a supervised predictive machine learning method, i.e., Decision Trees, is proposed for faster and accurate detection of botnet traffic. The advantage of using decision trees over other machine learning methodologies, is that it achieves accurate results with low computation time and power. Unlike deep learning methodologies, decision trees can provide visual representation of the decision making and detection process. This can be easily translated into explicit security policies in the IoT environment. In the experiments conducted, it can be clearly seen that decision trees can detect anomalous traffic with an accuracy of 99.997% and takes 59 seconds for training and 0.068 seconds for prediction, which is much faster than the state-of-art deep-learning based detector, i.e., Kitsune^[4]. Moreover, our results show that decision trees have an extremely low false positive rate of 0.019%. Using the 9 most important features, decision trees can further reduce the processing time while maintaining the accuracy. Hence, decision trees with important features are able to accurately and efficiently detect IoT botnets in real time and on a low performance edge device such as Raspberry Pi^[9].</p>

<div> <div> <div> <p>This thesis aims to introduce a system design that supports the user experience of outfit selection, storage, and matching. Clothes are indispensable items in daily human life. Purchasing one's wardrobe has become more affordable. This has allowed people to focus on purchasing more fashionable clothes. Garment shopping has even become a type of social and leisure activity. With the development of internet technology, shopping methods have changed dramatically. However, these seemingly convenient shopping methods also bring unavoidable problems, such as an inability to understand apparel companies' different size standards and the challenge of seeing the details of materials. On the other side, while overemphasizing the convenience of the shopping process, online companies have ignored people's clothes-wearing experience that is the most enjoyable and valuable for customers. This paper introduces an IoT (Internet of Things) design: "Closet Go" including a mobile application and a clip-able camera. "Closet Go" aims to improve customers' daily outfit selection experience by digitalizing their closets and conducting data analysis of customized dressing habits. In this thesis, I present the entire design process: user research, Ideation, UI/UX design, product development, and evaluation. In the research section, potential users were recruited for interviews to discover the current problems in acquiring, selecting, and matching outfits in daily life. The design process section introduces the design development progress and results via user flow, experience map, prototype, and user interface. Finally, the thesis concludes with a heuristics evaluation section that tests the design's usability and experience to refine the project. </p> </div> </div> </div>

Machine learning algorithms that run on Internet of Things sensory data requires high data quality to produce relevant output. By providing data cleaning at the edge, cloud infrastructures performing AI computations is relieved by not having to perform preprocessing. The main problem connected with edge cleaning is the dependency on unsupervised pre-processing as it leaves no guarantee of high quality output data. In this thesis an IoT gateway is extended to provide cleaning and live configuration of cleaning parameters before forwarding the data to a server cluster. Live configuration is implemented to be able to fit the parameters to match a time series and thereby mitigate quality issues. The gateway framework performance and used resources of the container was benchmarked using an MQTT stress tester. The gateway’s performance was under expectation. With high-frequency data streams, the throughput was below50%. However, these issues are not present for its Glava Energy Center connector, as their sensory data generates at a slower pace. / AI4ENERGY

Thesis(M.Sc. (Computer Science)) -- University of Limpopo, 2020 / Spectrum scarcity is one of the major problems affecting the advancement of wireless technology. The world is now entering into a new era called the “Fourth Industrial Revolution” and technologies like the Internet of Things (IoT) and blockchain are surfacing at a rapid pace. All these technologies and this new era need high speed network (Internet) connectivity. Internet connectivity is reliant on the availability of spectrum Channels. The Federal Communication Commission (FCC) has emphatically alluded on the urgency of finding quick and effective solutions to the problem of spectrum scarcity because the available spectrum bands are getting depleted at an alarming rate. Cognitive Radio Ad Hoc Networks (CRAHNs) have been introduced to solve the problem of spectrum depletion. CRAHNs are mobile networks which allow for two groups of users: Primary Users (PUs) and Secondary Users (SUs). PUs are the licensed users of the spectrum and SUs are the unlicensed users. The SUs access spectrum bands opportunistically by switching between unused spectrum bands. The current licensed users do not fully utilize their spectrum bands. Some licensed users only use their spectrum bands for short time periods and their bands are left idling for the greater part of time. CRNs take advantage of the periods when spectrum bands are not fully utilized by introducing secondary users to switch between the idle spectrum bands. The CRAHNs technology can be implemented in different types of routing environments including military networks. The military version of CRAHNs is called Military Cognitive Radio Ad Hoc Networks (MCRAHNs). Military networks are more complex than ordinary networks because they are subject to random attacks and possible destruction. This research project investigates the delays experienced in routing packets for MCRAHNs and proposes a new routing algorithm called Spectrum-Aware Transitive Multicasting On Demand Distance Vector (SAT-MAODV) which has been optimized for reducing delays in packet transmission and increasing throughput. In the data transmission process, there are several levels where delays are experienced. Our research project focuses on Routing Path (RP) delay, Spectrum Mobility (SM) delay and Node Relay (NR) delay. This research project proposes techniques for spectrum switching and routing called Time-Based Availability (TBA), Informed Centralized Multicasting (ICM), Node Roaming Area (NRA) and Energy Smart Transitivity (EST). All these techniques have been integrated into SAT-MAODV. SAT-MAODV was simulated and compared with the best performing algorithms in MCRHANs. The results show that SAT-MAODV performs better than its counterparts