Många samhällsnyttiga tjänster är idag manuella och med låg digitaliseringsgrad. Samhällsnyttiga tjänster finns i olika delar av samhället till exempel inom skola, äldreomsorg, vägnätet, vatten och avlopp och avfallshantering. Vissa av dessa tjänster är geografiskt spridda och behöver regelbundet övervakas eller hanteras på plats som till exempel vägnätet, vatten-och avlopp och avfallshantering. Detta sker förnärvarande till stor del manuellt och skulle kunna effektiviseras med hjälp av automatisering och digitalisering. Syftet med detta examensarbetet är att skapa förståelse för hur värdet av samhällsnyttiga tjänster skulle kunna höjas med hjälp av digitalisering och automatisering med IoT. Som en del i att samla in data för detta kommer en IT artefakt för hantering av hushållens sopkärl att tas fram. I examensarbetet har ett antal designprinciper framkommit som viktiga att ta hänsyn till vid design och utveckling av samhälssnyttiga tjänster. Dessa är energikonsumtion, informationssäkerhet, användaren och andra aktörer som medskapare av värde, robusthet, Öppen Data och API:er och IoT-lösningens arkitektur.

Denna rapport täcker arbetet och resultatet av en högnivåkontroller för en robotarmsombesitterintentionenattmonteraettcylinderformatverktygiettcirkulärt hål, så kallad peg-in-hole. Detta åstadkoms genom att implementera en kontroller som ett ROS paket. Styrning av roboten gjordes med hjälp av en impedans kontroller och kraftavläsning användes som externa sensorer för guidning av roboten. Utvecklingen gjordes både i simulering och på en riktig robot. Implemtationen visade sig vara lyckad att utföra uppgiften med hänsyn till de satta kraven. Förbättringspotentialen är bättre kraftavläsning samt ett tillägg av externa sensorer för en förbättrad avläsning av robotens omgivningen. / This thesis covers the work and result of a high level controller for a robot arm with the intention of inserting a cylinder shaped tool into a circular hole, a so called peg-in-hole insertion. This was done by implementing a controller as a ROS package. An impedance controller was used for controlling the robot and force feedback readings as external senses for guiding the robot tool. Development was done in simulation and on a real robot. The implementation proved to be successful at performing an insertion in regards to the set requirements, with the potential improvements being a better force reading as well as added external senses for improved environmental perception

Att ta fram fysiska komponenter kan både vara dyrt och medföra risker. Det finns risker att de fysiska komponenterna inte fungerar som det först var tänkt. Istället för att köpa in fysiska komponenter kan företag välja att utveckla simulerade modeller för att representerar dessa fysiska komponenter. De simulerade modellerna kan sedan användas i komplexa system och interagera med sin omgivande miljö. När man simulerar systemet på detta vis möjliggör man att testa systemet i ett tidigt stadie utan att behöva ta fram fysiska komponenter Denna rapport handlar om framtagandet av ett program som möjliggör användandet av simulerade modeller i HiQs befintliga produktionstestsystem HiMacs. HiMacs är ett industriellt system som används för testkörning och mätning av motorer och transmissioner. / Developing physical components can be both expensive and involve risks. There is a risk that the physical components will not work as expected. Instead of purchasing physical components, companies can choose to develop simulated models that represents these physical components. These simulated models can then be used in complex systems and interact with their surrounding environment. When simulating the system, it is possible to test the system at an early stage without having to develop physical components. This report is about the development of a program that enables the use of simulated models in HiQ's existing production test system HiMacs. HiMacs is an industrial system that is used for test driving and measuring of engines and transmissions.

Pre-programming a robot may be efficient to some extent, but since a human has code the robot it will only be as efficient as the programming. The problem can solved by using machine learning, which lets the robot learn the most efficient way by itself. This thesis is continuation of a previous work that covered the development of the framework ​Safe-To-Explore-State-Spaces​ (STESS) for safe robot manipulation. This thesis evaluates the efficiency of the ​Q-Learning with normalized advantage function ​ (NAF), a deep reinforcement learning algorithm, when integrated with the safety framework STESS. It does this by performing a 2D task where the robot moves the tooltip on a plane from point A to point B in a set workspace. To test the viability different scenarios was presented to the robot. No obstacles, sphere obstacles and cylinder obstacles. The reinforcement learning algorithm only knew the starting position and the STESS pre-defined the workspace constraining the areas which the robot could not enter. By satisfying these constraints the robot could explore and learn the most efficient way to complete its task. The results show that in simulation the NAF-algorithm learns fast and efficient, while avoiding the obstacles without collision. / Förprogrammering av en robot kan vara effektiv i viss utsträckning, men eftersom en människa har programmerat roboten kommer den bara att vara lika effektiv som programmet är skrivet. Problemet kan lösas genom att använda maskininlärning. Detta gör att roboten kan lära sig det effektivaste sättet på sitt sätt. Denna avhandling är fortsättning på ett tidigare arbete som täckte utvecklingen av ramverket Safe-To-Explore-State-Spaces (STESS) för säker robot manipulation. Denna avhandling utvärderar effektiviteten hos ​Q-Learning with normalized advantage function (NAF)​, en deep reinforcement learning algoritm, när den integreras med ramverket STESS. Det gör detta genom att utföra en 2D-uppgift där roboten flyttar sitt verktyg på ett plan från punkt A till punkt B i en förbestämd arbetsyta. För att testa effektiviteten presenterades olika scenarier för roboten. Inga hinder, hinder med sfärisk form och hinder med cylindrisk form. Deep reinforcement learning algoritmen visste bara startpositionen och STESS-fördefinierade arbetsytan och begränsade de områden som roboten inte fick beträda. Genom att uppfylla dessa hinder kunde roboten utforska och lära sig det mest effektiva sättet att utföra sin uppgift. Resultaten visar att NAF-algoritmen i simulering lär sig snabbt och effektivt, samtidigt som man undviker hindren utan kollision.

There has been a vast increase of users that use the internet in order to communicate and interact, and as a result, the amount of data created follows the same upward trend making data handling overwhelming. Users are often asked to submit their questions on various topics of their interest, and usually, that itself creates an information overload that is difficult to organize and process. This research addresses the problem of extracting information contained in a large set of questions by selecting the most representative ones from the total number of questions asked. The proposed framework attempts to find semantic similarities between questions and group them in clusters. It then selects the most relevant question from each cluster. In this way, the questions selected will be the most representative questions from all the submitted ones. To obtain the semantic similarities between the questions, two sentence embedding approaches, Universal Sentence Encoder (USE) and InferSent, are applied. Moreover, to achieve the clusters, k-means algorithm is used. The framework is evaluated on two large labelled data sets, called SQuAD and House of Commons Written Questions. These data sets include ground truth information that is used to distinctly evaluate the effectiveness of the proposed approach. The results in both data sets show that Universal Sentence Encoder (USE) achieves better outcomes in the produced clusters, which match better with the class labels of the data sets, compared to InferSent.

Today there are more cars on the road than ever before, the automotive industry is continually expanding and adapting to meet the need of new technologies. To improve complexity management and reduce engineering time and cost ​Automotive Open System Architecture​, better known as AUTOSAR, has been introduced which aims to standardize Electronic Control Units (​ECUs)​ . Today the AUTOSAR standardization is used for the automotive industry, the purpose of the thesis is to investigate whether the standard can be used for something which has no direct connection to the automotive industry. The work done in the report is giving background information about AUTOSAR and the project is using AUTOSAR to visualize real-time data from the web on a LED-sheet. In this project a physical visualization board has been created and the code was written within the integrated software development environment Arctic Studio and its tools, the visualization board will be used at the ARCCORE office in Linköping. / Idag finns det fler bilar på vägarna än någonsin tidigare, fordonsindustrin expanderar ständigt och anpassar sig efter behovet av ny teknik. För att förbättra komplexitets hanteringen och minska tillverkningstider och kostnader så har Automotive Open System Architecture, närmare känt som AUTOSAR, införts som har målet att standardisera elektroniska styrenheter (ECUn). Idag används AUTOSAR-standardiseringen för fordonsindustrin, vad projektet kommer att utforska är att se om man kan använda standarden för något som inte har någon direkt koppling till fordonsindustrin. Rapporten ger en förklaring om AUTOSAR, i projektet så används AUTOSAR för att visualisera realtidsdata från webben på en LED-karta. I detta projekt har en fysisk visualiseringstavla skapats där kod skrevs inom den integrerade mjukvaruutvecklingsmiljön Arctic Studio, visualiseringstavlan kommer att användas på ARCCOREs egna kontor i Linköping.

<p>QC 20150325</p>

Churn analysis can be understood as a problem of predicting and understanding abandonment of use of a product or service. Different industries ranging from entertainment to financial investment, and cloud providers make use of digital platforms where their users access their product offerings. Usage often leads to behavioural trails being left behind. These trails can then be mined to understand them better, improve the product or service, and to predict churn. In this thesis, we perform churn analysis on a reallife data set from a music streaming service, Spotify AB, with different signals, ranging from activity, to financial, temporal, and performance indicators. We compare logistic regression, random forest, along with neural networks for the task of churn prediction, and in addition to that, a fourth approach combining random forests with neural networks is proposed and evaluated. Then, a meta-heuristic technique is applied over the data set to extract Association Rules that describe quantified relationships between predictors and churn. We relate these findings to observed patterns in aggregate level data, finding probable explanations to how specific product features and user behaviours lead to churn or activation. For churn prediction, we found that all three non-linear methods performed better than logistic regression, suggesting the limitation of linear models for our use case. Our proposed enhanced random forest model performed mildly better than conventional random forest. / Churn analys kan förstås som ett tillvägagångssätt för att prediktera och förstå avslutad användning av en produkt eller tjänst. Olika industrier, som kan sträcka sig från underhållning till finansiell investering och molntjänsteleverantörer, använder digitala plattformar där deras användare har tillgång till deras produkter. Användning leder ofta till efterlämnande av beteendemönster. Dessa beteendemönster kan därefter utvinnas för att bättre förstå användarna, förbättra produkterna eller tjänsterna och för att prediktera churn. I detta arbete utför vi churn analys på ett dataset från en musikstreamingtjänst, Spotify AB, med olika signaler, som sträcker sig från aktivitet, till finansiella och temporala samt indikationer på prestanda. Vi jämför logistisk regression, random forest och neurala nätverk med uppgiften att utföra churn prediktering. Ytterligare ett tillvägagångssätt som kombinerar random forests med med neurala nätverk föreslås och utvärderas. Sedan, för att ta fram regler som är begripliga för beslutstagare, används en metaheuristisk teknik för datasetet, som beskriver kvantifierade relationer mellan prediktorer och churn. Vi sätter resultaten i relation till observerade mönster hos aggregerad data, vilket gör att vi hittar troliga förklaringar till hur specifika karaktärer hos produkten och användarmönster leder till churn. För prediktering av churn gav samtliga icke-linjära metoder bättre prestanda än logistisk regression, vilket tyder på begränsningarna hos linjära modeller för vårt användningsfall, och vår föreslagna förbättrade random forest modell hade svagt bättre prestanda än den konventionella random forest.

In this thesis we investigate two separate cryptographic notions: mix-nets and hierarchical group signatures. The former notion was introduced by Chaum (1981). The latter notion is introduced in this thesis, but it generalizes the notion of group signatures which was introduced by Chaum and Heyst (1991). Numerous proposals for mix-nets are given in the literature, but these are presented with informal security arguments or at best partial proofs. We illustrate the need for a rigorous treatment of the security mix-nets by giving several practical attacks against a construction of Golle et al. (2002). Then we provide the first definition of security of a mix-net in the universally composable security framework (UC-framework) introduced by Canetti (2001). We construct two distinct efficient mix-nets that are provably secure under standard assumptions in the UC-framework against an adversary that corrupts any minority of the mix-servers and any set of senders. The first construction is based on the El Gamal cryptosystem (1985) and is secure against a static adversary, i.e., an adversary that decides which parties to corrupt before the execution of the protocol. This is the first efficient UC-secure mix-net in the literature and the first sender verifiable mix-net that is robust. The second construction is based on the Paillier cryptosystem (1999) and secure against an adaptive adversary, i.e., an adversary that decides which parties to corrupt during the execution of the protocol. This is the first efficient adaptively secure mix-net in any model. An important subprotocol in the above constructions is a zero-knowledge proof of knowledge of a witness that a party behaves as expected. There are two known approaches for constructing such a protocol given by Neff (2002) and Furukawa and Sako (2002) respectively. We present a third independent approach. We introduce the notion of hierarchical group signatures. This is a generalization of group signatures. There are several group managers, and the signers and group managers are organized in a tree in which the signers are the leaves and the group managers are internal nodes. Given a signature, a group manager learns if it is an ancestor of the signer, and if so to which of its immediate subtrees the signer belongs, but it learns nothing else. Thus, the identity of the signer is revealed in a hierarchical way. We provide a definition of security of hierarchical group signatures and give two provably secure constructions. The first construction is secure under general assumptions. It is impractical and of purely theoretical interest. The second construction is provably secure under standard complexity assumptions and almost practical.

Research within the area of cryptography constitutes the core of this the- sis. In addition to cryptography, we also present results in peer-assisted streaming and web security. We present results on two specific cryptographic problems: broadcast encryption and secure multi-party computation. Broad- cast encryption is the problem of efficiently and securely distributing content to a large and changing group of receivers. Secure multi-party computation is the subject of how a number of parties can collaborate securely. All in all, this thesis spans from systems work discussing the Spotify streaming system with millions of users, to more theoretic, foundational results. Streaming is among the largest applications of the Internet today. On- demand streaming services allow users to consume the media content they want, at their convenience. With the large catalogs offered by many services, users can access a wide selection of content. Live streaming provides the means for corporations as well as individuals to broadcast to the world. The power of such broadcasts was shown in the recent (early 2011) revolts in Tunisia and Egypt, where protesters streamed live from demonstrations. To stream media to a large global audience requires significant resources, in particular in terms of the bandwidth needed. One approach to reduce the requirements is to use peer-to-peer techniques, where clients assist in distributing the media. Spotify is a commercial music-on-demand streaming system, using peer-to-peer streaming. In this thesis, we discuss the Spotify protocol and measurements on its performance. In many streaming systems, it is important to restrict access to content. One approach is to use cryptographic solutions from the area of broadcast encryption. Within this area, we present two results. The first is a protocol which improves the efficiency of previous systems at the cost of lowered secu- rity guarantees. The second contains lower-bound proofs, showing that early protocols in the subset cover framework are essentially optimal. Many streaming systems are web-based, where the user accesses content in a web browser. Apart from this usage of the web, subscriptions for streaming services are bought using a web browser. This means that to provide a secure streaming service, we must understand web security. This thesis contains a result on a new type of attack, using an old history detection vulnerability to time the execution of a redirect of a victim’s browser. Within the area of secure multi-party computation, this thesis has three contributions. Firstly, we give efficient protocols for the basic functions of summation and disjunction which adapt to the network they run on. Secondly, we provide efficient protocols for sorting and aggregation, by using techniques from the area of sorting networks. Finally, we prove a dichotomy theorem, showing that all functions with three distinct outputs are either maximally easy or maximally difficult with regards to the security provided. / QC 20110420