The increasing complexity of decision situations has required organizations to integrate more types of expertise and consider more criteria for effective group decision making. Researchers have begun to examine how computer based support in the form of a Group Decision Support System (GDSS) can enhance the process and outcomes of decision making groups. This dissertation investigated the impact of GDSS for strategic planning impact analysis. The GDSS was based on the Stakeholder Identification and Assumption Surfacing Model. A controlled laboratory experiment was used to compare the process and outcomes of 4-person groups which had GDSS support, comparable manual support, and no support. The experimental task was a policy statement requiring undergraduates to have a personal computer for admittance to a business college. Groups were asked to determine a list of the most critical stakeholders who would be impacted by the policy, and their assumptions regarding the policy statement. Measures were taken on decision outcomes (decision quality, decision time, and satisfaction with the outcomes) and decision process variables (quantity of unique alternatives, distribution of individual participation, and satisfaction with the process). Additionally, observational data was recorded through the use of videotape recordings of the sessions. The major findings of the study are: (1) Decision quality is enhanced when groups use a structured methodology; (2) Decision time was shortest in the unstructured groups, with GDSS groups finishing somewhat faster than manual structured groups; (3) Satisfaction with the outcomes was not different between structured and unstructured groups, however it was higher in the GDSS groups compared to the structured manual groups; (4) Quantity of unique alternatives was much higher in the groups using a structured methodology; (5) Distribution of individual participation was more equal in groups using a structured methodology; and (6) Satisfaction with the process was not different between structured and unstructured groups, however the GDSS groups were more satisfied than the structured manual groups.

Current research in screen design for human computer interaction has demonstrated that user task performance is influenced by placement, prompting methodology, and screen complexity. To assess the difference between a grouped item screen prompt and a series of single item screen prompts, a field experiment in a semiconductor manufacturing facility was designed. Subjects were randomly assigned to one of two groups to use a data entry system. Seven of the screen prompts differed between the two groups. During the four weeks of the study, a significant difference was measured between groups. The group screen users had lower task times for all four weeks. No significant correlation exists between work experience, performance review scores, or designated work shift. A strong negative correlation exists between frequency of system usage and task time. No difference was noted for measured errors. Subjective scores significantly favored the group screen design.

No description available.

The aim of this thesis is to explore the possibility of integrating an AC3 audio de- coding module into the company’s current product. Due to limited left resources on the FPGA chip in the company’s current product, the focus of this thesis is to be resource efficient. In this thesis, a system for AC3 audio decoding is designed and implemented. In order to use less logic on FPGA, PicoBlaze soft processor is used to control the whole processing flow. The system is designed and synthe- sized for a Spartan-6 FPGA which can be easily ported to the company’s current platform.

Interference is one of the most common causes of performance issues and reliability issues in wireless networks. Interference and background noise can be localized by performing spectrum analysis to measure amplitude and frequency characteristics of electromagnetic signals. Initially a theoretical description of spectrum analysis and site survey was performed. Manual spectrum analysis was also examined by performing practical measurements and creating graphs with Metageek Wi-Spy and Chanalyzer. Signals from 802.11 units were observed with spectrum analysis, and were analyzed according to their visual interference levels. This was compared to the actual interference impact of the units. This analysis showed that manual spectrum analysis could be hard to interpret when it comes to interference, as signals from units that looked harmful were in fact not. The channel access method of the units, in this case CSMA/CA, contributed to a low impact of interference, since the units waited for the medium to be free before sending, which reduced the probability of collisions. Cisco CleanAir technology has built-in spectrum analysis functionality, to detect, classify and avoid radio frequency interference. Dynamic interference detection and avoidance reduces costs in operating and network outages. This is made possible by Cisco Radio Resource Management (RRM) - a system consisting of multiple protocols that can dynamically change frequency channel and power. The protocols of RRM were examined and discussed relative to the lab measurement. The measurements showed that the Cisco Aironet 2700 access point was more robust against interference compared to a D-Link router for small offices and home use. The robustness of an access point depends on its wireless standard, modulation and dynamic management protocols. Wireless communication is a complex subject, where multiple factors has an impact on network performance. Therefore, it might not be appropriate to draw any hard conclusions from the laboratory measurements that were performed. For future studies, similar lab measurements should be carried out by creating interference from non-802.11 units to compare the interference impact from CSMA and non-CSMA communication. / Störningar är en av de vanligaste orsakerna till försämrad prestanda och tillförlitlighet i trådlösa nätverk. Frekvensstörningar och brus kan lokaliseras genom att utföra spektrumanalys för att mäta amplitud- och frekvensegenskaper hos elektromagnetiska signaler. Inledningsvis gjordes en teoretisk beskrivning av spektrumanalys och site survey. Manuell spektrumanalys undersöktes dessutom praktiskt genom laborationsmätningar och grafritning med Metageek Wi-Spy och dess mjukvara Chanalyzer. Signaler från 802.11-enheter betraktades med spektrumanalys, och analyserades utifrån hur störande dess signaler såg ut att vara. Detta jämfördes sedan mot den faktiska störningsinverkan enheterna hade. Analysen visade att manuellt genomförd spektrumanalys kan vara svårtolkad när det kommer till att identifiera störande moment, då signaler från enheter som såg ut att störa mycket i själva verket inte gjorde det. Enheternas kanalaccessmetod, i det här fallet CSMA/CA, bidrog till en låg störningsinverkan då enheterna väntade med att sända tills mediet var ledigt, vilket minskade risken för kollisioner. Ciscos nya CleanAir-teknologi har inbyggd spektrumanalysfunktionalitet, för att kunna upptäcka, klassificera och undvika radiofrekvensstörningar. Att störningar automatiskt kan upptäckas och åtgärdas sparar både driftskostnader och minimerar nätverksavbrott. Denna automatisering är möjlig tack vare Cisco Radio Resource Management (RRM) som är ett system bestående av flera protokoll, och innefattar bland annat funktioner för att byta kanal och ändra utsänd effekt automatiskt. De protokoll som ingår i RRM undersöktes och diskuterades sedan i förhållande till genomförda laborationsmätningar. Laborationsmätningar visade att Ciscos Aironet 2700-accesspunkt var mer robust mot störningar än en D-Link-router avsedd för hemmabruk och mindre kontorsmiljöer. En accesspunkts robusthet avgörs av dess trådlösa standard, modulationsteknik och dynamiska managementprotokoll. Trådlös kommunikation är ett komplext ämne, där många faktorer har betydelse för ett nätverks prestanda. Det är därför inte lämpligt att dra några konkreta slutsatser från de laborationsmätningar som genomfördes. För framtida arbeten bör liknande laborationer med störningar från icke-802.11-enheter genomföras för att jämföra störningsinverkan från CSMA- och icke-CSMA-kommunikation.

När en användare idag efterfrågar rendering av en kartvy i en applikation behöver en server först rendera en bild utifrån given geografisk data och därefter skicka bilden till klientens mobila enhet. Detta kan resultera i höga responstider, speciellt för användare som befinner sig i områden med bristfällig täckning. I denna studie utvärderas en alternativ lösning där rendering istället sker direkt på klientens enhet. En prototyp av en mobil kartapplikation med stöd för lokal rendering av geografisk rådata utvecklas och utvärderas utefter en konstant för acceptabel fördröjning vid visualisering av information. Resultatet av testerna visar att prototypens prestanda är beroende av mängden information som ska visas. För högre zoomnivåer ger prototypen ett tillfredsställande resultat, men vidare åtgärder krävs för de lägre nivåerna. De främsta utmaningarna vid utvecklingen av applikationen redovisas och förbättringsförslag för fortsatt utveckling framförs.

Teams consisting of only software developers will occasionally need to develop products that has to be easy to use. User Centered Design (UCD) is one approach to help increase the ease of use of a product and that can be incorporated into the teams’ traditional workflow when needed. The software developer team followed in this thesis had not tried to incorporate UCD into their current workflow when developing such products. So, this thesis looks at how a product designed with an agile/UCD approach differs from a product developed by the software development team’s traditional approach. The two products were designed to solve the same problem. The results show that the product developed with an agile/UCD-approach gave better usability rankings and better external appeal than the one created with the team’s traditional approach. On the other hand, traditional methods constitute a better choice for quick development of products for more technical user groups within the company for example.

A dissertation submitted to the Faculty of Engineering, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Master of Science in Engineering. Johannesburg, 1996. / This report examines the major factors influencing the development of embedded systems. The discussion covers, among other aspects, features of embedded systems that are common to software development in general, the main characteristics of embedded systems, some of the current development strategies, the mlcrocontroller market, language issues and development tools. Following the assessment study and a practical embedded system development, the report concludes that although: I embedded systems are typified by relatively few lines of codes, their complex nature necessitates the application of a disciplined development approach. The report presents a riskbased hardware-software co-design development approach in an attempt to reconcile the existing formal development models to the technical realities of embedded systems. A practical embedded system case study is presented, It involves the development of a microcontroller-based system for the three-phase inverter of an uninterruptible power supply (UPS).

Machine learning has been around for a long time, the applications range from a big variety of different subjects, everything from self driving cars to data mining. When a person takes a picture with its mobile phone it easily happens that the photo is a little bit crooked. It does also happen that people takes spontaneous photos with help of their phones, which can result in something irrelevant ending up in the corner of the image. This thesis combines machine learning with photo editing tools. It will explore the possibilities how machine learning can be used to automatically crop images in an aesthetically pleasing way and how machine learning can be used to create a portrait cropping tool. It will also go through how a straighten out function can be implemented with help of machine learning. At last, it is going to compare this tools with other software automatic cropping tools. / Maskinlärning har funnits en lång tid. Deras jobb varierar från flera olika ämnen. Allting från självkörande bilar till data mining. När en person tar en bild med en mobiltelefon händer det lätt att bilden är lite sned. Det händer också att en tar spontana bilder med sin mobil, vilket kan leda till att det kommer med något i kanten av bilden som inte bör vara där. Det här examensarbetet kombinerar maskinlärning med fotoredigeringsverktyg. Det kommer att utforska möjligheterna hur maskinlärning kan användas för att automatiskt beskära bilder estetsikt tilltalande samt hur maskinlärning kan användas för att skapa ett porträttbeskärningsverktyg. Det kommer även att gå igenom hur en räta-till-funktion kan bli implementerad med hjälp av maskinlärning. Till sist kommer det att jämföra dessa verktyg med andra programs automatiska beskärningsverktyg.

Organisationer är utsatta för ständiga säkerhetshot på internet och penetrationstester uppdagar hur sårbara nätverken är när mjuk- och hårdvara inte är uppdaterade. Uppdateringar i IT-sammanhang kallas ”patchar” och brukar generellt förbättra antingen funktioner eller säkerhet. Det finns en arbetsprocess inom IT kallad patch management, som ansvarar för hur mjukvara och annan utrustning uppdateras för att göra nätverket säkrare. Idag finns stora utmaningar i arbetsprocessen och denna studie undersöker hur arbetet kan effektiviseras. Ett problem historiskt sett har varit att det släppts för många patchar, vilket gjort det svårare för organisationer att hålla sig uppdaterade. Enligt rekommendationer från standardiseringsorgan ska patchar helst testas innan de implementeras, för att undvika eventuella följdproblem som kan uppstå. Genom intervjuer med personer som har ansvar för patch management, visades att det finns framgångsrika metoder för att hålla systemen uppdaterade, men delvis genom att bortse från vissa rekommenderade arbetsmetoder. Automatiserade verktyg underlättar processen till stor del men det finns delprocesser som ännu inte har blivit fullgott automatiserade. Test av patchar har närmast helt förbigåtts i organisationer vars nätverk är anslutna mot internet, därför att testprocessen i dagsläget tar för mycket tid. Slutsatsen som dragits är att det är att säkrare att snabbt lösa eventuella problem som uppstår på grund av en dålig patch, hellre än att testa patchar under långa perioder, eftersom nätverket är sårbart så länge ett känt säkerhetshål inte har täppts igen. / Organizations are exposed to constant security threats from the internet and penetration tests reveal just how vulnerable networks are when software and hardware patching aren’t up to date. Updates, known in IT as “patches”, usually enhances functions or security. Patch Management is the field in which anything related to patching of software and other various equipment falls under. As of today, Patch Management faces great challenges and the purpose of this study is to understand how the process can be made more efficient. Historically, a common issue has been the number of patch releases, which has made it cumbersome for organizations to stay up to date. Standardization bodies, such as IEC and NIST recommend that patches are tested in test environments before being installed to the production environment, to make sure no unintended consequences arise from faulty patches. Through interviews with professionals working in Patch Management, it became clear that there are ways to stay up to date, but partly through disregarding recommended best practice. Automated tools ease the Patch Management process to great extents but there are still areas that remain non-automated. The testing process has been largely ignored by organizations whose networks are connected to the internet, because said process is much too inefficient. Their answer to the problem of staying up to date is to solve problems quickly that arise through faulty patching, rather than test patches over longer periods of time. Their reasoning being that leaving known vulnerabilities unpatched is more damaging to the network.