Auditory overviews have the potential to improve the quality of auditory interfaces. However, in order to apply overviews well, we must understand them. Specifically, what are they and what is their impact? This thesis presents six characteristics that overviews should have. They should be a structured representation of the detailed information, define the scope of the material, guide the user, show context and patterns in the data, encourage exploration of the detail and represent the current state of the data. These characteristics are guided by a systematic review of visual overview research, analysis of established visual overviews and evaluation of how these characteristics fit current auditory overviews. The second half of the thesis evaluates how the addition of an overview impacts user interaction. While the overviews do not improve performance, they do change the navigation patterns from one of data exploration and discovery to guided and directed information seeking. With these two contributions, we gain a better understanding of how overviews work in an auditory interface and how they might be exploited more effectively.

Future 5G systems and beyond are expected to implement compact and versatile antennas in highly densifi ed millimetre-wave (MMW) wireless networks. This research emphasises on the realisation of 5G antennas provided with wide bandwidth, high gain, adaptable performance, preferably conformal implementation, and feasible bulk fabrication. Ka{band (26.5{40 GHz) is selected based on recent 5G standardisation, and novel antenna geometries are developed in this work on both rigid and flexible substrates by implementing advanced techniques of frequency reconfi guration, multiple-input-multiple- output (MIMO) assembly, as well as wideband and multiband antennas and arrays. Nove lMMW wideband antennas are presented for 5G and spatial diversity at the antenna front-ends is substantially improved by deploying wideband antennas in a MIMO topology for simultaneous multiple-channel communication. However, wideband operation is often associated with efficiency degradation, which demands a more versatile approach that allows the adaptable antenna to select the operating frequency. In this research, high performance recon figurable antennas are designed for frequency selection over Ka- {band. Also, an efficient and conformal antenna front-end solution is developed, which integrates both frequency recon guration and MIMO technology. Gain of the antenna is critically important for 5G systems to mitigate high propagation losses. Antenna design with both high gain and bandwidth is challenging as wideband antennas are traditionally gain-limited, while antenna arrays deliver high gain over a narrow bandwidth. An Enhanced Franklin array model is proposed in this thesis, which aggregates multiband response with high gain performance. Furthermore, novel flexible monopole antenna and array con gurations are realised to attain high gain profi le over the complete Ka{band. These proposed 5G antennas are anticipated as potential contribution in the progress towards the realisation of future wireless networks.

The Random Forests ensemble predictor has proven to be well-suited for solving a multitudeof different prediction problems. In this thesis, we propose an extension to the Random Forestframework that allows Random Forests to be constructed for multi-output decision problemswith arbitrary combinations of classification and regression responses, with the goal ofincreasing predictive performance for such multi-output problems. We show that our methodfor combining decision tasks within the same decision tree reduces prediction error for mosttasks compared to single-output decision trees based on the same node impurity metrics, andprovide a comparison of different methods for combining such metrics. / Program: Magisterutbildning i informatik

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Urban Studies and Planning, 1981. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ROTCH. / Bibliography: leaves BIB 1-BIB 6. / by F. Ron Jones. / Ph.D.

Thesis. 1976. M.S.--Massachusetts Institute of Technology. Dept. of Electrical Engineering and Computer Science. / Microfiche copy available in Archives and Engineering. / Includes bibliographical references. / by Antonio C. Gellineau. / M.S.

The supply chain risk management discipline studies how to address daily and extraordinary risks to avoid vulnerability and to guarantee production continuity. In the case of the electricity sector, the economic impact of interrupting the power supply is depicted by an indicator called the deficit cost, measured in monetary unit per electricity consumption unit (for example, in Brazil, R$/MWh). This value is commonly applied to cost-benefit analysis that results in decisions about maintenance and investment in the electricity system in the medium and long term, in addition to composing the short-term energy price. Most countries in the world have a thermal energy matrix, and cases in which interruption problems occur are mainly due to punctual failures in generation or transmission and last a few hours or days until maintenance is concluded. Brazil, however, has been strongly dependent on hydrological conditions ever since the main generation source became hydroelectric. Since restricted energy supply scenarios last longer, from weeks to months, a better measure for the electricity structural deficit impact is the economic loss in the deficit-affected regional production, translated as the GDP (gross domestic product) region. This dissertation estimates the Brazilian GDP marginal loss due to power deficits by applying an input-output (I/O) matrix analysis methodology and concludes that the officially adopted deficit cost is underestimated. / Sem resumo em português.

No description available.

Slow and deep breathing (SDB) has long been regarded as a nonpharmacological method for dealing with several physiological and emotional imbalances, and widely used for relaxation purposes. There is, however, limited understanding of the putative mechanisms by which SDB acutely impacts the cardiovascular and autonomic systems to elicit chronic adaptations. The present thesis explored how the manipulation of breathing pattern and intrathoracic pressure during SDB could further the understanding of the regulatory mechanisms that underpin the acute cardiovascular response to SDB. This thesis makes an original contribution to the existing knowledge by reporting a previously undescribed inversion of normal within-breath (inspiration vs. expiration) left ventricular stroke volume (LVSV) pattern for breathing frequencies < 8 breaths∙min-1. This finding might reflect the influence of a lag between enhanced right atrial filling and right ventricular stroke volume during inspiration, and its expression in left ventricular stroke volume; this lag results from the time required for blood to transit the pulmonary circulation. Furthermore, blood pressure variability (BPV) was reduced significantly at the lowest breathing frequencies, likely due to the involvement of baroreflex mediated responses. The pattern of responses was consistent with the buffering of respiratory-driven fluctuations in left ventricular cardiac output (Q̇) and arterial blood pressure (ABP) by within breath fluctuations in heart rate (fc), i.e., respiratory sinus arrhythmia (RSA) (Chapter 4). Chapter 5 demonstrated that magnifying negative intrathoracic pressure with inspiratory loading during SDB increased inspiratory pressure-driven fluctuations in LVSV and fc, and enhanced Q̇, independently of changes in VT and fR. The data support an important contribution to the amplification of RSA, during SDB, of previously underappreciated reflex, and/or 'myogenic', cardiac response mechanisms. The findings in Chapter 6 confirmed that inspiratory loading during SDB amplified the effects observed with un-loaded SDB (reported in chapter 5). In contrast, expiratory loading increased ABP and attenuated RSA, LVSV and Q̇ during SDB. A lower RSA for higher ABP, supports the presence of a formerly underappreciated contribution of sinoatrial node stretch to RSA, and throws into question the clinical benefits of expiratory resisted SDB, particularly in hypertensive populations. In conclusion, the findings of the present thesis provide novel information regarding the mechanisms contributing to acute cardiovascular response to SDB. These new insights may contribute to the development of more effective SDB interventions, geared towards maximising the perturbation to the cardiovascular control systems.

In wireless networks, the exponentially increasing demands for wireless services are encountered by the scarcity of the available radio resources. More bandwidth is required for not only accommodating the increasing number of users, but also to meet the requirements of the new services such as TV on demand, wireless gaming, and mobile Internet. Non-orthogonal multiple access (NOMA) has attracted a great attention recently due to its superior spectral efficiency (SE) over orthogonal multiple access and could play a vital role in improving the capacity of future networks. In particular, power based NOMA multiplexes the users in power domain via superposition coding (SC) and allows them to access the whole spectrum simultaneously while using successive interference cancellation (SIC) at the receiver side for signal detection. Since NOMA exploits the power domain for multiple access, power allocation is vital to achieve superior SE with NOMA. Resource allocation and its optimization are general methods used to further improve the NOMA based networks performance. In this thesis, the resource allocation in the downlink NOMA system is considered and optimized for different objective functions such as the sum rate and the energy efficiency (EE). In addition, the combination of NOMA and multiple antenna is considered using linear and non-linear precoders. In all the considered cases, suboptimal power allocation schemes are proposed and compared to the numerically obtained optimal one. Results confirm that NOMA outperforms OFDMA. It also support the effectiveness of the proposed schemes as compared to the existing ones and to the optimal one. The results also reveal that using multiple antennas with NOMA can significantly enhance the overall performance. Furthermore, a NOMA-multicell scenario is considered to test the proposed schemes under the effect of intercell interference (ICI). The results prove that the proposed methods effective as compared to the optimal one at a much lower complexity.

Zhang Ye. / "October 2001." / Thesis (Ph.D.)--Chinese University of Hong Kong, 2001. / Includes bibliographical references (p. 200-224). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Mode of access: World Wide Web. / Abstracts in English and Chinese.