Refine
Year of publication
Document Type
- Conference Proceeding (608) (remove)
Language
- English (463)
- German (144)
- Multiple languages (1)
Has Fulltext
- no (608) (remove)
Keywords
- 360-degree coverage (1)
- 3D Extended Object Tracking (EOT) (2)
- 3D ship detection (1)
- AAL (1)
- ADAM (1)
- AHI (1)
- Abrasive grain material (1)
- Academic german (1)
- Accelerometer (2)
- Accelerometer sensor (1)
Institute
- Fakultät Bauingenieurwesen (5)
- Fakultät Elektrotechnik und Informationstechnik (9)
- Fakultät Informatik (48)
- Fakultät Maschinenbau (7)
- Fakultät Wirtschafts-, Kultur- und Rechtswissenschaften (8)
- Institut für Angewandte Forschung - IAF (50)
- Institut für Optische Systeme - IOS (14)
- Institut für Strategische Innovation und Technologiemanagement - IST (27)
- Institut für Systemdynamik - ISD (61)
- Institut für Werkstoffsystemtechnik Konstanz - WIK (4)
Loch im Putz = alles neu?
(2016)
A method is investigated by which tight bounds on the range of a multivariate rational function over a box can be computed. The approach relies on the expansion of the numerator and denominator polynomials in Bernstein polynomials. Convergence of the bounds to the range with respect to degree elevation of the Bernstein expansion, to the width of the box and to subdivision are proven and the inclusion isotonicity of the related enclosure function is shown.
Ulrich Finsterwalder
(2016)
The business plan is one of the most frequently available artifacts to innovation intermediaries of technology-based ventures' presentations in their early stages [1]–[4]. Agreement on the evaluations of venturing projects based on the business plans highly depends on the individual perspective of the readers [5], [6]. One reason is that little empirical proof exists for descriptions in business plans that suggest survival of early-stage technology ventures [7]–[9]. We identified descriptions of transaction relations [10]–[13] as an anchor of the snapshot model business plan to business reality [13]. In the early-stage, surviving ventures are building transaction relations to human resources, financial resources, and suppliers on the input side, and customers on the output side of the business towards a stronger ego-centric value network [10]–[13]. We conceptualized a multidimensional measurement instrument that evaluates the maturity of this ego-centric value networks based on the transaction relations of different strength levels that are described in business plans of early-stage technology ventures [13]. In this paper, the research design and the instrument are purified to achieve high agreement in the evaluation of business plans [14]–[16]. As a result, we present an overall research design that can reach acceptable quality for quantitative research. The paper so contributes to the literature on business analysis in the early-stage of technology-based ventures and the research technique of content analysis.
Digital cameras are used in a large variety of scientific and industrial applications. For most applications the acquired data should represent the real light intensity per pixel as accurately as possible. However, digital cameras are subject to different sources of noise which distort the resulting image. Noise includes photon noise, fixed pattern noise and read noise. The aim of the radiometric calibration is to improve the quality of the resulting images by reducing the influence of the different types of noise on the measured data. In this paper, a new approach for the radiometric calibration of digital cameras using sparse Gaussian process regression is presented. Gaussian process regression is a kernel based supervised machine learning technique. It is used to learn the response of a camera system from a set of training images to allow for the calibration of new images. Compared to the standard Gaussian process method or flat field correction our sparse approach allows for faster calibration and higher reconstruction quality.
Measuring the natural frequency of buildings and bridges is a possibility to get information about the stiffness of the construction. Decreasing stiffness can be detected be repeatedly measurements. Damaged parts of the construction or too high wood moisture can be reasons for decreasing stiffness. The earlier the failure is detected the better is the chance to repair it with low costs. The method of monitoring by repeatedly measuring the natural frequency is applied at timber bridges, especially on footbridges. As damages due to high wood moisture cannot be seen easily, measuring the natural frequency is a good possibility to detect them and then to repair them. Equations to calculate the natural frequencies regarding the damaged parts are shown and applied to a simple supported beam.
Cognitive radio (CR) is a key enabler of wireless in industrial applications especially for those with strict quality-of-service (QoS) requirements. The cornerstone of CR is spectrum occupancy prediction that enables agile and proactive spectrum access and efficient utilization of spectral resources. Hidden Markov Models (HMM) provide powerful and flexible tools for statistical spectrum prediction. In this paper we introduce a HMM-based spectrum prediction algorithm for industrial applications that accurately predicts multiple slots in the future. Traditional HMM prediction approaches use two hidden states enabling the prediction of only one step ahead in the future. This one step is most often not enough due to internal hardware delays that render it outdated. We show in this work that extending the number of hidden states and formulating the prediction problem as a maximum likelihood (ML) classification approach enables a prediction span of multiple slots in the future even with fine spectrum sensing resolution. We verify the suitability of our approach to industrial wireless through extensive simulations that utilize a realistic measurement-based traffic model specifically tailored for industrial automotive settings.
Recent years have seen the proposal of several different gradient-based optimization methods for training artificial neural networks. Traditional methods include steepest descent with momentum, newer methods are based on per-parameter learning rates and some approximate Newton-step updates. This work contains the result of several experiments comparing different optimization methods. The experiments were targeted at offline handwriting recognition using hierarchical subsampling networks with recurrent LSTM layers. We present an overview of the used optimization methods, the results that were achieved and a discussion of why the methods lead to different results.