Search our Publications
Latest Publications
Marvin Struwe, Stephan Hasler, Ute Bauer-Wersing , "Combining Multiple Classifiers and Context Information for Detecting Objects under Real-world Occlusion Patterns", New Challenges in Neural Computation (NC2) Workshop, 2013.
AbstractCurrent state-of-the-art detection approaches reveal a strong degradation of performance with increasing occlusion of objects. In this paper we investigate di fferent strategies to improve detection of occluded objects based on the analytic feature framework presented and compare the results in a car detection task. Motivated by an analysis of annotated traffic scenes we fi rst describe a dedicated combination of classi fiers to deal with the pr...
Download Bibtex file
Tobias Kühnl, Franz Kummert, Jannik Fritsch , "Image-based Lane Level Positioning using Spatial Ray Features", Intelligent Vehicles Symposium (IV), 2013.
AbstractThis paper describes an approach for lane-level position estimation of a vehicle using only a single camera and no additional sensing equipment like, e.g., the typically employed IMU. The proposed method perceives a variety of local visual properties of the environment by means of base classifiers operating on patches extracted from monocular camera images, each represented in a metric confidence map obtained using inverse perspective mapping. Th...
Download Bibtex file
Nghia Le Minh, Yew Soon Ong, Stefan Menzel, Yaochu Jin, Bernhard Sendhoff , "Evolution by Adapting Surrogates", Evolutionary Computation Journal, vol. 21, no. 2, pp. 313-340, 2013.
AbstractTo deal with complex optimization problems plagued with computationally expensive fitness functions, the use of surrogates to replace the original functions within the evolutionary framework is becoming a common practice. However, the appropriate datacentric approximation methodology to use for the construction of surrogate model would depend largely on the nature of the problem of interest, which varies from fitness landscape and state of the ev...
Download Bibtex file
Stefan Klingelschmitt , "Autonomous learning for understanding traffic situations", Master thesis, TU Darmstadt, 2013.
AbstractThis thesis is concerned with developing and implementing a self-referential control architecture for autonomous learning. As current approaches towards situation understanding and scene analysis are based on visual methods or statistical learning techniques, they fail to provide a sufficient performance for proper scene understanding in more complex scenarios. Current driver assistance systems mainly provide comfort and simple safety functions l...
Download Bibtex file Per Mail Request
Antonello Ceravola , "Experience with Component Model at Honda", Object Management Group (OMG), 2013.
AbstractNo abstract...
Download Bibtex file Per Mail Request
Edgar Reehuis, Markus Olhofer, Bernhard Sendhoff, Thomas Bäck , "Novelty-guided Restarts for Diverse Solutions in Optimization of Airfoils", Evolve 2013, 2013.
AbstractA restart scheme is proposed that alternates between optimization on quality and novelty with respect to a domain-speci c distance measure, aimed at nding a diverse set of high-quality airfoil designs. Compared to the combined outcome of multiple standard optimization runs, the result set of the restart scheme is more diverse and it manages to nd solutions of strictly higher quality, while requiring only half of the budget of quality evaluation...
Download Bibtex file
Edgar Reehuis, Markus Olhofer, Bernhard Sendhoff, Thomas Bäck , "Learning-guided Exploration in Airfoil Optimization", International Conference on Intelligent Data Engineering and Automated Learning (IDEAL 2013), 2013.
AbstractA learning-based exploration approach is proposed to escape from the basins of attraction of converged-to optima, by selecting on what is termed the interestingness of a solution. This interestingness is based on the modeling error made by a surrogate model that is trained on all solutions encountered earlier during the search. Compared to multiple standard optimization runs, a learning-guided restart scheme that alternates between a quality opti...
Download Bibtex file
Federico Moro, Michael Gienger, Ambarish Goswami, Nikos Tsagarakis, Darwin Caldwell , "An Attractor-based Whole-Body Motion Control System for Humanoid Robots", IEEE RAS Int. Conf. on Humanoid Robots (Humanoids), 2013.
AbstractThis paper presents a novel whole-body torque-control concept for humanoid walking robots. The presented Whole-Body Motion Control (WBMC) system combines several unique concepts. First, a computationally efficient gravity compensation algorithm for floating-base systems is derived. Second, a novel balancing approach is proposed, which exploits a set of fundamental physical principles from rigid multi-body dynamics, such as the overall linear and ...
Download Bibtex file
Nikolas Hemion , "Integration of Internal Models based on Embodied Simulation in a Cognitive Architecture", Thesis, University Bielefeld, 2013.
AbstractBuilding robots with the ability to perform general intelligent action is a primary goal of artifi cial intelligence research. The traditional approach is to study and model fragments of cognition separately, with the hope that it will somehow be possible to integrate the specialist solutions into a functioning whole. However, while individual specialist systems demonstrate pro ficiency in their respective niche, current integrated systems remain...
Download Bibtex file Per Mail Request
Nikola Aulig and Markus Olhofer , "Evolutionary generation of neural network update signals for the topology optimization of structures", Genetic and Evolutionary Computation Conf. (GECCO Companion), pp. 213-214, 2013.
AbstractIn the adaptation of natural load bearing structures like bones and trees, regions subject to high physical loads accumulate structural material based on local stimuli, while it is reduced in others. This strategy can lead to efficient structures and has been modeled in the field of topology optimization. Instead of modeling the observed strategy we target the evolutionary process, which gave rise to theses strategies. We propose to use an evolut...
Download Bibtex file
1 ... 110 111 112 113 114 115 ... 183
Follow us
