Catálogo de publicaciones - libros

Interaction means to share a communicative space with others. Social interactions are reciprocally-oriented activities among currently present partners. An artificial system can be such a partner for humans. In this study, we investigate the effect of disturbance in human-robot interaction. Disturbance in communication is an attention shift of a partner caused by an external factor. In human-human interaction, people would cope with the problem to continue to communicate because they presuppose that the partner might get irritated and thereby shift his/her interactive orientation. Our hypothesis is that people reproduce a social attitude of reattracting the partner’s attention by varying their communication channels even toward a robot. We conducted an experiment of hybrid interaction between a human and a robot simulation and analyzed it from a sociological and an engineering perspective. Our qualitative analysis revealed that people established a communicative space with our robot and accepted it as a proactive agent.

This contribution addresses the generation of textual descriptions in several natural languages for evaluation of human behavior in video sequences. The problem is tackled by converting geometrical information extracted from videos of the scenario into predicates in fuzzy logic formalism, which facilitates the internal representations of the conceptual data and allows the temporal analysis of situations in a deterministic fashion, by means of Situation Graph Trees (SGTs). The results of the analysis are stored in structures proposed by the Discourse Representation Theory (DRT), which facilitate a subsequent generation of natural language text. This set of tools has been proved to be perfectly suitable for the specified purpose.

There are two significant challenges to standard approaches to detect humans through computer vision. First, scenarios when the poses and postures of the humans are completely unpredictable. Second, situations when there are many occlusions, i.e., only parts of the body are visible. Here a novel approach to perception is presented where a complete 3D scene model is learned on the fly to represent a 2D snapshot. In doing so, an evolutionary algorithm generates pieces of 3D code that are rendered and the resulting images are compared to the current camera picture via an image similarity function. Based on the feedback of this fitness function, a crude but very fast online evolution generates an approximate 3D model of the environment where non-human objects are represented by boxes. The key point is that 3D models of humans are available as code sniplets to the EA, which can use them to represent human shapes or portions of them if they are in the image. Results from experiments with real world data from a search and rescue application using a thermal camera are presented.

In content-based image retrieval we are faced with continuously growing image databases that require efficient and effective search strategies. In this context, shapes play a particularly important role, especially as soon as not only the overall appearance of images is of interest, but if actually their content is to be analysed, or even to be recognised. In this paper we argue in favour of numeric features which characterise shapes by single numeric values. Therewith, they allow compact representations and efficient comparison algorithms. That is, pairs of shapes can be compared with constant time complexity. We introduce three numeric features which are based on a qualitative relational system. The evaluation with an established benchmark data set shows that the new features keep up with other features pertaining to the same complexity class. Furthermore, the new features are well-suited in order to supplement existent methods.

Ellipse detection is very important in computer vision, object recognition, feature selection and so on. This paper proposes a new ellipse detection method using local information of edge points. It merges line segments using modified randomized Hough transform (RHT) that belongs to same ellipse. It is fast, correct and robust to noise because it detects ellipse using line segments that are constructed by local information of edge points. The proposed method in this paper can not apply only ellipse detection but line, circle and partially occluded elliptical object.

We present an approximation method that solves a class of Decentralized hybrid Markov Decision Processes (DEC-HMDPs). These DEC-HMDPs have both discrete and continuous state variables and represent individual agents with continuous measurable state-space, such as resources. Adding to the natural complexity of decentralized problems, continuous state variables lead to a blowup in potential decision points. Representing value functions as Rectangular Piecewise Constant (RPWC) functions, we formalize and detail an extension to the Coverage Set Algorithm (CSA) [1] that solves transition independent DEC-HMDPs with controlled error. We apply our algorithm to a range of multi-robot exploration problems with continuous resource constraints.

is a logic-based agent programming language and combines many important features from other dialects. One of the features of is to make use of decision-theoretic planning for specifying the behavior of an agent or robot. In this paper we show a method to reduce the planning time for decision-theoretic planning in the framework. Instead of planning policies on the fly over and over again, we calculate an abstract policy once and store it in a plan library. This policy can later be re-instantiated. With this plan library the on-line planning time can be significantly reduced. We compare computing policies on the fly with those stored in our plan library with examples from the robotic soccer domain. In the 2D soccer simulation league we show the significant speed-up when using our plan library approach. Moreover, the use of the plan library together with a suitable state space abstraction for the soccer domain makes it possible to apply macro-actions in an otherwise continuous domain.

This paper describes a system for the systematic construction and evaluation of planning strategies. It is based on a proper formal account of refinement planning and allows to decouple plan-deficiency detection, refinement computation, and search control. In adopting this methodology, planning strategies can be explicitly described and easily deployed in various system configurations.

We introduce novel domain-independent planning strategies that are applicable to a wide range of planning capabilities and methods. These so-called strategies are guided by information about current plan defects and solution options. The results of a first empirical performance evaluation are presented in the context of hybrid planning.

In two players try to connect opposing sides by placing pieces onto a rhombus-shaped board of hexagons. The game has a high strategic complexity and the number of possible board positions is larger than in . There are already some programs of recognizable strength, but which still play on a level below very strong human players. One of their major weaknesses is the time for evaluating a board.

In this work we apply machine learning for the computer player to improve his play by generating an fast evaluation function and lookup procedure for pattern endgame databases. The data structures used are neural networks for the evaluation of a position and limited branching trees to determine if a position can be classified as won or lost.

In this work, a new stochastic method for optimization problems is developed. Its theoretical bases guaranteeing the convergence of the method to a minimum of the objective function are presented, by using quite general hypotheses. Its application to recurrent discrete neural networks is also developed, focusing in the multivalued MREM model, a generalization of Hopfield’s. In order to test the efficiency of this new method, we study the well-known Traveling Salesman Problem. Experimental results will show that this new model outperforms other techniques, achieving better results, even on average, than other methods.