Catálogo de publicaciones - libros

A system for real-time traffic sign detection is described in this paper. The system uses restricted Hough transform for circumferences in order to detect circular signs, and for straight lines for triangular ones. Some results obtained from a set of real road images captured under both normal and adverse weather conditions are presented as well in order to illustrate the robustness of the detection system. The average processing time is 30 ms per frame, what makes the system a good approach to work in real time conditions.

The field of robotics is one of the most innovative in the last decade. We are moving now from conventional, unintelligent industrial robots to mobile, intelligent, cooperative robots. This new generation of robots opens a lot of new application fields. Some of them will be growing dramatically in the nearest future. Therefore in this paper the present state will be discussed, selected applications described and an outlook on future developments will be given.

Robot soccer has evolved into a very dynamic and competitive field within the last few years. Many different robot soccer leagues now exist, the league most strongly dedicated to entertainment and edutainment is currently FIRA MiroSot. It has now reached 11 vs. 11 robots sized 7.5 x 7.5 x 7.5 cm on a 440 x 280 cm field and is therefore the first robot soccer league to physically play 11 vs. 11 games. This paper aims to provide a short status report and introduction into current problems and possible solutions within the challenging areas of FIRA MiroSot robot soccer created by having 22 robots on the field.

We describe the key components of our game strategy for robot soccer [3] in detail. The game can be represented as a trajectory in so-called the virtual grid. The virtual grid generally allows us to reduce data volume for easy description of player motion and subsequently for controlling the game or for learning game strategies. The natural coordinate system is provided be accurate optical sensing of the subject position using lens optical transformations and the CCD camera. This natural coordinate system can be easily mapped to a virtual grid.

This article is a first approach to the use of algorithm to design a team of robotic soccer playing agents for . Rete algorithm is widely used to design rule-based expert systems. Robocup Soccer Server is a system that simulates 2D robotic soccer matches. The paper presents an architecture based on team architecture for Robocup Soccer Server simulation system. It generalizes the low-level information received by the agent as high-level soccer concepts. This way it can take advantage of expert system design techniques.

The objective of this contribution is to establish a 1-1 correspondence between an existing biological model of so-called tripartite synapses in biological brains and a corresponding autonomous multiagent system (MAS) in an unexplored environment. The logical part of the mMAS model is based on the concept of logical fiberings – systems of distributed logics for MAS. Two important notions, intention and rejection, will be subject of the development of a suitable mathematical formalization – the notion of space- and time dependent logical formulas will play a basic role. Our new general model for MAS is based on category theory and general categorical semantics. Of further basic interest in MAS is the issue of learning with respect to artificial neural network applications to agent systems and robotics. According to the cybernetic principle of feasibility, it is in the center of our work to achieve implementations of autonomous robots based on the proposed biomimetic MAS model.

In this paper we describe the design and construction of a distributed controller network for the humanoid robot ”Archie”. The mechanical design of the robot incorporates 6 DOF per leg, 6 DOF per arm and 3 DOF for the torso and the head respectively. The network consists of 3 subnetworks, each including a set of dedicated processing nodes interconnected by a CAN bus. This partitioning into levels of competence guarentees failure tolerance and reliability and minimizes communication overhead even on heavy load conditions.

This document presents the first operating version of CoolBOT, a component oriented software framework for programming robotic systems. CoolBOT has been designed having in mind the idea of programming by integrating software components, in order to reduce the developing effort typically invested when programming robots. CoolBOT also fosters some interesting features, such as asynchronous execution, asynchronous inter communication, data-flow-driven processing, and cognizant failure systems. A simple demonstrator illustrates the benefits of using the proposed approach.

This paper presents a new general method for obstacle representation in the configuration space (C-space) for redundant robots. The method is based on the analytical deconstruction of the C-space, i.e., the separated evaluation of the C-space portion contributed by the collisions of each link in the kinematic chain. The systematic application of a simple convolution of two functions describing each link and the workspace, respectively, is applied. In order to do that, the transformation of the workspace among reference systems located at one point of each link is needed; in this step a well-known and sound method is used.

It has become customary to use epistemic modal logic for the formal study of knowledge and beliefs of agents. Based on the concept of logical fibering, which is briefly summarized, we present so-called fibered global modal logics for the logical modeling of multiagent systems (MAS). Considering a simple multi-robot scenario we show that our global logical models allow to study the knowledge and beliefs of the agents, taking into account the communication between the agents.