Search

This paper presents a simulator for the behaviors of swarm robots based on a Dynamic Bayesian Network (DBN). Our task is to design each robot's controller which enables the robot to patrol as many regions as possible without collisions. As the first step, we use two swarm robots, each of which has two motors each of which is connected to a wheel and three distance-measurement sensors. To design the controllers of these robots, we must determine several parameters such as the motor speed and thresholds of the three sensors. The simulator is used to reduce the number of real experiments in deciding values of such parameters. We fist performed measurement experiments for our real robots in order to get probabilistic data of the DBN. The simulator based on the DBN revealed appropriate values of a threshold parameter and interesting phase transitions of their behaviors in terms of the values.

This paper treats interactive applications run on a digital TV. Digital TV services include data broadcasting services. Especially, in this paper, the authors propose a framework for developing interactive applications of data broadcast contents integrated with web services on a digital TV. The digital TV has become one of the terminal devices based on the recent ICT (Information Communication Technologies) supporting various intellectual human activities because data broadcast contents are very similar to web contents and can provide any information with multimedia data on a TV screen. The use of a digital TV is also significant to solve the digital divide problem because elder people are used to TV rather than PC. However, there are few data broadcast contents although there are a lot of various types of applications on the Web. Therefore, the integration of data broadcast contents with web services is effective way to enhance application fields of data broadcasting services. In this paper, the authors propose the application framework for that and also show several interactive applications to clarify its usefulness.

Recent advances in computer hardware technology enable us to handle 3D multimedia data more easily and many 3D Computer Graphics (CG) contents have been created and stored for various application fields. In this situation, we need any 3D multimedia data search system that allows us to efficiently retrieve our required data. The authors have proposed a 3D multimedia data search system using stochastic Attributed Relational Graph (ARG) matching. However, there is a problem that stochastic ARG matching needs huge calculation cost. To reduce such a cost, the authors propose new matching algorithm, called stochastic Attributed Relational Tree (ART) matching because calculation cost of tree matching is less than that of graph matching. The authors applied stochastic ART matching method for 3D model search and obtained better performance rather than those of the conventional matching methods besides stochastic ARG matching method.

This paper proposes a motion retrieval system using Interactive Evolutionary Computation (IEC) based on Genetic Algorithm (GA) and motion features defined based on Laban Movement Analysis (LMA) used for the similarity calculation of motions in the system. The proposed IEC-based motion retrieval system allows the user to retrieve motions similar to his/her required motions easily and intuitively only through the evaluation repeatedly performed by scoring satisfaction points to retrieved motions without entering any search queries. The authors newly define LMA-based motion features to represent them as genes of GA used for the similarity calculation in the system. This paper also clarify that the LMA-based motion features are available as similarity features of motions by showing results of analyzing them using SOM visualization.

Formation control is an important research topic in swarm robotics, as it enables each swarm robot to concentrate on its task by relying on others [1,2]. This paper is a brief summary of our endeavor for constructing low-cost swarm robots that march in column formation [4]. n = 5 swarm robots move in file, except the leading one each follows another one and all of them avoid perimeters of an indoor arena. We describe detailed settings of the control program for further studies. Our work focuses on real robots and is thus considered to be complementary to theoretical and simulation studies, cf. [3].

This paper treats a 3D multimedia data retrieval system based on a stochastic Attributed Relational Graph (ARG) matching method. ARG can be used to represent the structural feature of data, so using ARG as feature representation of 3D multimedia data, it is possible to accurately retrieve the user required data similar to his/her query in terms of its component structure. In this paper, the authors propose a 3D multimedia data search system consisting of an image/video-scene search system, a 3D model search system and a motion data search system. Each search system works commonly with a stochastic ARG matching engine. Experimental results show that the proposed system can effectively retrieve 3D multimedia data similar to the user's query in terms of component structure of the data. Especially, its 3D model search system and motion search system have better experimental results than that of other search systems.

In this chapter, we describe a novel design method of swarm robots based on the dynamic Bayesian network. Recently, an increasing attention has been paid to swarm robots due to their scalability, flexibility, cost-performance, and robustness. Designing swarm robots so that they exhibit intended collective behaviors is considered as the most challenging issue and so far ad-hoc methods which heavily rely on extensive experiments are common. Such a method typically faces a huge amount of data and handles them possibly using machine learning methods such as clustering.We argue, however, that a more principled use of data with a probabilistic model is expected to lead to a reduced number of experiments in the design and propose the fundamental part of the approach. A simple but a real example using two swarm robots is described as an application.

A person identification system based on fast face learning of lifting wavelet filters is proposed. The real power of our system lies in fast learning of lifting wavelet filters adaptive to facial parts such as eyes, nose and lips, in a set of training faces. In our system, free parameters in the lifting filter are learned fast by using Newton’s method. The learned parameters are memorized in a database together with the training faces. The lifting filters with the learned parameters in the database are applied to each of video frames which contain faces of a person, and the faces are detected by measuring some kind of distance. A person whose face is detected in a maximum number of frames is identified as a target person. To realize fast face detection, the learned filters are applied only to the skin areas separated from background by using color segmentation. Simulation results show that our person identification algorithm is accurate and fast.

This paper describes a new method for finding portions having the same feature in target signals or images from a time series or a reference image. The new method uses an integer-type Haar lifting wavelet transform. Free parameters contained in this transform are learned by using training signals or images. The advantage of this method is to be able to find portions having the same feature in the targets, and to realize robust extraction due to rounding-off arithmetic in the trained transform. In simulations, we show how well the method finds geomagnetic sudden commencements from time series of geomagnetic horizontal components, and extracts facial images from a snapshot.

This paper proposes a method for detecting geomagnetic sudden commencement (SC) from a geomagnetic horizontal (H) component by using lifting wavelet filters. Lifting wavelet filters are biorthogonal wavelet filers containing free parameters. Our method is to learn such free parameters based on some training signals which contain the SC. The learnt wavelet filters have the feature of training signals. Applying such wavelet filters to the test signals, we can detect the time when SC phenomena occurred.

Recognition of moving objects is one of the most important problems in computer vision. This problem has many applications such as individual recognition, gesture recognition and lip reading. Murase and Sakai [1] proposed a method using parametric eigenspace representation to realize efficient recognition of moving objects. This method is independent of the applications and requires a little time to recognize the moving objects. However, it is needed to normalize the position and size of silhouettes extracted from moving images