Yoichi Motomura

A probabilistic model for appearance-based robot localization

Abstract In this paper we present a method for an appearance-based modeling of the environment of a mobile robot. We describe the task (localization of the robot) in a probabilistic framework. Linear image features are extracted using a Principal Component Analysis. The appearance model is represented as a probability density function of the image feature vector given the location of the robot. We estimate this density model from the data with a kernel estimation method. We show how the parameters of the model influence the localization performance. We also study how many features and which features are needed for good localization.

Jijo-2: an office robot that communicates and learns

Describes how the authors have combined speech recognition, dialogue management, and statistical learning procedures to develop Jijo-2; an office robot that can communicate with humans and learn about its environment.

Supervised dimension reduction of intrinsically low-dimensional data

High-dimensional data generated by a system with limited degrees of freedom are often constrained in low-dimensional manifolds in the original space. In this article, we investigate dimension-reduction methods for such intrinsically low-dimensional data through linear projections that preserve the manifold structure of the data. For intrinsically one-dimensional data, this implies projecting to a curve on the plane with as few intersections as possible. We are proposing a supervised projection pursuit method that can be regarded as an extension of the single-index model for nonparametric regression. We show results from a toy and two robotic applications.

Context-Aware Preference Model Based on a Study of Difference between Real and Supposed Situation Data

We propose a novel approach for constructing statistical preference models for context-aware recommender systems. To do so, one of the most important but difficult problems is acquiring sufficient training data in various contexts/situations. Particularly, some situations require a heavy workload to set them up or to collect subjects under those situations. To avoid this, often a large amount of data in a supposed situation is collected, i.e., a situation where the subject pretends/imagines that he/she is in a specific situation. Although there may be difference between the preference in the real situation and the supposed situation, this has not been considered in existing researches. Here, to study the difference, we collected a certain amount of corresponding data. We asked subjects the same question about preference both in the real and the supposed situation. Then we proposed a new model construction method using a difference model constructed from the correspondence data and showed the effectiveness through the experiments.

A concept of needs-oriented design and evaluation of assistive robots based on ICF

In the robotics community, a great number of assistive robots for elderly and handicapped people have been developed in the past few decades. However, very few of them became commercially available. It is often claimed that the major problems for the commercialization of robotic technologies are the “cost” and the “safety.” However we believe that the mismatch of “needs in daily lives” and “seeds in the technologies” is also a major problem. In this paper, we describe our novel ideas on the development of assistive robots which fit the real needs of users based on ICF (International Classification of Functioning, Disability and Health), which is a part of the WHO Family of International Classifications for describing whole activities of a person in daily lives. By utilizing ICF, the development process of assistive robots — analyzing and discovering needs in daily lives, designing robots and evaluating the products — will be achieved in an objective manner.

Detection and separation of speech event using audio and video information fusion and its application to robust speech interface

A method of detecting speech events in a multiple-sound-source condition using audio and video information is proposed. For detecting speech events, sound localization using a microphone array and human tracking by stereo vision is combined by a Bayesian network. From the inference results of the Bayesian network, information on the time and location of speech events can be known. The information on the detected speech events is then utilized in the robust speech interface. A maximum likelihood adaptive beamformer is employed as a preprocessor of the speech recognizer to separate the speech signal from environmental noise. The coefficients of the beamformer are kept updated based on the information of the speech events. The information on the speech events is also used by the speech recognizer for extracting the speech segment.

Combining probabilistic map and dialog for robust life-long office navigation

A design of mobile robot for robust life-long navigation in office environment is proposed and evaluated. The key idea is combining probabilistic map and dialog with humans for reducing the location uncertainty. Bayesian inference with the map represented by probabilistic automata is used in order to reduce the number of queries and to evaluate the success rate of planned paths. We experimentally implemented the design using a simple Bayesian network with continuous nodes and demonstrated its effectiveness in a real environment.

Supervised linear feature extraction for mobile robot localization

We are seeking linear projections of supervised high-dimensional robot observations and an appropriate environment model that optimize the robot localization task. We show that an appropriate risk function to minimize is the conditional entropy of the robot positions given the projected observations. We propose a method of iterative optimization through a probabilistic model based on kernel smoothing. To obtain good starting optimization solutions we use canonical correlation analysis. We apply our method on a real experiment involving a mobile robot equipped with an omnidirectional camera in an office setup.

Fusion of audio and video information for detecting speech events

In this paper, a method of detecting speech events in a multiple-sound-source condition us- ing sound and vision information is proposed. Detec- tion of speech event is an important issue for automatic speech recognition operated in a real environment. Fur- thermore, as stated an this paper, the performance of sound source separation using adaptive beamforming is greatly improved by knowing when and where the target speech event occurs. For this purpose, sound localiza- tion using a microphone array and human tracking by stereo vision is combined by a Bayesian network. From the inference results of the Bayesian network, the in- formation on time and location of speech events can be known in a multiple-sound-source condition. Results of an off-line experiment an a real environment with TV and music interference are shown.

Generative user models for adaptive information retrieval

For information retrieval (IR) tasks, user models are used to estimate users true intention and demand. Unfortunately, most user models are constructed in a specialized form that is not applied to other systems or domains. This specialization makes it difficult to share user models as common resources for developing information retrieval systems and for researching cognitive characteristics in various users. In order to solve this problem, we need a general user modeling method. A user model based on a probabilistic framework is proposed. We call this model a generative user model. The generative user model represents users mental depth by latent (hidden) variables. It also has visible variables that mean word set and qualifier of each word as a subjective probability distribution. The model can handle uncertainty of the users subjectivity by a probabilistic framework. Recent statistical studies for such latent models give a learning algorithm. Our generative user model can be constructed from a dataset taken by information retrieval tasks. As an example, we also introduce two different kinds of information retrieval systems, ART MUSEUM (Multimedia Database with Sense of Color and Construction upon the Matter of ART) and DSIU (Decision Support for Internet Users). The generative user model is applied to these systems. The properties of the model and interactive learning mechanism are shown.