川嶋 宏彰

To design interactive systems that proactively assist users' decision making, the users' gaze information is an important cue for the system to estimate users' selection criteria. Users sometimes change selection criteria while browsing content. Therefore, temporal changes of those criteria need to be traced from gaze data in short time scales. In this paper, we propose an approach to detecting users' distinctive browsing periods with its appropriate time-scale by leveraging multiscale exact tests so that the system can trace temporal changes of selection criteria. We demonstrate the applicability of the proposed method through a toy example and experiments.

With the increase in demand flexibility and the rapid introduction of uncontrollable renewable power sources, effective schemes for managing the power usage of end-users are required. In this line, we propose an energy management framework for coordinating the power usage of communities of networked agents. More specifically, communities (groups of loads) coordinate to minimize their aggregated power imbalance, while taking into account each community's objectives and constraints, as well as the preferred power usage pattern of each end-user. For having a robust coordination that can work under unexpected events, we propose to assign the agents to communities using a measure of the flexibility of sets of agents. The coordination framework builds on the alternating directions method of multipliers (ADMM), algorithm that is used to implement a distributed coordination using a hierarchical architecture. While the distributed coordination allows to manage the power usage of each end-user, the hierarchical architecture enables the integration, in a single framework, of energy management problems that would be otherwise handled independently. We illustrate and analyze the coordination framework using three simulated scenarios. (C) 2015 Elsevier Ltd. All rights reserved.

Eye movements are an important cue to understand consumer decision processes. Findings from existing studies suggest that the consumer decision process consists of a few different browsing states such as screening and evaluation. This study proposes a hidden Markov-based gaze model to reveal the characteristics and temporal changes of browsing states in catalog browsing situations. Unlike previous models that employ a heuristic rule-based approach, our model learns the browsing states in a bottom-up manner. Our model employs information about how often a decision maker looks at a selected item (the item finally selected by a decision maker) to identify the browsing states. We evaluated our model using eye tracking data in catalog browsing and confirmed our model can split decision process into meaningful browsing states. Finally, we propose an estimation method of browsing states that does not require the information of the selected item for applications such as an interactive decision support.

Studies on gaze analysis have revealed some of the relationships between viewers' gaze and their internal states (e.g., interests and intentions). However, understanding content browsing behavior in uncontrolled environments is still challenging because human gaze can be very complex; it is affected not only by viewers' states but also by the spatio-semantic structures of visual content. This study proposes a novel gaze analysis framework which introduces the content creators' point of view to understand the meaning of browsing behavior. Visual content such as web pages, digital articles and catalogs are comprised of structures intentionally designed by content creators, which we refer to as designed structure. This paper focuses on two design factors of designed structure: spatial structure of content elements (content layout), and their relationships such as "being in the same group". The framework was evaluated with an experiment involving 12 participants, wherein the participant's state was estimated from their gaze behavior. The results from the experiment show that the use of design structure improved estimation accuracies of user states compared to other baseline methods.

This paper explores how haptic interfaces should be designed to enable effective human-swarm interactions. When a single operator is interacting with a team of mobile robots, there are certain properties of the team that may help the operator complete the task at hand if these properties were fed back via haptics. However, not all team-level properties may be particularly well-suited for haptic feedback. In this paper, characteristics that make a property of a multi-agent system appropriate for haptic feedback are defined. The focus here is on leader-follower networks, in which one robot, the so-called leader, is controlled via an operator with a haptic device, whereas the remaining robots, the so-called followers, are tasked with maintaining distances between one another. Multi-agent manipulability, a property which describes how effective the leader is at controlling the movement of the followers, is proposed as one such appropriate property for haptic feedback in a human-swarm interaction scenario. Manipulability feedback is implemented using a PHANTOM Omni haptic joystick and experiments in which a team of mobile robots is controlled via a human operator with access to this feedback show that this is viable in practice.

This paper introduces the notion of manipulability to mobile, multi-agent networks as a tool to analyze the instantaneous effectiveness of injecting control inputs at certain, so-called leader nodes in the network. Effectiveness is interpreted to characterize how the movements of the leader nodes translate into responses among the remaining follower nodes. This notion of effectiveness is a function of the interaction topologies, the agent configurations, and the particular choice of inputs used to influence the network. In fact, classic manipulability is an index used in robotics to analyze the singularity and efficiency of configurations of robot-arm manipulators. To define similar notions for leader follower networks, we use a rigid-link approximation of the follower dynamics and, under this assumption, we prove that the instantaneous follower velocities can be uniquely determined from that of the leaders', which allows us to define a meaningful and computable manipulability index for the leader follower networks. This paper examines the property of the proposed index in simulation and with real mobile robots, and demonstrates how the index can be used to find effective interaction topologies. (C) 2014 Elsevier Ltd. All rights reserved.

The development of energy management systems (EMS) has attracted increasing attention during the last years. One of the main goals of EMS is to balance the power usage and generation, while also maintaining the quality of life (QoL) of the users. In this paper, a distributed coordination framework for on-line scheduling of appliances and control of the aggregated power demand of households is proposed. Each household consists of a set of appliances, and each appliance is modeled using a probabilistic generative model of its power usage profile, which can be updated through the day to indicate changes in user preferences. The coordination framework is formulated as a receding horizon distributed optimization, where the households' QoL, and the deviation from a scheduled usage are taken into account. The implemented distributed optimization can be seen as a negotiation among the households and a coordinator: the coordinator seeks to balance the aggregated power consumption by minimizing the deviation from a scheduled aggregated power usage, while each household tries not to deviate much from its preferred usage pattern.

When we are watching videos, there are spatiotemporal gaps between where we look (points of gaze) and what we focus on (points of attentional focus), which result from temporally delayed responses or anticipation in eye movements. We focus on the underlying structure of those gaps and propose a novel learning-based model to predict where humans look in videos. The proposed model selects a relevant point of focus in the spatiotemporal neighborhood around a point of gaze, and jointly learns its salience and spatiotemporal gap with the point of gaze. It tells us "this point is likely to be looked at because there is a point of focus around the point with a reasonable spatiotemporal gap." Experimental results with a public dataset demonstrate the effectiveness of the model to predict the points of gaze by learning a particular structure of gaps with respect to the types of eye movements and those of salient motions in videos. © 2013 Information Processing Society of Japan.

This paper proposes a distributed scheduling algorithm of demands of multiple households/consumers in order to achieve a power balancing in total. Assume that an autonomous energy management system is installed in each household and that those households are capable of communicating with an aggregator. Then, it becomes possible to negotiate via the autonomous energy management systems to find an agreement point that takes both the users' demands and the aggregator's objective into account. The key aspect of the algorithm is that it enables us to encapsulate the particular control and objective in each household, and realizes the negotiation based on power profiles. We also show the proposed framework can be smoothly integrated with a probabilistic generative model of power profiles.

This paper unifies the previously defined instantaneous measures of responsiveness in multi-agent systems, namely those of stiffness and manipulability. These concepts are used for characterizing how a networked multi-agent system responds to perturbation of agents or to exogenous movements of its leader nodes. The resulting unified notion of a system's responsiveness provides us with a precise characterization of the effect of different choices of leader nodes and interaction topologies have on how easy or hard it is to control the network.

This paper presents a novel framework to interpret eye movements using semantic relations and spatial layouts of displayed contents, i.e., the designed structure. We represent eye movements in a multi-scale, interval-based manner and associate them with various semantic relations derived from the designed structure. In preliminary experiments, we apply the proposed framework to the eye movements when browsing catalog contents, and confirm the effectiveness of the framework via user-state estimation. © 2012 ACM.

In this paper, we address the problem of selecting leaders in a network by investigating how much instantaneous impact the leaders have on the remaining agents. As a measurement of the influence of leaders' inputs, we exploit the notion of manipulability, which is recently developed for leader-follower networks driven by a state-dependent weighted consensus equation. This paper first extends the manipulability index in order to measure the influence of leaders' inputs on the network centroid. We then demonstrate in simulation how the manipulability index is suitable for selecting effective leaders.

This paper applies a general computational technique for optimal control of switched-mode hybrid systems, recently developed by the authors, to voltage-regulation problems in a step-down DC-DC converter. Unlike existing techniques that are based on model-predictive control and a specific algebraic structure of the problem, the algorithms presented here are based on gradient descent with Armijo step size, and consequently can incorporate time-dependent state constraints in a natural way. The approach proposed in this paper is complementary to the extant MPC-based techniques, and it appears to compare favorably with some of the established works. Two problems are being addressed: one concerns pulse-width modulation and computes the optimal duty ratio, and the other computes an optimal switching schedule without a fixed cycle time.

This paper concerns the problem of optimal switching control in voltage converter circuits, where the objective is to minimize a cost-performance function comprised of the sum of a tracking-related measure and the switching energy. Most of the existing approaches to optimal switching are based on continuous-parameter optimization and optimal control techniques, which are mostly suitable to continuous-parameter functions such as tracking-related performance metrics. On the other hand, the switching-energy cost performance is inherently a discontinuous function dependent on the number of switchings, and hence its inclusion in the problem often is done in ad-hoc ways. This paper explores a systematic approach to optimizing performance - energy tradeoffs by extending an algorithm for optimizing tracking, developed by the authors, to include the energy performance via an averaging technique. The problem is posed in the setting of Pulse Width Modulation, and the controlled variables are the cycle time and duty ratios at each cycle. Extensive simulation results suggest the potential generality of the proposed approach. © 2012 IFAC.

We present a method to analyze a relationship between eye movements and saliency dynamics in videos for estimating attentive states of users while they watch the videos. The multi-mode saliency-dynamics model (MMSDM) is introduced to segment spatio-temporal patterns of the saliency dynamics into multiple sequences of primitive modes underlying the saliency patterns. The MMSDM enables us to describe the relationship by the local saliency dynamics around gaze points, which is modeled by a set of distances between gaze points and salient regions characterized by the extracted modes. Experimental results show the effectiveness of the proposed model to classify the attentive states of users by learning the statistical difference of the local saliency dynamics on gaze-paths at each level of attentiveness. © 2012 ACM.

The spatio-temporal correlation analysis between visual saliency and eye movements is presented for the estimation of the mental focus toward videos. We extract spatio-temporal dynamics patterns of saliency areas from the videos, which we refer to as saliency-dynamics patterns, and evaluate eye movements based on their correlation with the saliency-dynamics patterns in view. Experimental results using TV commercials demonstrate the effectiveness of the proposed method for the mental-focus estimation. © 2012 Information Processing Society of Japan.

We introduce the notion of manipulability to leader-follower networks as a tool to analyze how effective inputs injected at a leader node are in terms of their impact on the movements of the follower nodes, as a function of the interaction topologies and agent configurations. Classic manipulability is an index used in robotics for analyzing the singularity and efficiency of configurations of robot-arm manipulators. To define similar notions for leader-follower networks, we use a rigid-link approximation of the follower dynamics and under this assumption, we prove that the instantaneous follower velocities can be uniquely determined by that of the leader's, which allows us to define a meaningful manipulability index of the leader-follower networks.

People are being inundated under enormous volumes of information and they often dither about making the right choices from these. Interactive user support by information service system such as concierge services will effectively assist such people. However, human-machine interaction still lacks naturalness and thoughtfulness despite the widespread utilization of intelligent systems. The system needs to estimate user's interest to improve the interaction and support the choices. We propose a novel approach to estimating the interest, which is based on the relationship between the dynamics of user's eye movements, i.e., the endogenous control mode of saccades, and machine's proactive presentations of visual contents. Under a specially-designed presentation phase to make the user express the endogenous saccades, we analyzed the timing structures between the saccades and the presentation events. We defined resistance as a novel time-delay feature representing the duration a user's gaze remains fixed on the previously presented content regardless of the next event. In experimental results obtained from 10 subjects, we confirmed that resistance is a good indicator for estimating the interest of most subjects (75% success in 28 experiments on 7 subjects). This demonstrated a higher accuracy than conventional estimates of interest based on gaze duration or frequency.