Voice Activity Projection (VAP) has been actively studied to enable natural turn-taking in spoken dialogue systems, relying primarily on acoustic features. Visual cues such as head movements are also known to contribute to turn-taking prediction; however, camera-based approaches are affected by placement and lighting conditions and are not always reliably available to dialogue systems. As a camera-independent approach for directly capturing head motion, earable devices offer a promising solution. In this study, we propose Sensor-Augmented VAP, a framework that integrates in-ear inertial measurement unit (IMU) signals with a pre-trained VAP model via a lightweight residual fusion module. To validate our proposed method, we collected a dataset pairing conversational audio with in-ear IMU data, comprising 12 dyadic Japanese dialogues recorded using microphones and earbuds. Experiments in speaker-independent and speaker-dependent settings demonstrate that IMU fusion consistently improves weighted F1 score for shift detection and reduces VAP loss over the audio-only baseline. These results confirm that head-motion cues are effective for enhancing turn-taking prediction.

This study proposes a new deep learning method for reconstructing depth images of moving objects within a specific area using Wi-Fi channel state information (CSI). The Wi-Fibased depth imaging technique has novel applications in domains such as security and elder care. However, reconstructing depth images from CSI is challenging because it is difficult to learn the mapping function between CSI and depth images, both of high-dimensionalities. To address the challenge, we propose Wi-Depth. The main idea behind the design of Wi-Depth is that a depth image of a moving object can be decomposed into three core components: the shape, depth, and position of the target. Therefore, in the depth-image reconstruction task, Wi-Depth simultaneously estimates the three core pieces of information as auxiliary tasks in our proposed VAE-based teacher-student architecture, enabling it to output images with the consistency of a correct shape, depth, and position. In addition, the design of Wi-Depth is based on our idea that this decomposition efficiently takes advantage of the fact that shape, depth, and position relate to primitive information inferred from CSI such as angle-of-arrival, time-of-flight, and Doppler frequency shift.

Location-based games (LBGs) such as Pokémon Go integrate gameplay with real-world locations, creating novel player experiences. Motivated by the idea that new gameplay forms can emerge by rethinking the game field, we have proposed WWW-based games (WBGs) as a new approach to implementing location-based gaming within the WWW. WBGs utilize URLs and web content as game elements, allowing players to interact with web environments in new ways. To explore its feasibility and impact, we developed Text Monster, a game in which players capture monsters inhabiting web pages and compete for control over websites. User studies with two versions of the game demonstrated that WBG is technically feasible and engaging, as players voluntarily participated and exhibited strategic behaviors. Our findings suggest that WBG can influence web usage patterns and introduce new possibilities for interactive web-based experiences.

Accurate measurement of propeller rotation speed is essential for stable drone control. While various sensors have been employed for this purpose, their cost, size, and additional weight pose challenges, particularly for small indoor drones. This paper presents a method for estimating propeller rotational speed using Wi-Fi Channel State Information (CSI), which captures fine-grained wireless signal fluctuations and is obtainable on many commercial Wi-Fi devices. The proposed approach applies resampling and short-time Fourier transform (STFT) to the amplitude component of CSI to estimate propeller rotational speed. In this study, experiments were conducted using both fixed propellers and a hovering drone in indoor environments. The results revealed that antenna placement and polarization alignment significantly affect the accuracy of rotation speed estimation. With a tilted antenna configuration, the estimation error rate remained below 3.16% in all STFT windows (each 0.5 -second estimation interval). Furthermore, even with a horizontally placed antenna, selecting dominant subcarriers through signal processing reduced the estimation error rate to below 4% in over 95% of STFT windows. In addition, by evaluating three distances ($2 \mathrm{m}, 4 \mathrm{m}$, and 8 m) and varying the antenna elevation from 0° to 90°, we identified optimal orientations that reduced RMSE to as low as 23.60 RPM, achieving a mean error of only 1.24% at 8 m. These findings demonstrate the feasibility of noncontact propeller rotational speed estimation using Wi-Fi CSI and suggest its potential for real-time control applications.

The disinfection of environmental surfaces is an effective countermeasure for COVID-19. In this paper, we use a thermographic camera and lightweight background image processing and propose a method that detects and visualizes the places touched by a person. Our method will support effective disinfection activities.

Sensor networks allow us to collect data, such as camera images, over a wide area. Understanding the sensing area by aggregating and processing the data from multiple sensors is promising. Deep Learning (DL) is a powerful method for interpreting data. However, communication cost on the sensor network and computational cost on the server are two substantial problems of aggregating and processing data from multiple sensors. We therefore propose divided processing of the DL between a server and a powerful Edge-Heavy Sensor (EHS). In our study, we reduced transmission data to twelve times lower than the amount of raw input data while maintaining a 4.5% decrease in the DL's recognition accuracy.

The datafication of spatio-temporal city-wide events is one essential factor for smart management of the city. For this purpose, the combination of real-time event detection on edge sensor nodes mounted public vehicles, and event accumulation on a server is one realistic and efficient solution. We can analyze the accumulated data to understand complex phenomena occurring in entire the city. In this paper, we introduce a novel datafication procedure of city-wide events by sensor mounted garbage trucks and evaluated the preliminary implementation of event detection system on actual vehicle-mounted sensors.

This paper proposes a new method that makes it easy for us to construct a positioning model for device-free passive indoor localization by using model transfer techniques. With device-free passive indoor positioning, a wireless sensor network is used to detect the movement of a person based on the fact that RF signals transmitted between a transmitter and a receiver are affected by human movement. However, because device-free passive indoor positioning relies on machine learning techniques, we must collect labeled training data at many training points in an end user's environment. This paper proposes a method that transfers a signal strength model used for locating a person obtained in another environment (source environment) to the end user environment. With the transferred models, we can construct a positioning model for the end user environment inexpensively. Our evaluation showed that our method achieved almost the same positioning performance as a supervised method that requires labeled training data obtained in an end user's environment.

This article proposes a new concept for a context-aware Web search method that automatically retrieves a webpage related to the daily activity that a user currently is engaged in and displays the page on nearby Internet-connected home appliances such as televisions. For example, when a user is washing a coffeemaker, a webpage is retrieved that includes tips such as “cleaning a coffee maker with vinegar removes stains well,” and the page is displayed on a nearby appliance. In this article, we design and implement a Web search method that employs ubiquitous sensors to monitor a user's daily life. Our proposed method automatically searches for a webpage related to a daily activity by using a query constructed from the use of daily objects employed in the activity that is detected with object-attached sensors. We evaluate the search method with real datasets collected from vast numbers of sensors and achieve very accurate webpage retrieval. We then investigate the usefulness and effectiveness of a daily life Web search with Wizard-of-Oz (WOz)-like experiments. We confirm that the presentation of webpages related to daily activities improves participants' future daily lives and triggers communication among the participants in the experiment.

The new concept proposed in this paper is a query free web search that automatically retrieves a web page including information related to the daily activity that we are currently engaged in for automatically displaying the page on Internet-connected domestic appliances around us such as televisions. When we are washing a coffee maker, for example, a web page is retrieved that includes tips such as `cleaning a coffee maker with vinegar removes stains well.' A method designed on the basis of this concept automatically searches for a web page by using a query constructed from the use of ordinary household objects that is detected by sensors attached to the objects. An in-situ experiment tests a variety of IR techniques and the experiment confirmed that our daily activities can produce related web pages with high accuracy.

Accessing information when we are on the move is a key feature if mobile computing environments, and using an optical see-through head mounted display (HMD) is one of the most suitable ways to do this. Although the HMD can display information without interfering with the user’s view, when the sight behind the display is too complex or too bright, the information displayed can bee very difficult to see. To solve this problem, we have created a way of laying out information for the optical see-through HMD. The ideal area for displaying information is determined by evaluating the sight image behind the HMD captured by a pantoscopic camera mounted on it. Moreover, if there is no suitable area for displaying information, our method select involves using the sight image around users use to the ideal direction and instructing them to face the direction. Our method displays information to ideal areas.

Motorbike racing is one of the most popular motorsports and many fans visit circuits to watch races. However, since audiences and pit crews can only obtain limited information, it is difficult for them to get detailed information about teams except for high ranking teams and few popular teams. To solve this problem, we have already presented an information browsing system for pit crews who used wearable computers. This paper describes an extension to the system. We also applied our system to racing audiences. Since our system can provide various information such as detailed ranking tables and favorite teams, it can make supporting the teams more advantageous and entertain audiences much more.

This paper proposes new concept of query free web search for daily living. We ordinarily benefit from additional information about our daily activities that we are currently engaged in. When washing a coffee maker, for example, we receive the benefit if we obtain such information as 'cleaning a coffee maker with vinegar removes its stain well.' Our proposed method automatically searches for a web page including such information relates to an activity of daily living when the activity is performed. We assume that wireless sensor nodes are attached to daily objects to detect object use; our method makes a query from the names of objects which are used. Then, the method retrieves a web page relates to the activity of daily living by using the query.

In the near future, small ubiquitous devices will be embedded into most of artifacts to support human dailylife. In such environments, we need multiple types of human-computer interfaces for enjoying ubiquitous services. This paper describes several ubiquitous gadgets; ubiquitous chip, Pin&Play, and VCC, to provide flexible ubiquitous services and user interfaces. The integrative use of these devices achieves the physical and intuitive interaction between users and ubiquitous services.

Because of advances in information technologies, car navigation systems have come into widespread use as useful tools to guide drivers where they want to go. Conventional car navigation systems present the most suitable route according to a destination input into the system. However, since the required operation to input the destination costs so much, users do not usually use car navigation systems for daily driving. In this paper, to exploit the effective functions of car navigation systems, we propose a new system that automatically predicts user purpose and destination. The proposed car navigation system presents various information based on predicted purpose without interaction from users.