Humanophilic systems Lab. researches on cyber-physical systems (CPS) that snuggle up to human life by combining sensors (IoT), machine learning, networks, and applications. Primarily, we are focusing on human activity recognition using sensors (IoT), machine learning (AI). However, we cover various topics widely. Sometimes, we develop novel sensors (Hardware) and applications. In recent years, we start focusing on research related to human behavior change by information technologies as the next research beyond human activity recognition.
Sensing & Network
Activity Recognition by tiny-wearable sensor SenStick
SenStick is the platform including the tiny BLE-enable sensing board (eight typical MEMS sensors are embedded), peripheral software, and 3D case data. It was developed to focus on data analysis, which is the essence of IoT research, and ready to use upon purchase without having to do electronic work. Anyone can easily measure data by using an iOS / Android application or Node.js library. The unusual shape, small and stick-shaped, is suitable for making our belongings such as glasses, chopsticks, and canes a sensor. A rechargeable LiPo battery is used instead of coin batteries (Competetive product usually adopts a coin battery). The BLE chip is equipped with Nordic’s nRF52, which allows advanced data processing with Cortex-M4F. Unlike a general BLE sensor, a large-capacity flash memory (32 Mbytes) is embedded on the board, so continuous logging is possible without a smartphone. SenStick has released circuit diagrams, firmware, and peripheral software, all as university research results, and it is possible to rewrite the firmware and develop original sensor boards.
Masashi Takata, Yugo Nakamura, Yohei Torigoe, Manato Fujimoto, Yutaka Arakawa, Keiichi Yasumoto, “Strikes-Thrusts Activity Recognition Using Wrist Sensor Towards Pervasive Kendo Support System,” WristSense’19 – Workshop on Sensing Systems and Applications using Wrist Worn Smart Devices, March 2019.
Yugo Nakamura, Yutaka Arakawa, and Keiichi Yasumoto, “SenStick: A Rapid Prototyping Platform for Sensorizing Things,” The Ninth International Conference on Mobile Computing and Ubiquitous Networking (ICMU 2016), 4-6 Oct. 2016.
Context recognition by an eye tracker (Pupil/Tobii)
There is research to sense eye movements in order to recognize human context. Through gaze and pupil diameter, it is possible to unravel information about what you are interested in and how much you understand. Measuring devices are called eye trackers, and they are commercially available from ones attached to personal computers to wearable ones. Using these devices, we are developing things such as measuring interest during sightseeing and a dynamic textbook creation support system linked to the line of sight.
Yuki Matsuda, Dmitrii Fedotov, Yuta Takahashi, Yutaka Arakawa, Keiichi Yasumoto, Wolfgang Minker, “EmoTour: Estimating Emotion and Satisfaction of Users Based on Behavioural Cues and Audio-Visual Data” MIDPI Sensors 2018, 18(11), 3978. (Impact factor: 2.475)
https://doi.org/10.3390/s18113978
Dmitrii Fedotov, Yuki Matsuda, Yuta Takahashi, Yutaka Arakawa, Keiichi Yasumoto, Wolfgang Minker, “Towards Real-Time Contextual Touristic Emotion and Satisfaction Estimation with Wearable Devices,” IEEE International Conference on Pervasive Computing and Communications (PerCom 2019), Demo, March 2019.
Shoya Ishimaru, Ko Watanabe, Nicolas Großmann, Carina Heisel, Pascal Klein, Yutaka Arakawa, Jochen Kuhn and Andreas Dengel, “Demonstration of HyperMind Builder: Pervasive User Interface to Create Intelligent Interactive Documents,” Ubicomp/ISWC 2018, Oct. 2018.
Urban sensing by a participatory mobile sensing platform ParmoSense
There is still a lot of city information that cannot be searched using services like Google Maps, such as park playground equipment information and collection and delivery times of each postbox. As a method for efficiently sensing such urban data, there is a concept of participatory mobile sensing that utilizes users’ smartphones. Of course, just handing out the app doesn’t help with sensing, so it is necessary to incorporate rewards and gamification to motivate participants. Therefore, we have developed ParmoSense as a participatory mobile sensing platform with gamification functions.
Yutaka Arakawa and Yuki Matsuda, “[Invited Paper] Gamification mechanism for enhancing a participatory urban sensing: survey and practical results,” Journal of Information Processing, Vol.24, No.1, pp.31–38, Jan. 2016.
Yoshitaka Ueyama, Morihiko Tamai, Yutaka Arakawa, and Keiichi Yasumoto, “Gamification-Based Incentive Mechanism for Participatory Sensing”, IEEE CROWDSENSING 2014 (First International Workshop on Crowdsensing Methods, Techniques, and Applications) in conjunction with IEEE PerCom 2014, pp.98–103, Mar. 24, 2014.
Activity recognition by WiFi signal analytic
Activity recognition by Acoustic sound analysis
Activity recognition by a smartphone and smartwatch
Urban sensing by 5G and LoRaWAN
Urban Sensing by social data analysis
Energy-harvest Human Activity recognition
Communication skill sensing by analyzing a micro gesture and facial expression
HumanoPhillic system in a work place
Estimating Human Psychological State (Stress, QoL, Work Engagement)
We focus on estimating human psychological states, such as stress and QoL (Quality of Life). Under the work style reform, Workers stress checks have become a social issue, and the goal is to make it easier to know your condition with smartphones and wearable devices. So far, paper questionnaires such as WHOQOL-BREF have been used for QoL surveys, and when analyzing the correlations between the results and sensor data, it has become clear that it can be recognized to some extent by machine learning. In the future, we will respond to experiments at actual companies and through questionnaires (work engagement, etc.)
Chishu Amenomori, Teruhiro Mizumoto, Hirohiko Suwa, Yutaka Arakawa, Keiichi Yasumoto, “A Method for Simplified HRQOL measurement by Smart Devices,” 7th EAI International Conference on Wireless Mobile Communication and Healthcare (MobiHealth 2017), November 2017.
It is known that office workers’ oversitting can cause back pain and stiff shoulders. Therefore, it is important to adjust the chair to suit your body shape, and various high-performance chairs are on the market. However, in fact, there is no quantitative measurement of what kind of posture when working sat on the chair or whether you are sat in a good posture. Therefore, we have developed a chair that can always sense posture.
Yasuhiro Otoda, Teruhiro Mizumoto, Yutaka Arakawa, Chihiro Nakajima, Mitsuhiro Kohana, Motohiro Uenishi, and Keiichi Yasumoto, “Census: Continuous Posture Sensing Chair for Office Workers,” 2018 IEEE International Conference on Consumer Electronics (ICCE), January 2018.
By collating information about where and how many times you are in an office building, you can objectively grasp the number of times you go to a smoking area and the number of meetings, and reflect and improve efficiency. There are various methods for estimating indoor locations, but in any case, power sources may be a problem. Therefore, we propose a place estimation method that does not require charging by using energy harvesting elements such as solar cells both as a power source and sensor. This paper was accepted as a full paper at the top conference IEEE PerCom2019.
Yoshinori Umetsu, Yugo Nakamura, Yutaka Arakawa, Manato Fujimoto, Hirohiko Suwa, “EHAAS: Energy Harvesters As A Sensor for Place Recognition on Wearables,” IEEE International Conference on Pervasive Computing and Communications (PerCom 2019), Full paper, March 2019.
Yutaka Arakawa, Yoshinori Umetsu, Yugo Nakamura, Manato Fujimoto, Hirohiko Suwa, Keiichi Yasumoto, “Poster: Feasibility Study toward a Battery-free Place Recognition System based on Solar Cells,” Ubicomp/ISWC 2018, Oct. 2018.
Like Apple Watch’s activity reminder, it is becoming a reality that information devices work on people and change their future behavior. These information devices that can change their behavior are sometimes referred to as Digital Medicine for lifestyle-related disease prevention and are expected to spread to society in the future. We are developing digital signage that actively talks to people passing by, and is conducting empirical research on whether these can stimulate behavioral changes in daily dialogue.
Zhihua Zhang, Yuta Takahashi, Manato Fujimoto, Yutaka Arakawa, Keiichi Yasumoto, “Investigating Effects of Interactive-signage-based Stimulation for Promoting Behavior Change,” Computational Intelligence Journal (Conditionally accepted)
Zhihua Zhang, Yutaka Arakawa, Harri Oinas-kukkonen, “Design of Behavior Change Environment with Interactive Signage Having Active Talk Function,” PerPersuasion’19 – 1st International Workshop on Pervasive Persuasive System for Behavior Change, March 2019.
Zhihua Zhang, Yuta Takahashi, Manato Fujimoto, Keiichi Yasumoto, and Yutaka Arakawa, “Investigating User Reactions to Interactive-signage-based Stimulation Toward Behavior Change,” 11th International Conference on Mobile Computing and Ubiquitous Network (ICMU), Oct. 2018.
Walking support system by unconscious speed control based on predicted future heart rate
The number of people walking is increasing due to health-consciousness. However, if you do not walk at an appropriate pace, there is a risk of increasing the burden on the heart. Therefore, we developed a walking support system that predicts the future. heartbeat before walking from the past exercise history and the gradient information of the route, and presents the optimal route based on that prediction. We have developed a system that can change the pace of pedestrians unconsciously by exquisitely changing the speed of the music.
Atsushi Otsubo, Hirohiko Suwa, Yutaka Arakawa, Keiichi Yasumoto, ”BeatSync: Walking Pace Control through Beat Synchronization between Music and Walking,” IEEE International Conference on Pervasive Computing and Communications (PerCom 2019), Demo, March 2019.
Shogo Maenaka, Hirohiko Suwa, Yutaka Arakawa, and Keiichi Yasumoto, “Heart Rate Prediction for Easy Walking Route Planning,” SICE Journal of Control, Measurement, and System Integration, Vol.11, No.4, pp.284–291, July 2018.
Shogo Maenaka, Hirohiko Suwa, Yutaka Arakawa, and Keiichi Yasumoto, “Recommending Optimal Route for Walking Support based Heart Rate Prediction,” The Second International Workshop on Smart Sensing Systems (IWSSS’17) , August 2017.
