My research theme is to clarify what can be realized with sensors.
Sensors are a device or equipment that detects changes in the physical state of a phenomenon or object, such as the magnitude of force, distance or luminance, and converts them into signals and data as outputs. It may sound “difficult” to many readers.
Conventionally, some sensors were built primarily into factory machinery for machine control, labor saving and quality control. Others were embedded in equipment and products to enable automation as well as to enhance usability and safety. Both types of sensors represent key technologies that contributed greatly to boosting the competitiveness of the industrial products of post-war Japan. However, their functions were largely limited and we seldom found sensors close or familiar to us.
In recent years, however, sensors rapidly became familiar to us thanks to the spread of smartphones. With a global annual shipment of nearly 600 million units, this small device, in fact, houses various types of sensors.
For example, when a smartphone is tilted horizontally, the screen display will rotate horizontally according to the movement of the body. This is enabled by a gyroscope sensor, which measures angular velocity (the rotation speed of an object, or changes in the angle of an object per unit of time).
A proximity sensor, unlike a mechanical switch that passes or halts an electric current by physical contact, is capable of switching the electric current status simply by an object coming close to the sensor. This technology is used to prevent smartphones from malfunctioning in call mode: The touch panel display will automatically turn off when user’s ear comes close to the device.
In addition, many sensors, including an ambient light sensor, acceleration sensor, gravity sensor, barometric pressure sensor, temperature sensor, are packed into such a small device. We are benefiting from functions offered by various sensors through smartphone “apps.”
As indicated above, today, the great number of various kinds of sensors, to an extent unimaginable some years ago, are built in smartphones and carried around by people.
In line with the rapid penetration of smartphones, research and development efforts are underway at an accelerated pace, aiming at developing high-performance sensors with high functionality. Sensor application technologies have a great potential.
The widespread use of smartphones and wearable devices creates “a situation in which many people carry around high-performance sensors at all times,” concurrently with “a situation such high-performance sensors are always connected with a network.” By connecting sensors fitted to equipment, social infrastructure (buildings, roads, railways, etc.) and people (smartphones, wearable devices, etc.) with the network, this sensor network can collect and analyze vast amounts of data, and serve as a platform for providing solutions to a wide range of issues in society.
For example, a weather information service company utilizes reports and location data from members’ smartphones, as well as images of the sky photographed with their smartphone cameras, to provide weather forecasts. The company collects additional data that cannot be acquired with its own observation network, by relying on the overwhelmingly large number of smartphones.
Technical considerations are being made on a system that utilizes sensors installed on bridges and tunnels for continuous monitoring for cracks and any other abnormalities. Such a system will issue an alert to prompt on-site inspection when such abnormalities are detected, and in a time of disaster, it can be used for the unmanned, real-time determination of damage to the infrastructure.
Moreover, the sensor network should be instrumental to addressing the issues of increasing medical costs and aging society. The Japanese government is promoting home-based medical care in an effort to reduce medical benefit expenditures. In this light, wireless sensing is regarded as assisting the management of people’s health conditions. Basically, the use of wearable sensors is contemplated to collect temperature, heart rate and other biometric data. A Japanese manufacturer has succeeded in the commercialization of a wrist band type device housing a heart rate sensor and an acceleration sensor, for the purpose of health management. By using the heart rate and acceleration sensors, a wearer’s daily activities including sleep, walking and indoor walking as well as physical condition can be accumulated as time-series data. If unusual patterns are detected by comparing the current data with past data, the system judges them as a deterioration of the wearer’s physical condition and an immediate response can be made, such as reporting it to a medical institution.
As exemplified above, sensor networks are expected to be utilized for the resolution of various social problems and the creation of new businesses.
Against this backdrop, our research group is currently working on “sensor technology that facilitates person-to-person communication.”
For example, we are exploring the application of the technology to meetings. Meetings are an important means of decision making and consensus building in many different groups, including workplaces, schools, families and regional communities. However, I believe that many people find it difficult to induce all participants to proactively share their views and to promptly coordinate different opinions.
To conduct active meetings, it is necessary to employ objective indices that assess the activity levels of a meeting. With these indices, it is possible to give advice on how to encourage active meetings through comparing and analyzing activity levels of different meetings. Based on this assumption, our lab has defined indices of activity levels of meetings and studied a system capable of determining if a meeting is successful or not.
In this research, we use sensors to measure voice data during a meeting and analyze which of the participants was talking. Then we quantify the activity level of the meeting by using the following three indices.
(1) Equitability in discussion: An active meeting requires that members participate equitably in a discussion. Focusing on the duration of speech by each member, this index is to measure if all members equitably make statements during the meeting.
(2) Dominance of discussion: To facilitate an active discussion, the presence of a dominant person or discussion moderator (leader) is essential. Focusing on the frequency of remarks by each member, the person who makes remarks most frequently is regarded as the leader of the discussion.
(3) Mediation in discussion: It is necessary for the leader to give everyone an equal opportunity to make a statement in order to solicit members’ opinions. Here, by analyzing who makes a statement right after the remarks of the leader, we measure whether or not the leader gives everyone an equal opportunity to voice his/her opinion.
KAIHUI is the system we use to analyze the data above and determine the indices of an activity level of a meeting. Using KAIHUI, our lab measured the activity levels of six meetings, which differ in (1) gender of participants; (2) meeting themes; and (3) how the discussion was proceeded.
This research has just begun. Going forward, we will need to devise indices focused on quality aspects of meetings by incorporating elements such as participants’ statement contents, facial expressions and gestures. As we proceed with this study, we may be able to see a “facilitator robot” in the future. Such robots would urge participants to make a statement, coordinate the course of a discussion, confirm that the participants have a shared understanding on a matter and support their consensus building and mutual understanding, while occasionally making witty jokes.
Sensor technology has evolved by making it possible to measure what was previously unmeasurable. Through future research, even “happiness and comfort levels” may become measurable. The evolution of sensor technology offers immense possibilities.
(This column is as of 2016.)
