ABC2022にて「A Method for Estimating the Number of Steps Taken Using a BLE Beacon Attached to the Soles of Footwear」というタイトルで発表をしてきました.
発表概要
The increase in life expectancy fueled by the development of medical technology and the aging of the world’s populations have created a labor shortage in the nursing care industry in Japan.
In addition to supporting the daily lives of seniors, elderly care facilities monitor and manage the amount of physical activity of their occupants to maintain their health.
To reduce the burden of such work, activity amount estimation has been proposed using acceleration sensors, smartphones, and cameras, although it has not been widely used due to installation costs.
Therefore, this study proposes a method that estimates the number of steps taken using BLE beacons as sensors to achieve a low-cost activity-amount estimation.
A BLE beacon is attached to the soles of footwear, and the number of steps taken by a person is estimated from the received radio wave strength that changes with walking movements.
In an evaluation experiment, participants walked 100 steps at an arbitrary speed, and then the actual number of steps was compared with the estimated number.
We found that the estimation accuracy was 91.6%.
報告と感想
こんにちは,M2の大鐘です.
10月27日(木)~10/29日(土)の日程でイギリスのロンドンにて開催された
ABC2022
において,研究発表を行ってきましたので報告させていただきます.
今回の発表は「A Method for Estimating the Number of Steps Taken Using a BLE Beacon Attached to the Soles of Footwear」というタイトルで,BLEビーコンを用いた歩数推定に関する研究を発表させていただきました.
海外での発表が行えたのは研究室としては約3年ぶりであり,僕自身も始めての海外だったので期待と不安が入り混じる中での発表でした.
どの参加者も英語のレベルが高く質疑応答も難なくこなしており,過去に参加した国際学会・ワークショップと比較しても一番レベルが高く感じました.
自分の発表では発表自体は問題なくできましたが,質疑応答では質問内容が理解できずかなりボロボロになりました.
英語力は大事ですね…
ですが,イギリスまで来られて特に大きなトラブルなく帰って来れたので,全体的にはとても楽しめることができました!
IWIN2022にて「Implementation of spatio-temporal fencing for crowd sensing in a smartphone applications」を発表してきました.
この発表で「IWIN2022 Best Presentation Awards」を受賞しました
発表概要
Crowdsensing platforms are available to solve the problems of dedicated system development and operating costs associated with crowdsensing.
Among them, there are several crowdsensing platforms that provide incentive elements to motivate and maintain users.
We build a crowdsensing platform (Lavlus) to mitigate disincentive factors that can be used in conjunction with incentive factors.
This study implements a smartphone application for Lavlus.
The Lavlus smartphone app needs to have the following functions: sensing project download, sensing request notification, automatic sensing, and sensor data upload.
Among them, sensing request notification and automatic sensing are based on Spatio-temporal fencing.
Geofencing used latitude and longitude, which are easy for users to visually recognize, but GPS is unstable indoors and in areas with many buildings.
Therefore, a margin is provided on the geofence to determine whether the user has entered or exited the geofence with certainty.
Lavlus may also target specific facilities, such as amusement parks or factories, as actual use cases.
In this case, the geofence may be arbitrarily polygonal.
We then draw a circle with the user at its center and generate geofencing points at eight points on the circumference of the circle.
If one or more of the generated points are inside the geofence, the system determines that it is about to enter space-time and issues a notification whether or not it will cooperate with sensing.
If all eight points are inside, the application judges that it has definitely entered the area and performs automatic sensing.
If all eight points are outside, the application judges that it has definitely exited and terminates sensing.
This allows proper geofencing even when the geofence is arbitrarily polygonal.
In this paper, we implemented these functions and verified their operation.
As a result, the geofence worked properly for arbitrary polygons.