@article{oai:ipsj.ixsq.nii.ac.jp:00217743,
 author = {村井, 大地 and 浦野, 健太 and 望月, 祐洋 and 米澤, 拓郎 and 西田, 純二 and 河口, 信夫 and Daichi, Murai and Kenta, Urano and Masahiro, Mochizuki and Takuro, Yonezawa and Junji, Nishida and Nobuo, Kawaguchi},
 issue = {2},
 journal = {情報処理学会論文誌デジタルプラクティス（TDP）},
 month = {Apr},
 note = {Wi-Fiパケットセンサは，スマートフォンから送信されるプローブ要求を収集するセンサであり，人数・移動経路の両者を推定・分析できる．しかし，プローブ要求は電波信号であるため，周囲の環境により送信範囲が変化する．特に，都市型・屋外型施設では，プローブ要求が同時に複数のセンサで観測される可能性が高く，また，施設外からの信号も多く観測されるため，施設内での人流分析に大きな影響が出る．本稿では大規模環境を屋内と屋外，都市型と郊外型の2つの軸で分類した．Wi-Fiパケットセンサが不得意とする名古屋市にある都市型・屋外型施設である東山動植物園を対象に入園者推定を行った．また，推定した入園者のデータをもとに，各ゲートの人数推定と入園者の移動経路推定を行い評価した．結果として，滞在時間と移動数を用いることで，1年間の入園者数の傾向を捉えることができた．また，より正確な人数を推定するには，天気や曜日など，他のセンサデータと組み合わせる重要性が示唆された．また各ゲートでの人数推定では，人の流量や歩行速度によって精度が変わり，特に，人の流量が少なく，歩行速度が速い場所では，ユニークアドレスのみでは推定が困難であることが示唆された．移動経路推定では，1分ごとにRSSI値が最大のデータを用いることで，園内でのもっともらしい移動経路を作成することができた．, The Wi-Fi packet human flow sensor is a sensor that collects probe requests sent from smartphones, and can analyze number of people estimation and tracing. However, since the probe request is a radio signal, the transmission range changes depending on the surrounding environment. Especially in urban outdoor facilities, probe requests are likely to be observed by multiple sensors at the same time, and many signals from outside the facility are also observed, which greatly affects the analysis of human flow in the facility. In this paper, we classified large-scale environments into two axes: indoor and outdoor, and urban and suburban. In addition, we estimated the number of visitors to the Higashiyama Zoo and Botanical Garden in Nagoya, which are urban and outdoor facilities where Wi-Fi packet sensors have difficulty with analysis. As a result, we were able to capture the trend of the number of visitors during a year by using the time spent and the number of movements. In order to estimate the number of visitors more accurately, it is important to combine other sensor data such as weather and day of the week. The accuracy of estimating the number of people at each gate depended on the flow rate of people and their walking speed, suggesting that it was difficult to estimate the number of people at each gate using only unique addresses, especially when the flow rate was low and the walking speed was high. In the estimation of movement paths, we were able to create a plausible movement path in the park by using the data with the largest RSSI value per minute.},
 pages = {12--21},
 title = {大規模屋外施設におけるWi-Fiパケットセンサへの影響と利活用の検証},
 volume = {3},
 year = {2022}
}