中央研究院 資訊科學研究所

活動訊息

友善列印

列印可使用瀏覽器提供的(Ctrl+P)功能

TIGP (SNHCC) -- Crowd Footprints: Location-less Crowd Mobility Analytics using Wireless Traces of Mobile Devices

:::

TIGP (SNHCC) -- Crowd Footprints: Location-less Crowd Mobility Analytics using Wireless Traces of Mobile Devices

  • 講者巫芳璟 教授 (TU Dort­mund Uni­ver­sity)
    邀請人:TIGP (SNHCC)
  • 時間2021-09-27 (Mon.) 14:00 – 16:00
  • 地點僅提供視訊
線上串流

To join the talk on 【Webex】, or open Webex and enter Meeting number: 2515 893 0140 and Password: iis0927

摘要

Human mobility is an important key to many promising applications especially for the sectors of smart cities and smart environments. As advanced sensing, and computing, communication technologies are rapidly developed in the past decades, locations of crowds can be collected and further analyzed to boost these applications. Typical mobility analytics highly relies on location information such as outdoor GPS coordinates, indoor relative coordinates, or coordinates on images. However, outdoor absolute or indoor relative location information is not always available in many places, and taking images of people compromises personal privacy that is restricted in many countries. Therefore, to understand crowd mobility, social relationships among them, and their social behavior, this talk will review location-less crowd mobility analytics based on proximity sensing technology to estimate crowd sizes, infer mobility groups, monitor passenger flows, and detect human queues in crowds. The key idea is to extract important features behind wireless packets of mobile devices and correlate them to crowd mobility behavior. To verify the feasibility of the proposed location-less approaches, we conduct experiments in both indoor and outdoor places. Comprehensive experiments are conducted in indoor environments to verify the detection results of human queues and mobility groups. Moving toward more dynamic and uncertain outdoor places, we launched large-scale pilot studies in the Wellington Railway Station and Re:START Mall in New Zealand to showcase the crowd estimation results. In addition, we conducted experiments on the hanging railway system in TU Dortmund to showcase passenger flow mentoring.