This article investigates new steps for the implementation and the utilization of presence indicators to identify displacement activities of a person and to model their life habits. This modeling is based on preliminary measures of displacement rate variations over a period of time. A possible time slot division of activities is highlighted by the analysis of these measures, slight variations of slot boundaries could appear from day to day, from person to person. On this basis, we show how to build three new indicators by working from time slot to time slot, in order to detect alerts or drifts from “normal” behavior as soon as possible and in a more reliable way. These indicators include start and end times of time slots, displacement rate and duration of each time slot. The algorithm we propose has been tested in real situations to show its use and relevance. Results are finally integrated in a more ambitious process of detection and automatic decision-making support through the conception of a web interface.

本文研究了用于实施和利用在场指示器的新步骤，以识别一个人的流离失所活动并为他们的生活习惯建模。该建模基于一段时间内位移率变化的初步度量。通过对这些措施的分析，强调了活动的可能时隙划分，时隙边界的细微变化每天可能在每个人之间出现。在此基础上，我们展示了如何通过在每个时隙之间工作来建立三个新指标，以便尽快以一种更可靠的方式检测警报或偏离“正常”行为。这些指标包括时隙的开始和结束时间，位移率和每个时隙的持续时间。我们提出的算法已在实际情况下进行了测试，以显示其用途和相关性。最终，通过Web界面的概念将结果集成到一个更具雄心的检测和自动决策支持过程中。