PhD Electrical and Electronic Engineering

我正在研究如何使用无线电信号，特别是WiFi，来检测生命体征，提供室内位置信息，并对活动进行分类。与其他传感器（如蓝牙可穿戴设备）的数据一起，我将利用这些信息来研究行为模式，以及这些行为模式如何与人的可追踪变化相关。为了自动化实验数据的注释，将使用激光雷达、雷达和商业室内定位系统（如Pozyx）中的一种或组合，其中的数量将足够大，使得任何手动工作都过于繁琐和耗时。我目前正在研究和应用这些不同传感器的数据提取和融合技术。在一个更大规模的实验中，没有地面真实标签，传感器将被信任到不同的准确度，我将研究如何最好地转移在监督情况下完成的学习，并将推理问题转化为强化学习“语言”。为了实现这一点，我将学习生成模型以及记忆。

I am studying how to use radio signals, specifically focusing on WiFi, to detect vital signs, provide location information indoors, and to classify activity. Together with data for other sensors, such as bluetooth wearable devices, I will be using this information to study behavioural patterns and how these can be related to traceable alterations in the person. To automate annotation of experimental data, of which there will be an amount large enough to make any manual effort too tedious and time consuming, either of, or a combination of, lidar, radar, and commercial indoor positioning systems (such as Pozyx), will be used. I am currently studying and applying techniques for data extraction and fusion from these various sensors. In a bigger-scale experiment, where there will be no ground truth labels, and the sensors will be trusted to varying degrees of accuracy, I will look into how to best transfer the learning done in the supervised case and into casting the inference problem into reinforcement learning "language". To achieve this, I will learn about generative models as well as memory.