Real-time identifying human activities by using Internet of things

Project title:

Supervisors:

Professor Guojun Lu (Principal)
Dr Jiangang Ma
Professor Manzur Murshed
Associate Professor Shyh Wei Teng

Contact person and email address:

Professor Guojun Lu: guojun.lu@federation.edu.au

Project description:

The aim of this project is to develop a novel sensor network-related system for automated human activity discovery and monitoring. With recent developments in cheap sensor and networking technologies, it has become possible to develop a wide range of practical applications, including the scenarios of communications, surveillance and remote controls such as using internet of things (IoT) monitoring elderly in hospital or homes, traffic monitoring, intruder mining on specific areas such as borders, etc.

Currently most of existing approaches typically exploit supervised learning algorithms. However, conventional technology is neither effective nor very practical in real-world situations.

In this project, you’ll develop innovative techniques that automatically discover, identify, and track an individual’s frequent routine activities in real-time without any manual annotation of activity data. In particular, the proposed project will use low-cost radio-frequency identification (RFID) sensor network technology and networked systems for human activity recognition. First, you will develop a software platform that includes RFID tag, readers and related software. The platform uses the signal from RFID readers for communication and obtaining its operating energy. Second, you will design unsupervised algorithms and models to automatically discover interesting user activity patterns from RFID and sensor data streams. For this purpose, you’ll combine machine learning approaches like pattern mining and clustering techniques with novel deep learning algorithms to discover activity patterns.