Sensor-based Human Activity Recognition

Sensor-based systems and technologies for human activity recognition have recently gained a momentous interest from research and industry communities. Accelerated by the rapid advancement in Internet of Things, wearable and wireless sensor technology, there has been a continuous growth of research innovation and real-world applications including remote patient monitoring, elderly monitoring, daily-life activity monitoring, accident monitoring (fall detection), vital signs monitoring, sleep monitoring, fitness/wellbeing tracking, and so on.

Key research areas related to sensor-based activity recognition include machine/deep learning, pattern recognition, digital signal processing, context-aware computing, data management and processing, information fusion and remote sensing.

Research areas

• Machine learning methods including deep learning, transfer learning, reinforcement learning, and hybrid semi-supervised learning
• Online and active learning for sensor stream with noisy data
• Behavior profiling and anomaly detection
• Device-free indoor localisation and tracking
• Non-contact sensing methods based on IoT sensors, passive infrared sensors, wifi, radar or visible light
• Non-contact vital signs monitoring using UWB or mmWave radar
• Human activity recognition for smart/remote healthcare
• Computational intelligence methods for privacy-aware and secured data management, processing, and analysis
• Case studies and technical evaluation of existing human activity recognition methods and systems
• Any topics related to human activity recognition and its applications (further reading)

Current projects

• Deep learning methods for IoT sensor-based activity recognition in a multi-resident environment
• Non-intrusive, device-free indoor localisation and tracking of multiple people
• Radar-based human detection and activity recognition based on mmWave sensors
• Radar-based vital signs monitoring based on mmWave sensors
• Intelligent data management and computational offloading in IoT, Fog, and Edge computing

Research networks

• Professor Michael Sheng (Macquarie University, Australia)
• Professor Lulu Wang (Shenzhen Technology University, China)
• Professor Kevin Sowerby (University of Auckland)
• Dr Saeed Rehman (Flinders University, Australia)
• Dr Kevin Ellyett (University of Auckland, Auckland District Health Board)
• Dr Andrew Austin (University of Auckland)
• Dr Veronica Joachim (University of Otago)
• Associate Professor Boon-Chong Seet (AUT)
• Professor Edmund Lai (AUT)
• Professor Andrew Lowe (AUT)
• Professor Peter Chong (AUT)
• SmartLife NZ
• Rush Digital
• Northland Innovation