Two interesting talks by Prof Karl Woodbridge


In October, we will have two interesting talks by Prof Karl Woodbridge of University College London (see more information here), organised by the SA Chapter of IEEE AESS. Prof Woodbridge has previously presented talks to the Radar Remote Sensing Group at UCT, and we are very pleased that he is visiting us again this year.

The first talk on 11 October 2016 will be held in the Anatomy Building on Medical Campus, while the second talk on 12 October will be held in the Menzies Building on Upper Campus.


Talk 1: Utilising In-home WiFi for Activity Recognition in an Ambient Assisted Living Environment

Where: Seminar Room, Level 7, Anatomy Building, Medical Campus

When: 11 October, 11 am – 12 pm


Ambient Assistant Living (AAL) is a concept whereby a range of e-healthcare sensors are employed to monitor elderly and disabled people for alert situations such as falls or prolonged periods of inactivity. We have developed a novel technology using in-home WiFi signals for non-obtrusive activity recognition using the micro-Doppler signature from human subjects for detection and classification. In this presentation our software defined radio based passive WiFi system hardware and software design will be outlined and the deployment and operation of the system within a dedicated domestic e-healthcare housing unit described. Real time monitoring results will be presented and the benefits and challenges of rolling out this technology widely for AAL will be discussed.


Talk 2: Development of a Real Time Passive WiFi Radar for Security and Healthcare Applications

Where: Seminar room 7.04, Menzies Building, UCT

When: 12 October, 3-4 pm


This presentation describes the hardware and signal processing development of a passive radar system designed to detect human activity. The system leverages the increasing availability of ambient WiFi transmissions to detect and classify a wide range of motions using Doppler returns from the subject. The software defined radio based demo system and real time batch processing methods developed for human body motion sensing will be described and challenges discussed. The demo system has been tested in a wide range of applications and examples of real time Doppler results will be presented in the security and healthcare areas. Preliminary micro-Doppler signature classification and tracking filter design and test results will also be described and the future development of the system towards a versatile low cost sensor for the above markets discussed.