Combined video and radar signal processing for fall detection

点击翻译按钮获取中文摘要

Improved safety and quality of life for elderly people and home-bound patients is the motivation for this research. Future home health care systems will be able to compile information about the patients they monitor and produce reliable alarms when an emergency situation, such as fall, occurs. A common problem with available home monitoring systems is frequent false alarms, which can result in desensitization to alarms and missed alarms. Current approaches for addressing this problem include using different types and placement of sensors, signal processing approaches to extract more useful data before making decisions, and the development of various classification methods to generate alarms. In this project, we will work with Remotronic to determine the accuracy of their fall-detection solution based on off-the-shelf cameras and to compare the accuracy against the radar based solution. A camera-based solution has advantage because cameras are widely available and inexpensive. Radar-based solutions can detect falls through walls and are capable of extracting breathing and heartbeat signals. The result of this research will be a proposal of a solution that will detect emergency events, such as falls, with higher accuracy, and differentiate these events from false alarms.