Predicting and Preventing Dangerous Events via Video Surveillance Using a Robotic Platform

Abstract

Our goal is to improve and extend our earlier, prototypical work on event-detection platforms, which was able to take immediate action by means of a robotic platform, when a potentially dangerous event was detected. Here we go further by enhancing and extending the system in order to not only detect instances of potentially dangerous events, but to try to predict them before they actually occur. This allows the robot to carry out specific preventive actions to stop them happening in the first place.