An Autonomous Meeting Assistant
This project was the final assignment of the final year Intelligent Robotics module. In a team of four, we worked to design a system which would, given a map of the space it was working in, find an empty meeting room from one of the designated rooms, find a meeting attendee and then guide them to the meeting room. We used depth data from a Kinect combined with a model of face sizes at various distances to identify faces, and a probabilistic road map to navigate the task space. We had developed MCL localisation nodes in previous assignments, but based on experimental results we chose instead to use the standard ROS AMCL implementation. The system was implemented using ROS Java, and used the OpenNI framework to interface with the Kinect. We received a small prize for the “Best Robot” on the course.