Further development of a robotic museum guide

Abstract

This report presents the design and implementation of a software system for robotic museum guide. This system has been implemented on MAVEN and it includes 5 behavioural modes. They are “greeting”, “wait”, “scan for user”, “facing centroid of audience” and “facing a random audience” modes. In “greeting” mode, the robot docks at the charging station and waits for visitors to arrive. During this mode, MAVEN will give its greeting speech once it detects a visitor entering through the door. The user will be able to select which station they would like to visit using a menu provided on a touch screen device. In the “wait” mode, MAVEN would navigate to the various stations with obstacle avoidance capability. When MAVEN reaches the specified station, it activates the “Scan for user”, “face centroid of audience” and “face a random audience” modes. These modes make use of a Kinect sensor to track and face the audience to simulate eye contact with the audience while giving the speech. In the “scan the user” mode, the robot turns and searches for a member of audience. Once MAVEN detects someone, it will begin the speech for the station where it is situated. In the “face the centroid of audience” mode, the robot turns to face and track the centroid of the audience. Finally in the “face a random audience” mode, the robot randomly selects a member of the audience and track him or her. MAVEN performed very well during the tests. It is able to select from a database of speeches and present it at each station. MAVEN can also navigate to each station with obstacle avoidance reliably.