Autonomous human tracking of multiple robotic lamps

This paper proposes an automatic lighting system using multiple 3-DOF robotic lamps. The robotic lamp is a spherical type parallel mechanism with a unique forward kinematic solution. In order to minimize the dynamic load, three motors are installed at the base frame and a counter balancing design is performed. Using the coordinate transformation between the environment coordinate and the robot coordinate, the position data of the human is converted to three output angles of the robotic lamp. Positions of people are measured by a laser scanner, and in order to track the positions of people accurately by multiple robotic lamps, the Kalman filter algorithm and a data association algorithm were applied. In experiment, three robotic lamps were employed to track three people independently. For each lamp, two tilting angles are controlled for a localized light and the zoom mechanism is employed to control the intensity of the light.