유비쿼터스센서 기반 다중 조명 로봇 시스템

Abstract

In this thesis, automatic lighting systems are proposed using a 3-DOF robotic lamp that uses the structure of a spherical-type 3-DOF parallel mechanism. The spherical type 3-DOF robotic lamp is compact and suitable for installing at the ceiling. Out of 3-DOFs of the robotic lamp, 2-DOFs are tilting motion to direct a person, and the rest 1-DOF is a zoom-in and – out motion to control the intensity of a localized light. If the 3-DOF robotic lamp performs tilting motion to illuminate a person, more than two motors are operated because of guide linkages and a bevel gear set in the parallel linkage. Therefore, the robotic lamp can create a large force to support a heavy lamp. There are two versions of the 3-DOF robotic lamp in this thesis. Both of the two versions have three subsystems: the base frame, the parallel linkage, and the lamp module. The first version of the robotic lamp has two screws in the lamp module that controls zoom-in and – out motion, so the robotic lamp can handle the inner case that has a heavy lens. On the other hand, the second version of the robotic lamp has one screw in the lamp module, so the lamp module can be designed thinner so that the workspace is bigger. Unlike the first version, the second version of the robotic lamp has a parallel linkage in the upper hemisphere and has a weight at the end of the guide axis. Because of this special structure, the lamp is operated close to the base frame, and the robotic lamp can control a heavy lamp with less power due to the counter balancing. There are three automatic lighting systems. In the automatic lighting system using Zigbee and PIR sensors, a person is detected by nine PIR sensors, and the position data is transferred by Zigbee that makes wireless sensor network. A 3-DOF robotic lamp lightens the person in the detecting area, and the zoom-in and – out motion performs according to the position of the person. In the automatic lighting system using two laser scanners, two laser scanners installed in the test environment to avoid a blind spot measure the positions of people. And two robotic lamps track and illuminate each person continuously in the scanning using the positions of persons and the coordinate transformation. In the automatic lighting system using the multiple robotic lamps, one laser scanner detects people in the scanning area, and to solve the blind spot problem due to using only one laser scanner, the Kalman filter algorithm is applied to this lighting system. Therefore, even though there are obstacles in the scanning area, multiple robotic lamps track and illuminate people continuously using the Kalman filter and the data association algorithm.