Koninklijke Philips Electronics N.V. 2011

avoidance is an essential part of robotic navigation in which the robot, based on some global path planning strategy, has to reach a certain goal position. In robotic navigation the problem of obstacle avoidance refers to the determination of a local navigation strategy. Here, the navigation strategy is based on the detection of obstacles which is made depending on the local environment of the robot. This thesis presents the development and implementation of a novel obstacle avoidance system in which the detection of obstacles is based on the use of computer vision implemented using the principle of structured light based depth map creation. This allows the detection of an obstacle of at least 20mm x 18mm at the projection distance of 675mm. The obstacle avoidance system consists of a robotic platform called Create from iRobot with a vision system, made up of a camera-projector pair, all controlled by an embedded computing unit called BeagleBoard. The BeagleBoard, based on inputs from the visio