Development of Rough Terrain Mobile Robot by Self-Position Estimation Used Single Marker and Outer Product
In the present study, a mobile robot control strategy of rough terrain capable of position and direction estimation by a single marker was proposed, and the effect was performed. To realize wheel / caterpillar mobile robot movement on the rough terrain such as house farm, two major problems (1) the precise position estimation and (2) the direction estimation are important before the development of the control theory. Generally, the position / direction have been estimated by the wheel rotation (odometry) and gyro or acceleration sensor's output integration, but there are many problems such as many impacts from the ground and the wheel slipping on the rough ground to the precise estimation. The aim of this study is to develop a method (algorithm) to estimate the mobile robot direction using a single marker placed at the top of the robot by changing the front movement and the target direction movement periodically. The reason using a single marker is that it is difficult to process to find and analyze positions of the many markers attached on the robot in the camera image from a long distance, and the positions of many markers have been unstable by the tilting of the robot, agricultural product's shadow, getting dirty of the body and so on. Our proposed method would be effective in the difficult situation to find or analyze the marker position of the robot surfaces by the reason of camera position and the environment, and it could realize wheels / caterpillar type mobile robot movement on the muddy situation. Keywords - Odometry,Caterpillar Mobile Robot,Outer Product, Marker.