Motion control of an omnidirectional climbing robot based on dead reckoning method

作者：Tavakoli Mahmoud^*; Lopes Pedro; Sgrigna Lucio; Viegas Carlos

来源：Mechatronics, 2015, 30: 94-106.

DOI：10.1016/j.mechatronics.2015.06.003

• Omnidirectional climbing robot
• Optical flow sensor
• Motion control
• Localization

While omnidirectional wheels enable a holonomic drive and a good maneuverability, the slippage of the wheels as an inherent characteristic of the omnidirectional wheels prevents using rotary shaft encoders as a reliable source of data for the robot's odometry. When installed on a climbing robot, omnidirectional wheels may suffer from additional slippage on the surface. In a previous study, we described how the resulting vibration decreases the trajectory following accuracy of the robot, and why rotary encoders, as the most popular dead reckoning method cannot be used. In this paper, we address this problem by integration of low cost and light weight exteroceptive sensors, i.e. an accelerometer and an optical flow sensor. The Omniclimber climbing robot was used as the testing platform in this study. Omniclimbers are omnidirectional climbing robots that can climb and navigate over flat and curved structures. We attempt to compensate the errors due to the wheel slippage through closing the position control loop without significantly increasing the robot's weight, cost and complexity of the robot. We also integrated an algorithm which corrects the robot kinematics on the curved structures based on the curvature diameter and the robot's heading angle. Taking advantage of these sensors and algorithms we could make remarkable improvements on the path following accuracy of the Omniclimbers, which is presented in this article.

• 出版日期2015-9