ABSTRACT
Accurate positioning is an essential task in mobile robot control to ensure safe operation of the robots. Pose feedback is widely used in the control strategy to achieve accurate tracking performance. The most direct approach to obtaining pose information is through positioning systems, e.g., a global navigation satellite system (GNSS) [Huang et al., 2018]. Since this approach relies on the availability of satellite signals, it is not applicable in scenarios involving indoor or underground operations. Alternatively, the robot’s pose can be estimated using different sensor systems and localization techniques [Brossard et al., 2018, Biswas and Veloso, 2010]. However, the accuracy of the estimated pose depends on the sensor models, the sensor setup, and the localization algorithms. Instead of using pose feedback, visual servoing [Chaumette and Hutchinson, 2006], especially image-based visual servoing (IBVS), offers another solution based on direct use of the images captured by a camera to calculate the control actions.
