No abstract

No abstract

Positioning accuracy in robotics is a key issue for the manufacturing process. One of the possible ways to achieve high accuracy is the implementation of machine learning (ML), which allows robots to learn from their own practical experience and find the best way to perform the prescribed operation. Usually, accuracy improvement methods cover the generation of a positioning error map for the whole robot workspace, providing corresponding correction models. However, most practical cases require extremely high positioning accuracy only at a few essential points on the trajectory. This paper provides a methodology for the online deep Q-learning-based approach intended to increase positioning accuracy at key points by analyzing experimentally predetermined robot properties and their impact on overall accuracy. Using the KUKA-YouBot robot as a test system, we perform accuracy measurement experiments in the following three axes: (i) after a long operational break, (ii) using different loads, and (iii) at different speeds. To use this data for ML, the relationships between the robot’s operating time from switching on, load, and positioning accuracy are defined. In addition, the gripper vibrations are evaluated when the robot arm moves at various speeds in vertical and horizontal planes. It is found that the robot’s degrees of freedom (DOFs) clearances are significantly influenced by operational heat, which affects its static and dynamic accuracy. Implementation of the proposed ML-based compensation method resulted in a positioning error decrease at the trajectory key points by more than 30%.

No abstract

Increasing the imaging rate of atomic force microscopy (AFM) without impairing of the imaging quality is a challenging task, since the increase in the scanning speed leads to a number of artifacts related to the limited mechanical bandwidth of the AFM components. One of these artifacts is the loss of contact between the probe tip and the sample. We propose to apply an additional nonlinear force on the upper surface of a cantilever, which will help to keep the tip and surface in contact. In practice, this force can be produced by the precisely regulated airflow. Such an improvement affects the AFM system dynamics, which were evaluated using a mathematical model that is presented in this paper. The model defines the relationships between the additional nonlinear force, the pressure of the applied air stream, and the initial air gap between the upper surface of the cantilever and the end of the air duct. It was found that the nonlinear force created by the stream of compressed air (aerodynamic force) prevents the contact loss caused by the high scanning speed or the higher surface roughness, thus maintaining stable contact between the probe and the surface. This improvement allows us to effectively increase the scanning speed by at least 10 times using a soft (spring constant of 0.2 N/m) cantilever by applying the air pressure of 40 Pa. If a stiff cantilever (spring constant of 40 N/m) is used, the potential of vertical deviation improvement is twice is large. This method is suitable for use with different types of AFM sensors and it can be implemented practically without essential changes in AFM sensor design.