Associated to the weld quality, the weld bead geometry is one of the most important parameters in welding processes. It is a significant requirement in a welding project, especially in automatic welding systems where a specific width, height, or penetration of weld bead is needed. This paper presents a novel technique for real-time measuring of the width and height of weld beads in gas metal arc welding (GMAW) using a single high-speed camera and a long-pass optical filter in a passive vision system. The measuring method is based on digital image processing techniques and the image calibration process is based on projective transformations. The measurement process takes less than 3 milliseconds per image, which allows a transfer rate of more than 300 frames per second. The proposed methodology can be used in any metal transfer mode of a gas metal arc welding process and does not have occlusion problems. The responses of the measurement system, presented here, are in a good agreement with off-line data collected by a common laser-based 3D scanner. Each measurement is compare using a statistical Welch’s t-test of the null hypothesis, which, in any case, does not exceed the threshold of significance level α = 0.01, validating the results and the performance of the proposed vision system.

One of the greatest challenges in today's industrial robotics is the development of off-line programming systems that allow drastic reduction in robots' reprogramming time, improving productivity. The article purpose is to pave the way to the construction of generic calibration systems easily adapted to any type of robot, regardless their application, such as modular robots and robot controllers specifically designed for non-standard applications. A computer system was built for developing and implementing a calibration system that involves the joint work of computer and measurement systems. Each step of this system's development is presented together with its theoretical basis. With the development of a remote maneuvering system based on ABB S3 controller experimental tests have been carried out using an IRB2000 robot and a measurement arm (ITG ROMER) with 0.087 mm of position measurement accuracy. The robot model used by its controller was identified and the robot was calibrated and evaluated in different workspaces resulting in an average accuracy improvement from 1.5 mm to 0.3 mm

This paper presents and discusses a method to calibrate a specially built laser triangulation sensor to scan and map the surface of hydraulic turbine blades and to assign 3D coordinates to a dedicated robot to repair, by welding in layers, the damage on blades eroded by cavitation pitting and/or cracks produced by cyclic loading. Due to the large nonlinearities present in a camera and laser diodes, large range distances become difficult to measure with high precision. Aiming to improve the precision and accuracy of the range measurement sensor based on laser triangulation, a calibration model is proposed that involves the parameters of the camera, lens, laser positions, and sensor position on the robot arm related to the robot base to find the best accuracy in the distance range of the application. The developed sensor is composed of a CMOS camera and two laser diodes that project light lines onto the blade surface and needs image processing to find the 3D coordinates. The distances vary from 250 to 650 mm and the accuracy obtained within the distance range is below 1 mm. The calibration process needs a previous camera calibration and special calibration boards to calculate the correct distance between the laser diodes and the camera. The sensor position fixed on the robot arm is found by moving the robot to selected positions. The experimental procedures show the success of the calibration scheme.

This paper presents and discusses the results of an ongoing R&D project aiming to design and build a fully automated prototype of a specialized spherical robotic welding system for repairing hydraulic turbine surfaces eroded by cavitation pitting and/or cracks produced by cyclic loading. The system has an embedded vision sensor built to acquire range images by laser scanning over the blade's surface and produce 3D models to locate the damaged spots to be registered in a 3D coordinate system into the robot controller, enabling the robot to repair the flaws automatically by welding in layers. The paper is focused on the robot kinematic model and describes an iterative algorithm to process the inverse kinematics with only five degrees-of-freedom. The algorithm makes use of data collected from a vision sensor to ensure that the welding gun axis is perpendicular to the blade's surface. Besides this, it proposes a modelling and optimization mathematical routine for more efficient robot calibration, which can be used with any type of robot. This robot calibration optimization scheme finds the optimal error parameter vector based on the condition number of the manipulator transformation composed from the partial derivatives of the error parameters. Experimental results proved both the iterative algorithm to perform the inverse kinematics and the technique to optimize robot calibration to be very efficient.

This research introduces a new method to estimate the position of a robot’s Tool Center Point (TCP) using Inertial Measurement Units (IMUs), sensor fusion and Artificial Neural Networks (ANNs). The objective is to make an accurate estimate of TCP navigation, using the signals from an IMU as resources of a neural network capable of predicting the position. Considering that the IMU sensors suffer noise in the measurements and the noise progresses over time, this proposal employs a technique that eliminates the filtering step, and the process is done internally by the network. The work employs a non-parametric approach to reset the reference dynamically, minimize noise from sensors, and converge positioning to a nominal result. This method offers a solution for fast, cheap, and efficient robot calibration. The work does not want to replace current techniques but to introduce a new design to the literature. The concept does not require sophisticated mechanical parts and the production line to be idle during the calibration process, and the results show that the developed technique can accurately predict the TCP position with millimeter errors and in real-time. The study also implemented the concept with other neural networks, for which it used a smaller set of data in an attempt to reduce training time. The research used the Multilayer Perceptron and XGBRegressor networks to test the approach introduced with others algorithms. Different applications that need real-time positioning can benefit from the proposal.