Zaojun Fang

A Vision-Based Self-Tuning Fuzzy Controller for Fillet Weld Seam Tracking

In this paper, an intelligent seam-tracking system is developed for fillet weld workpiece. Structured light vision is employed to detect the deviation of the welding torch. Two stripe lines are formed on the surface of the fillet weld workpiece by the structured laser light. Their intercepts are selected as the image feature, which can result in an approximately uncoupled relationship between the image feature and the movement of adjustment mechanism. A two-step feature-extraction method is designed to robustly detect the two lines in image. In order to get a favorable tracking precision in the welding process, a new self-tuning fuzzy controller is designed. Its input and output scaling factors are tuned according to its active working region. Moreover, a supervisory level is introduced into the control system to limit the output pulses in order to guarantee steady and safe tracking. Finally, experiments are well conducted to verify the performance of the proposed system and methods.

Visual Measurement of the Racket Trajectory in Spinning Ball Striking for Table Tennis Player

A high-speed monocular vision system is established to track the racket trajectory for estimating the spinning ball initiated by a human. A new fast and robust method is proposed to extract the feature sequence of a rackets pose during the process of striking. Considering the complexity of the working environment and fast movement of the racket, a novel corner extraction algorithm with good robustness and high efficiency is presented, which mainly consists of two parts, i.e., target segmentation and line detection. Then PnP-based algorithm is adopted to get the rough pose of the racket, which is then optimized by orthogonal iteration algorithm to guarantee the orthogonality of the rackets orientation matrix. In order to classify the spinning mode of the ball with a higher accuracy, the rackets movement is represented by the sequence of normal vector of the rackets surface. Experiments are thoroughly conducted to verify the effectiveness of the proposed method.

Trajectory Prediction of Spinning Ball Based on Fuzzy Filtering and Local Modeling for Robotic Ping–Pong Player

In this paper, a novel prediction method of the striking position is proposed for a robotic ping-pong player. In order to remove the noise involved in the coordinates of ping-pong ball, a new nonlinear filter based on fuzzy logic approach is presented. Then, least square method (LSM) is utilized to compute the initial flying and rotational velocities based on the filtered positions of the ball. The impact between a ping-pong ball and the table is studied, and the analytic model that represents the relationship between the velocities before and after rebound is developed. Based on statistical analysis, a memory-based local modeling approach is presented to obtain a more accurate velocity after rebound. The succeeding trajectory is predicted according to the initial state of the ball and the flying and rebound models. The striking position can be obtained from the predicted trajectory. Experiments are well conducted and verify that sufficient precision of the striking position have been achieved with the proposed method.

Vision-Based Alignment Control for Grating Tiling in Petawatt-Class Laser System

Large-sized diffraction gratings are required for the compressor of chirped-pulse-amplification systems. However, producing single large grating is of great difficulty, which makes grating tiling an alterative solution. This paper presents a vision-based alignment control method for grating tiling in petawatt-class laser system. Tiling errors with six degrees of freedom are grouped into three ones, and a decoupled vision control method is proposed. The piston error is measured from the far-field focal-spot image. To determine the piston error, the energy ratio between the main and secondary spots in the image is computed, and then a mapping formulation is determined to compute the piston from the energy ratio. To guarantee precision and reliability, a closed-loop piston control system with clustering error filtering and output verification is designed. The angular error is composed of two parts, i.e., tip and tilt errors, which is measured from the near-field interference fringe image. To reduce the influence of airflow on the angular measurement, a novel image quality estimation method is designed using the average crest-to-trough difference. Then, Fourier fringe pattern analysis is applied to the left and right parts of the image to reconstruct the wavefront, based on which a closed-loop grating angular control system is designed. Finally, experiments are well conducted in a tritiled gratings to verify the performance and precision of the proposed grating tiling system.

Vision Based Starting Position Recognition and Positioning Control for Thin Steel Sheet Welding Robot

For the thin steel plate butt welding of containers, a visual control method of welding robot is proposed to recognize starting position and drive the weld torch to accurately track the starting position. Firstly, an iterative circle Hough transformation is presented, which can automatically obtain the region of interest including the torch characteristic point. In the second iteration, the torch characteristic point is recognized and extracted, robustly and quickly. Then the seam is recognized and its characteristic point is extracted by using line Hough transformation. Furthermore, an image space based visual control scheme of starting position is designed, and the accurate starting position control is achieved. The experimental results show that the proposed scheme can extract the torchs and seams characteristic point coordinates, accurately and robustly. Moreover, the torch can be driven to the starting position quickly and accurately by using the visual control. Finally, test results in the container factory also verify the effectiveness of these methods, and the satisfactory results are obtained.

Vision tracking system for narrow butt seams with CO2 gas shielded arc welding

An image-based visual system is proposed for narrow butt welding seam tracking in container manufacturing. A smart camera is used as the sensor to measure the welding seam. The feature extraction algorithm is designed based on the smart camera, which is used to compute the error between the welding torch and the welding seam. The reference feature is determined with many frames of images based on their statistical property. The current estimated feature is checked based on the reference feature. The abnormal features caused by strong noises such as welding arc light, splashes and fog are rejected. The normal features are used for welding seam tracking. A compact vision system for tracking seams is presented, which employs a programmable controller to control a stepper motor to eliminate the error detected by the smart camera. Experiments are well conducted to verify the effectiveness of the proposed system and method.

Image-based visual seam tracking system for fillet joint

In this paper a compact vision-based seam tracking system is designed for fillet joint. The main consideration in developing this system is its applicability and reliability. Thus programmable logical controller (PLC) is used as the controller and smart camera as the vision sensor. To get reliable image features, laser structured light is used. Besides its hardware structure, the image processing and control algorithm are designed elaboratively. The intercepts of the two lines in the stripe image are selected as the image features to get an approximately uncoupled relationship between the change of image features and the movement of two axes in two directions. New edge detector with fine noise reduction and robust line extraction method are proposed. In addition, the controller is carefully constructed with input filtering and output verification. Finally, a series of experiments were well conducted to verify the performance of the proposed system.

Modeling the stroke process in table tennis robot using neural network

To hit incoming balls back to a desired position, it is a key factor for table tennis robot to get racket parameters accurately. For modeling the stroke process, a novel model is built based on multiple neural networks. The input data for neural networks are the ball velocity differences during the stroke, and racket parameters are the output data. To reduce the influences from the invalid data, a neural network based on each empirical data is established. The training data are clustered based on the empirical data. The way of choosing a neural network to compute the racket parameters depends on the comparison between the new coming data and the empirical data. Moreover, a novel way based on a binocular vision system to verify the stroke model is proposed. Experimental results have showed that the stroke model created via the proposed method is applicable and the verification method is effective.

A simple technique for structured light calibration in welding robots

Structured light calibration is one of the most crucial parts in a visual sensing system of welding robots. And the main aspect of the structured light calibration is the accuracy. Besides, a practical and simplistic calibration approach is preferable. In this paper, we propose a simple technique for structured light calibration. The vertical and horizontal lines are added to the planar object to facilitate points detection. The information from the camera calibration is fully used to detect the points on the light stripe. Then parameters of the structured light plane are estimated based on three non-collinear points definition. It could be said that the computation in this study is straight forward. Furthermore, it does not require any additional equipment. Thus, it is a truly simple and practical calibration method. According to the experimental results, the calibration errors are less than 0.2 mm. It shows that our proposed method is acceptable to measurement system of welding robots.

A novel measurement and control method for automatic plastering machine

Plastering machine is a type of construction machine which is greatly promoted in the building industry. Work efficiency is improved by using plastering machine and the labor intensity is reduced. To improve the quality of plastering using the plastering machine, a new measurement and control method is proposed in this paper. This method utilizes the laser sensor to measure the distances between the plaster board and the wall, calculates the errors of the bracing pieces perpendicularity and the deflection angle of plaster board, and compensates the error by trajectory planning. In addition, the principle of laser triangulation is analyzed and the types of measurement, including direct and oblique measurements, are also illustrated. The oblique measurement based on laser triangulation is adopted in the system due to its higher resolution and good feasibility. According to the space geometry relation, the errors are calculated accurately. Furthermore, a scheme of error compensation is addressed based on a two nested loops structure. Each of these loops implements a Proportional Integral (PI) controller. With the compensation scheme, the trajectory of plaster board can be determined and tracked to verify the plaster results. The theoretical analysis is validated by an extensive set of numerical simulations which shows the performance of the proposed control system.