Yaonan Wang

Intelligent hybrid control strategy for trajectory tracking of robot manipulators

We address the problem of robust tracking control using a PD-plus-feedforward controller and an intelligent adaptive robust compensator for a rigid robotic manipulator with uncertain dynamics and external disturbances. A key feature of this scheme is that soft computer methods are used to learn the upper bound of system uncertainties and adjust the width of the boundary layer base. In this way, the prior knowledge of the upper bound of the system uncertainties does need not to be required. Moreover, chattering can be effectively eliminated, and asymptotic error convergence can be guaranteed. Numerical simulations and experiments of two-DOF rigid robots are presented to show effectiveness of the proposed scheme.

On-line detection of foreign substances in glass bottles filled with transfusion solution through computer vision

In this paper, a real time computer vision system was developed to detect a glass bottle filled with transfusion solution on a packing line. The theory, the mechanical structure, and the detection algorithm are described. Foreign substances mixed in transfusion solution are distinguished from bottle side wall defects by a relative movement between a resting bottle and the inside moving solution. Foreign objects moving down in the bottle are first detected with background subtracting, where a normalized cross correlation is employed to identify two frames between which no moving objects overlap and the static back-ground is obtained through an inter-frame difference. In the second stage, falling objects are recognized through an adapted mean-shift tracker, in which an adaptive searching window is used to only find falling objects. The system was tested on 250-ml transfusion bottles. The result shows that it can effectively detect the small visible foreign substances fulfilling the requirements of the application..

A new delay-dependent H∞ control criterion for a class of discrete-time systems with time-varying state delays and input delays

This paper deals with delay-dependent H∞ control for discrete-time systems with time-varying state delays and input delays. A new finite-sum inequality is first established to derive a delay-dependent condition, under which the resulting closed-loop system via a state feedback is asymptotically stable with a prescribed H∞ noise attenuation level. Moreover, a modified iterative algorithm according to the previously published algorithm of Ghaoui and Oustry (IEEE Trans. Autom. Control, 1997, 42(8), 1171–1176) involving convex optimization is proposed to obtain a suboptimal H∞ controller. Two numerical examples are presented which show the effectiveness of the proposed method.

Kernel based incremental learning Isomap algorithm

Isomap is one of widely-used low-dimensional embedding methods. However, in many scenarios, the data come sequentially and the effect of the data is accumulated. Isomap algorithms have no the ability of new data be added for all data need to be available when estimates the geodesic distances. In this paper we propose an incremental learning isomap algorithm, which take the approximate geodesic distance matrix as a kernel matrix then projection is converted to solve a kernel eigenvalues problem. We have no need to reconstructing the neighborhood graph for those incremental samples. Experiments results on Swiss datasets and Helix datasets show that our algorithm is more effective than isomap method.

The 3D Path Planning Based on A* Algorithm and Artificial Potential Field for the Rotary-Wing Flying Robot

An efficient fusion algorithm for the rotary-wing flying robot is presented for solving path planning problem in the 3D mountain environment. This fusion algorithm combines A*algorithm with artificial potential field method both improved and optimized for 3D environment. Firstly, A*algorithm is adopted to plan the preliminary path in the environmental model established by grid method. This A* algorithm is optimized in the heuristic function according to the 3D environment. Then, we select and save the key nodes in the preliminary path. Finally artificial potential field algorithm is used to do the path smoothing processing on the basis of these key nodes. We carried out the simulation in MATLAB of the proposed algorithm, single A*algorithm and artificial potential field method respectively in the 3D map consisting of random irregular surface and defined peak. The simulation results indicates that the fusion algorithm is feasible in 3D path planning and superior in smoothness performance, which can satisfy the property of real-time tracking for rotary-wing flying robots.

Sensing incline terrain for mobile robot autonomous navigation under unknown environment

In order to correctly sense incline terrain, its geometrical calculated model is analyzed. Based on the change trend of distance between the mobile robot and slope, their relative position can be determined. Then a novel method which takes the use of the powerful nonlinearity approach capability of RBF network is introduced to estimate the slope of the terrain with respect to the robots current angular tilt. The experimental results show that the proposed approach can be used for mobile robot incline terrain perception and has the merits of easy, trustiness and robustness.

A Real-Time Detection Framework for On-Tree Mango Based on SSD Network.

On-tree fruit detection in orchards is important for yield estimation, mapping and automatic harvesting in modern agriculture. This paper proposes a real-time detection framework for on-tree mango based on SSD (Single shot Multi Box Detector) network, a state-of-the-art object detection algorithms based on deep learning. The mango image dataset used in this paper was gathered from outdoor mango orchards. Firstly, the dataset was annotated and converted to a trainable dataset for SSD network. Secondly, the author designed new sampling strategies and image distortions at the image pre-processing stage to optimize data augmentation techniques. Moreover, the default box proposal methods of SSD network were improved by redesigning the shapes of default boxes on multiple feature maps according to our own dataset. Finally, to explore which classification network is most suitable for mango detection, an experiment was presented to compare the detection performance of SSD network with the VGG16 and ZFNet as base network respectively. Almond dataset was also used to verify our proposed method. Experimental results demonstrated that, with optimization of data augmentation techniques and default box proposals, our improved VGG16-based SSD network can achieve higher performance than Faster R-CNN in on-tree mango detection, with F1 score of 0.911 at 35 FPS for 400 × 400 input image, which is a real-time detection.

Design of 3D Visualization System Based on VTK Utilizing Marching Cubes and Ray Casting Algorithm

Medical image visualization, also known as medical image 3D reconstruction technique, is a theory, method and technique which use computer graphics, image processing, computer vision and human-computer interaction techniques to reconstruct 3D stereoscopic image from the received 2D medical images (such as MRI, CT, etc.),then display it on-screen and interact with it. Currently, medical image 3D visualization technology is a hot in the medical image processing, and it is of great significance in theory research and clinical applications. This paper mainly introduces the key techniques in the medical image 3D visualization and designs a system of the medical image 3D visualization based on VTK. The system integrates Marching Cubes and Ray Casting visualization algorithm, implements the function of DICOM format of medical images file read,3D reconstruction and display. The system can be applied to medical diagnosis, medical education and medical research, etc.

The Quality Detection of Surface Defect in Dispensing Dack-End Based on HALCON

This paper proposes a quality detection method of LED surface defects in dispensing back-end based on computer vision library HALCON. The key steps mainly include image preprocessing, image matching and defect detection, etc. Firstly, to choose the matching area automatically according to the regional characteristics of the LED chip. Secondly, image matching with standard template. Finally, complete defect detection using the improved matching algorithm. The method can select matching area automatically from the collection of image with CMOS. It can complete image matching quickly and accurately, and detect the defect effectively.

The intelligent multimodal pressure control of diamond synthesis process for cubic hinge press

In order to improve the quality and yield of synthetic diamond using Y-500 cubic hinge press, a novel intelligent control in synthesis process was studied deeply in this paper. After both classical and new technical automatic synthesis processes were analyzed, the characteristics and current situation of control of pressure plant were discussed. Then an intelligent multimodal pressure control method is proposed. The proposed method divides the precise control process of rising and holding pressure into three phases according to the characteristics of dynamic information, and the compound control is used. Further more, the switch control, the fuzzy control and the proportional plus intermittent intelligent integral control are applied to different phase respectively, and the rules of online expert segmentation identification and switching are established, thus, the control mode can change with the segment and perform smoothly. Finally the proposed method is applied to control system of Y-500 cubic hinge press, and excellent pressure tracking performance is achieved, thus its feasibility is proved too.