Shenyang Institute of Automation,Chinese Academy Of Sciences
Not a member yet
23582 research outputs found
Sort by
Numerical simulation of the stability of water fiber-optic in water jet-guided laser machining
No full textWater jet-guided laser machining is a new compound machining technology, which has been widely used in many fields due to its better processing effect. In this technology, the coupling of laser beam and micro-water jet directly determines the machining effect, and the prerequisite for successful coupling is the steady flow of the water jet, so ensuring the stability of the micro-water jet is the key to the stable machining of water jet-guided laser. Therefore, it is of great significance to studying the stability of the water fiber-optic in water jet-guided laser processing. In this paper, aiming at the problem that the stability of the water fiber-optic is difficult to control, a finite element model of the water fiber-optic is established. The convection model is vortex gas-phase flow "enveloped" water fiber-optic which is used to explain the interaction mechanism, and the flow field distribution of gas-phase flow and water fiber-optic convection was obtained. The results show that water fiber-optic is refined under the constraint of gas-phase flow, and the maximum processing distance can increase by three times. At the same time, the gas-phase flow can accelerate the removal of processing debris, and the processing accuracy and efficiency are improved
Spatial curing growth mechanism and defect control of alumina green bodies manufactured by stereo-lithography
No full textUnderstanding the spatial curing growth mechanism and defect formation of green bodies is of great significance for the manufacture of precision and high-performance ceramic products. In this work, the effects of different powder volume fractions and average particle sizes on the spatial curing growth mechanism and defect regulation of alumina green bodies in stereo-lithography were discussed by combining novel mathematical theories and experiments. We found the spatial curing growth characteristics and defect shapes of green bodies in the mathematical model, and obtained the distribution patterns of the beam zone, scattering zone, inadequate curing area and overlap area, as well as their relationship with defect evolution. Furthermore, we verified these characteristics of green bodies through experiments and found that these characteristics could be improved by the most optimal experimental parameters based on mathematical theories. In particular, the print layer thickness could selectively modify the curing shape and improve the printing condition
Research on battery array pose measurement technology based on multi-line structured light
No full text获取机械手与电池阵列平面相对位姿是机械手对电池阵列平面自动操作的前提。针对多片状拼接特点的非合作电池阵列测量目标，提出了一种基于多线结构光的电池阵列平面位姿测量方法。该测量法采用多线结构光获取电池阵列上激光光条的图像，在图像中提取光条中心直线并进行短直线信息融合，依据不同光条之间的几何约束关系识别出直线组，进而基于直线组上的特征点求解相对位姿。在图像处理过程中，为快速获取高精度光条中心直线，提出了一种基于主成分分析与RANSAC相结合的光条中心直线提取方法。经实验验证，所提光条中心直线提取方法，精度与Steger算法相近，速度约为Steger算法的3倍；所提短直线信息融合方法能够改善位姿求解精度；所提测量方法求解的相对位置误差小于4.6mm，相对姿态角误差小于2°，满足实际工程应用需求。</p
Dsa-PAML: a parallel automated machine learning system via dual-stacked autoencoder
No full textFinding a high-performance machine learning pipeline (ML pipeline) for a supervised learning task takes much time. It requires many choices, including preprocessing datasets, selecting algorithms, tuning hyperparameters, and ensembling candidate models. With increasing pipelines arises a combination explosion problem. This work presents a new automated machine learning (AutoML) system called Dsa-PAML to address this challenge by recommending, training, and ensembling suitable models for supervised learning tasks. Dsa-PAML is a parallel automated system based on a dual-stacked autoencoder (Dsa). Firstly, meta-features of datasets and ML pipelines are used to alleviate cold-start recommendation problems. Secondly, a novel dual-stacked autoencoder is used to simultaneously learn the latent features of datasets and ML pipelines, efficiently learning collaborations of both datasets and ML pipelines and recommending suitable ML pipelines for a new dataset. Thirdly, Dsa-PAML can train the recommended ML pipelines on the new dataset in a parallel method, which substantially reduces the time complexity of the proposed method. Finally, a parallel selective ensemble system is embedded into Dsa-PAML. It selects base models from candidate ML pipelines according to their runtime, classification performance, and diversity on the validation set, enhancing Dsa-PAML’s stability for most datasets. Amounts of experiments on 30 UCI datasets show that our approach outperforms current state-of-the-art methods.</p
PRECISE SEGMENTATION and MEASUREMENT of INCLINED FISH’S FEATURES BASED on U-NET and FISH MORPHOLOGICAL CHARACTERISTICS
No full textAccurate measurement of fish’s features is of great significance for breeding management and decision-making. Fish body area, body length, and body width are important features for judging the growth status of fish in smart aquaculture. These features can be used as an important reference for bait feeding, fishing, and classification. In view of the fact that fish body is usually inclined on actual production line, this research proposes a scheme based on U-Net and fish morphological characteristics to segment and precisely measure the features of inclined fish. Firstly, the data set is processed and expanded through data enhancement such as contrast transformation and rotation transformation. This operation can simulate the real shooting environment and enhance robustness of the training model. Secondly, U-Net is introduced. Using the expanded training set to generate a segmentation model. Trained model is used to segment the test samples to generate accurate segmented images and output fish body area. Finally, by combining fish morphological characteristics, the inclined angle of the fish body is determined. After rotation correction, circumscribed rectangle method is adopted to obtain the body length and width of fish in the image. The experimental results show that using the proposed scheme, the mIoU of test set is as high as 0.974, the relative error of average fish body area is only 1.25%, the relative error of average fish body length is only 0.65%, and the relative error of average fish body width is only 0.84%. Compared with traditional circumscribed rectangle method, the relative error of body length is reduced by 5.25%, and the relative error of body width is reduced by 39.87%.</p
Variable-gain control for continuum robots based on velocity sensitivity
No full textKinematic control for continuum robots usually involves an inverse model to provide actuator positions according to the desired end-tip position, as well as a servo controller at the actuator level. The resulting control performance of a continuum robot is then related to its kinematic characteristics that vary at different configurations. In this paper, a kinematic model for a typical rod-driven continuum robot is presented. Following this, a kinematic parameter, velocity sensitivity, is proposed to evaluate the kinematic characteristics of the continuum robot, indicating the contribution of the individual actuators to the instant movement of the end-tip when tracking a given path. Next, a variable gain control strategy is presented to tune the servo controller with respect to the varying velocity sensitivity along the path, reducing the fluctuation of the tracking errors in real time. The simulated and experimental results show that the presented methods can effectively smooth the movement of the continuum robot over its workspace by considering the coordination between the kinematic and servo controllers
Research on the integrated manipulator of point cloud measurement and precise cutting for waste nuclear tank
No full textPurpose Nuclear waste tanks need to be cut into pieces before they can be safely disposed of, but the cutting process produces a large amount of aerosols with radiation, which is very harmful to the health of the operator. The purpose of this paper is to establish an intelligent strategy for an integrated robot designed for measurement and cutting, which can accurately identify and cut unknown nuclear waste tanks and realize autonomous precise processing. Design/methodology/approach A robot system integrating point cloud measurement and plasma cutting is designed in this paper. First, accurate calibration methods for the robot, tool and hand-eye system are established. Second, for eliminating the extremely scattered point cloud caused by metal surface refraction, an omnidirectional octree data structure with 26 vectors is proposed to extract the point cloud model more accurately. Then, a minimum bounding box is calculated for limiting the local area to be cut, the local three-dimensional shape of the nuclear tank is fitted within the bounding box, in which the cutting trajectories and normal vectors are planned accurately. Findings The cutting precision is verified by changing the tool into a dial indicator in the simulation and the experiment process. The octree data structure with omnidirectional pointing vectors can effectively improve the filtering accuracy of the scattered point cloud. The point cloud filter algorithm combined with the structure calibration methods for the integrated measurement and processing system can ensure the final machining accuracy of the robot. Originality/value Aiming at the problems of large measurement noise interference, complex transformations between coordinate systems and difficult accuracy guarantee, this paper proposes structure calibration, point cloud filtering and point cloud-based planning algorithm, which can greatly improve the reliability and accuracy of the system. Simulation and experiment verify the final cutting accuracy of the whole system.</p
Neural Adaptive Command Filtered Control for Cooperative Path Following of Multiple Underactuated Autonomous Underwater Vehicles Along One Path
No full textThe aim of this article is to develop a new cooperative path following control scheme for a team of autonomous underwater vehicles (AUVs) which track one curve with nonlinear uncertainties. Individual path-following controllers are designed to guarantee that each AUV meets the desired tracking performance. For the cooperative path following, a containment control approach is incorporated into the design, where each AUV is forced to evenly disperse on a path that is parameterized by a continuous variable over a communication network. The main features of the paper that unlike the existing designs are that: first, by employing the command filtered control technique, the assumption of second-order derivative of the reference path is removed. Second, a globally uniformly ultimately bounded (GUUB) path following control structure is proposed that enables two kinds of controllers to switch under different cases. Third, coordination between multiple AUVs is achieved through a containment design, and a path variable containment cooperative path following controller is derived that enables multiple AUVs to be evenly dispersed and guided by the virtual leaders. Finally, the stability analysis and simulation example are given to verify the proposed control strategy.</p
蓬勃发展的空间机器人技术与应用
No full text随着人类太空探索的深入,机器人已经在空间各领域得到了广泛应用,尤其是为载人航天、月球与深空探测、小天体采样返回等航天任务提供了重要的支撑。空间机器人可代替或协助人类在空间环境中进行各种作业,在成本、效率、安全等方面具有突出优势,但空间的大温差、高真空、强辐射、大时延、微重力等苛刻环境,以及航天器的质量、体积、功耗、寿命、可靠性等因素的约束。</p
Speeding up the Topography Imaging of Atomic Force Microscopy by Convolutional Neural Network
No full textAtomic force microscopy (AFM) provides unprecedented insight into surface topography research with ultrahigh spatial resolution at the subnanometer level. However, a slow scanning rate has to be employed to ensure the image quality, which will largely increase the accumulated sample drift, thereby, resulting in the low fidelity of the AFM image. In this paper, we propose a fast imaging method which performs a complete fast Raster scanning and a slow μ-path subsampling together with a deep learning algorithm to rapidly produce an AFM image with high quality and small drift. A supervised convolutional neural network (CNN) model is trained with the slow μ-path subsampled data and its counterpart acquired with fast Raster scan. The fast speed acquired AFM image is then inputted to the well-trained CNN model to output the high quality one. We validate the reliability of this method using a silicon grids sample and further apply it to the fast imaging of a vanadium dioxide thin film. The results demonstrate that this method can largely improve the imaging speed up to 10.3 times with state-of-the-art imaging quality, and reduce the sample drift by 8.9 times in the multiframe AFM imaging of the same area. Furthermore, we prove that this method is also applicable to other scanning imaging techniques such as scanning electrochemical microscopy.</p