CN107378780A - A kind of robot casting grinding adaptive approach of view-based access control model system - Google Patents
A kind of robot casting grinding adaptive approach of view-based access control model system Download PDFInfo
- Publication number
- CN107378780A CN107378780A CN201710588436.0A CN201710588436A CN107378780A CN 107378780 A CN107378780 A CN 107378780A CN 201710588436 A CN201710588436 A CN 201710588436A CN 107378780 A CN107378780 A CN 107378780A
- Authority
- CN
- China
- Prior art keywords
- casting
- robot
- grinding
- time
- vision
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000005266 casting Methods 0.000 title claims abstract description 171
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 24
- 238000000034 method Methods 0.000 claims abstract description 72
- 230000000877 morphologic effect Effects 0.000 claims abstract description 24
- 230000003746 surface roughness Effects 0.000 claims abstract description 17
- 230000000007 visual effect Effects 0.000 claims abstract description 17
- 238000005259 measurement Methods 0.000 claims abstract description 16
- 238000004458 analytical method Methods 0.000 claims abstract description 8
- 238000000605 extraction Methods 0.000 claims abstract description 7
- 238000012876 topography Methods 0.000 claims abstract description 7
- 238000005498 polishing Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 238000013519 translation Methods 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 6
- 238000013135 deep learning Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 4
- 230000007547 defect Effects 0.000 claims description 3
- 238000005286 illumination Methods 0.000 claims description 3
- 239000000463 material Substances 0.000 claims description 3
- 239000013598 vector Substances 0.000 claims description 3
- 238000007517 polishing process Methods 0.000 claims 1
- 238000012512 characterization method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000004512 die casting Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000003754 machining Methods 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B24—GRINDING; POLISHING
- B24B—MACHINES, DEVICES, OR PROCESSES FOR GRINDING OR POLISHING; DRESSING OR CONDITIONING OF ABRADING SURFACES; FEEDING OF GRINDING, POLISHING, OR LAPPING AGENTS
- B24B51/00—Arrangements for automatic control of a series of individual steps in grinding a workpiece
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Finish Polishing, Edge Sharpening, And Grinding By Specific Grinding Devices (AREA)
Abstract
本发明提供了基于视觉系统的机器人铸件打磨自适应方法,包括如下步骤,铸件表面形貌特征机器视觉提取;利用刻画铸件表面粗糙度方法,将铸件高度的一阶矩特征、二阶矩特征、三阶矩特征和四阶矩特征分别表征了铸件表面平均偏移程度、标准差、扭曲程度与峰度;根据铸件表面粗糙度,结合视觉测量与定位方法、运动学分析方法、力控制方法与位置控制伺服方法,生成视觉自适应机器人打磨参数、并控制打磨执行元件进行打磨;根据打磨过程中,工件表面实时的动态形态特征参数与目标形态特征参数实时比对,直至工件表面的粗糙度达到目标形态特征参数。本发明可以通过阶矩刻画铸件表面形态特征获取的工件表面质量能更全面、更精细地刻画铸件表面形态特征。
The invention provides an adaptive method for grinding robot castings based on a vision system, which includes the following steps: machine vision extraction of surface topography features of castings; using the method of describing the surface roughness of castings, the first-order moment feature, second-order moment feature, and The third-order moment feature and the fourth-order moment feature characterize the average deviation degree, standard deviation, twist degree and kurtosis of the casting surface respectively; according to the surface roughness of the casting, combined with visual measurement and positioning method, kinematic analysis method, force control method and The position control servo method generates the visual adaptive robot grinding parameters and controls the grinding actuators for grinding; according to the real-time dynamic morphological characteristic parameters of the workpiece surface during the grinding process and the target morphological characteristic parameters, it is compared in real time until the roughness of the workpiece surface reaches Target morphological feature parameters. The invention can describe the surface morphology characteristics of the castings more comprehensively and finely by describing the surface morphology characteristics of the castings through step moments.
Description
技术领域technical field
本发明涉及智能制造领域或者打磨机器人领域,尤其涉及一种基于视觉系统的机器人铸件打磨自适应方法。The invention relates to the field of intelligent manufacturing or the field of grinding robots, in particular to an adaptive method for grinding robot castings based on a vision system.
背景技术Background technique
工业机器人在打磨行业中的应用是近年来机器人应用范围里一种新型的生产工艺,对机器人控制系统、打磨设备解决方案都提出了更高的要求,工业机器人与数控机床相比加工精度等级要差很多,如何能提高机器人在打磨行业中的加工精度及加工效率、质量涉及到方方面面的技术问题及工艺方法。The application of industrial robots in the grinding industry is a new type of production process in the scope of robot applications in recent years. It puts forward higher requirements for robot control systems and grinding equipment solutions. Compared with CNC machine tools, industrial robots have higher machining accuracy levels. How to improve the processing accuracy, processing efficiency and quality of robots in the grinding industry involves all aspects of technical issues and process methods.
现有的机器人打磨系统依据力传感器获取的力参数,采用控制伺服补偿电机、弹性机构等,实现机器人打磨精度的控制,2014年,曹金学等提出了公开号为CN104149028A的一种高精度机器人打磨系统及其控制方法,通过标定系统提高打磨精度;2016年,梁英等提出了公开号为CN105773368A的一种力控打磨装置及应用其的打磨机器人,以提高打磨精度;2017年,张弓等提出了公开号为CN106425790A的一种压铸件多机器人协同打磨装置及方法,实现了多机器人协同打磨。当打磨的对象为缸体等大型铸件时,单一的标定系统、基于力传感器的补偿与多机器人协同并不能保证打磨铸件的质量,当打磨后的铸件存在质量问题时,不能及时检测并重新打磨。The existing robot grinding system uses the force parameters obtained by the force sensor to control the servo compensation motor, elastic mechanism, etc., to realize the control of robot grinding precision. In 2014, Cao Jinxue et al. proposed a high-precision robot grinding system with the publication number CN104149028A and its control method, and improve the grinding precision through the calibration system; in 2016, Liang Ying et al. proposed a force-controlled grinding device with the publication number CN105773368A and a grinding robot using it to improve the grinding accuracy; in 2017, Zhang Gong et al. proposed The publication number is CN106425790A, a multi-robot collaborative grinding device and method for die castings, which realizes multi-robot collaborative grinding. When the object to be polished is a large casting such as a cylinder block, a single calibration system, compensation based on a force sensor and multi-robot collaboration cannot guarantee the quality of the polished casting. When there is a quality problem in the polished casting, it cannot be detected and re-polished in time. .
发明内容Contents of the invention
针对现有技术中存在不足,本发明提供了一种基于视觉系统的机器人铸件打磨自适应方法,采用阶矩刻画铸件表面形态特征获取的工件表面质量能更全面、更精细地刻画铸件表面形态特征。Aiming at the deficiencies in the prior art, the present invention provides an adaptive method for robot casting grinding based on a vision system. The surface quality of the workpiece obtained by describing the surface morphological characteristics of the casting with steps and moments can describe the surface morphological characteristics of the casting more comprehensively and finely. .
本发明是通过以下技术手段实现上述技术目的的。The present invention achieves the above-mentioned technical purpose through the following technical means.
一种基于视觉系统的机器人铸件打磨自适应方法,包括如下方法:An adaptive method for grinding robot castings based on a vision system, including the following methods:
S01:铸件表面形貌特征机器视觉提取;S01: Machine vision extraction of casting surface topography features;
S02:铸件表面形貌特征多阶矩刻画:利用刻画铸件表面粗糙度方法,将铸件高度的一阶矩特征、二阶矩特征、三阶矩特征和四阶矩特征分别表征了铸件表面平均偏移程度、标准差、扭曲程度与峰度;S02: Multi-order moment characterization of casting surface topography features: Using the method of describing the surface roughness of castings, the first-order moment features, second-order moment features, third-order moment features and fourth-order moment features of the casting height are used to characterize the average deviation of the casting surface. degree of shift, standard deviation, degree of distortion and kurtosis;
S03:根据铸件表面粗糙度,结合视觉测量与定位方法、运动学分析方法、力控制方法与位置控制伺服方法,生成视觉自适应机器人打磨参数、并控制打磨执行元件进行打磨;S03: According to the surface roughness of the casting, combined with visual measurement and positioning methods, kinematic analysis methods, force control methods and position control servo methods, generate visual adaptive robot grinding parameters, and control the grinding actuators for grinding;
S04:根据打磨过程中,工件表面实时的动态形态特征参数与目标形态特征参数实时比对,直至工件表面的粗糙度达到目标形态特征参数。S04: According to the grinding process, the real-time dynamic morphological characteristic parameters of the workpiece surface are compared with the target morphological characteristic parameters in real time until the roughness of the workpiece surface reaches the target morphological characteristic parameters.
进一步,所述的步骤S01具体为:采用固定在铸件表面上方的CCD摄像机来摄取光照图像,光源从不同方向照射铸件表面光度立体视觉,所述光源方向向量不少于5个;在每次获取铸件表面图像过程中,光源在同一平面环内均匀分布置,每次只开启一个光源,且依次开启,得到光照铸件表面图像,使用多光源光度立体视觉方法得到铸件形貌。Further, the step S01 specifically includes: using a CCD camera fixed above the surface of the casting to capture the illumination image, the light source illuminates the surface of the casting from different directions for photometric stereoscopic vision, and the direction vectors of the light source are not less than 5; In the process of casting surface imaging, the light sources are evenly distributed in the same plane ring, and only one light source is turned on at a time, and turned on sequentially to obtain the illuminated casting surface image, and the casting morphology is obtained by using the multi-light source photometric stereo vision method.
进一步,所述的多光源光度立体视觉方法具体为:重建铸件三维表面,得到t时刻铸件表面图像It;从铸件表面图像It得到位置(i,j)t处的高度为z(i,j)t,分析得到t时刻要处理缺陷区域的长Lt和宽Wt。Further, the multi-light source photometric stereoscopic vision method is specifically: rebuilding the three-dimensional surface of the casting, and obtaining the casting surface image I t at time t ; obtaining the height at the position (i, j) t from the casting surface image I t is z(i, j) t , the length L t and width W t of the defect area to be processed at time t are obtained by analysis.
进一步,所述的步骤S02具体为:通过机器人控制系统对铸件表面图像It分析,得到分解后的毛刺、孔眼、裂纹、与铸件整体表面粗糙情况;利用刻画铸件表面粗糙度方法,分析铸件高度的一阶矩、二阶矩、三阶矩和四阶矩特征,图像的矩特征作为磨削控制器的输入刻画了加工铸件表面的知识特征。Further, the step S02 specifically includes: analyzing the surface image I t of the casting through the robot control system to obtain decomposed burrs, holes, cracks, and the roughness of the overall surface of the casting; using the method of describing the surface roughness of the casting to analyze the height of the casting The first-order moment, second-order moment, third-order moment and fourth-order moment feature of the image, the moment feature of the image is used as the input of the grinding controller to describe the knowledge characteristics of the processed casting surface.
进一步,所述刻画铸件表面粗糙度方法具体为:铸件表面图像It任意一点(i,j)t的高度值为z(i,j)t,铸件的粗糙度对应铸件高度的各阶矩:Further, the method for describing the surface roughness of the casting is specifically : the height value of any point (i, j) t on the surface image I of the casting is z(i, j) t , and the roughness of the casting corresponds to each step of the casting height:
铸件高度的一阶矩表征了铸件表面平均偏移程度,值越大,则平均偏移量越大,铸件越不平整;First moment of casting height Characterizes the average deviation degree of the casting surface, the larger the value, the greater the average deviation, and the more uneven the casting;
铸件的二阶矩表征了铸件表面各点高度的标准差,即铸件表面的粗糙程度,值越大,铸件表面越粗糙;The second moment of the casting Characterizes the standard deviation of the height of each point on the casting surface, that is, the roughness of the casting surface, the larger the value, the rougher the casting surface;
铸件的三阶矩表征了铸件表面的扭曲程度,值越大,铸件表面扭曲程度越高;The third moment of the casting Characterizes the degree of distortion of the casting surface, the larger the value, the higher the degree of distortion of the casting surface;
铸件的四阶矩表征了铸件表面的峰度值,峰度越高,铸件表面越粗糙;The fourth moment of the casting Characterizes the kurtosis value of the casting surface, the higher the kurtosis, the rougher the casting surface;
则铸件表面粗糙度为:Fflat-t={M1(i,j)t,M2(i,j)t,M3(i,j)t,M4(i,j)t},St为铸件表面的知识特征:St=It(Lt,Wt,Fflat-t)=Lt*Wt*Fflat-t。Then the casting surface roughness is: F flat-t = {M 1 (i,j) t ,M 2 (i,j) t ,M 3 (i,j) t ,M 4 (i,j) t }, S t is the knowledge feature of the casting surface: S t =I t (L t , W t , F flat-t )=L t *W t *F flat-t .
进一步,所述S03步骤中的视觉测量与定位方法为:用视觉测量与定位系统控制器将输入的铸件表面的粗糙度转换为输出的磨削机器人砂轮t时刻的平移速度vr(t)、砂轮的转动速度ωr(t)、加工铸件的深度d(t)和机器人关节转角qr(t);视觉测量与定位系统控制器由深度学习神经网络N完成,通过反复的正向训练学习与反向反馈学习,建立工件整体表面粗糙情况与工件加工砂轮的平移速度vr(t)、转动速度ωr(t)与加工工件的进给深度d(t)控制因子的关联关系,(vr(t),ωr(t),d(t))=N{M1(i,j)t,M2(i,j)t,M3(i,j)t,M4(i,j)t}。Further, the visual measurement and positioning method in the step S03 is: using the visual measurement and positioning system controller to convert the input roughness of the casting surface into the output translational velocity v r (t) of the grinding wheel of the grinding robot at time t, The rotation speed ω r (t) of the grinding wheel, the depth d(t) of the processed casting, and the joint rotation angle q r (t) of the robot; the vision measurement and positioning system controller is completed by the deep learning neural network N, which learns through repeated forward training And reverse feedback learning, establish the relationship between the overall surface roughness of the workpiece and the translational speed v r (t), rotational speed ω r (t) of the workpiece processing grinding wheel and the control factor of the workpiece feed depth d(t), ( v r (t), ω r (t), d(t))=N{M 1 (i,j) t ,M 2 (i,j) t ,M 3 (i,j) t ,M 4 ( i, j) t }.
进一步,所述S03步骤中的运动学分析方法为:利用在t时刻的视觉测量与定位系统控制器的机器人臂末端速度vp(t)与位置p(t),通过运动学模块得到修正后的机器人关节转角qr-n(t)与机器人关节角速度 Further, the kinematics analysis method in the step S03 is: using the velocity vp(t) and the position p (t) of the end of the robot arm of the controller of the visual measurement and positioning system at time t, after being corrected by the kinematics module The robot joint rotation angle q rn (t) and the robot joint angular velocity
进一步,所述S03步骤中的力控制方法为:通过力控制模块修正机器人关节角速度得到修正后的机器人关节角速度 Further, the force control method in the S03 step is: correcting the angular velocity of the robot joint through the force control module Get the corrected angular velocity of the robot joints
进一步,所述S03步骤中的位置控制伺服方法为:通过位置伺服模块将修正后的机器人关节转角qr-n(t)、修正后的机器人关节角速度机器人反馈的关节转角qm(t)、关节角速度转换为机器人实时关节驱动力矩τ(t);机器人臂末端速度与砂轮的平移速度一致:Further, the position control servo method in the S03 step is: through the position servo module, the corrected robot joint rotation angle q rn (t), the corrected robot joint angular velocity Joint rotation angle q m (t) and joint angular velocity fed back by the robot Converted to the real-time joint drive torque τ(t) of the robot; the end speed of the robot arm is consistent with the translation speed of the grinding wheel:
vr(t)、vp(t)为t时刻砂轮的平移速度与机器人臂末端速度:vr(t)=vp(t)=F1(St)=k1St;v r (t), v p (t) are the translation speed of the grinding wheel and the end speed of the robot arm at time t: v r (t) = v p (t) = F 1 (S t ) = k 1 S t ;
ωr(t)为t时刻砂轮的转动速度:ωr(t)=F2(St)=k2St;ω r (t) is the rotation speed of the grinding wheel at time t: ω r (t) = F 2 (S t ) = k 2 S t ;
d(t)为t时刻砂轮的进给深度:d(t)=F3(St)=k3St;d(t) is the feed depth of the grinding wheel at time t: d(t)=F 3 (S t )=k 3 S t ;
其中:系数k1、k2和k3依据铸件材质确定。Among them: the coefficients k 1 , k 2 and k 3 are determined according to the material of the casting.
进一步,所述的步骤S04具体为:设M1、M2、M3、M4分别为铸件目标形态特征的一、二、三、四阶矩标准特征值;分别为动态形态特征参数M1(i,j)t、M2(i,j)t、M3(i,j)t、M4(i,j)t在t时刻平均值,若且且且则在t时刻的表面铸件质量合格,铸件当前表面完成打磨,进入下一个表面的打磨重复步骤S01-S04;其中,δ1、δ2、δ3、δ4分别为质量合格铸件阶矩标准特征值允许误差范围;Further, the step S04 is specifically: set M 1 , M 2 , M 3 , and M 4 as standard eigenvalues of the first, second, third, and fourth moments of the target morphological characteristics of the casting, respectively; are the average values of dynamic morphological characteristic parameters M 1 (i,j) t , M 2 (i,j) t , M 3 (i,j) t , and M 4 (i,j) t at time t, respectively, if and and and Then the quality of the surface casting at time t is qualified, the current surface of the casting is polished, and the next surface is polished and the steps S01-S04 are repeated; where δ 1 , δ 2 , δ 3 , and δ 4 are the standard characteristics of the step moments of the quality-qualified casting Value allowable error range;
若不满足,则表示在t时刻的表面铸件质量不合格,则t=kT,且k=k+1,其中k为打磨次数;T为从视觉特征提取到完成一次打磨需要的时间;重复步骤S01-S04。If it is not satisfied, it means that the surface casting quality at time t is unqualified, then t=kT, and k=k+1, where k is the number of times of polishing; T is the time required from visual feature extraction to finishing a polishing; repeat steps S01-S04.
本发明的有益效果在于:The beneficial effects of the present invention are:
1.本发明所述的基于视觉系统的机器人铸件打磨自适应方法,采用多光源光度立体视觉方法获取铸件三维形态能更高精度地测定加工工件表面质量。1. The vision system-based robot casting grinding adaptive method of the present invention adopts the multi-light source photometric stereo vision method to obtain the three-dimensional shape of the casting, which can measure the surface quality of the processed workpiece with higher accuracy.
2.本发明所述的基于视觉系统的机器人铸件打磨自适应方法,采用阶矩刻画铸件表面形态特征获取的工件表面质量能更全面、更精细地刻画铸件表面形态特征。2. The vision system-based robot casting grinding self-adaptive method described in the present invention, the surface quality of the workpiece obtained by using step moments to describe the surface morphological characteristics of the casting can describe the surface morphological characteristics of the casting more comprehensively and finely.
3.本发明所述的基于视觉系统的机器人铸件打磨自适应方法,采用深度学习神经网络控制器完成视觉测量与定位,本方法集打磨与质量检测于一体,能自适应地完成工件打磨,视觉测量与定位系统控制器由深度学习神经网络实现,深度学习神经网络通过反复的正向训练学习与反向反馈学习,建立工件整体表面粗糙情况与工件加工砂轮的平移速度、转动速度与加工工件的进给深度等控制因子的关联关系,进而完成机器人铸件实时自适应打磨,保证具有大型特征的缸体铸件的打磨质量,当打磨后的铸件存在质量问题时,实时地检测并重新打磨,直到铸件打磨质量合格。3. The robot casting grinding adaptive method based on the vision system of the present invention uses a deep learning neural network controller to complete visual measurement and positioning. This method integrates grinding and quality inspection, and can adaptively complete workpiece grinding. The controller of the measurement and positioning system is realized by a deep learning neural network. The deep learning neural network establishes the roughness of the overall surface of the workpiece and the translational speed and rotation speed of the workpiece processing grinding wheel through repeated forward training learning and reverse feedback learning. The relationship between control factors such as feed depth, and then complete the real-time adaptive grinding of robot castings to ensure the grinding quality of cylinder castings with large features. Polishing quality is acceptable.
附图说明Description of drawings
图1为本发明所述的基于视觉系统的机器人铸件打磨自适应方法的流程图。FIG. 1 is a flow chart of the vision system-based adaptive grinding method for robot castings according to the present invention.
图2为本发明所述的多光源光度立体视觉方法的示意图。Fig. 2 is a schematic diagram of the multi-light source photometric stereoscopic vision method of the present invention.
图3为本发明所述的机器人视觉自适应机器人打磨控制示意图。Fig. 3 is a schematic diagram of robot vision adaptive robot grinding control according to the present invention.
图中:In the picture:
1-1:CCD摄像机;1-2:光源;1-3:铸件。1-1: CCD camera; 1-2: light source; 1-3: casting.
具体实施方式detailed description
下面结合附图以及具体实施例对本发明作进一步的说明,但本发明的保护范围并不限于此。The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.
如图1所示,利用一种基于视觉系统的机器人铸件打磨自适应方法,自动打磨重量为38kg的HT300缸体铸件,铸件体积为400mm×320mm×253mm,具体步骤如下:As shown in Figure 1, using an adaptive method for grinding robot castings based on the vision system, the HT300 cylinder casting with a weight of 38kg is automatically polished, and the volume of the casting is 400mm×320mm×253mm. The specific steps are as follows:
S01:铸件表面形貌特征机器视觉提取:S01: Machine vision extraction of casting surface topography features:
加工工件为4缸缸体,采用固定在缸体上方的Panasoinc WV-CP410/G型CCD摄像机来摄取光照图像,摄像机的焦距f=16mm,摄像机到缸体的距离u=745mm,通过CCD摄像机1-1来摄取光照图像,光源(1-2)从不同方向照射铸件表面光度立体视觉,如图2所示,所述光源(1-2)方向向量分别为S1=(65,-320,480),S2=(295,-195,480),S3=(272,74,480),S4=(155,218,480),S5=(-188,230,480),S6=(-130,120,480);在每次获取铸件表面图像过程中,光源(1-2)在同一平面环内均匀分布置,每次只开启一个光源,且依次开启,得到光照铸件表面图像,使用多光源光度立体视觉方法得到铸件形貌。The processed workpiece is a 4-cylinder cylinder block, and the Panasoinc WV-CP410/G CCD camera fixed above the cylinder block is used to capture the light image. The focal length of the camera is f=16mm, and the distance from the camera to the cylinder block is u=745mm. -1 to pick up the illumination image, the light source (1-2) irradiates the casting surface photometric stereoscopic vision from different directions, as shown in Figure 2, the direction vector of the light source (1-2) is respectively S1=(65,-320,480 ), S2=(295,-195,480), S3=(272,74,480), S4=(155,218,480), S5=(-188,230,480), S6=(-130, 120, 480); in the process of obtaining the surface image of the casting each time, the light sources (1-2) are evenly distributed in the same plane ring, and only one light source is turned on at a time, and turned on sequentially to obtain the surface image of the illuminated casting, using multiple light sources The photometric stereo vision method was used to obtain the casting morphology.
多光源光度立体视觉方法具体为:重建铸件三维表面,t时刻铸件表面图像It,位置(i,j)t处的高度为z(i,j)t,t时刻要处理缺陷区域的长Lt和宽Wt。The multi-light source photometric stereo vision method is as follows: reconstruct the three-dimensional surface of the casting, cast surface image I t at time t, the height at position (i,j) t is z(i,j) t , and the length L of the defect area to be processed at time t t and width W t .
S02:铸件表面形貌特征多阶矩刻画:通过机器人控制系统对铸件表面图像It分析,得到分解后的毛刺、孔眼、裂纹、与铸件整体表面粗糙情况;利用刻画铸件表面粗糙度方法,分析铸件高度的一阶矩、二阶矩、三阶矩和四阶矩特征,图像的矩特征作为磨削控制器的输入刻画了加工铸件表面的知识特征。S02: Multi-moment characterization of casting surface topography features: analyze the casting surface image I t through the robot control system, and obtain the decomposed burrs, holes, cracks, and overall surface roughness of the casting; use the method of depicting the surface roughness of the casting to analyze The first-order moment, second-order moment, third-order moment and fourth-order moment feature of the casting height, and the moment feature of the image is used as the input of the grinding controller to describe the knowledge characteristics of the processed casting surface.
所述刻画铸件表面粗糙度方法具体为:铸件表面图像It任意一点(i,j)t的高度值为z(i,j)t,铸件的粗糙度对应铸件高度的各阶矩:The method for describing the surface roughness of the casting is specifically : the height value of any point (i, j) t on the surface image I of the casting is z(i, j) t , and the roughness of the casting corresponds to each order moment of the casting height:
铸件高度的一阶矩表征了铸件表面平均偏移程度,值越大,则平均偏移量越大,铸件越不平整;First moment of casting height Characterizes the average deviation degree of the casting surface, the larger the value, the greater the average deviation, and the more uneven the casting;
铸件的二阶矩表征了铸件表面各点高度的标准差,即铸件表面的粗糙程度,值越大,铸件表面越粗糙;The second moment of the casting Characterizes the standard deviation of the height of each point on the casting surface, that is, the roughness of the casting surface, the larger the value, the rougher the casting surface;
铸件的三阶矩表征了铸件表面的扭曲程度,值越大,铸件表面扭曲程度越高;The third moment of the casting Characterizes the degree of distortion of the casting surface, the larger the value, the higher the degree of distortion of the casting surface;
铸件的四阶矩表征了铸件表面的峰度值,峰度越高,铸件表面越粗糙;The fourth moment of the casting Characterizes the kurtosis value of the casting surface, the higher the kurtosis, the rougher the casting surface;
则铸件表面粗糙度为:Fflat-t={M1(i,j)t,M2(i,j)t,M3(i,j)t,M4(i,j)t},Then the casting surface roughness is: F flat-t = {M 1 (i,j) t ,M 2 (i,j) t ,M 3 (i,j) t ,M 4 (i,j) t },
St为铸件表面的知识特征:St=It(Lt,Wt,Fflat-t)=Lt*Wt*Fflat-t。S t is the knowledge feature of the casting surface: S t =I t (L t , W t , F flat-t )=L t *W t *F flat-t .
S03:如图3所示,根据铸件表面粗糙度,结合视觉测量与定位方法、运动学分析方法、力控制方法与位置控制伺服方法,生成视觉自适应机器人打磨参数、并控制打磨执行元件进行打磨;S03: As shown in Figure 3, according to the surface roughness of the casting, combined with the visual measurement and positioning method, kinematics analysis method, force control method and position control servo method, the visual adaptive robot grinding parameters are generated, and the grinding actuator is controlled for grinding ;
视觉测量与定位方法为:用视觉测量与定位系统控制器将输入的铸件表面的粗糙度转换为输出的磨削机器人砂轮t时刻的平移速度vr(t)、砂轮的转动速度ωr(t)、加工铸件的深度d(t)和机器人关节转角qr(t);视觉测量与定位系统控制器由深度学习神经网络N完成,通过反复的正向训练学习与反向反馈学习,建立工件整体表面粗糙情况与工件加工砂轮的平移速度vr(t)、转动速度ωr(t)与加工工件的进给深度d(t)控制因子的关联关系,(vr(t),ωr(t),d(t))=N{M1(i,j)t,M2(i,j)t,M3(i,j)t,M4(i,j)t}。The visual measurement and positioning method is as follows: use the visual measurement and positioning system controller to convert the input roughness of the casting surface into the output translational velocity v r (t) of the grinding wheel of the grinding robot at time t, and the rotational speed of the grinding wheel ω r (t ), the depth d(t) of the processed casting and the robot joint rotation angle q r (t); the visual measurement and positioning system controller is completed by the deep learning neural network N, and the workpiece is established through repeated forward training learning and reverse feedback learning The relationship between the roughness of the overall surface and the translational speed v r (t), rotational speed ω r (t) of the workpiece processing grinding wheel and the control factor of the workpiece feed depth d(t), (v r (t), ω r (t), d(t))=N{M 1 (i,j) t , M 2 (i,j) t , M 3 (i,j) t ,M 4 (i,j) t }.
运动学分析方法为:利用在t时刻的视觉测量与定位系统控制器的机器人臂末端速度vp(t)与位置p(t),通过运动学模块得到修正后的机器人关节转角qr-n(t)与机器人关节角速度 The kinematics analysis method is as follows: using the vision measurement and positioning system controller’s end speed v p (t) and position p (t) of the robot arm at time t, the corrected robot joint rotation angle q rn (t ) and the robot joint angular velocity
力控制方法为:通过力控制模块修正机器人关节角速度得到修正后的机器人关节角速度 The force control method is: correct the angular velocity of the robot joint through the force control module Get the corrected angular velocity of the robot joints
位置控制伺服方法为:通通过位置伺服模块将修正后的机器人关节转角qr-n(t)、修正后的机器人关节角速度机器人反馈的关节转角qm(t)、关节角速度转换为机器人实时关节驱动力矩τ(t);机器人臂末端速度与砂轮的平移速度一致:The position control servo method is: through the position servo module, the corrected robot joint rotation angle q rn (t), the corrected robot joint angular velocity Joint rotation angle q m (t) and joint angular velocity fed back by the robot Converted to the real-time joint drive torque τ(t) of the robot; the end speed of the robot arm is consistent with the translation speed of the grinding wheel:
vr(t)、vp(t)为t时刻砂轮的平移速度与机器人臂末端速度:vr(t)=vp(t)=F1(St)=k1St;v r (t), v p (t) are the translation speed of the grinding wheel and the end speed of the robot arm at time t: v r (t) = v p (t) = F 1 (S t ) = k 1 S t ;
ωr(t)为t时刻砂轮的转动速度:ωr(t)=F2(St)=k2St;ω r (t) is the rotation speed of the grinding wheel at time t: ω r (t) = F 2 (S t ) = k 2 S t ;
d(t)为t时刻砂轮的进给深度:d(t)=F3(St)=k3St;d(t) is the feed depth of the grinding wheel at time t: d(t)=F 3 (S t )=k 3 S t ;
其中:系数k1、k2和k3依据铸件材质确定。Among them: the coefficients k 1 , k 2 and k 3 are determined according to the material of the casting.
S04:根据打磨过程中,工件表面实时的动态形态特征参数与目标形态特征参数实时比对,直至工件表面的粗糙度达到目标形态特征参数。与目标铸件表面期望的几何形貌相比较,得到实时的打磨特征参数,依据实时的几何形貌,自适应地完成铸件的打磨。S04: According to the grinding process, the real-time dynamic morphological characteristic parameters of the workpiece surface are compared with the target morphological characteristic parameters in real time until the roughness of the workpiece surface reaches the target morphological characteristic parameters. Compared with the expected geometric shape of the target casting surface, the real-time grinding characteristic parameters are obtained, and the grinding of the casting is completed adaptively according to the real-time geometric shape.
具体为:设M1、M2、M3、M4分别为铸件目标形态特征的一、二、三、四阶矩标准特征值,其中M1=0.3μm,M2=0.003μm,M3=0.01μm,M4=0.03μm;分别为动态形态特征参数M1(i,j)t、M2(i,j)t、M3(i,j)t、M4(i,j)t在t时刻平均值,若且且且则在t时刻的表面铸件质量合格,铸件当前表面完成打磨,进入下一个表面的打磨重复步骤S01-S04;其中,δ1、δ2、δ3、δ4分别为质量合格铸件阶矩标准特征值允许误差范围,本案中δ1=0.05μm,δ2=0.005μm,δ3=0.005μm,δ4=0.005μm;Specifically: Let M 1 , M 2 , M 3 , and M 4 be the standard eigenvalues of the first, second, third, and fourth moments of the target morphological characteristics of the casting, where M 1 =0.3 μm, M 2 =0.003 μm, and M 3 = 0.01 μm, M4 = 0.03 μm ; are the average values of dynamic morphological characteristic parameters M 1 (i,j) t , M 2 (i,j) t , M 3 (i,j) t , and M 4 (i,j) t at time t, respectively, if and and and Then the quality of the surface casting at time t is qualified, the current surface of the casting is polished, and the next surface is polished and the steps S01-S04 are repeated; where δ 1 , δ 2 , δ 3 , and δ 4 are the standard characteristics of the step moments of the quality-qualified casting Value allowable error range, in this case δ 1 =0.05μm, δ 2 =0.005μm, δ 3 =0.005μm, δ 4 =0.005μm;
若不满足,则表示在t时刻的表面铸件质量不合格,则t=kT,且k=k+1,其中k为打磨次数;T为从视觉特征提取到完成一次打磨需要的时间;重复步骤S01-S04。If it is not satisfied, it means that the surface casting quality at time t is unqualified, then t=kT, and k=k+1, where k is the number of times of polishing; T is the time required from visual feature extraction to finishing a polishing; repeat steps S01-S04.
在本案例中,采用自适应打磨方法,加工次数有了较大幅度的减少,同时有效加工次数增加了,在30件缸体的加工中,自适应打磨方法加工结果中,只有1件加工不合格;而在这之前,30件缸体的加工中,有3-5件不合格。In this case, using the adaptive grinding method, the number of processing has been greatly reduced, and the number of effective processing has increased. In the processing of 30 cylinder blocks, only 1 of the processing results of the adaptive grinding method was not correct. Qualified; before that, 3-5 of the 30 cylinder blocks were unqualified.
所述实施例为本发明的优选的实施方式,但本发明并不限于上述实施方式,在不背离本发明的实质内容的情况下,本领域技术人员能够做出的任何显而易见的改进、替换或变型均属于本发明的保护范围。The described embodiment is a preferred implementation of the present invention, but the present invention is not limited to the above-mentioned implementation, without departing from the essence of the present invention, any obvious improvement, replacement or modification that those skilled in the art can make Modifications all belong to the protection scope of the present invention.
Claims (10)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710588436.0A CN107378780B (en) | 2017-07-19 | 2017-07-19 | A kind of robot casting grinding adaptive approach of view-based access control model system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710588436.0A CN107378780B (en) | 2017-07-19 | 2017-07-19 | A kind of robot casting grinding adaptive approach of view-based access control model system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107378780A true CN107378780A (en) | 2017-11-24 |
CN107378780B CN107378780B (en) | 2019-01-08 |
Family
ID=60339998
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710588436.0A Active CN107378780B (en) | 2017-07-19 | 2017-07-19 | A kind of robot casting grinding adaptive approach of view-based access control model system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107378780B (en) |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108608296A (en) * | 2018-06-15 | 2018-10-02 | 芜湖泓鹄材料技术有限公司 | Mould cast grinding apparatus |
CN108637860A (en) * | 2018-04-18 | 2018-10-12 | 武汉理工大学 | High ferro white body automation wire-drawing frame and method based on Robot Hand-eye control |
CN109605140A (en) * | 2018-12-25 | 2019-04-12 | 珞石(山东)智能科技有限公司 | Based on machine vision and have the function of that the cutter of six shaft mechanical arm of power control puts the first edge on a knife or a pair of scissors method |
CN109732625A (en) * | 2019-03-15 | 2019-05-10 | 珠海格力电器股份有限公司 | Industrial robot flexible polishing method and system based on machine vision |
CN110091221A (en) * | 2019-05-13 | 2019-08-06 | 成都工业学院 | A kind of die surface processing method |
CN110405559A (en) * | 2019-08-09 | 2019-11-05 | 珠海心怡科技有限公司 | A kind of metope intelligence sanding and polishing machine in robot |
CN110587484A (en) * | 2019-07-29 | 2019-12-20 | 苏州超徕精工科技有限公司 | Device and method for predicting removal effect in polishing process in real time |
CN111331435A (en) * | 2020-03-27 | 2020-06-26 | 中冶赛迪工程技术股份有限公司 | An intelligent surface grinding process and production line for an alloy medium and heavy plate |
CN111468989A (en) * | 2020-03-30 | 2020-07-31 | 黄河水利职业技术学院 | Five-axis linkage numerical control manipulator polishing control system and method |
CN111604719A (en) * | 2020-06-02 | 2020-09-01 | 湖北大学 | An adaptive, high-efficiency, and large-removal longitudinal grinding method for cylindrical grinding |
CN111906788A (en) * | 2020-08-12 | 2020-11-10 | 湖北民族大学 | Bathroom intelligent polishing system based on machine vision and polishing method thereof |
CN112345539A (en) * | 2020-11-05 | 2021-02-09 | 菲特(天津)检测技术有限公司 | Aluminum die casting surface defect detection method based on deep learning |
CN113021082A (en) * | 2019-12-24 | 2021-06-25 | 沈阳智能机器人创新中心有限公司 | Robot casting polishing method based on teleoperation and panoramic vision |
CN114310962A (en) * | 2022-01-18 | 2022-04-12 | 太原科技大学 | An intelligent robot communication control system and method suitable for grinding |
US20220126319A1 (en) * | 2019-02-05 | 2022-04-28 | 3M Innovative Properties Company | Paint repair process by scenario |
CN115026660A (en) * | 2022-08-11 | 2022-09-09 | 昆山市恒达精密机械工业有限公司 | CCD-based grinding process intelligent control method and system |
CN115100211A (en) * | 2022-08-29 | 2022-09-23 | 南通电博士自动化设备有限公司 | Intelligent regulation and control method for surface polishing speed of metal plate by robot |
CN115338709A (en) * | 2022-10-18 | 2022-11-15 | 徐州艾奇川自动化设备有限公司 | Numerical control machining intelligent monitoring control system based on industrial intelligence |
CN116967939A (en) * | 2023-09-22 | 2023-10-31 | 大儒科技(苏州)有限公司 | Special force control system for grinding and polishing |
CN116993230A (en) * | 2023-09-26 | 2023-11-03 | 山东省智能机器人应用技术研究院 | Machine polishing operation quality evaluation system |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2003062752A (en) * | 2001-08-20 | 2003-03-05 | For-A Co Ltd | Finishing device |
CN103831695A (en) * | 2014-03-28 | 2014-06-04 | 中国科学院自动化研究所 | Large free-form surface robot polishing system |
CN105196148A (en) * | 2015-09-25 | 2015-12-30 | 广东省自动化研究所 | Self-adaption polishing and grinding system with intelligent feed and discharge function |
CN205520871U (en) * | 2015-09-25 | 2016-08-31 | 广东省自动化研究所 | Intelligence polishing system of polishing based on vision sensor |
CN205703645U (en) * | 2016-04-13 | 2016-11-23 | 广州文冲船厂有限责任公司 | A kind of robot sanding apparatus |
-
2017
- 2017-07-19 CN CN201710588436.0A patent/CN107378780B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2003062752A (en) * | 2001-08-20 | 2003-03-05 | For-A Co Ltd | Finishing device |
CN103831695A (en) * | 2014-03-28 | 2014-06-04 | 中国科学院自动化研究所 | Large free-form surface robot polishing system |
CN105196148A (en) * | 2015-09-25 | 2015-12-30 | 广东省自动化研究所 | Self-adaption polishing and grinding system with intelligent feed and discharge function |
CN205520871U (en) * | 2015-09-25 | 2016-08-31 | 广东省自动化研究所 | Intelligence polishing system of polishing based on vision sensor |
CN205703645U (en) * | 2016-04-13 | 2016-11-23 | 广州文冲船厂有限责任公司 | A kind of robot sanding apparatus |
Cited By (30)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108637860A (en) * | 2018-04-18 | 2018-10-12 | 武汉理工大学 | High ferro white body automation wire-drawing frame and method based on Robot Hand-eye control |
CN108608296A (en) * | 2018-06-15 | 2018-10-02 | 芜湖泓鹄材料技术有限公司 | Mould cast grinding apparatus |
CN108608296B (en) * | 2018-06-15 | 2024-04-26 | 芜湖泓鹄材料技术有限公司 | Die casting polishing equipment |
CN109605140A (en) * | 2018-12-25 | 2019-04-12 | 珞石(山东)智能科技有限公司 | Based on machine vision and have the function of that the cutter of six shaft mechanical arm of power control puts the first edge on a knife or a pair of scissors method |
US20220126319A1 (en) * | 2019-02-05 | 2022-04-28 | 3M Innovative Properties Company | Paint repair process by scenario |
US12275038B2 (en) * | 2019-02-05 | 2025-04-15 | 3M Innovative Properties Company | Paint repair process by scenario |
CN109732625B (en) * | 2019-03-15 | 2020-11-27 | 珠海格力电器股份有限公司 | Industrial robot flexible polishing method and system based on machine vision |
CN109732625A (en) * | 2019-03-15 | 2019-05-10 | 珠海格力电器股份有限公司 | Industrial robot flexible polishing method and system based on machine vision |
CN110091221A (en) * | 2019-05-13 | 2019-08-06 | 成都工业学院 | A kind of die surface processing method |
CN110587484A (en) * | 2019-07-29 | 2019-12-20 | 苏州超徕精工科技有限公司 | Device and method for predicting removal effect in polishing process in real time |
CN110405559A (en) * | 2019-08-09 | 2019-11-05 | 珠海心怡科技有限公司 | A kind of metope intelligence sanding and polishing machine in robot |
CN113021082B (en) * | 2019-12-24 | 2022-06-07 | 沈阳智能机器人创新中心有限公司 | Robot casting polishing method based on teleoperation and panoramic vision |
CN113021082A (en) * | 2019-12-24 | 2021-06-25 | 沈阳智能机器人创新中心有限公司 | Robot casting polishing method based on teleoperation and panoramic vision |
CN111331435A (en) * | 2020-03-27 | 2020-06-26 | 中冶赛迪工程技术股份有限公司 | An intelligent surface grinding process and production line for an alloy medium and heavy plate |
CN111468989A (en) * | 2020-03-30 | 2020-07-31 | 黄河水利职业技术学院 | Five-axis linkage numerical control manipulator polishing control system and method |
CN111468989B (en) * | 2020-03-30 | 2021-08-24 | 黄河水利职业技术学院 | A five-axis linkage numerical control manipulator polishing control system and method |
CN111604719B (en) * | 2020-06-02 | 2021-08-10 | 湖北大学 | Self-adaptive high-efficiency large-grinding-amount longitudinal grinding method for cylindrical grinding |
CN111604719A (en) * | 2020-06-02 | 2020-09-01 | 湖北大学 | An adaptive, high-efficiency, and large-removal longitudinal grinding method for cylindrical grinding |
CN111906788B (en) * | 2020-08-12 | 2021-09-14 | 湖北民族大学 | Bathroom intelligent polishing system based on machine vision and polishing method thereof |
CN111906788A (en) * | 2020-08-12 | 2020-11-10 | 湖北民族大学 | Bathroom intelligent polishing system based on machine vision and polishing method thereof |
CN112345539A (en) * | 2020-11-05 | 2021-02-09 | 菲特(天津)检测技术有限公司 | Aluminum die casting surface defect detection method based on deep learning |
CN114310962A (en) * | 2022-01-18 | 2022-04-12 | 太原科技大学 | An intelligent robot communication control system and method suitable for grinding |
CN115026660A (en) * | 2022-08-11 | 2022-09-09 | 昆山市恒达精密机械工业有限公司 | CCD-based grinding process intelligent control method and system |
CN115100211B (en) * | 2022-08-29 | 2022-11-18 | 南通电博士自动化设备有限公司 | Intelligent regulation and control method for surface polishing speed of metal plate by robot |
CN115100211A (en) * | 2022-08-29 | 2022-09-23 | 南通电博士自动化设备有限公司 | Intelligent regulation and control method for surface polishing speed of metal plate by robot |
CN115338709A (en) * | 2022-10-18 | 2022-11-15 | 徐州艾奇川自动化设备有限公司 | Numerical control machining intelligent monitoring control system based on industrial intelligence |
CN116967939A (en) * | 2023-09-22 | 2023-10-31 | 大儒科技(苏州)有限公司 | Special force control system for grinding and polishing |
CN116967939B (en) * | 2023-09-22 | 2023-12-26 | 大儒科技(苏州)有限公司 | Special force control system for grinding and polishing |
CN116993230A (en) * | 2023-09-26 | 2023-11-03 | 山东省智能机器人应用技术研究院 | Machine polishing operation quality evaluation system |
CN116993230B (en) * | 2023-09-26 | 2023-12-15 | 山东省智能机器人应用技术研究院 | Machine polishing operation quality evaluation system |
Also Published As
Publication number | Publication date |
---|---|
CN107378780B (en) | 2019-01-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107378780A (en) | A kind of robot casting grinding adaptive approach of view-based access control model system | |
CN103325106B (en) | Based on the Moving Workpieces method for sorting of LabVIEW | |
CN109571152B (en) | Automatic workpiece polishing method based on offline programming | |
Huo et al. | Model-free adaptive impedance control for autonomous robotic sanding | |
Schneider et al. | Stiffness modeling of industrial robots for deformation compensation in machining | |
CN107192331A (en) | A kind of workpiece grabbing method based on binocular vision | |
CN113146172A (en) | Multi-vision-based detection and assembly system and method | |
CN109013405A (en) | It is a kind of independently detected with cast(ing) surface and substandard products sorting function robot system | |
CN111645074A (en) | Robot grabbing and positioning method | |
CN109940606B (en) | Robot guiding system and method based on point cloud data | |
Kosler et al. | Adaptive robotic deburring of die-cast parts with position and orientation measurements using a 3D laser-triangulation sensor | |
CN109648202B (en) | Additive manufacturing forming precision control method for non-flat surface autonomous recognition robot | |
CN113267452A (en) | Engine cylinder surface defect detection method and system based on machine vision | |
Hsu et al. | Development of a faster classification system for metal parts using machine vision under different lighting environments | |
CN110394802B (en) | Polishing robot and position compensation method | |
CN113500593A (en) | Method for grabbing designated part of shaft workpiece for loading | |
Tsarouchi et al. | Vision system for robotic handling of randomly placed objects | |
CN116079732A (en) | Cabin assembly method based on laser tracker and binocular vision mixed guidance | |
CN108107842A (en) | Robot polishing track evaluation method based on power control | |
KR102096897B1 (en) | The auto teaching system for controlling a robot using a 3D file and teaching method thereof | |
CN116237933A (en) | Digital twin accurate repair equipment and repair method for surface defects of workpiece | |
CN110281152B (en) | A robot constant force grinding path planning method and system based on online trial touch | |
CN205765272U (en) | Processing device with vision measurement function | |
CN105867300A (en) | Reverse remachining method for large forming welded part with complex contour | |
CN113414784A (en) | Robot micro-assembly grabbing system based on deep reinforcement learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |