CN103056759B - Robot grinding system based on feedback of sensor - Google Patents

Abstract

The invention discloses and relates to a robot grinding system based on feedback of a sensor. The robot grinding system based on the feedback of the sensor is capable of detecting an contour of a workpiece and regulating grinding tracks in real time and comprises a robot (1), a workpiece contour detection unit (2), an electro-spindle unit (3), a grinding wheel wear detection unit (4), a grinding force detection unit (5) and a system control host (6). The robot grinding system based on the feedback of the sensor achieves automatic grinding on workpiece surface by detecting the grinding force between a grinding wheel and a workpiece, the three-dimensional measuring data of the outline of the workpiece and wear degree of the grinding wheel and controlling feed speed and rotating speed of the grinding wheel through a control algorithm.

Description

A kind of robot grinding system based on sensor feedback

Technical field

The present invention relates to a kind of robot grinding system for grinding work piece, belong to robot grinding manufacture field.

Background technology

The range of application of grinding is widely used, and the quality of grinding often determines the final mass of product.Along with the extensive use of the materials such as special mild steel, aluminium alloy or nickel alloy in the fields such as Aero-Space, national defence, electric power, boats and ships, such as aero-engine, large-size steam turbine, gas turbine, marine propeller, blade of wind-driven generator etc. adopt above-mentioned material manufacture in a large number, propose challenging requirement to grinding-polishing technology.

Traditional large-scale workpiece grinding mainly contains several modes such as artificial grinding, special purpose machine tool grinding and Digit Control Machine Tool grinding.The mode workload of artificial grinding is large, efficiency is low, and the quality of work pieces process can not be guaranteed; Special purpose machine tool versatility is bad, is mainly used in the production process of parts in enormous quantities; Digit Control Machine Tool cost is higher.All do not have the process control for stock removal in these grinding process, grinding process mainly improves the roughness of workpiece.In addition, homogeneity and the quality thereof of the inefficiency of artificial grinding-polishing, grinding and polishing can not ensure, the bad environments such as the dust in grinding place are serious to the health hazard of workman.Therefore, develop automatic grinding polishing machine robot system and there is important application value.

Compared with Digit Control Machine Tool, grinding machine people has the higher flexibility workpiece of grinding different size, different size and unlike material (can), better versatility and adaptability (by researching and developing new software module, robot can be made in complex work environment to complete grinding task efficiently).Robot grinding technology becomes one of main development direction ensureing grinding quality, raising grinding efficiency at present.

Publication number is the Chinese invention patent (a kind of grinding process position-stance error automatic adjusting method and system) of CN101462255A, by position and the attitude information of automatic detection device on-line real-time measuremen workpiece, feed back to control system analysis, determine the Position and orientation parameters of workpiece, then by position and the attitude of executing agency's adjustment workpiece.Publication number is each stage that the patent of invention (robot grinding method based on robot learning) of CN101738981A works in abrasive band, grinding is carried out to the workpiece of unlike material, obtain the contact force of workpiece and emery wheel, the curvature of workpiece grinding face and stock removal, process velocity.Utilize initial data, adopt the method for machine learning, carry out the Dynamic Modeling of grinding system, and according to the measurement data of current working condition, set up current robot self-adapting power model, the grinding track of optimization robot.Publication number is that type grinding is repaiied in the automation that the patent of invention (a kind of method for shaping, grinding and processing abrasive band based on standard workpiece) of CN101462248 adopts the mode of sbrasive belt grinding to realize complex profile workpiece, the method generates each grinding parameter automatically according to the stock removal of each point of input, then according to these parameters, actual machining path is generated; Grinding unit adopts belt grinding machine, and it can accept the instruction of robot, completes type of the repairing grinding of workpiece.

Above-mentioned method mainly adopts off-line to set up grinding parameter, and generate robot grinding track, control completes according to grinding track and repaiies type grinding.In grinding process, rub the abrasion of grinding wheel caused, and vibrates the workpiece position error etc. caused, all can produce the grinding deviation of robot.If carry out repairing type grinding according to the grinding track of segregation reasons, type amount repaiied by the workpiece probably cannot removing setting.

Summary of the invention

(1) technical problem that will solve

Technical problem to be solved by this invention is that existing robot grinding system can not be revised the workpiece position error in grinding process and cause reaching the problem that type amount repaiied by predetermined workpiece

(2) technical scheme

Ask for solving above-mentioned technology, the present invention proposes a kind of robot grinding system, for carrying out grinding to workpiece, comprise Systematical control main frame, robot, the emery wheel of the spindle motor unit being installed on robot end and the end being installed on spindle motor unit, described Systematical control main frame is by controlling the rotation to emery wheel of spindle motor unit, and the translation of control machine robot end to emery wheel has carried out the grinding process to workpiece, described robot grinding system also comprises workpiece profile detecting unit, it is for detecting the profile of workpiece, produce workpiece profile data and flow to Systematical control main frame, described Systematical control main frame produces rotary speed data and feed speed data according to these workpiece profile data, is inputed to spindle motor unit and robot respectively, to control the grinding process of described emery wheel for workpiece.

(3) beneficial effect

Robot grinding system of the present invention, by adopting force snesor information feedback control robot and electro spindle, regulates velocity of rotation and the feed speed of emery wheel, can grinding dissimilar, the workpiece of complex profile.

Accompanying drawing explanation

Fig. 1 is the structural representation of an embodiment of the robot grinding system based on force feedback of the present invention;

Fig. 2 is arbitrfary point P on reconstruct surface of the work of the present invention _ithree-dimensional coordinate (x _i, y _i, z _i) schematic diagram;

Fig. 3 is the schematic diagram of grinding angle α;

Fig. 4 is of the present invention by m experimental calculation grinding COEFFICIENT K _tand K _nflow chart;

Fig. 5 is the control structure figure of the Systematical control main frame 6 of the robot grinding system based on force feedback of the present invention;

Figure 6 shows that the operational flowchart of grinding trajectory planning module 61 of the present invention;

Figure 7 shows that the operational flowchart of control of grinding force module 62 of the present invention.

Detailed description of the invention

For making the object, technical solutions and advantages of the present invention clearly understand, below in conjunction with specific embodiment, and with reference to accompanying drawing, the present invention is described in further detail.

Fig. 1 is the structural representation of an embodiment of the robot grinding system based on force feedback of the present invention.As shown in Figure 1, this system is used for carrying out grinding to workpiece 7, and system comprises robot 1, workpiece profile detecting unit 2, spindle motor unit 3, abrasion of grinding wheel detecting unit 4, grinding force detecting unit 5 and Systematical control main frame 6.

Wherein, robot 1 for installing spindle motor unit 3, and controls spindle motor unit 3 and carries out translation, to arrive the precalculated position of workpiece 7 being carried out to grinding.Robot 1 can adopt existing robot, the robot IRB4400 of such as ABB AB.

Workpiece profile detecting unit 2, for detecting the profile of workpiece, produces workpiece profile data and flows to Systematical control main frame 6.The present invention preferably adopts binocular camera (two cameras) to detect the reference position of surface of the work, and utilizes the two-dimensional workpiece image of two camera Real-time Collections, adopts three-dimensional reconstruction method to build three-dimensional workpiece profile data.As shown in Figure 1, workpiece profile detecting unit 2 comprises two cameras 21,22, constitutes binocular camera.

Spindle motor unit 3 comprises electro spindle 31, emery wheel 32, electro spindle controller 33 and electro spindle driver 34; Electro spindle 31 is installed on the end of work robot 1, and emery wheel 32 is installed on the end of electro spindle 31.Electro spindle controller 33, for receiving the rotary speed data sent by Systematical control main frame 6, is delivered to electro spindle driver after being converted into drive singal; Electro spindle driver 34 controls electro spindle 31 according to this drive singal and reaches predetermined velocity of rotation.Thus, spindle motor unit 3 can drive emery wheel 32 pairs of workpiece 7 to carry out grinding under the control of Systematical control main frame 6.Electro spindle 31 can adopt HCS180-24000/41 type electro spindle.Electro spindle controller 33 can adopt LNCT520i spindle controller, and electro spindle driver 34 can adopt the raw EV6000 frequency converter of Ai Mo.

Abrasion of grinding wheel detecting unit 4, for detecting the degree of wear of emery wheel 32, obtains abrasion of grinding wheel level data and is transferred to Systematical control main frame 6.Abrasion of grinding wheel detecting unit 4 can adopt laser vision sensor 41.

Grinding force detecting unit 5, for detecting the grinding force between workpiece 7 and emery wheel 32, produces grinding force data and flows to Systematical control main frame 6.Grinding force detecting unit 5 comprises six-dimension force sensor 51 and data acquisition module 52.Six-dimension force sensor 41 is arranged on the end of robot 1, for detecting the grinding force in grinding process between workpiece 7 and emery wheel 32.Data acquisition module 52, for gathering tangential component and the normal component of grinding force, produces grinding force data and is input in Systematical control main frame 6.

Systematical control main frame 6 has carried out the grinding process to workpiece 7 for the rotation and robot 1 translation of end to emery wheel 32 controlling spindle motor unit 3 pairs of emery wheels 32.Specifically, Systematical control main frame 6 produces rotary speed data and feed speed data according to workpiece profile data, abrasion of grinding wheel level data and grinding force data, inputed to electro spindle controller 33 and the robot 1 of spindle motor unit 3 respectively, to control the grinding process of emery wheel 32 for workpiece 7.

Specifically, Systematical control main frame comprises three inputs and two outputs.As shown in Figure 1, input 641 connects the output of sand workpiece profile detecting unit 2; Input 642 connects the output of abrasion of grinding wheel detecting unit 4; Input 643 connects the output of grinding force detecting unit 5.Output 651 connects the input of robot 1; Output 653 connects the input of spindle motor unit 3.The output 651 of the input connected system main control system 6 of robot 1, for the feed speed data that receiving system main control system 6 sends, thus drives emery wheel translation.

In addition, Systematical control main frame 6 comprises grinding trajectory planning module 61, Robot calibration module 62 and control of grinding force module 63.Function and the operation of modules will describe in detail hereinafter.

Reconstruct workpiece profile

The present invention is preferably and adopts binocular camera as workpiece profile detecting unit 3, with the 3 d measurement data of three-dimensionalreconstruction workpiece profile as previously mentioned.Illustrate the seat calibration method of arbitrfary point Pi on three-dimensionalreconstruction surface of the work below:

Fig. 2 is arbitrfary point P on reconstruct surface of the work _ithree-dimensional coordinate (x _i, y _i, z _i) schematic diagram.As shown in Figure 2, the present invention adopts the method for different surface beeline to solve P _icoordinate.Arbitrfary point P on three-dimensionalreconstruction surface of the work _ithe job step of coordinate as follows:

Steps A 1: the two-dimensional pixel coordinate (u setting up camera 21 ^l, v ^l) and three dimensional space coordinate (x ^l, y ^l, z ^l) between relation, shown in (1):

$\left[\begin{array}{c}{\mathrm{su}}^L\\ {\mathrm{sv}}^L\\ s^L\end{array}\right]=\left[\begin{array}{cccc}{m}_{11}^L& m_{12}^L& m_{13}^L& m_{14}^L\\ m_{21}^L& m_{22}^L& m_{23}^L& m_{24}^L\\ m_{31}^L& m_{32}^L& m_{33}^L& m_{34}^L\end{array}\right]\left[\begin{array}{c}x\\ y\\ z\\ 1\end{array}\right]---\left(1\right)$

Parameter m in formula (1) _ij(i=1,2,3; J=1,2,3,4) can obtain in camera calibration process, s is proportionality coefficient.

Steps A 2: the two-dimensional pixel coordinate (u setting up camera 22 ^r, v ^r) and three dimensional space coordinate (x ^r, y ^r, z ^r) between relation, shown in (2):

$\left[\begin{array}{c}{\mathrm{su}}^R\\ {\mathrm{sv}}^R\\ s^R\end{array}\right]=\left[\begin{array}{cccc}{m}_{11}^R& m_{12}^R& m_{13}^R& m_{14}^R\\ m_{21}^R& m_{22}^R& m_{23}^R& m_{24}^R\\ m_{31}^R& m_{32}^R& m_{33}^R& m_{34}^R\end{array}\right]\left[\begin{array}{c}x\\ y\\ z\\ 1\end{array}\right]---\left(2\right)$

Steps A 3: camera 21 and 22 is collection point P simultaneously _iimage.In the image of camera 21 and in the image of camera 22, identify respectively and calculate a P _iimage coordinate in two images: be (u in the image coordinate of camera 21 _i ^r, v _i ^r), be (u in the image coordinate of camera 22 _i ^l, v _i ^l).

Steps A 4: by coordinate [(u _i ^r, v _i ^r), (u _i ^l, v _i ^l)] substitute in formula (1) and formula (2), elimination factor s, obtains about a P _icoordinate (x _i, y _i, z _i) equation.This equation solved is such as formula shown in (3):

$\left\{\begin{array}{c}(u_i^L{m}_{31}^L-m_{11}^L)x_i+(u_i^L{m}_{32}^L-m_{12}^L)y_i+(u_i^L{m}_{33}^L-m_{13}^L)z_i=m_{14}^L-u_i^L{m}_{34}^L\\ (v_i^L{m}_{31}^L-m_{21}^L)x_i+(v_i^L{m}_{32}^L-m_{22}^L)y_i+(v_i^L{m}_{33}^L-m_{23}^L)z_i=m_{24}^L-v_i^L{m}_{34}^L\\ (u_i^R{m}_{31}^R-m_{11}^R)x_i+(u_i^R{m}_{32}^R-m_{12}^R)y_i+(u_i^R{m}_{33}^R-m_{13}^R)z_i=m_{14}^R-u_i^R{m}_{34}^R\\ (v_i^R{m}_{31}^R-m_{21}^R)x_i+(v_i^R{m}_{32}^R-m_{22}^R)y_i+(v_i^R{m}_{33}^R-m_{23}^R)z_i=m_{24}^R-v_i^R{m}_{34}^R\end{array}\right.---\left(3\right)$

Steps A 5: incite somebody to action formula simultaneous solution equation group above, obtain a P _ithree-dimensional coordinate (x _i, y _i, z _i).In order to reduce the error solved, utilizing least square method to solve transfiniting of simultaneous above and determining homogeneous equation group, obtain a little 23, namely put P _ithree-dimensional coordinate (x _i, y _i, z _i).

Under ideal model, the straight line 211 that formula (3) defines and straight line 221 meet at 1: 23.If when the model of camera 21 and 22 is not complete perspective model, and there is noise during imaging, straight line 211 and straight line 221 are different surface beelines, thus can not meet at a bit.Thus, we also can use the mid point 24 of different surface beeline common vertical line to approach P _i.

Detect abrasion of grinding wheel degree

As previously mentioned, abrasion of grinding wheel detecting unit 3 adopts laser vision sensor to detect abrasion of grinding wheel degree.In grinding process, every the set time, utilize abrasion of grinding wheel detecting unit 3 to detect the degree of wear of emery wheel 32, build the real-time characteristic profile of wheel face.

Adopt laser vision sensor to detect abrasion of grinding wheel degree, first utilize laser sensor to detect the surface of emery wheel in real time, by three-dimensional reconstruction method, the feature of real-time reconstruction wheel face.The real-time characteristic profile of reconstruction and the feature contour preset are compared, calculates the wear extent of emery wheel 32, produce abrasion of grinding wheel level data.Laser vision sensor comprises binocular camera and line structure light source.Adopt line-structured light triangulation to rebuild the real-time characteristic profile of wheel face, key step comprises:

Step B1: irradiate testee with line-structured light.When the reference stripe of structured light projects testee surface, because body surface is uneven, striped there occurs distortion, and this distortion contains the three-dimensional information of object surface shape.

Step B2: taken by the subject image of the different parts of structured light by binocular camera.According to the Mathematical Modeling of corresponding relation between distortion striped and object surface shape, go out the three-dimensional information of object surface shape from the shape of stripes information inference after distortion.

Detect grinding force

Grinding force detecting unit 5, by being arranged on the six-dimension force sensor 41 of robot 1 end, detecting the grinding force between workpiece 7 and emery wheel 32 in grinding process, obtains grinding force data.By detecting the tangential component F of the grinding force obtained _twith normal component F _n, Systematical control main frame 6, by controlling rotation and the translation of spindle motor unit 3 and industrial machine 1, regulates velocity of rotation and the feed speed of emery wheel 32, to be removed the unnecessary part on workpiece 7 surface by grinding process.The tangential component F of the grinding force of grinding force _twith normal component F _nwith the relation of the velocity of rotation of emery wheel, feed speed and grinding removal amount as shown in the formula expression:

On surface of the work, the tangential grinding force of unit width is:

$F_t=K_t\frac{V_W}{V_C}\left(\frac{1}{2}{\sin }^2\left(\mathrm{\α}\right)\right)---\left(1\right)$

On surface of the work, the normal grinding force of unit width is:

$F_n=K_n\frac{V_W}{V_C}\left(\frac{1}{2}{\sin }^2\left(\mathrm{\α}\right)\right)---\left(2\right)$

In formula, V _cthe velocity of rotation of emery wheel, V _wit is the feed speed of emery wheel; K _tand K _nbe grinding coefficient, can calculate by experiment; α is used to the grinding angle representing grinding removal amount SG.Be illustrated in figure 3 the schematic diagram of grinding angle α.

Fig. 4 is by m experimental calculation grinding COEFFICIENT K _tand K _nstep:

In steps A 1: the i-th experiment, the velocity of rotation that Systematical control main frame 6 controls emery wheel is V _ci, feed speed is V _wi, grinding force detecting unit 5 measures the tangential component F of grinding force by six-dimension force sensor 41 _tiwith normal component F _ni;

Steps A 2: measure grinding removal amount SG by workpiece profile detecting unit 2 _i.

Steps A 3: repeat, in steps A 1 and A2, to finish m experiment, obtain training data set { (V _c1, V _w1, F _t1, F _n1, α ₁) ..., (V _cm, V _wm, F _tm, F _nm, α _m).

Steps A 4: based on m group training data, calculates optimum grinding COEFFICIENT K _tand K _n, computing formula is:

$K_t=\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{F}_{\mathrm{ti}}\frac{{2V}_{\mathrm{Ci}}}{V_{\mathrm{Wi}}{\sin }^2\left({\mathrm{\α}}_i\right)}$

$K_n=\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{F}_{\mathrm{ni}}\frac{{2V}_{\mathrm{Ci}}}{V_{\mathrm{Wi}}{\sin }^2\left({\mathrm{\α}}_i\right)}$

Systematical control

Be illustrated in figure 5 the control structure figure of Systematical control main frame 6.As shown in Figure 5, Systematical control main frame comprises grinding trajectory planning module 61, Robot calibration module 62 and control of grinding force module 63.Wherein, grinding trajectory planning module 61, for the design data according to the workpiece profile data measured in real time and workpiece profile, obtains the grinding track of surface of the work, then plans grinding motion track; Grinding motion track for setting up the coordinate conversion relation between robot coordinate system and workpiece coordinate system, and is converted to the end movement track of robot by Robot calibration module 62; Control of grinding force module 63 is for according to described grinding force data, abrasion of grinding wheel data, workpiece profile data and location of workpiece data, calculate feed speed and the velocity of rotation of emery wheel, and velocity of rotation data are transferred in described spindle motor unit (3), described feed speed is transferred in robot (1).

Be illustrated in figure 7 the algorithm flow of grinding trajectory planning module 61:

Step D1: input workpiece profile from workpiece profile detecting unit 2 and detect data acquisition system { SR _i(i=1,2 ..., n; N is the quantity of data).

Step D2: store workpiece profile design data set { SI in Systematical control main frame 6 in advance _i(i=1,2 ..., n; N is the quantity of data).

Step D3:{SR _iand { SI _idata carry out man-to-man comparison, difference is exactly the removal amount set { SG of surface of the work _i(i=1,2 ..., n; N is the quantity of data).

Step D4: according to removal amount set { SG _iand the best grinding force of emery wheel interval, calculate back and forth back and forth grinding number of times t, and the secondary removal amount set { SG of jth (1≤j≤t) _ij.T is for meeting t _min≤ t≤t _maxan integer, wherein

$t_{\min }=\arcsin \sqrt{2\·\frac{\min \left(F_t\right)\min \left(V_C\right)}{K_t\max \left(V_W\right)}}$

$t_{\max }=\arcsin \sqrt{2\·\frac{\max \left(F_t\right)\max \left(V_C\right)}{K_t\min \left(V_W\right)}}$

Min (F _t) and max (F _t) be the minimum tangential grinding force in the best grinding force interval of emery wheel and maximum tangential grinding force.Min (V _c) and max (V _c) be the minimum rotation speed of emery wheel and maximum rotational speed.Min (V _w) and max (V _w) be the minimum feed speed of emery wheel and maximum feed speed.

Step D5: select grinding track: if the surface of workpiece is plane, then adopt the grinding track of cycloidal type; If surface of the work is surface of revolution, then adopt the grinding track of fractal scanning formula; If surface of the work is other complex profile, then adopt scan line formula grinding track.

The job step of Robot calibration module 62 describes as follows:

Step e 1: place tactile sensing element and optical sensing element on workpiece, and robot end install touching or stimulating apparatus to determine the position relationship between robot end and workpiece;

Step e 2: obtain the coordinate conversion relation between robot coordinate system and workpiece coordinate system.

Step e 3: according to the coordinate conversion relation between robot coordinate system and workpiece coordinate system, is converted to the end movement track of robot by the grinding track obtained in grinding trajectory planning module 61.

Be illustrated in figure 7 the algorithm flow in control of grinding force module 63:

Step F 1: in the contactless constraint space of emery wheel 32 and workpiece 7, namely when emery wheel 32 does not contact workpiece 7, in order to make up because abrasion of grinding wheel, trade union college position deviation cause the position deviation of robot 1 end orbit and surface of the work, online adaptive Compensating Robot 1 end orbit, enables emery wheel keep in touch with surface of the work.

Step F 2: in the contiguity constraint space of emery wheel 32 and workpiece 7, namely when emery wheel 32 grinding work piece 7, based on the grinding force data (feedback signal of six-dimension force sensor 51) that grinding force detecting unit 5 inputs, the emery wheel 32 that the end of control 1 is installed and the contact force of workpiece 7, concrete steps comprise:

Step F 21: according to the secondary removal amount set { SG of jth (1≤j≤t) _i} _j, and the current velocity of rotation of emery wheel and feed speed, calculated in jth time grinding process by formula (1) and formula (2), expect tangential grinding force and expect normal grinding force.

Step F 22: according to actual tangential grinding force and the difference expecting tangential grinding force, the difference of actual normal grinding force and expectation normal grinding force, adoption rate-derivative-integral (PID) control algolithm, regulates the velocity of rotation V of emery wheel _cwith feed speed V _w.

Step F 23: repeat step F 21 and step F 22, until the number of times t completing needs back and forth grinding back and forth.

Above-described specific embodiment; object of the present invention, technical scheme and beneficial effect are further described; be understood that; the foregoing is only specific embodiments of the invention; be not limited to the present invention; within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (9)

1. a robot grinding system, for carrying out grinding to workpiece (7), comprise Systematical control main frame (6), robot (1), be installed on the spindle motor unit (3) of robot (1) end and be installed on the emery wheel (32) of end of spindle motor unit (3), described Systematical control main frame (6) is by controlling the rotation to emery wheel (32) of spindle motor unit (3), and the translation of control (1) end to emery wheel (32) has come, to the grinding process of workpiece (7), to it is characterized in that:

Described robot grinding system also comprises workpiece profile detecting unit (2) and abrasion of grinding wheel detecting unit (4), wherein,

Described workpiece profile detecting unit (2), for detecting the profile of workpiece (7), produces workpiece profile data and flows to Systematical control main frame (6);

Described abrasion of grinding wheel detecting unit (4) is for detecting the degree of wear of emery wheel (32), the real-time characteristic profile of wheel face is rebuild by the degree of wear, this real-time characteristic profile and the feature contour preset are compared, calculate the wear extent of emery wheel (32), produce abrasion of grinding wheel level data and be transferred to Systematical control main frame (6);

Described Systematical control main frame (6) produces rotary speed data and feed speed data according to these workpiece profile data, inputed to spindle motor unit (3) and robot (1) respectively, to control the grinding process of described emery wheel (32) for workpiece (7), meanwhile, rotary speed data and feed speed data are produced according to this abrasion of grinding wheel level data.

2. robot grinding system as claimed in claim 1, it is characterized in that, described workpiece profile detecting unit (2) comprises two cameras (21,22), the two-dimensional workpiece image of this two cameras (21,22) Real-time Collection, adopts three-dimensional reconstruction method to build three-dimensional workpiece profile data.

3. robot grinding system as claimed in claim 1, it is characterized in that, described abrasion of grinding wheel detecting unit (4) is laser vision sensor, and it comprises binocular camera and line structure light source, and rebuilds the real-time characteristic profile of wheel face by line-structured light triangulation.

4. robot grinding system as claimed in claim 1, it is characterized in that, also comprise grinding force detecting unit (5), it is for detecting the grinding force between workpiece described in described grinding process (7) and described emery wheel (32), produces grinding force data and flows to Systematical control main frame (6);

Described Systematical control main frame (6) also produces rotary speed data and feed speed data according to these grinding force data.

5. robot grinding system as claimed in claim 4, it is characterized in that, described grinding force detecting unit (5) comprises six-dimension force sensor (51) and data acquisition module (52),

Described six-dimension force sensor (41) is arranged on the end of described robot (1), for detecting the grinding force between workpiece in grinding process (7) and emery wheel (32);

Described data acquisition module (52), for gathering tangential component and the normal component of grinding force, produces grinding force data and is input in Systematical control main frame (6).

6. robot grinding system as claimed in claim 4, it is characterized in that, described Systematical control main frame (6) comprises grinding trajectory planning module (61), Robot calibration module (62) and control of grinding force module (63), wherein

Described grinding trajectory planning module (61), for the design data according to the workpiece profile data measured in real time and workpiece profile, obtains the grinding track of surface of the work, planning grinding motion track;

Grinding motion track for setting up the coordinate conversion relation between robot coordinate system and workpiece coordinate system, and is converted to the end movement track of robot by described Robot calibration module (62);

Described control of grinding force module (63) is for according to described grinding force data, abrasion of grinding wheel data, workpiece profile data and location of workpiece data, calculate feed speed and the velocity of rotation of emery wheel, and velocity of rotation data are transferred in described spindle motor unit (3), described feed speed is transferred in robot (1).

7. robot grinding system as claimed in claim 6, it is characterized in that, the algorithm flow of grinding trajectory planning module (61) is:

Step D1: detect data acquisition system { SR from workpiece profile detecting unit (2) input workpiece profile _i, wherein i=1,2 ..., n, n are the quantity of data;

Step D2: store workpiece profile design data set { SI in Systematical control main frame (6) in advance _i, wherein i=1,2 ..., n, n are the quantity of data;

Step D3:{SR _iand { SI _idata carry out man-to-man comparison, calculate its difference, this difference is exactly the removal amount set { SG of surface of the work _i, wherein, i=1,2 ..., n, n are the quantity of data;

Step D4: according to removal amount set { SG _iand the best grinding force of emery wheel interval, calculate back and forth back and forth grinding number of times t, and jth time removal amount set { SG _i} _j, t is for meeting t _min≤ t≤t _maxan integer, wherein 1≤j≤t,

$t_{\min }=\arcsin \sqrt{2\·\frac{\min \left(F_t\right)\min \left(V_C\right)}{K_t\max \left(V_W\right)}}$

$t_{\max }=\arcsin \sqrt{2\·\frac{\max \left(F_t\right)\max \left(V_C\right)}{K_t\min \left(V_W\right)}},$

Min (F _t) and max (F _t) be the minimum tangential grinding force in the best grinding force interval of emery wheel and maximum tangential grinding force, min (V _c) and max (V _c) be the minimum rotation speed of emery wheel and maximum rotational speed, min (V _w) and max (V _w) be the minimum feed speed of emery wheel and maximum feed speed;

Step D5: select grinding track: if the surface of workpiece is plane, then adopt the grinding track of cycloidal type; If surface of the work is surface of revolution, then adopt the grinding track of fractal scanning formula; If surface of the work is other complex profile, then adopt scan line formula grinding track.

8. robot grinding system as claimed in claim 6, it is characterized in that, the algorithm flow in described control of grinding force module (63) is:

Step F 1: in the contactless constraint space of emery wheel (32) and workpiece (7), namely when emery wheel (32) does not contact workpiece (7), in order to make up because abrasion of grinding wheel, trade union college position deviation cause the position deviation of robot (1) end orbit and surface of the work, online adaptive Compensating Robot (1) end orbit, enables emery wheel keep in touch with surface of the work;

Step F 2: in the contiguity constraint space of emery wheel (32) and workpiece (7), namely when emery wheel (32) grinding work piece (7), based on the grinding force data that grinding force detecting unit (5) inputs, the contact force of the emery wheel (32) that the end of control (1) is installed and workpiece (7).

9. robot grinding system as claimed in claim 8, it is characterized in that, described step F 2 comprises:

Step F 21: according to jth time removal amount set { SG _i} _j, and the current velocity of rotation of emery wheel and feed speed, calculate in jth time grinding process, expect tangential grinding force and expect normal grinding force, wherein 1≤j≤t;

Step F 22: according to actual tangential grinding force and the difference expecting tangential grinding force, the difference of actual normal grinding force and expectation normal grinding force, adoption rate-derivative-integral (PID) control algolithm, regulates the velocity of rotation V of emery wheel _cwith feed speed V _w;

Step F 23: repeat step F 21 and step F 22, until the number of times t completing needs back and forth grinding back and forth.