CN101524842A - Industrial robot calibration method - Google Patents

Abstract

The invention discloses an industrial robot calibration method. Joints of an industrial robot can move on axes J1, J2, J3, J4, J5 and J6. The method comprises the following steps: taking an initial test point C0 as a reference point, and establishing an initial point-based test coordinate system {C} with a laser tracker; determining three points with included angles over 30 degrees between the axes J1, J2, J3, J4, J5 and J6 and a lower axis respectively in the {C}, tracking curve arc with the laser tracker, and recording position coordinates measured from clockwise direction and counterclockwise direction; calculating the included angles and distances between the axes, and comparing the included angles and the distances with theoretical parameters of robot to realize calibration. The invention overcomes disadvantages of the prior art (needing manual programming and low video camera tracking accuracy), and provides the industrial robot calibration method which has high accuracy, simple steps and convenient use and can be used widely in the automatic equipment field.

Description

Industrial robot calibration method

Technical field

The present invention relates to a kind of calibration steps of equipment, be specifically related to the calibration steps of robot.

Background technology

Industrial robot, be divided into general and special-purpose, refer generally to be used for machinery manufacturing industry and replace the people to finish having work in enormous quantities, the high-quality requirement, as spot welding, arc-welding in the industry automatic production lines such as automobile making, motorcycle manufacturing, naval vessel manufacturing, some household appliances (television set, refrigerator, washing machine), chemical industry, spray paint, the robot of operations such as the carrying of cutting, electronic assemblies and logistics system, packing, piling.Industrial robot is realized the full automation of production process easily, improves production efficiency of products and quality, but industrial robot in the space accurately the location be the basic demand that it can normal reliable work.All need its reference point is imitated standard when generally speaking, industrial robot dispatches from the factory or when changing heavy mechanical equipment.In on August 13rd, 2008 disclosed CN200810008968.3 patent application document, disclose a kind of calibration steps of robot.This method need be carried out manual programming, adopts the tracking of video camera, and precision is not high.

Summary of the invention

The present invention has overcome the deficiencies in the prior art, and a kind of high-precision industrial robot calibration method is provided.。

For solving above-mentioned technical problem, the present invention by the following technical solutions:

Industrial robot calibration method, industrial robot joint can move on J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, a J6 axle, uses laser tracker system to measure, and carries out according to following flow process:

Step 1 respectively with J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, J6 axle each, keep other zero-bit attitudes, from negative spacing move to just spacing, again by just spacing move to negative spacing;

Step 2 is X under angular coordinate, Y, and each coordinate of Z, robot begin with the arm end by zero-bit to serve as the control object, walk along rectangular co-ordinate that stroke is greater than 1 meter, move negative direction by rectangular co-ordinate to positive direction, return losing side;

It a bit is the test initial point that step 3 is set, and is designated as C0; With C0 is reference point, with the test coordinate system { C} of laser tracker foundation based on initial point;

Step 4 respectively with J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, J6 axle each, keep other zero-bit attitudes, from negative spacing move to just spacing, again by just spacing move to negative spacing, with laser tracker test machine robot end ring flange center at the test coordinate system { coordinate among the C};

Step 5 is { among the C}, respectively with J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, J6 axle each, to keep other zero-bit attitudes, from negative spacing move to just spacing, negative spacingly between its axle and lower shaft, determine 3 points by just spacing moving to again, angle between 3 is followed the tracks of its curve circular arc greater than 30 degree with laser tracker, notes the position coordinates that clockwise direction records and counterclockwise records to the position coordinates that records;

Step 6 calculates J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, each shaft centre line of J6 axle in the { direction among the C} with the data input that measures;

Step 7 is calculated J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, each rotary middle point of J6 axle;

Step 8 is calculated the mean value of each central point, obtains J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, each shaft centre line of J6 axle in the { position among the C};

Angle and distance in step 9 computer memory between J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, each axis of J6 axle;

The data that step 10 obtains above processing and the theoretical parameter value of robot compare, for calibration is done by robot.

Further technical scheme is that step 3 repeatedly repeats repeatedly to calibrate to step 10.

Further technical scheme is that step 6 to step 9 is calculated with the MATHLAB mathematical software.

Compared with prior art, the invention has the beneficial effects as follows the measuring accuracy height, step is simple, can easily finish the calibration of industrial robot.

The specific embodiment

In the calculating of the present invention, adopt and calculate as drag.

Model one takes following method to determine that any one rotation central axis line of robot is in { the direction among the C}.

Model is set up: with the J1 axle is example: measure data DA1=[X1 Y1 Z1]; Because X1=[x1 x2x3 ... xn] '; Y1=[y1 y2 y3 ... yn] '; Z1=[z1 z2 z3 ... zn] '; In DA1, appoint and get three line data as unit sample.If existing: SDA1=[L1 L2 L3] '; Wherein L1=[x1 y1z 1]; L2=[x2y2z2]; L3=[x3y3z3];

The conformational space vector:

L2L1＝OL1-OL2＝[(x1-x2)(y1-y2)(z1-z2)]；

L2L3＝OL3-OL2＝[(x3-x2)(y3-y2)(z3-z2)]；

Ask the direction of vertical two vectors:

ADIR＝L2L1×L2L3＝|i j k |

|(x1-x2)(y1-y2)(z1-z2)|

|(x3-x2)(y3-y2)(z3-z2)|；

＝adiri+adirj+adirk；

Model two takes following method to find out each rotary middle point.

Ask the central point that revolves:

Model is set up: with the J1 axle is example: utilize the data L2L1 in the model one, L2L3 asks the mid point M21 of two vectors, M23;

M21＝(20L1+OL2)/2＝(mx21?my21?mz21)；M23＝(20L3+OL2)/2＝(mx23?my23?mz23)；

Ask the overwinding central point and perpendicular to the plane of revolving:

Known plane normal direction is L2L1, L2L3 and mistake M21, and the M23 point is so plane P 21 is

(x1-x2)(x-mx21)+(y1-y2)(y-my21)+(z1-z2)(z-mz21)＝0；

Plane p23 is: (x3-x2) (x-mx23)+(y3-y2) (y-my23)+(z3-z2) (z-mz23)=0;

Ask the intersection and two on two planes to revolve the intersection point on plane, place:

Solving equation group: (x1-x2) (x-mx21)+(y1-y2) (y-my21)+(z1-z2) (z-mz21)=0;

(x3-x2)(x-mx23)+(y3-y2)(y-my23)+(z3-z2)(z-mz23)＝0；

adiri(x-x1)+adirj(y-y1)+adirk(z-z1)＝0；

Obtain AR (x y z);

Model three: take following method to ask the mean value of each central point, obtain each shaft centre line in the { position of C}.

Model is set up: the result of combination model one and model two obtains the direction and the position of rotation central axis line.

DIR=[ADIR1 ADIR2 ADIR3 is gone in data statistics ... ] ';

R＝[AR1AR2AR3…]’；

Average by row: with attached DIRM, the RM of giving of data;

Each rotation that model four takes following method to utilize above model to try to achieve, push away angle and the distance between each axis in the space.

Model is set up: ask the angle between two axial lines

Normal line vector: s1=(m1, n1, p1); S2=(m2, n2, p2);

Angle: $\cos \mathrm{\θ}\frac{|m_1{m}_2+n_1{n}_2+p_1{p}_2|}{\sqrt{{m_1}^2+{n_1}^2+{p_1}^2}g\sqrt{{m_2}^2+{n_2}^2+{p_2}^2}};$

If: cos θ=0 explanation two axial lines is vertical, and cos θ=1 o'clock explanation two axial lines is parallel;

Calculate the distance of two axial lines in the space

Known two axial lines L1, the direction vector s1 of L2, s2 and a process point d1 (x1 y1 z1), d2 (x2 y2);

When cos θ=0, be plane Z1 through L1, and normal is s2; Z1:s2* ([x y z]-d1) '=0;

Be plane Z2 through L2, and normal is s1; Z2:s1* ([x y z]-d2) '=0;

Find intersection po1: simultaneous solution L1, Z2;

Find intersection po2: simultaneous solution L2, Z1;

Ask the distance D 12=|po2-po1| between po1 and po2;

Checking $\frac{{\mathrm{po}}_1{\mathrm{po}}_2}{\left|{\mathrm{po}}_1{\mathrm{po}}_2\right|}=s_1\×s_2;$

When cos θ=1, be plane Z1 through d1, and normal is s1; Z1=s1* ([x y z]-d1);

Find intersection po1: simultaneous solution L1, Z1;

Find intersection po2: simultaneous solution L2, Z1;

Ask the distance D 12=|po2-po1| between po1 and po2;

Checking $\frac{{\mathrm{po}}_1{\mathrm{po}}_2}{\left|{\mathrm{po}}_1{\mathrm{po}}_2\right|}\×s_1=\frac{{\mathrm{po}}_1{\mathrm{po}}_2}{\left|{\mathrm{po}}_1{\mathrm{po}}_2\right|}\×s_2;$

When 0＜cos θ＜1, be L1, it is n that L2 does multiplication cross.Get sn1 with s1 and n multiplication cross again, doing normal direction is the plane Z1 of sn11 and process L1.On L1, look for 1 d1;

Both Z1:s1 * s2 * s1 ([x y z]-d1) '=0; In like manner get sn2 with s2 and n multiplication cross, doing normal direction is the plane Z2 of sn2 and process L2.On L2, look for 1 d2;

Both Z2:s1 * s2 * s2 ([x y z]-d2) '=0;

Find intersection po1: simultaneous solution L1, Z2;

Find intersection po2: simultaneous solution L2, Z1;

Ask the distance D 12=|po2-po1|| between po1 and po2;

Model five: take following method when the special data situation occurring, can subsidiary go out the approximation of robot connecting rod distance.

Model is set up: after finishing model 1 ~ 3, each normal and rotary middle point are done down record.When measuring axis spacing when being 0, need to consider another kind of method.

At first be plane ZJ1, do multiplication cross, obtain with J1 normal multiplication cross again: N=SJ1 * SJ2 * SJ1 with the normal of J1 axle and the normal of J2 axle; Cross the J1 axis and do the plane that normal is N.

ZJ1：N*([x?y?z]-DJ1)’＝0；

Axis devious and ZJ1 plane simultaneous were asked the intersection point of face, and wherein the distance between the intersection point can be approximately the robot length of connecting rod.

Claims (3)

1, industrial robot calibration method, industrial robot joint can move on J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, a J6 axle, uses laser tracker system to measure, and it is characterized in that according to following flow process:

Step 1 respectively with J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, J6 axle each, keep other zero-bit attitudes, from negative spacing move to just spacing, again by just spacing move to negative spacing;

Step 2 is X under angular coordinate, Y, and each coordinate of Z, robot begin with the arm end by zero-bit to serve as the control object, walk along rectangular co-ordinate that stroke is greater than 1 meter, move negative direction by rectangular co-ordinate to positive direction, return losing side;

It a bit is the test initial point that step 3 is set, and is designated as C0; With C0 is reference point, with the test coordinate system { C} of laser tracker foundation based on initial point;

Step 4 respectively with J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, J6 axle each, keep other zero-bit attitudes, from negative spacing move to just spacing, again by just spacing move to negative spacing, with laser tracker test machine robot end ring flange center at the test coordinate system { coordinate among the C};

Step 5 is { among the C}, respectively with J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, J6 axle each, to keep other zero-bit attitudes, from negative spacing move to just spacing, move to negative spacingly by just spacing again, determine 3 points between its axle and lower shaft, the angle between 3 is greater than 30 degree, follow the tracks of its curve circular arc with laser tracker, note the position coordinates that clockwise direction records and counterclockwise record to the position coordinates that records;

Step 6 calculates J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, each shaft centre line of J6 axle in the { direction among the C} with the data input that measures;

Step 7 is calculated J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, each rotary middle point of J6 axle;

Step 8 is calculated the mean value of each central point, obtains J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, each shaft centre line of J6 axle in the { position among the C};

Angle and distance in step 9 computer memory between J1 axle, J2 axle, J3 axle, J4 axle, J5 axle, each axis of J6 axle;

The data that step 10 obtains above processing and the theoretical parameter value of robot compare, for calibration is done by robot.

2, industrial robot calibration method according to claim 1 is characterized in that described step 3 repeatedly repeats repeatedly to calibrate to step 10.

3, industrial robot calibration method according to claim 2 is characterized in that described step 6 to step 9 calculates with the MATHLAB mathematical software.