CN105081524B - In welding process, the online dynamic programming of track follows the tracks of collaborative control method with welding bead - Google Patents

Abstract

The invention provides the online dynamic programming of track in a kind of welding process and follow the tracks of collaborative control method with welding bead, belong to mobile welding robot technical field.The present invention abundant integrated structure light sensors region is ahead of electric arc and realizes the detection of front, weld zone seam welds groove and arc sensor obtains the advantage of welding torch pose in real time, realizes, based on welding bead Isometric Approximation thought, the collaborative uneoupled control that the online dynamic programming of robot body track is followed the tracks of with welding bead.The present invention can improve mobile welding robot speed of welding stability in large-scale component complicated track weld seam welding process and welding bead tracking accuracy, improve joint trajectory local deep camber and the quality of weld seam molding in knuckle region, can be applicable to, in the mobile robot welding process of the equipment manufacturing such as boats and ships, the energy and track traffic, be particularly suited for the welding occasion of large turn or deep camber curved welding seam.

Description

In welding process, the online dynamic programming of track follows the tracks of collaborative control method with welding bead

Technical field

The invention belongs to mobile (formula) Technology of Welding Robot field.Relate to the online dynamic programming of track in a kind of welding process Follow the tracks of collaborative control method with welding bead, can be widely applied to the aspects such as mobile robot automatic welding.

Background technology

Large-scale component welding occasion in the fields such as resource equipment, heavy-duty machinery, shipbuilding, mobile welding robot is because of it The advantages such as grand movement ability become the effective ways of the space curve welding seam automatic welding solving large scale, complexity.Pacified It is loaded on airborne actuator devices on the robot body of mobile welding robot (such as mechanical arm, rectangular coordinate system Formula X-Y Balladeur train etc.) impulse stroke limited, need to combine robot body and adjust its position in real time with attitude to ensure welding torch centering weldering Seam center, i.e. carries out weld job under the robot body Collaborative Control with airborne actuator.

Existing mobile welding robot mainly uses the weld seam tracking sensor of single type, passes including arc sensor, vision Sensor two class: use the mobile welding robot of arc sensor only detect molten bath zone and therefrom obtain welding deviation information, machine Device human body adjusts position according to the actual correction behavior of welding deviation information or airborne actuator in real time that arc sensor obtains Appearance, it is desirable to airborne actuator has bigger effective impulse stroke；And, the pursuit movement of welding bead is easily subject to by robot body The electromechanical inertial impact of mobile welding robot system causes controlling delayed, particularly in deep camber region, curved welding seam track local Mobile welding robot is needed to perform the wide-angle turning motion region with tracking seam track with knuckle region etc., because being difficult in time Adjust pose, easily produce and partially weld defect, or because speed of welding change causes appearance of weld bad；Use the shifting of vision sensor Dynamic welding robot disturbs in order to avoid strong arc light, and generally and electric arc melting bath region exists a spacing in vision sensor detection region From, it is difficult to real-Time Compensation causes the change of position while welding and size because of welding base metal temperature distortion.

In sum, prior art is all difficult to take into account mobile welding robot in the complicated track weld seam welding process of large scale Dynamic response and static accuracy, and be difficult to solve the technology that couples with airborne actuator motor control of robot body motion and ask Topic.

Summary of the invention

It is an object of the invention to overcome the deficiencies in the prior art, it is proposed that in a kind of welding process the online dynamic programming of track with Welding bead follows the tracks of collaborative control method, in order to ensure while realizing the welding bead accurate tracking in the welding of large scale complicated track weld seam Quality of weld seam molding.

To achieve these goals, the present invention takes techniques below scheme:

In a kind of welding process, the online dynamic programming of track follows the tracks of collaborative control method with welding bead, comprises the following steps:

1) mobile welding robot system coordinate system is set up, including: basis cartesian coordinate system O_xyz, robot body coordinate It is M_xyz, airborne actuator coordinate system U_xyzWith structured light sensor coordinate system L_xyz, wherein, airborne actuator coordinate system U_xyzWith structured light sensor coordinate system L_xyzX-axis forward consistent with the direction of robot body straight forward, z-axis forward with The direction of ground normal is consistent, and y-axis forward is determined by lefft-hand rule, airborne actuator coordinate system U_xyzInitial point be positioned at airborne Actuator and the fixed connection place of robot body；

2) when welding arc swinging, to rotate or swings-rotary composite movement form scanning groove time, employing structured light sensor Extract the positional information of front, weld zone seam welds groove characteristic point i, including bevel for welding characteristic point i with welding torch at structure light Sensor coordinate system L_xyzOffset distance E on middle y-axis direction_iWith bevel for welding characteristic point in structured light sensor coordinate system L_xyzIn coordinate, use arc sensor to gather arc energy signal in real time simultaneously, use integration differential method, characteristic harmonics Method or extremum method extract the real-time welding torch posture information in the arc energy signal after signal filtering with processing and amplifying, by airborne Actuator adjusts welding torch pose according to acquired real-time welding torch posture information, adjusts the same of welding torch pose in airborne actuator Time arc sensor start the arc energy signals collecting in next cycle, so circulate, i.e. realize welding bead and follow the tracks of；

3) according to bevel for welding characteristic point in structured light sensor coordinate system L_xyzIn coordinate, carry out coordinate transform to basis flute card That coordinate system O_xyz, it is thus achieved that the coordinate G of bevel for welding characteristic point i_i=(x_i,y_i)；

4) according to step 2) and 3) solve the coordinate of a series of bevel for welding characteristic point successively, it is thus achieved that a bevel for welding feature Point coordinates sequence { G₁,G₂,...,G_NAnd an offset distance array { E₁,E₂,...,E_N, wherein

$N=\frac{\mathrm{\λ}}{{\mathrm{vT}}_{op}}+1$

In formula: λ is that structured light sensor is preposition in the distance of welding torch, and v is speed of welding, T_opFor bevel for welding characteristic point position Confidence breath extracting cycle；

5) use SPL to be fitted bevel for welding characteristic point coordinate sequence calculating, obtain lopcus function to be welded S (x) also carries out derivation, solves in bevel for welding characteristic point coordinate sequence each bevel for welding characteristic point at track letter to be welded Tangent slope k on number S (x)_i, i=1,2 ..., N, obtain comprising N number of bevel for welding characteristic point position auto-control P_i=[x_i y_i arctan(k_i)], i=1,2 ..., the set of N；

6) set threshold value σ withCompare, and set threshold value ζ and compare with Δ E, whenOr During Δ E ＞ ζ, robot body performs continuous path Motion trajectory；WhenOr during Δ E ＜-ζ, Robot body performs around car body center pivot turn trajectory planning；WhenAnd During-ζ≤Δ E≤ζ, robot body performs point-to-point craspedodrome trajectory planning, wherein:

$\left\{\begin{array}{c}\mathrm{\&Delta;}E=E_N-E_1\\ \mathrm{\&sigma;}=L_m-L_s+\frac{1}{2}{L}_a,\frac{1}{2}{L}_a<L_m<L_a\\ \mathrm{\&zeta;}=\frac{1}{4}{L}_a\end{array}\right.$

In formula: E₁And E_NIt is respectively the 1st and n-th bevel for welding characteristic point with welding torch in structured light sensor coordinate system L_xyzOn middle y-axis direction offset distance, L_sFor welding torch in airborne actuator coordinate system U_xyzIn y-coordinate, L_aFor Airborne actuator is in airborne actuator coordinate system U_xyzMaximum functional stroke on middle y-axis direction, L_mFor airborne execution machine Structure is in airborne actuator coordinate system U_xyzDefault impulse stroke on middle y-axis direction；

7) robot body is according to continuous path Motion trajectory, arrive around car body center pivot turn trajectory planning or point Point craspedodrome trajectory planning performs pose and adjusts, and repeats step 2 simultaneously)～6), carry out trajectory planning next time, i.e. realize In welding process, the online dynamic programming of track follows the tracks of collaborative control with welding bead.

In technique scheme, step 6) described in continuous path Motion trajectory, the method for employing is:

According to step 5) described in the set of bevel for welding characteristic point position auto-control, use following formula:

$\left[\begin{array}{c}{x}_{pi}\\ y_{pi}\\ {\mathrm{\&theta;}}_{pi}\end{array}\right]=\left[\begin{array}{c}{x}_i+\frac{1}{2}{L}_{a}sin\left(arctan\right(\frac{dS\left(x_i\right)}{dx}\left)\right)\\ y_i-\frac{1}{2}{L}_a\cos \left(arctan\right(\frac{dS\left(x_i\right)}{dx}\left)\right)\\ \arctan \left(\frac{dS\left(x_i\right)}{dx}\right)\end{array}\right],i=1,2,...,N$

Computing machine human body is the position auto-control of i-th on robot body movement locus: U_pi:[x_pi,y_pi,θ_pi], (i=1,2 ..., N), wherein: S (x_i) it is the function upper i-th of lopcus function S (x) to be welded Value, x_pi,y_piIt is respectively on robot body movement locus at i-th in basis cartesian coordinate system O_xyzIn coordinate, θ_pi For robot body azimuth of i-th on robot body movement locus, use SPL to robot body Coordinate (the x all put on movement locus_pi,y_pi), i=1,2 ..., N carries out interpolation calculation, generates continuous and smooth robot Body movement locus；

In technique scheme, step 6) described in around car body center pivot turn trajectory planning, the method for employing be:

Control robot body left and right sides driving wheel, make driving wheel velocity magnitude equal, in opposite direction, it is achieved machine Human body is around the pivot turn of car body center, robot body angle of turn θ_REmploying following formula calculates:

${\mathrm{\&theta;}}_R=arctan\left(\frac{dS\left(x_N\right)}{dx}\right)-arctan\left(\frac{dS\left(x_1\right)}{dx}\right)$

In technique scheme, step 6) described in point-to-point craspedodrome trajectory planning, the method for employing is:

Keep robot body left and right sides driving wheel velocity-stabilization, make robot body long along current direction displacement Degree is λ, and translational speed size keeps constant.

Welding torch of the present invention uses the rotary welding torch of magnetic control, machinery rotating type welding torch, automatically controlled swing type welding torch or mechanical swinging type The one of welding torch.

The present invention has the following advantages and the technique effect of salience: fully integrated structure light sensors region is ahead of electricity Arc realizes the seam welds groove detection of front, weld zone and arc sensor obtains the advantage of welding torch pose in real time by robot Body and the collaborative uneoupled control of airborne actuator, it is possible to effectively solve mobile welding robot bent to the complex space of large scale Cause that tracing control is delayed, be difficult to take into account dynamic response asks with technology such as static accuracies because of electromechanical inertial during wire bonding seam welding Topic；Improve speed of welding stability and the welding track smoothness of mobile welding robot, particularly can improve curved welding seam Track deep camber region, local and the quality of weld seam molding in weld seams shaped zigzag line knuckle region.Large-scale component complicated track is being welded by the method In seam automatic welding, relatively traditional method has clear superiority, has important actual application value, can be widely applied to boats and ships, heavy type In the mobile robot welding process of the equipment manufacturing such as machinery, the energy and track traffic, be particularly suited for large-size cylinder body girth joint, The large turns such as gantry crossbeam side weld seam, bucket-wheel stacker reclaimer bucket wheel weld seam, locomotive central sill and side structure string beam weld seam or big The welding occasion of curvature curve weld seam.

Accompanying drawing explanation

Fig. 1 is that in welding process, the online dynamic programming of track follows the tracks of collaborative control method FB(flow block) with welding bead.

Fig. 2 is robot body kinematics model schematic diagram of the present invention.

Fig. 3 is continuous path Motion trajectory schematic diagram of the present invention.

In Fig. 3-Fig. 5: a, b are respectively the two ends of seam track.

Fig. 4 is that the present invention is around car body center pivot turn trajectory planning schematic diagram.

Fig. 5 is point-to-point craspedodrome trajectory planning schematic diagram of the present invention.

Fig. 6 is the mobile welding robot system structure principle schematic realizing the method for the invention.

In figure: 1 moves robot；2 structured light sensor；3 image pick-up cards；4 welding torches；5 Hall elements Or signal transmitting device；6 signal acquisition process modules；7 industrial computers；8 motion control cards；9 airborne actuators；10— Movement Controller of Mobile Robot；11 sources of welding current；12 workpiece.

Detailed description of the invention

With embodiment, the principle of the invention and work process are described in further details below in conjunction with the accompanying drawings.

Fig. 1 show the online dynamic programming of track in the welding process of the present invention and follows the tracks of collaborative control method flow chart element with welding bead Figure, including following step:

1) mobile welding robot system coordinate system is set up, including: basis cartesian coordinate system O_xyz, robot body Coordinate system M_xyz, airborne actuator coordinate system U_xyzWith structured light sensor coordinate system L_xyz, wherein, airborne execution Mechanics coordinates U_xyzWith structured light sensor coordinate system L_xyzThe direction of x-axis forward and robot body straight forward Unanimously, z-axis forward is consistent with the direction of ground normal, and y-axis forward is determined by lefft-hand rule, and airborne actuator is sat Mark system U_xyzInitial point be positioned at the fixed connection place of airborne actuator and robot body, robot body of the present invention exists Basis cartesian coordinate system has at the displacement of XOY plane and two degree of freedom of angle of rotating about the z axis；

Setting up the kinematics model of robot body of the present invention, the method for employing is: due to machine of the present invention Human body uses with side wheel parallel drive, when carrying out two dimensional surface motion pose and calculating by robot body model simplification For two-wheeled model, and ignore squeegee action and car body longitudinal sliding motion.As in figure 2 it is shown, robot body moves from p point Moving to p ' place, R is radius of turn, and b is left and right wheels spacing, and θ, θ ', Δ θ are respectively robot body initially side Parallactic angle, back track parallactic angle and the angle turned over, inscribe robot body center in basis cartesian coordinate time (k+1) It is O_xyzIn pose (x^k+1,y^k+1,θ^k+1) be calculated by following formula (1) or following formula (2),

$\left\{\begin{array}{c}{x}^{k+1}=x^k+\frac{r}{2}{\&Integral;}_{T_k}^{T_{k+1}}({{\mathrm{\&omega;}}_r}^k+{{\mathrm{\&omega;}}_l}^k){\mathrm{cos\&theta;}}^{k}dt\\ y^{k+1}=y^k+\frac{r}{2}{\&Integral;}_{T_k}^{T_{k+1}}({{\mathrm{\&omega;}}_r}^k+{{\mathrm{\&omega;}}_l}^k){\mathrm{sin\&theta;}}^{k}dt\\ {\mathrm{\&theta;}}^{k+1}={\mathrm{\&theta;}}^k+\frac{r}{b}{\&Integral;}_{T_k}^{T_{k+1}}|{{\mathrm{\&omega;}}_r}^k-{{\mathrm{\&omega;}}_l}^k|dt\end{array}\right.---\left(1\right)$

$\left\{\begin{array}{c}{x}^{k+1}=x^k+{\mathrm{\&pi;rcos\&theta;}}^k{\&Integral;}_{T_k}^{T_{k+1}}({n_l}^k+{n_r}^k)dt\\ y^{k+1}=y^k+{\mathrm{\&pi;rsin\&theta;}}^k{\&Integral;}_{T_k}^{T_{k+1}}({n_l}^k+{n_r}^k)dt\\ {\mathrm{\&theta;}}^{k+1}={\mathrm{\&theta;}}^k+\frac{2\mathrm{\&pi;}r}{b}{\&Integral;}_{T_k}^{T_{k+1}}|{n_r}^k-{n_l}^k|dt\end{array}\right.---\left(2\right)$

In formula: T is the period of motion, r is radius of wheel, ω_lAnd ω_rIt is respectively left and right wheel angles speed, n_lAnd n_rIt is respectively left and right wheels Rotating speed；

2), in the present embodiment, use structured light sensor to tilt to be incident upon on seam welds groove by structure light, work as welding arc Time under additional transverse rotating magnetic field controls with rotary motion form scanning groove, gather the structure comprising groove geometry feature Striations image, uses adaptive threshold fuzziness, medium filtering, refinement to process image with the step of slope calculations successively With feature extraction, obtain the positional information of bevel for welding characteristic point i, pass at structure light with welding torch including bevel for welding characteristic point i Sensor coordinate system L_xyzOffset distance E on middle y-axis direction_iWith bevel for welding characteristic point in structured light sensor coordinate system L_xyz In coordinate；When welding arc scans groove with rotary motion form under additional transverse rotating magnetic field controls, arc sensor Begin with Hall element and gather welding current signal, use Butterworth low pass ripple and wavelet filtering to welding current signal Carry out pretreatment and realize filtering of interference signal, and carry out signal amplification, use integration differential method, characteristic harmonics method or extremum method Extracting welding current signal after treatment and carry out real-time welding torch posture information, airborne actuator is according to acquired real-time welding torch Posture information adjusts welding torch pose so that it is alignment Weld pipe mill and maintain constant torch height, adjusts in airborne actuator While welding torch pose, arc sensor starts the arc energy signals collecting in next cycle, so circulates, and i.e. realizes welding bead automatic Follow the tracks of；

3) according to bevel for welding characteristic point in structured light sensor coordinate system L_xyzIn coordinate, carry out coordinate transform to basis flute card That coordinate system O_xyz, it is thus achieved that the coordinate G of bevel for welding characteristic point i_i=(x_i,y_i)；

4) according to step 2) and 3) solve the coordinate of a series of bevel for welding characteristic point successively, it is thus achieved that a bevel for welding feature Point coordinates sequence { G₁,G₂,...,G_NAnd an offset distance array { E₁,E₂,...,E_N, wherein

$N=\frac{\mathrm{\&lambda;}}{{\mathrm{vT}}_{op}}+1---\left(3\right)$

In formula: λ is that structured light sensor is preposition in the distance of welding torch, according to welding bead trajectory geometry feature and welding procedure Requirement determines, sets λ as the value between 30mm to 50mm, and v is speed of welding, T_opFor bevel for welding characteristic point Positional information extracting cycle；

5) SPL (Catmull-Rom curve) is used to be fitted bevel for welding characteristic point coordinate sequence calculating, Obtain lopcus function S (x) to be welded and carry out derivation, solving each bevel for welding in bevel for welding characteristic point coordinate sequence special Levy the tangent slope k a little on lopcus function S (x) to be welded_i, i=1,2 ..., N, obtain comprising N number of bevel for welding feature Point position auto-control P_i=[x_i y_i arctan(k_i)], i=1,2 ..., the set of N；

6) set threshold value σ withCompare, and set threshold value ζ and compare with Δ E, whenOr During Δ E ＞ ζ, robot body performs the turning motion planning of side, welding torch place, i.e. continuous path Motion trajectory；WhenOr during Δ E ＜-ζ, robot body performs the turning motion planning of side, non-welding torch place, i.e. around car Body center pivot turn trajectory planning；WhenAnd during-ζ≤Δ E≤ζ, robot body performs point To a craspedodrome trajectory planning；Wherein

$\left\{\begin{array}{c}\mathrm{\&Delta;}E=E_N-E_1\\ \mathrm{\&sigma;}=L_m-L_s+\frac{1}{2}{L}_a,\frac{1}{2}{L}_a<L_m<L_a\\ \mathrm{\&zeta;}=\frac{1}{4}{L}_a\end{array}\right.---\left(4\right)$

In formula: E₁And E_NIt is respectively the 1st and n-th bevel for welding characteristic point with welding torch in structured light sensor coordinate system L_xyzIn On y-axis direction offset distance, L_sFor welding torch in airborne actuator coordinate system U_xyzIn y-coordinate, L_aHold for airborne Row mechanism is in airborne actuator coordinate system U_xyzMaximum functional stroke on middle y-axis direction, L_mFor airborne actuator at machine Carry actuator coordinate system U_xyzDefault impulse stroke on middle y-axis direction, according to actual welding structure and airborne actuator Size sets；

7) robot body is according to continuous path Motion trajectory, around car body center pivot turn trajectory planning or point-to-point Craspedodrome trajectory planning performs pose and adjusts, and repeats step 2 simultaneously)～6), carry out trajectory planning next time, i.e. realize weldering In termination process, the online dynamic programming of track follows the tracks of collaborative control with welding bead.

Step 6) described in continuous path Motion trajectory, the method for employing is:

According to step 5) described in the set of bevel for welding characteristic point position auto-control, use following formula:

$\left[\begin{array}{c}{x}_{pi}\\ y_{pi}\\ {\mathrm{\&theta;}}_{pi}\end{array}\right]=\left[\begin{array}{c}{x}_i+\frac{1}{2}{L}_{a}sin\left(arctan\right(\frac{dS\left(x_i\right)}{dx}\left)\right)\\ y_i-\frac{1}{2}{L}_a\cos \left(arctan\right(\frac{dS\left(x_i\right)}{dx}\left)\right)\\ \arctan \left(\frac{dS\left(x_i\right)}{dx}\right)\end{array}\right],i=1,2,...,N---\left(5\right)$

Computing machine human body is the position auto-control of i-th on robot body movement locus U_pi:[x_pi,y_pi,θ_pi], (i=1,2 ..., N), wherein: S (x_i) it is the function upper i-th of lopcus function S (x) to be welded Value, x_pi,y_piIt is respectively on robot body movement locus at i-th in basis cartesian coordinate system O_xyzIn coordinate, θ_pi For robot body azimuth of i-th on robot body movement locus, use SPL (Catmull-Rom Curve) to the coordinate (x all put on robot body movement locus_pi,y_pi), i=1,2 ..., N carries out interpolation calculation, generates Continuous and smooth robot body movement locus；Differential speed control method is used to control robot left and right wheels so that it is along path node Complete continuous path motion, as it is shown on figure 3,

U_p1:[x_p1,y_p1,θ_p1] it is robot body starting point position auto-control on robot body movement locus,

U_pN:[x_pN,y_pN,θ_pN] it is robot body terminating point position auto-control on robot body movement locus,

U_pi:[x_pi,y_pi,θ_pi], (i=2,3 ..., N-1) it is robot body path node position on robot body movement locus Appearance matrix；

Step 6) described in around car body center pivot turn trajectory planning, the method for employing be:

As shown in Figure 4, control robot body left and right sides driving wheel, make driving wheel velocity magnitude equal, direction phase Instead, it is achieved robot body is around the pivot turn of car body center, and makes robot body central speed direction start in turning At the end of with consistent with institute matching seam track tangential direction respectively, robot body angle of turn θ_RUse following formula meter Calculate:

${\mathrm{\&theta;}}_R=arctan\left(\frac{dS\left(x_N\right)}{dx}\right)-arctan\left(\frac{dS\left(x_1\right)}{dx}\right)---\left(6\right)$

Step 6) described in point-to-point craspedodrome trajectory planning, the method for employing is:

As it is shown in figure 5, U_p1For the initial pose of described robot body, U_pNFor described robot body object pose, Keep robot body left and right sides driving wheel velocity-stabilization, make robot body a length of along current direction displacement λ, translational speed size keeps constant.

Realize the mobile welding robot system of method described in the present embodiment, as shown in Figure 6, it is characterised in that: it includes a shifting Mobile robot (the most described robot body) 1, one structured light sensor 2 (includes light source, video camera, filter, in figure all Be not drawn into), image pick-up card 3, arc sensor (including welding torch 4, Hall element or a signal transmitting device 5), One signal acquisition process module 6, industrial computer 7, motion control card 8, a set of airborne actuator 9 (include balladeur train, cunning Block, motor, motor driver), Movement Controller of Mobile Robot 10, source of welding current 11.Described airborne actuator is fixed It is arranged in described mobile robot 1.The end of described airborne actuator 9 side is fixing with described welding torch 4 to be connected.Described One interface of the source of welding current 11 is connected with workpiece 12 through described Hall element or signal transmitting device 5.The described source of welding current 11 Another interface be connected with described welding torch 4, formed welding current loop.Described Hall element or signal transmitting device 5 defeated Going out end and connect the input of described signal acquisition process module 6, the outfan of described signal acquisition process module 6 connects described work Another input of control machine 7.Described structured light sensor 2 is fixedly installed on the side of described welding torch 4, described knot through a support The outfan of structure optical sensor 2 connects the input of described image pick-up card 3, and the outfan of described image pick-up card 3 connects institute State an input of industrial computer 7.The input of the one outfan described motion control card 8 of connection of described industrial computer 7 controls it and enters Row work, the outfan of described motion control card 8 connects described airborne actuator 9.Another outfan of described industrial computer 7 Connect the input of described Movement Controller of Mobile Robot 10 to control it and be operated, the output of described Movement Controller of Mobile Robot 10 End connects described mobile robot 1.

Described mobile robot uses wheeled mobile robot, mobile platform or the one of mobile trolley, it is characterised in that: energy Enough realize the pivot turn that radius of turn is zero, the present embodiment uses ActivMedia company real with Stanford University SIR Test the Pioneer3-AT robot car that room is developed jointly.

Described welding torch uses the rotary welding torch of magnetic control, machinery rotating type welding torch, automatically controlled swing type welding torch or mechanical swinging type The one of welding torch, uses the rotary welding torch of magnetic control in the present embodiment.

Described industrial computer according to actual needs can be by the one generation in single-chip microcomputer, DSP, PLC, ARM, FPGA or computer Replace.

Claims (5)

1. in a welding process, the online dynamic programming of track follows the tracks of collaborative control method with welding bead, it is characterised in that described side Method comprises the steps:

1) mobile welding robot system coordinate system is set up, including: basis cartesian coordinate system O_xyz, robot body coordinate It is M_xyz, airborne actuator coordinate system U_xyzWith structured light sensor coordinate system L_xyz, wherein, airborne actuator coordinate system U_xyzWith structured light sensor coordinate system L_xyzX-axis forward consistent with the direction of robot body straight forward, z-axis forward with The direction of ground normal is consistent, and y-axis forward is determined by lefft-hand rule, airborne actuator coordinate system U_xyzInitial point be positioned at airborne Actuator and the fixed connection place of robot body；

2) when welding arc swinging, to rotate or swings-rotary composite movement form scanning groove time, employing structured light sensor Extract the positional information of front, weld zone seam welds groove characteristic point i, including bevel for welding characteristic point i with welding torch at structure light Sensor coordinate system L_xyzOffset distance E on middle y-axis direction_iWith bevel for welding characteristic point in structured light sensor coordinate system L_xyzIn coordinate, use arc sensor to gather arc energy signal in real time simultaneously, use integration differential method, characteristic harmonics Method or extremum method extract the real-time welding torch posture information in the arc energy signal after signal filtering with processing and amplifying, by airborne Actuator adjusts welding torch pose according to acquired real-time welding torch posture information, adjusts the same of welding torch pose in airborne actuator Time arc sensor start the arc energy signals collecting in next cycle, so circulate, i.e. realize welding bead and follow the tracks of；

3) according to bevel for welding characteristic point in structured light sensor coordinate system L_xyzIn coordinate, carry out coordinate transform to basis flute Karr coordinate system O_xyz, it is thus achieved that the coordinate G of bevel for welding characteristic point i_i=(x_i,y_i)；

4) according to step 2) and 3) solve the coordinate of a series of bevel for welding characteristic point successively, it is thus achieved that a bevel for welding feature Point coordinates sequence { G₁,G₂,...,G_NAnd an offset distance array { E₁,E₂,...,E_N, wherein

$N=\frac{\mathrm{\&lambda;}}{{\mathrm{vT}}_{op}}+1$

In formula: λ is that structured light sensor is preposition in the distance of welding torch, and v is speed of welding, T_opFor bevel for welding characteristic point position Confidence breath extracting cycle；

5) use SPL to be fitted bevel for welding characteristic point coordinate sequence calculating, obtain lopcus function to be welded S (x) also carries out derivation, solves in bevel for welding characteristic point coordinate sequence each bevel for welding characteristic point at track letter to be welded Tangent slope k on number S (x)_i, i=1,2 ..., N, obtain comprising N number of bevel for welding characteristic point position auto-control P_i=[x_i y_i arctan(k_i)], i=1,2 ..., the set of N；

6) set threshold value σ withCompare, and set threshold value ζ and compare with Δ E, whenOr During Δ E ＞ ζ, robot body performs continuous path Motion trajectory；WhenOr during Δ E ＜-ζ, machine Device human body performs around car body center pivot turn trajectory planning；WhenAnd during-ζ≤Δ E≤ζ, Robot body performs point-to-point craspedodrome trajectory planning；Wherein

$\left\{\begin{array}{c}\mathrm{\&Delta;}E=E_N-E_1\\ \mathrm{\&sigma;}=L_m-L_s+\frac{1}{2}{L}_a,\frac{1}{2}{L}_a<L_m<L_a\\ \mathrm{\&zeta;}=\frac{1}{4}{L}_a\end{array}\right.$

In formula: E₁And E_NIt is respectively the 1st and n-th bevel for welding characteristic point with welding torch in structured light sensor coordinate system L_xyzIn On y-axis direction offset distance, L_sFor welding torch in airborne actuator coordinate system U_xyzIn y-coordinate, L_aHold for airborne Row mechanism is in airborne actuator coordinate system U_xyzMaximum functional stroke on middle y-axis direction, L_mFor airborne actuator at machine Carry actuator coordinate system U_xyzDefault impulse stroke on middle y-axis direction；

7) robot body is according to continuous path Motion trajectory, arrive around car body center pivot turn trajectory planning or point Point craspedodrome trajectory planning performs pose and adjusts, and repeats step 2 simultaneously)～6), carry out trajectory planning next time, i.e. realize In welding process, the online dynamic programming of track follows the tracks of collaborative control with welding bead.

In a kind of welding process, the online dynamic programming of track follows the tracks of collaborative control method with welding bead, It is characterized in that: step 6) described in continuous path Motion trajectory, the method for employing is:

According to step 5) described in the set of bevel for welding characteristic point position auto-control, use following formula:

$\left[\begin{array}{c}{x}_{pi}\\ y_{pi}\\ {\mathrm{\&theta;}}_{pi}\end{array}\right]=\left[\begin{array}{c}{x}_i+\frac{1}{2}{L}_{a}sin\left(arctan\right(\frac{dS\left(x_i\right)}{dx}\left)\right)\\ y_i-\frac{1}{2}{L}_a\cos \left(arctan\right(\frac{dS\left(x_i\right)}{dx}\left)\right)\\ \arctan \left(\frac{dS\left(x_i\right)}{dx}\right)\end{array}\right],i=1,2,...,N$

Computing machine human body is the position auto-control of i-th on robot body movement locus: U_pi:[x_pi,y_pi,θ_pi], (i=1,2 ..., N), wherein: S (x_i) it is the function upper i-th of lopcus function S (x) to be welded Value, x_pi,y_piIt is respectively on robot body movement locus at i-th in basis cartesian coordinate system O_xyzIn coordinate, θ_pi For robot body azimuth of i-th on robot body movement locus, use SPL to robot body Coordinate (the x all put on movement locus_pi,y_pi), i=1,2 ..., N carries out interpolation calculation, generates continuous and smooth robot Body movement locus.

In a kind of welding process the most according to claim 1, the online dynamic programming of track follows the tracks of collaborative controlling party with welding bead Method, it is characterised in that: step 6) described in around car body center pivot turn trajectory planning, the method for employing be:

Control robot body left and right sides driving wheel, make driving wheel velocity magnitude equal, in opposite direction, it is achieved machine Human body is around the pivot turn of car body center, robot body angle of turn θ_REmploying following formula calculates:

${\mathrm{\&theta;}}_R=arctan\left(\frac{dS\left(x_N\right)}{dx}\right)-arctan\left(\frac{dS\left(x_1\right)}{dx}\right).$

In a kind of welding process the most according to claim 1, the online dynamic programming of track follows the tracks of collaborative controlling party with welding bead Method, it is characterised in that: step 6) described in point-to-point craspedodrome trajectory planning, the method for employing is:

Keep robot body left and right sides driving wheel velocity-stabilization, make robot body along the current direction a length of λ of displacement, Translational speed size keeps constant.

In a kind of welding process the most according to claim 1, the online dynamic programming of track follows the tracks of collaborative controlling party with welding bead Method, it is characterised in that: described welding torch uses the rotary welding torch of magnetic control, machinery rotating type welding torch, automatically controlled swing type welding torch or machinery The one of swing type welding torch.