A multi-layer and multi-path trajectory planning method based on welding robot

Abstract

The invention relates to the field of industrial robot automation technology, and in particular to a multi-layer multi-pass trajectory planning method based on a welding robot, comprising measuring a weld angle size and height to obtain measurement parameters; determining the number of welding layers, the number of passes and process parameters based on the measurement parameters to obtain determination parameters; teaching a starting point and an end point of a reference weld based on the determination parameters; determining a coordinate system according to the starting point and the end point; determining a multi-layer multi-pass planning direction according to the coordinate system; planning a weld path according to the set number of layers and passes and the weld angle size and height based on the multi-layer multi-pass planning direction; converting a quasi-weld start point and an end point in a user coordinate system to a point position in the coordinate system to obtain a conversion start point and a conversion end point; offsetting the conversion start point and the conversion end point based on the weld path to obtain an offset point; and converting the offset point to the weld after the offset in the user coordinate system, thereby solving the problem that the trajectory calculation method of the existing welding robot is relatively cumbersome.

Description

Technical Field

[0001] The invention relates to the technical field of industrial robot automation, and in particular to a multi-layer and multi-path trajectory planning method based on a welding robot. Background Art

[0002] With the development of society, machines have replaced manual labor in various automated operations and have developed rapidly. In the field of medium and thick plates, multi-layer and multi-pass welding is a commonly used welding method. Compared with conventional welding methods, it has more advantages in terms of heat input and deformation.

[0003] The multi-layer and multi-pass welding robots currently on the market generally set the offset of each layer and each pass in advance to calculate the trajectory, which makes the setting more complicated. Summary of the invention

[0004] The purpose of the present invention is to provide a multi-layer and multi-path trajectory planning method based on a welding robot, aiming to solve the problem that the trajectory calculation method of the existing welding robot is relatively complicated.

[0005] To achieve the above object, the present invention provides a multi-layer and multi-path trajectory planning method based on a welding robot, comprising the following steps:

[0006] S1 measures the size and height of the welding angle to obtain the measurement parameters;

[0007] S2 determines the number of welding layers, the number of passes and process parameters based on the measured parameters to obtain determined parameters;

[0008] S3 teaches the starting point and the end point of the reference weld based on the determined parameters;

[0009] S4 determines a coordinate system according to the starting point and the end point;

[0010] S5 determines the multi-layer multi-channel planning direction according to the coordinate system;

[0011] S6 plans the weld path based on the multi-layer and multi-pass planning direction according to the set number of layers and passes and the size and height of the weld angle;

[0012] S7 converts the quasi-weld start point and end point in the user coordinate system to points in the coordinate system to obtain a conversion start point and a conversion end point;

[0013] S8 offsets the conversion starting point and the conversion end point based on the weld path to obtain offset points;

[0014] S9 converts the offset point to the weld after welding offset in the user coordinate system.

[0015] Wherein, the method further comprises:

[0016] S10 repeats steps S8-S9 until the welding of the left and right welds is completed.

[0017] Wherein, the planning of the weld path based on the multi-layer and multi-pass planning direction according to the set number of layers and passes and the size and height of the weld angle includes:

[0018] Based on the multi-layer and multi-pass planning direction, equal area planning and posture planning are performed according to the set number of layers and passes and the size and height of the weld angle to obtain the weld path.

[0019] The equal area planning includes:

[0020] Total area S = H*L / 2.

[0021] The area of ​​each weld bead Sa = S / N.

[0022] Height of each weld bead

[0023]

[0024] The last width of each layer

[0025] W i =S a / h i

[0026] Then the first weld Y of each layer is:

[0027] Y i0 =Y0+H i-1 *tan(θ / 2)

[0028] Each weld Y is:

[0029] Y ij =Y i0 +n*W i

[0030] The last Y is

[0031] Y0-H i-1 *tan(θ / 2)

[0032] Height of each floor

[0033] Z i =H i-1

[0034] In the above formula, the total number of channels is N, the current number of channels is n, and the number of channels in the 0-i layer is A. i , the total height is H, the weld angle size is L, and the angle is θ.

[0035] Wherein, the posture planning includes:

[0036] The weld angle formed by each weld pass is β

[0037] The attitude deviation is

[0038] γ=β / 2-atan(W l / h i ).

[0039] The present invention provides a multi-layer multi-pass trajectory planning method based on a welding robot. The method measures the size and height of the welding angle to obtain measurement parameters; determines the number of welding layers, the number of passes and the process parameters based on the measurement parameters to obtain determination parameters; teaches the starting point and the end point of the reference weld based on the determination parameters; determines the coordinate system according to the starting point and the end point; determines the multi-layer multi-pass planning direction according to the coordinate system; plans the weld path according to the set number of layers and passes and the size and height of the welding angle based on the multi-layer multi-pass planning direction; converts the quasi-weld starting point and the end point in the user coordinate system to the point position in the coordinate system to obtain the conversion starting point and the conversion end point; offsets the conversion starting point and the conversion end point based on the weld path to obtain the offset point; converts the offset point to the weld offset in the user coordinate system. Through the above steps, in practical applications, for a weld, only the size of the welding angle to be achieved by welding needs to be entered, and the multi-layer multi-pass trajectory planning can be completed. The problem that the trajectory calculation method of the existing welding robot is relatively troublesome is solved.

[0040] Beneficial effects:

[0041] 1. The present invention proposes a multi-layer and multi-channel path planning method for a welding robot, which solves the problem that the offset of each layer and each channel needs to be manually set, and is more convenient to use.

[0042] 2. The present invention solves the problem of needing to calibrate the workpiece coordinate system in advance by calculating the weld coordinate system through the reference weld.

[0043] 3. The present invention determines the planning direction through a coordinate system to ensure that the flat angle welds are welded from bottom to top. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 It is a schematic diagram of the location planning method.

[0046] Figure 2 It is a schematic diagram of the posture planning method.

[0047] Figure 3 The present invention provides a flowchart of a multi-layer and multi-channel trajectory planning method based on a welding robot. DETAILED DESCRIPTION

[0048] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0049] See also Figures 1 to 3 The present invention provides a multi-layer and multi-path trajectory planning method based on a welding robot, comprising the following steps:

[0050] S1 measures the size and height of the welding angle to obtain the measurement parameters;

[0051] Specifically, measure the size L and height H of the weld fillet.

[0052] In addition, before measuring the weld angle, the method further includes: obtaining the shape of the welded object; determining the surface curvature according to the shape; selecting a clamping tool according to the surface curvature; and clamping and fixing the welded object with the clamping tool. By clamping and fixing the welded object, when measuring the weld angle of the welded object, the welded object is displaced by external force, thereby affecting the accuracy of the weld angle measurement.

[0053] The obtaining of the shape of the welding object comprises: photographing the welding object to obtain a photographed image;

[0054] Adjust the exposure, clarity, planning, noise removal, etc. of the captured image to obtain a high-definition image;

[0055] The high-definition image is input into a trained feature extraction model to extract features and obtain the shape of the welded object.

[0056] The construction method of the feature extraction model includes: obtaining a training picture set, inputting the training picture set into an initial model, extracting a number of local features from the training picture by a local feature extraction layer in the initial model and outputting the local features to a feature enhancement layer in the initial model, performing feature enhancement on the local features based on an attention mechanism by the feature enhancement layer to obtain a number of enhanced local features and outputting the local features to a global feature extraction layer in the initial model, performing feature aggregation based on the enhanced local features by the global feature extraction layer to obtain a global feature and outputting the global feature to a classifier in the initial model, obtaining predicted classification parameters of the training picture according to the global features by the classifier, adjusting the model parameters of the initial model according to the label parameters and predicted classification parameters of the training picture to obtain a trained classification model, constructing a feature extraction model for extracting the global features of the picture to be identified based on the local feature extraction layer and the global feature extraction layer in the classification model, thereby realizing the construction of a feature extraction model that can be used to extract the global features of the picture to be identified, and then in an application, the global features can be obtained by inputting the picture to be identified into the feature extraction model. It can be seen that this realizes an end-to-end global feature extraction scheme, which helps to effectively align local features with global features, and feature extraction and feature aggregation are jointly implemented, thereby improving the feature extraction efficiency and reducing the algorithm complexity compared to the algorithm in the prior art that performs feature extraction and feature aggregation separately; at the same time, by utilizing the attention mechanism to enhance local features, it is possible to suppress irrelevant background interference in local features, weaken the influence of background noise and local occlusion on features, and further improve the feature representation ability and richness of representation.

[0057] S2 determines the number of welding layers, the number of passes and process parameters based on the measured parameters to obtain determined parameters;

[0058] Specifically, determine the number of welding layers, the number of passes N and process parameters.

[0059] S3 teaches the starting point and the end point of the reference weld based on the determined parameters;

[0060] Specifically, teach the starting and ending points of the reference weld.

[0061] S4 determines a coordinate system according to the starting point and the end point;

[0062] Specifically, the coordinate system C is determined according to the starting point and the end point of the reference weld.

[0063] S5 determines the multi-layer multi-channel planning direction according to the coordinate system;

[0064] Specifically, the multi-layer and multi-channel planning directions are determined according to the coordinate system C.

[0065] S6 plans the weld path based on the multi-layer and multi-pass planning direction according to the set number of layers and passes and the size and height of the weld angle;

[0066] Specifically, based on the multi-layer and multi-pass planning direction, equal area planning and posture planning are performed according to the set number of layers and passes and the size and height of the weld angle to obtain the weld path.

[0067] The equal area planning includes:

[0068] Total area S = H*L / 2.

[0069] The area of ​​each weld bead Sa = S / N.

[0070] Height of each weld bead

[0071]

[0072] The last width of each layer

[0073] W i =S a / h i

[0074] Then the first weld Y of each layer is:

[0075] Y i0 =Y0+H i-1 *tan(θ / 2)

[0076] Each weld Y is:

[0077] Y ij =Y i0 +n*W i

[0078] The last Y is

[0079] Y0-H i-1 *tan(θ / 2)

[0080] Height of each floor

[0081] Z i =H i-1

[0082] In the above formula, Figure 1 As shown, it is the cross section of the weld, the total number of passes is N, the current number of passes is n, and the number of passes in the 0-i layer is A i , the total height is H, the weld angle size is L, and the angle is θ. O is the origin of the coordinate system established with the weld direction in the X direction.

[0083] The posture planning includes:

[0084] The weld angle formed by each weld pass is β

[0085] The attitude deviation is

[0086] γ=β / 2-atan(W l / h i ).

[0087] like Figure 2 As shown in the figure, each weld bead forms a new weld angle in the manner of a dotted line, and the weld angle formed is β

[0088] S7 converts the quasi-weld start point and end point in the user coordinate system to points in the coordinate system to obtain a conversion start point and a conversion end point;

[0089] S8 offsets the conversion starting point and the conversion end point based on the weld path to obtain offset points;

[0090] S9 converts the offset point to the weld after welding offset in the user coordinate system.

[0091] S10 repeats steps S8-S9 until the welding of the left and right welds is completed.

[0092] The present invention discloses a multi-layer and multi-channel trajectory planning method based on a welding robot. First, a training picture set is obtained and input into an initial model, so that a local feature extraction layer in the initial model extracts a plurality of local features from the training picture and outputs the extracted local features to a feature enhancement layer in the initial model, so that the feature enhancement layer performs feature enhancement on the plurality of local features based on an attention mechanism to obtain a plurality of enhanced local features and outputs the enhanced local features to a global feature extraction layer in the initial model, so that the global feature extraction layer performs feature aggregation based on the plurality of enhanced local features to obtain a global feature and outputs the global feature to a classifier in the initial model, so that the classifier obtains predicted classification parameters of the training picture according to the global features, adjusts the model parameters of the initial model according to the label parameters and the predicted classification parameters of the training picture, obtains a trained classification model, and constructs a feature extraction model for extracting the global features of the picture to be identified based on the local feature extraction layer and the global feature extraction layer in the classification model, thereby realizing the construction of a feature extraction model that can be used to extract the global features of the picture to be identified.

[0093] Next, the welding object is photographed to obtain a photographed image; the photographed image is adjusted for exposure, clarity, planning, noise removal, etc. to obtain a high-definition image; the high-definition image is input into a trained feature extraction model for feature extraction to obtain the shape of the welding object. The surface curvature is determined based on the shape; a clamping tool is selected based on the surface curvature; and the welding object is clamped and fixed using the clamping tool.

[0094] Then, the size and height of the weld angle of the welded object after clamping are measured to obtain the measurement parameters; the number of welding layers, the number of passes and the process parameters are determined based on the measurement parameters to obtain the determination parameters; the starting point and the end point of the reference weld are taught based on the determination parameters; the coordinate system is determined according to the starting point and the end point; the multi-layer and multi-pass planning direction is determined according to the coordinate system; the weld path is planned according to the set number of layers and passes and the size and height of the weld angle based on the multi-layer and multi-pass planning direction; the quasi-weld starting point and end point in the user coordinate system are converted to the point position in the coordinate system to obtain the conversion starting point and the conversion end point; the conversion starting point and the conversion end point are offset based on the weld path to obtain the offset point; the offset point is converted to the weld after welding offset in the user coordinate system.

[0095] Through the above steps, in practical applications, for a weld, the present invention only needs to input the required weld angle size and complete multi-layer multi-pass trajectory planning, thereby solving the problem that the trajectory calculation method of the existing welding robot is relatively complicated.

[0096] The entire calculation process and calculation method flow of the present invention is a common calculation method for industrial robots and is applicable to the path planning of all welding robots on the market.

[0097] Beneficial effects:

[0098] 1. The present invention proposes a multi-layer and multi-channel path planning method for a welding robot, which solves the problem that the offset of each layer and each channel needs to be manually set, and is more convenient to use.

[0099] 2. The present invention solves the problem of needing to calibrate the workpiece coordinate system in advance by calculating the weld coordinate system through the reference weld.

[0100] 3. The present invention determines the planning direction through a coordinate system to ensure that the flat angle welds are welded from bottom to top.

[0101] What is disclosed above is only a preferred embodiment of a multi-layer and multi-path trajectory planning method based on a welding robot of the present invention. Of course, this cannot be used to limit the scope of rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above embodiments are implemented, and equivalent changes made according to the claims of the present invention are still within the scope of the invention.

Claims

1. A multi-layer and multi-path trajectory planning method based on a welding robot, characterized in that: The following steps are involved: S1 measures the size and height of the welding angle to obtain the measurement parameters; S2 determines the number of welding layers, the number of passes and process parameters based on the measured parameters to obtain determined parameters; S3 teaches the starting point and the end point of the reference weld based on the determined parameters; S4 determines a coordinate system according to the starting point and the end point; S5 determines the multi-layer multi-channel planning direction according to the coordinate system; S6 plans the weld path based on the multi-layer and multi-pass planning direction according to the set number of layers and passes and the size and height of the weld angle; S7 converts the quasi-weld start point and end point in the user coordinate system to points in the coordinate system to obtain a conversion start point and a conversion end point; S8 offsets the conversion starting point and the conversion end point based on the weld path to obtain offset points; S9 converts the offset point to the weld after welding offset in the user coordinate system; The planning of the weld path based on the multi-layer and multi-pass planning direction according to the set number of layers and passes and the size and height of the weld angle includes: Based on the multi-layer and multi-pass planning direction, equal area planning and posture planning are performed according to the set number of layers and passes and the size and height of the weld angle to obtain a weld path; The equal area planning includes: Total area S = H*L / 2; The area of ​​each weld bead Sa = S / N; Height of each weld bead The last width of each layer W i =S a / h i Then the first weld Y of each layer is: Y i0 =Y0+H i-1 *time(θ / 2) Each weld Y is: Y ij =Y i0 +n*W i The last Y is Y=Y0-H i-1 *time(θ / 2) Height of each floor Z i =H i-1 In the above formula, the total number of channels is N, the current number of channels is n, and the number of channels in the 0-i layer is A. i , the total height is H, the welding angle size is L, and the angle is θ; The posture planning includes: The weld angle formed by each weld pass is β The attitude deviation is γ=β / 2-atan(W i / h i )。 2. The multi-layer and multi-path trajectory planning method based on a welding robot according to claim 1, characterized in that: The method further comprises: S10 repeats steps S8-S9 until the welding of the left and right welds is completed.

Images