Automobile door cover installation control method and device, computer equipment, medium and product

Abstract

The invention relates to an automobile door cover installation control method and device, computer equipment, a storage medium and a computer program product. The method comprises the steps of determining door cover information of a to-be-installed door cover in response to an installation instruction for the to-be-installed door cover; determining a target installation and adjustment model matched with the door cover information from a plurality of trained candidate installation and adjustment models; under the condition that the door cover to be installed is controlled to move to the initial installation and adjustment position, the first gap surface difference of the door cover to be installed at the initial installation and adjustment position is determined; inputting the first gap surface difference into a target installation and adjustment model, and determining an updated installation and adjustment position of the to-be-installed door cover based on the output of the target installation and adjustment model; and if the second gap surface difference of the to-be-installed door cover at the updated installation and adjustment position meets the installation condition, the to-be-installed door cover is installed at the updated installation and adjustment position. By adopting the method, the installation and adjustment accuracy and the installation and adjustment success rate in the door cover installation process can be improved, and then the door cover installation efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of automobile production technology, and in particular to an automobile door cover installation control method, device, computer equipment, storage medium and computer program product. Background Art

[0002] With the rapid development of production automation and intelligent technology, the use of artificial intelligence and visual algorithms to guide robots to manufacture and assemble parts has become a major trend in the industry. Compared with traditional manual manufacturing, automated and intelligent automobile manufacturing technology has the advantages of high efficiency and low cost, and the automatic assembly function of installing body parts such as doors and hoods on the body also has such advantages.

[0003] When using the automatic assembly function, people have found that if only a fixed track is used to install the door cover, it is easy to cause uneven gaps and surface differences between the door cover and the body, affecting the overall appearance of the vehicle. In order to further eliminate process errors, a three-two-one adjustment algorithm is currently usually used to automatically adjust the automatic assembly function to ensure that the installed door cover can meet the requirements of beauty and balance in terms of gaps and surface differences with various parts of the body.

[0004] Although the 321 adjustment algorithm has the advantages of fast cycle and high efficiency, the technical principle of the algorithm is to first determine the parallelism between the door cover and the vehicle body through three measuring points, then control the rotation and height clearance of the door cover on the parallel plane through two measuring points, and finally determine the left and right clearance of the door cover through one measuring point. This means that when moving the subsequent steps, the gap and face difference of the adjusted measuring points referenced by the previous steps may be affected, making the entire door cover still deviate from the ideal position, resulting in a low success rate of installation and adjustment, and still requiring manual intervention. Therefore, the currently commonly used automatic adjustment scheme has the problem of low adjustment accuracy and reduced success rate of installation and adjustment. Summary of the invention

[0005] Based on this, it is necessary to provide a car door cover installation control method, device, computer equipment, computer readable storage medium and computer program product to address the above technical problems, which can improve the installation accuracy and success rate of the door cover installation process, thereby improving the door cover installation efficiency.

[0006] In a first aspect, the present application provides a method for controlling the installation of a vehicle door cover, the method comprising:

[0007] In response to an installation instruction for a door cover to be installed, determining door cover information of the door cover to be installed;

[0008] Determine a target assembly model adapted to the door cover information from a plurality of trained candidate assembly models;

[0009] Under the condition that the door cover to be installed is controlled to move to the initial adjustment position, a first gap face difference of the door cover to be installed at the initial adjustment position is determined; the training sample of the target adjustment model includes sample gap face differences corresponding to the sample door cover represented by the door cover information at multiple sample adjustment positions; the sample adjustment position is determined based on the initial adjustment position;

[0010] Inputting the first gap flushness into the target assembly model, and determining an updated assembly position of the door cover to be installed based on an output of the target assembly model;

[0011] If the second gap face difference of the door cover to be installed at the update adjustment position meets the installation condition, the door cover to be installed is installed at the update adjustment position.

[0012] In one embodiment, the door cover information includes the vehicle type to which the door cover to be installed belongs and the door cover position identifier of the door cover to be installed in the vehicle type;

[0013] The step of determining a target assembly model adapted to the door cover information from among the trained multiple candidate assembly models includes:

[0014] According to the vehicle type to which the door cover to be installed belongs, determining a plurality of selected installation and adjustment models adapted to the vehicle type from a plurality of candidate installation and adjustment models;

[0015] Based on the door cover position identifier of the door cover to be installed in the vehicle model, a target installation and adjustment model adapted to the door cover position identifier is determined from a plurality of the selected installation and adjustment models.

[0016] In one embodiment, the method further comprises:

[0017] For each sample door cover, an initial adjustment position and an initial adjustment model of the sample door cover are obtained; the initial adjustment position is the adjustment position of the sample door cover when it is in an ideal adjustment state;

[0018] Determine multiple sample adjustment positions according to the initial adjustment position and a preset training sample expansion method;

[0019] Controlling the sample door cover to move to each of the sample adjustment positions respectively, and determining the sample gap face difference of the sample door cover at each of the sample adjustment positions;

[0020] Determining a training sample of the sample door cover based on each of the sample gap face differences and each of the sample installation and adjustment positions;

[0021] Model training is performed on the initial assembly model based on the training samples to obtain a trained target assembly model.

[0022] In one embodiment, the determining of multiple sample adjustment positions according to the initial adjustment position and a preset training sample expansion method includes:

[0023] Taking the initial adjustment position as the center position, dividing the sample generation space for the sample door cover according to the preset sample extension range;

[0024] Based on a preset anchor point establishment method, multiple anchor points are established on each spatial axis of the sample generation space;

[0025] Determine each of the anchor points and the initial adjustment position as the sample generation reference points of the sample door cover;

[0026] For each of the sample generation reference points, a plurality of sample adjustment positions are generated for the sample generation reference point based on a Gaussian distribution scheme, taking the sample generation reference point as a center point.

[0027] In one embodiment, the method further comprises:

[0028] If the second gap flushness of the door cover to be installed does not meet the installation condition, the first gap flushness is updated using the second gap flushness to obtain an updated first gap flushness;

[0029] Based on the updated first gap and flush difference, the step of inputting the first gap and flush difference into the target adjustment model is returned.

[0030] In one embodiment, the method further comprises:

[0031] In response to an installation completion event for the door cover to be installed, determining an installation surface difference of the installed door cover and a surface difference of a suction leveling point of the installed door cover in a suction leveling state;

[0032] Determine the corrected adjustment position of the installed door cover according to the difference information between the installation surface difference and the suction level point surface difference and the corrected rotation dimension of the installed door cover;

[0033] Performing adjustment correction on the installed door cover based on the corrected adjustment position;

[0034] The installation adjustment re-measurement result of the installed door cover is determined according to the third gap face difference of the installed door cover after the installation adjustment correction is completed.

[0035] In a second aspect, the present application further provides a vehicle door cover installation control device, the device comprising:

[0036] An instruction response module, configured to determine door cover information of the door cover to be installed in response to an installation instruction for the door cover to be installed;

[0037] A model determination module, used to determine a target assembly model adapted to the door cover information from a plurality of trained candidate assembly models;

[0038] A gap and flushness determination module is used to determine a first gap and flushness of the door cover to be installed at the initial adjustment position when controlling the door cover to be installed to move to the initial adjustment position; the training samples of the target adjustment model include sample gap and flushness corresponding to the sample door cover represented by the door cover information at multiple sample adjustment positions; the sample adjustment position is determined based on the initial adjustment position;

[0039] An installation position updating module, used for inputting the first gap flushness into the target installation model, and determining an updated installation position of the door cover to be installed based on an output of the target installation model;

[0040] The door cover installation module is used to install the door cover to be installed at the update adjustment position if the second gap surface difference of the door cover to be installed at the update adjustment position meets the installation condition.

[0041] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0042] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0043] In a fifth aspect, the present application also provides a computer program product, including a computer program, which implements the steps of the above method when executed by a processor.

[0044] The above-mentioned automobile door cover installation control method, device, computer equipment, storage medium and computer program product can first determine a target installation model that is adapted to the door cover information from multiple trained candidate installation models when performing automated installation control on the door cover to be installed based on the door cover information of the door cover to be installed. Since the training samples of the target installation model include sample gap face differences corresponding to the sample door cover represented by the door cover information at multiple sample installation positions, and the sample installation position is determined based on the initial installation position, the door cover to be installed can be controlled to move to the initial installation position, and a first gap face difference of the door cover to be installed at the initial installation position is determined, and the first gap face difference is input into the target installation model. The target installation model can directly determine the installation information required for the door cover installation based on the first gap face difference, and determine the updated installation position of the door cover to be installed based on the output of the target installation model. If the second gap face difference of the door cover to be installed at the updated installation position meets the installation condition, the door cover to be installed is installed at the updated installation position. By pre-training a target installation and adjustment model that is compatible with the door cover to be installed, the gap and surface differences of all test points can be considered and balanced simultaneously during automatic adjustment, and the updated installation and adjustment position of the door cover to be installed can be directly determined without the need for staged operations to make adjustments. This effectively improves the installation and adjustment accuracy and success rate during the door cover installation process, reduces the configuration workload and complexity of the installation and adjustment process, and thereby improves the door cover installation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 An application environment diagram of a vehicle door cover installation control method in one embodiment;

[0046] Figure 2 A schematic diagram of a flow chart of a method for controlling the installation of a vehicle door cover in one embodiment;

[0047] Figure 3 A schematic flow chart of a method for controlling the installation of a vehicle door cover in another embodiment;

[0048] Figure 4 It is a schematic diagram of a flow chart of determining multiple sample adjustment positions according to an initial adjustment position and a preset training sample expansion method in an embodiment;

[0049] Figure 5 A schematic flow chart of a method for controlling the installation of a vehicle door cover in another embodiment;

[0050] Figure 6 A schematic flow chart of a method for controlling the installation of a vehicle door cover in another embodiment;

[0051] Figure 7 It is a structural block diagram of a vehicle door cover installation control device in one embodiment;

[0052] Figure 8FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0054] The automobile door cover installation control method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the control platform 102 communicates with the server 104 and the control terminal 106 through the network respectively. The data storage system can store the data that the control platform 102 needs to process. The data storage system can be integrated on the control platform 102, or it can be placed on the cloud or other network servers. When it is necessary to install the door cover to be installed, the control terminal 106 can trigger the installation instruction for the door cover to be installed, and the control platform 102 responds to the installation instruction for the door cover to be installed, determines the door cover information of the door cover to be installed, and determines the target adjustment model adapted to the door cover information from the trained multiple candidate adjustment models, and then controls the corresponding control terminal 106 to move the door cover to be installed to the initial adjustment position, and determines the first gap face difference of the door cover to be installed at the initial adjustment position, wherein the training samples of the target adjustment model include the sample gap face differences corresponding to the sample door cover represented by the door cover information at multiple sample adjustment positions, and the sample adjustment position is determined based on the initial adjustment position. The first gap face difference is input into the target assembly model, and based on the output of the target assembly model, the updated assembly position of the door cover to be installed is determined. If the second gap face difference of the door cover to be installed at the updated assembly position meets the installation conditions, the corresponding control terminal 106 is controlled to install the door cover to be installed at the updated assembly position.

[0055] Among them, the control platform 102 is a software platform for executing control logic operations, generating and issuing control instructions. The control platform 102 can perform control logic operations or processing based on the collected or acquired information, generate corresponding control instructions and issue them to the corresponding control terminal 106 for processing. It can be understood that the control platform 102 can be integrated on the user terminal of the technician, or on the server 104. The user terminal can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers and portable wearable devices. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.

[0056] The control terminal 106 may be a terminal device on the production line that is controlled by the control platform 102 and actually performs adjustment and installation operations. For example, on the door cover installation production line, the control terminal 106 may include but is not limited to mobile control equipment, information collection equipment, etc., wherein the mobile control equipment is a device that can grab the door cover and drive the door cover to move, such as a robot, an intelligent robotic arm, etc. The information collection equipment is a device that collects information about the position of the door cover with the vehicle body during the installation process. For example, the information collection equipment may include a camera, a ranging sensor, etc. The control platform 102 may generate corresponding control instructions according to the actual control situation and send them to the corresponding control terminal 106, and control the control terminal 106 based on the control instructions.

[0057] In one embodiment, Figure 2 As shown, a method for controlling the installation of a car door cover is provided, and the method is applied to Figure 1 Taking the control platform 102 in the example as an example, the following steps are included:

[0058] S202: In response to an installation instruction for a door cover to be installed, determine door cover information of the door cover to be installed.

[0059] Among them, the door cover to be installed is a door cover component that needs to be installed on the car body, such as the car door, hood, trunk lid, etc. The installation instruction for the door cover to be installed is a command signal for executing the control platform to install the door cover to be installed on the car body, which can be triggered by the technician based on the user terminal or by a custom event. For example, when the door cover to be installed is transported to the installation station, the technician selects the door cover to be installed through the user terminal, and clicks the installation control to trigger the generation of the installation instruction for the door cover to be installed. For another example, the technician can pre-set the installation events of various parts of the car body in the platform program of the control platform based on the production task information. The installation event can be a time event, such as installing the door cover to be installed at a certain time, thereby triggering the installation instruction for the door cover to be installed.

[0060] The door cover information of the door cover to be installed is information data used to characterize the configuration of the door cover to be installed. For example, the door cover information may include but is not limited to the vehicle model to which the door cover to be installed belongs, the door cover position identification in the vehicle model, the door cover specification parameters and other information data.

[0061] Specifically, the control platform determines the door cover information of the door cover to be installed in response to the installation instruction for the door cover to be installed.

[0062] In some of the embodiments, the installation instruction carries the door cover information of the door cover to be installed, and the control platform can directly obtain the door cover information of the door cover to be installed according to the installation instruction.

[0063] In some of the embodiments, the installation instruction carries a door cover identification of the door cover to be installed, and the control platform can obtain the door cover information of the door cover to be installed from the server based on the door cover identification of the door cover to be installed.

[0064] S204, determining a target assembly model adapted to the door cover information from the multiple trained candidate assembly models.

[0065] Among them, the candidate installation model is a model obtained after installation training for sample door covers corresponding to different door cover information. Designers can perform installation training for sample door covers corresponding to different door cover information in advance, and train the initial installation model based on the training samples collected during the installation training to obtain the candidate installation model corresponding to the door cover information. It can be understood that the specific number of candidate installation models can be determined according to the number of door cover information types corresponding to the door covers that the control platform can support installation. The same door cover information can be classified into the same category, and the initial state model can be any deep learning model.

[0066] In some of the embodiments, the candidate assembly model may be trained by a designer on a control platform.

[0067] In some embodiments, in order to improve the versatility of the control platform, the candidate installation and adjustment model can also be uploaded to the control platform by a technician through a user terminal. When the user needs to use the control platform to install the door cover to be installed corresponding to the newly added door cover information, the technician can upload the candidate installation and adjustment model corresponding to the newly added door cover information to the control platform for the control platform to call at any time.

[0068] Specifically, after obtaining the door cover information of the door cover to be installed, the control platform can determine a target assembly model that is compatible with the door cover information from a plurality of trained candidate assembly models according to the door cover information.

[0069] In some of the embodiments, the door cover information may include a door cover identification of the door cover to be installed. The control platform is pre-configured with the correspondence between each door cover identification and each candidate assembly model. After obtaining the door cover information of the door cover to be installed, the correspondence between each door cover identification and each candidate assembly model can be searched according to the door cover identification existing in the door cover information, and the target assembly model matching the door cover identification can be determined from multiple candidate assembly models.

[0070] S206, determining a first gap face difference of the door cover to be installed at the initial adjustment position while controlling the door cover to be installed to move to the initial adjustment position.

[0071] The training samples of the target assembly model include sample gap face differences corresponding to the sample door cover represented by the door cover information at multiple sample assembly positions. The sample assembly position is determined based on the initial assembly position.

[0072] The sample door cover can be understood as the standard door cover corresponding to the door cover information. In actual production, due to process differences, even if the door cover information is exactly the same, there will be some slight differences between the door cover to be installed obtained in actual production and the standard door cover. Therefore, even if a fixed installation trajectory is obtained through sample door cover simulation, when the fixed installation trajectory is used to control the operation of the door cover to be installed, there will still be gaps and surface differences between the door cover and the body, affecting the overall appearance of the vehicle.

[0073] Gap and flush are information data used to characterize the gap and flushness between the door cover and the vehicle body. Gap refers to the deviation of the minimum distance between the door cover and the vehicle body, and flushness refers to the difference in height between the door covers. Gap and flushness determine the flatness and aesthetics of the vehicle body. Sample gap and flushness is the gap and flushness corresponding to the sample door cover being moved to the sample adjustment position, and the first gap and flushness is the gap and flushness determined when the door cover to be installed is moved to the initial adjustment position.

[0074] In some of the embodiments, the control platform can directly obtain the gap face difference uploaded by the technician. For example, the technician can determine the gap face difference of the door cover to be installed through the gap face difference determination system, and then upload the determined gap face difference to the control platform to provide data support for the control process of the control platform.

[0075] In some of the embodiments, the control platform can collect the gap and face difference when the door cover is in the installation position through a pre-configured information collection device. Specifically, after determining that the door cover has moved to the installation position, the control platform can control the information collection device to collect point cloud information when the door cover is in the installation position, perform information analysis and recognition based on the point cloud information, and determine the gap and face difference when the door cover is in the installation position. It can be understood that the door cover can include a door cover to be installed and a sample door cover, and the installation position can include an initial installation position, an updated installation position, or a sample installation position.

[0076] Among them, the installation position refers to the debugging position used when controlling the installation of the door cover. During the installation process, by determining the gap and face difference of the door cover when it is in the debugging position, the relative position information between the current door cover and the vehicle body can be understood, thereby providing a data basis for whether the door cover can be installed in the future. The sample installation position is the training installation position determined during training, and the initial installation position is the installation position of the sample door cover in the ideal installation state. It can be understood that the initial installation position can be determined by the technician based on the experimental data and uploaded to the control platform. Each door cover information corresponds to its own initial installation position.

[0077] In some of the embodiments, the installation position may refer to the position of the door cover to be installed during the installation process. Since in the automatic installation process, the movement of the door cover to be installed is assisted by a corresponding mobile control terminal, such as a robot. Therefore, the initial installation position of the door cover to be installed corresponds one-to-one to the terminal position of the mobile control terminal. For example, the position coordinates of the initial installation position of the door cover to be installed correspond one-to-one to the robot coordinates of the grasping robot. If you want to control the door cover to be installed to move to the initial installation position, the control platform only needs to generate a control instruction to instruct the mobile control terminal to move according to the terminal position information of the mobile control terminal corresponding to the initial installation position, and then move the door cover to be installed to the corresponding initial installation position.

[0078] Specifically, the control platform can determine the initial installation position matched by the door cover information based on the door cover information of the door cover to be installed, and then instruct the corresponding control terminal to move the door cover to be installed to the initial installation position. When it is determined that the door cover to be installed is moved to the initial installation position, the control platform can determine the first gap face difference of the door cover to be installed at the initial installation position.

[0079] S208, inputting the first gap flushness into the target assembly model, and determining an updated assembly position of the door cover to be installed based on an output of the target assembly model.

[0080] The updated installation position refers to a theoretical installation position determined after analysis of the target installation model. Installing the door cover to be installed at the updated installation position can theoretically meet the user's installation requirements, that is, meet the installation conditions.

[0081] Specifically, after obtaining the first gap flushness, the controller can input the first gap flushness into the target assembly model. The target assembly model can analyze the difference between the first gap flushness and the ideal gap flushness of the door cover to be installed based on the first gap flushness, and determine the adjustment information required to adjust the first gap flushness to the theoretical gap flushness. The control platform can obtain the output of the target assembly model, and determine the updated assembly position of the door cover to be installed based on the output of the target assembly model.

[0082] In some embodiments, the output of the target assembly model is the position information of the mobile control terminal corresponding to the updated assembly position of the door cover to be installed, for example, the position coordinates of the robot corresponding to the updated assembly position.

[0083] In some of the embodiments, the output of the target assembly model may be position difference information between an initial assembly position and an updated assembly position. After the control platform obtains the position difference information, it may determine an updated assembly position of the door cover to be installed based on the position difference information and the initial assembly position.

[0084] S210: If the second gap face difference of the door cover to be installed at the updated installation position meets the installation condition, the door cover to be installed is installed at the updated installation position.

[0085] The installation condition is a preset condition for determining whether the door cover to be installed can be installed at the updated installation position. If the second gap face difference of the door cover to be installed at the updated installation position meets the installation condition, it means that installing the door cover to be installed at the updated installation position can meet the user's installation requirements.

[0086] In some of the embodiments, the installation condition may be that the second gap flushness is the same as the theoretical gap flushness.

[0087] In some embodiments, the installation condition may be that the difference between the second gap flushness and the theoretical gap flushness meets a preset allowable difference range, for example, the difference value of the gap flushness is within the range of ±0.5 mm.

[0088] Specifically, after determining the update adjustment position of the door cover to be installed, the control platform can determine the second gap face difference of the door cover to be installed at the update adjustment position, and then call the preset installation conditions to determine whether the second gap face difference meets the installation conditions; if the second gap face difference meets the installation conditions, the door cover to be installed is installed at the update adjustment position.

[0089] In the above embodiment, when the door cover to be installed is automatically installed and controlled, a target installation model that is compatible with the door cover information can be first determined from multiple trained candidate installation models based on the door cover information of the door cover to be installed. Since the training samples of the target installation model include sample gap face differences corresponding to the sample door cover represented by the door cover information at multiple sample installation positions, and the sample installation position is determined based on the initial installation position, the door cover to be installed can be controlled to move to the initial installation position, and a first gap face difference of the door cover to be installed at the initial installation position is determined, and the first gap face difference is input into the target installation model. The target installation model can directly determine the installation information required for the door cover installation based on the first gap face difference, and determine the updated installation position of the door cover to be installed according to the output of the target installation model. If the second gap face difference of the door cover to be installed at the updated installation position meets the installation condition, the door cover to be installed is installed at the updated installation position. By pre-training a target installation and adjustment model that is compatible with the door cover to be installed, the gap and surface differences of all test points can be considered and balanced simultaneously during automatic adjustment, and the updated installation and adjustment position of the door cover to be installed can be directly determined without the need for staged operations to make adjustments. This effectively improves the installation and adjustment accuracy and success rate during the door cover installation process, reduces the configuration workload and complexity of the installation and adjustment process, and thereby improves the door cover installation efficiency.

[0090] In some other embodiments, the automobile door cover installation control method further includes: if the second gap flushness of the door cover to be installed does not meet the installation condition, the first gap flushness is updated using the second gap flushness to obtain an updated first gap flushness, and based on the updated first gap flushness, the step of inputting the first gap flushness into the target assembly model is returned.

[0091] Among them, if the second gap face difference does not meet the installation conditions, it can be considered that installing the door cover to be installed in the updated adjustment position will not meet the user's installation requirements. For example, there is a difference between the second gap face difference and the theoretical gap face difference, or the degree of difference between the second gap face difference and the theoretical gap face difference exceeds the allowable difference range.

[0092] Specifically, if the second gap face difference of the door cover to be installed does not meet the installation conditions, it means that although the updated adjustment position that can theoretically meet the user's installation requirements has been determined, in the actual control process, it may be due to control errors or other influencing factors, resulting in the adjustment process not achieving the expected results. In order to meet the user's installation requirements, the control platform can use the second gap face difference to update the first gap face difference to obtain an updated first gap face difference, and then return to the step of inputting the first gap face difference into the target adjustment model based on the updated first gap face difference until the second gap face difference meets the installation conditions.

[0093] In the above embodiment, if one adjustment is unsuccessful, the control platform can perform adjustment control on the door cover to be installed again in a cyclic adjustment manner until it is determined that the door cover to be installed can meet the user's installation requirements when installed in the updated adjustment position. The door cover can then be installed, which can effectively reduce the situation where the adjustment accuracy is reduced due to control deviation or other influencing factors.

[0094] In some of the embodiments, in order to avoid some extreme situations, such as the updated adjustment position of the door cover to be installed continuously failing to meet user needs due to program failure or hardware failure, the designer can pre-set a cycle count threshold. Before each execution of the adjustment cycle, the control platform first determines whether the number of adjustment cycles of the door cover to be installed has reached the cycle count threshold. If it has reached it, it is determined that there is an installation abnormality in the door cover to be installed, and an abnormal prompt message can be generated to notify the technician to conduct abnormality troubleshooting.

[0095] In the automotive field, different car models have different corresponding body contours and door cover installation requirements. For the same car model, door covers in different positions may also have different corresponding door cover installation requirements. For example, the installation requirements for left and right doors may not be the same as those for the hood or trunk lid.

[0096] Based on this, in some embodiments, the door cover information includes the vehicle type to which the door cover to be installed belongs and the door cover position identifier of the door cover to be installed in the vehicle type. S204, determining a target assembly model adapted to the door cover information from multiple trained candidate assembly models, including: according to the vehicle type to which the door cover to be installed belongs, determining multiple selected assembly models adapted to the vehicle type from multiple candidate assembly models. Based on the door cover position identifier of the door cover to be installed in the vehicle type, determining a target assembly model adapted to the door cover position identifier from multiple selected assembly models.

[0097] Among them, the model of the door cover to be installed is the model information of the car corresponding to the door cover to be installed. It can be understood that the car model can be classified and determined according to the structure, purpose, function, brand and other factors of the car. Different models have different corresponding door cover installation requirements, that is, each model has its own door cover installation requirements. Take the face difference of the trunk cover as an example. The installation requirement of the face difference of the trunk cover of some models is 0mm, while the installation requirement of the face difference of the trunk cover of some models is 0.5mm, etc.

[0098] The door cover position identification is information data used to characterize the body installation position corresponding to the door cover to be installed in the vehicle model to which it belongs. The body installation position may include the left front door, right front door, left rear door, right rear door, engine hood, trunk lid, etc.

[0099] The selected installation model is a candidate installation model corresponding to the vehicle model. It can be understood that each vehicle model may correspond to multiple selected installation models, and the number of selected installation models is equal to the number of door covers that need to be installed on the vehicle model.

[0100] Specifically, the control platform can determine multiple selected assembly models adapted to the vehicle type from multiple candidate assembly models according to the vehicle type to which the door cover to be installed belongs. Then, based on the door cover position identifier of the door cover to be installed in the vehicle type, determine the target assembly model adapted to the door cover position identifier from the multiple selected assembly models.

[0101] In some of the embodiments, a first mapping relationship for characterizing the correspondence between each vehicle model and each candidate assembly model is pre-stored in the control platform. The control platform can search for the first mapping relationship according to the vehicle model to which the door cover to be installed belongs, so as to determine multiple selected assembly models suitable for the vehicle model, and each selected assembly model corresponds to a different door cover position identifier.

[0102] In some of the embodiments, a second mapping relationship for characterizing the correspondence between each door cover position identifier and each selected assembly model is pre-stored in the control platform. After determining multiple selected assembly models, the control platform can search for the second mapping relationship based on the door cover position identifier of the door cover to be installed, so as to determine the target assembly model that matches the door cover position identifier of the door cover to be installed.

[0103] In the above embodiment, by pre-training corresponding installation and adjustment models for different vehicle models and door covers at different positions in different vehicle models, the universality of the control platform in the door cover installation scenario and the accuracy of door cover installation control for different door covers can be effectively improved.

[0104] In the entire scheme of automobile door cover installation control, the target assembly model has an important influence on the assembly accuracy, and the training process of the target assembly model will directly affect the actual use process of the target assembly model. Based on this, the training process of the target assembly model will be explained through several embodiments below.

[0105] In some embodiments, Figure 3 As shown, the automobile door cover installation control method also includes:

[0106] S302: For each sample door cover, an initial installation position and an initial installation model of the sample door cover are obtained.

[0107] The initial adjustment position is the adjustment position of the sample door cover in an ideal adjustment state. The initial adjustment model is an untrained deep learning model, and the initial adjustment model can be pre-configured in the control platform by the designer.

[0108] Specifically, when training the model, for each door cover position in each vehicle model, the designer will prepare a corresponding sample door cover and a sample vehicle body corresponding to the sample door cover. For each sample door cover, the control platform can obtain the initial adjustment position and initial adjustment model of the sample door cover.

[0109] S304, determining multiple sample adjustment positions according to the initial adjustment position and a preset training sample expansion method.

[0110] Among them, the training sample expansion method is a sample generation method used to generate multiple sample adjustment positions based on the initial adjustment position. It can be understood that the designer can write the training sample expansion method according to the actual sample generation requirements, and set the training sample expansion method in the control platform, so that the control platform can generate training samples according to an initial adjustment position.

[0111] In some embodiments, the training sample expansion method may be random expansion, that is, taking the initial adjustment position as the center, randomly selecting adjustment positions from the periphery as sample adjustment positions.

[0112] In some of the embodiments, in order to reduce the complexity of training and the degree of resource consumption, the training sample expansion method can be to randomly select a first number of multiple sample adjustment positions within a preset spatial range centered on the initial adjustment position, and randomly select a second number of multiple sample adjustment positions outside the preset spatial range centered on the initial adjustment position, wherein the first number is greater than the second number. By generating more sample adjustment positions in an area closer to the initial adjustment position and generating fewer sample adjustment positions in an area farther from the initial adjustment position, the training complexity and the resource consumption of the training process can be reduced while ensuring that all fine adjustments around the initial adjustment position can be trained and that training points farther from the initial adjustment position will not be discrete during adjustment.

[0113] S306, respectively controlling the sample door cover to move to each sample installation and adjustment position, and determining the sample gap surface difference of the sample door cover at each sample installation and adjustment position.

[0114] Specifically, the control platform can control the sample door cover to move to each sample adjustment position respectively. For each sample adjustment position, the control platform will determine the sample gap surface difference of the sample door cover at the sample adjustment position, and obtain the sample gap surface difference of the sample door cover at each sample adjustment position.

[0115] The mobile control terminal is used as a robot, and the sample installation position is {X1, X2, ..., X i For example, the control platform can control the robot to move to the robot coordinate position corresponding to each sample installation position, and then control the information acquisition device to collect the point cloud information of the sample door cover and the sample body, and obtain the point cloud information corresponding to each sample installation position. Based on each point cloud information, the gap surface difference values ​​{Y1, Y2, ..., Y i}.

[0116] S308, determining training samples of sample door covers based on the gap and flushness of each sample and the adjustment position of each sample.

[0117] Specifically, the control platform can bind each sample gap face difference with its corresponding sample installation position to obtain multiple groups of samples, and determine the multiple groups of samples as training samples of the sample door cover.

[0118] S310, performing model training on the initial assembly model based on the training samples to obtain a trained target assembly model.

[0119] Specifically, after obtaining the training sample of the sample door cover, the control platform can perform model training on the initial assembly model based on the training sample and the preset model training method, iteratively optimize the model parameters in the initial assembly model, and obtain the target assembly model after training. It can be understood that the preset model training method can be any training method of a deep learning model.

[0120] In some of the embodiments, the preset model training method may be a stochastic gradient descent method. By using the stochastic gradient descent method for model training, the influence of noise during the training process can be effectively suppressed, thereby improving the accuracy of model training.

[0121] In the above embodiment, by generating training samples of sample door covers according to the initial installation position and the preset training sample expansion method, the number of samples during model training can be effectively increased, thereby improving the training accuracy of the target installation model, and providing accurate installation tools for the subsequent actual installation process of the door cover.

[0122] Further, in some embodiments, Figure 4 As shown, S304, according to the initial adjustment position and the preset training sample expansion method, multiple sample adjustment positions are determined, including:

[0123] S402, taking the initial adjustment position as the center position, dividing the sample generation space for the sample door cover according to the preset sample extension range.

[0124] The preset sample extension range is a range parameter used to limit the size of the sample generation space, which can be preset by the designer according to the actual training requirements. The preset sample extension range can include an axis length threshold on each spatial axis with the initial adjustment position as the center position.

[0125] Specifically, the control platform can call a preset sample extension range, take the initial installation position as the center position, and divide the sample generation space for the sample door cover according to the preset sample extension range.

[0126] In some of the embodiments, the sample generation space is a six-dimensional space, and the six dimensions are three translation dimensions and three selection dimensions.

[0127] S404: Based on a preset anchor point establishment method, multiple anchor points are established on each spatial axis of the sample generation space.

[0128] Among them, the anchor point is the reference point used to generate the adjustment position of each sample. The anchor point is related to the initial adjustment position, and each sample adjustment position is related to each anchor point. That is, by setting the anchor point, each sample adjustment position can be associated with the initial adjustment position to avoid the discreteness of the sample adjustment position during adjustment.

[0129] The preset anchor point establishment method is a preset method for establishing anchor points in the sample generation space. The preset anchor point establishment method can be generated by designers according to actual needs. For example, the preset anchor point establishment method can be an axis equalization anchor point establishment method. Taking the sample generation space as a six-dimensional space as an example, the control platform can establish an anchor point at one-third and two-thirds of each space axis, for a total of 32 anchor points.

[0130] Specifically, the control platform can establish multiple anchor points on each spatial axis of the sample generation space based on a preset anchor point establishment method.

[0131] S406: Determine each anchor point and the initial adjustment position as a sample generation reference point of the sample door cover.

[0132] Specifically, the control platform can uniformly determine the determined anchor points and initial installation positions as sample generation reference points of the sample door cover.

[0133] For example, after the control platform sets up an anchor point at one-third and two-thirds of each spatial axis, these 32 anchor points and the initial adjustment position can be unified as sample generation reference points, and the sample door cover includes 33 sample generation reference points.

[0134] S408 , for each sample generation reference point, taking the sample generation reference point as the center point, and generating a plurality of sample adjustment positions for the sample generation reference point based on a Gaussian distribution scheme.

[0135] Among them, the Gaussian distribution scheme refers to a point generation scheme that uses Gaussian distribution to generate sample adjustment positions. Its principle is to generate a series of random points based on the probability density function of Gaussian distribution. Gaussian distribution is a continuous probability distribution with a bell-shaped shape, which is determined by two parameters: mean and standard deviation. The mean represents the center position of the distribution, while the standard deviation represents the degree of dispersion of the data.

[0136] Specifically, for each sample generation reference point, with the sample generation reference point as the center point, a plurality of sample adjustment positions are randomly generated for the sample generation reference point based on a Gaussian distribution scheme.

[0137] Taking the example of a sample door cover with 33 sample generation reference points, the control platform can randomly use a sample generation reference point as the center each time and use Gaussian distribution to generate a certain sample installation position.

[0138] In the above embodiment, a plurality of anchor points related to the initial adjustment position are first set on each spatial axis in the sample generation space, and then a plurality of sample adjustment positions are generated based on each anchor point and the initial adjustment position using a Gaussian distribution scheme. By using this mixed Gaussian distribution scheme, a plurality of sample adjustment positions are generated based on the initial adjustment position, which ensures that small fine-tuning around the initial adjustment position can be trained, and also ensures that the boundary points of the training space will not be discrete during adjustment due to being too far away from the training point, thereby effectively improving the rationality of sample sampling of the target adjustment model, and thereby improving the training accuracy of the model.

[0139] In order to be able to check the final effect of the installation, in some embodiments, such as Figure 5 As shown, the automobile door cover installation control method also includes the following steps:

[0140] S502, in response to an installation completion event for the door cover to be installed, determining the installation surface difference of the installed door cover and the surface difference of the suction leveling point of the installed door cover in a suction leveling state.

[0141] Among them, the installation completion event is an event triggered after the door cover to be installed is installed. For example, after the screws of the door cover to be installed are tightened, the technician can determine that the door cover to be installed is installed, and trigger the installation completion event based on the user terminal. It can be understood that after the door cover to be installed is installed, it can be called an installed door cover, that is, the installed door cover and the door cover to be installed are the same door cover in different states. The installation face difference of the installed door cover is the face difference information between the installed door cover and the vehicle body.

[0142] Among them, the suction leveling state is the surface difference information between the installed door cover and the vehicle body when the suction leveling process is performed on the installed door cover. It is understandable that after the door cover is installed, the corresponding mobile control terminal will release the control of the door cover, for example, the robot will release the gripper. At this time, the door cover may change the position of the door, or the door may naturally sag without supporting the door in the natural state, resulting in a difference between the placement position of the door cover and the actual placement position of the door cover when leaving the factory. If it is directly re-measured, that is, the installation effect is checked, it is easy to cause measurement errors. Therefore, in order to make the door cover position during re-measurement consistent or approximately consistent with the actual door cover placement position when leaving the factory, the control platform needs to perform suction leveling on the installed door cover, that is, use suction leveling equipment, such as a cylinder to pull the door cover to the actual door cover placement position when leaving the factory, or to the door cover placement position close to the actual factory for measurement.

[0143] The suction flat point is the measurement point that the user hopes the ideal surface difference between the door cover and the vehicle body in the factory state corresponds to. The suction flat point surface difference is the ideal surface difference between the door cover and the vehicle body in the factory state.

[0144] Specifically, the control platform can respond to the installation completion event for the door cover to be installed, control the information collection device to collect information on the installed door cover, determine the installation surface difference of the installed door cover, and simultaneously obtain the suction leveling point surface difference of the installed door cover in the suction leveling state. It can be understood that the suction leveling point surface difference can be preset by the designer according to the actual factory requirements of the user and stored in the control platform.

[0145] S504, determining the corrected installation position of the installed door cover according to the difference information between the installation surface difference and the suction level point surface difference and the corrected rotation dimension of the installed door cover.

[0146] Among them, the corrected rotation dimension of the installed door cover is the dimension in which the installed door cover can be rotated when performing the adjustment operation during the retest. After the door cover is installed, it will not be able to be rotated and adjusted in any dimension as in the adjustment process before installation, but can only be rotated in the corresponding rotatable dimension during normal use. For example, in the case where the door cover is the left or right door of the vehicle body, its rotatable dimension is consistent with the rotation dimension of the door when it is normally opened and closed.

[0147] When correcting the installation position is to compensate for the difference information, the control platform needs to control the target installation position of the installed door cover rotational adjustment in the correction rotation dimension.

[0148] Specifically, after determining the installation surface difference and the suction leveling point surface difference of the installed door cover, the control platform can first compare the installation surface difference and the suction leveling point surface difference to determine the difference information between the installation surface difference and the suction leveling point surface difference, and determine the corrected installation position of the installed door cover based on the difference information between the installation surface difference and the suction leveling point surface difference and the corrected rotation dimension of the installed door cover.

[0149] S506: Correct the installation of the installed door cover based on the corrected installation position.

[0150] Specifically, the control platform can control the correction rotation dimension corresponding to the rotation of the installed door cover, move to the correction adjustment position, and complete the adjustment correction of the installed door cover.

[0151] S508, determining the installation and adjustment re-measurement result of the installed door cover according to the third gap face difference of the installed door cover after the installation and adjustment correction is completed.

[0152] Specifically, after the control platform completes the installation and adjustment correction of the installed door cover, it can determine the third gap face difference of the installed door cover after the installation and adjustment correction. The specific method for determining the gap face difference has been described in detail in the previous text and will not be repeated here. After obtaining the third gap face difference, the control platform can determine the installation and adjustment re-measurement result of the installed door cover based on the third gap face difference.

[0153] In some of the embodiments, the control platform can directly determine the third gap face difference as the installation and adjustment re-measurement result of the installed door cover, thereby providing a data basis for the subsequent installation and adjustment optimization process of the control platform.

[0154] In some of the embodiments, the control platform can determine the difference information between the third gap face difference and the suction level point face difference, evaluate the installation process of the door cover based on the difference information between the third gap face difference and the suction level point face difference, and obtain the installation and retest results of the installed door cover.

[0155] In the above embodiment, by retesting the installed door cover after installation, the installation effect of the installed door cover can be effectively fed back, providing a data basis for optimization of subsequent installation processes.

[0156] In some embodiments, Figure 6 As shown, a method for controlling the installation of automobile door covers is provided, which can be summarized into three parts, namely, the model training part, the adjustment part, and the retest part. Among them, the training part refers to learning the relationship between the current gap face difference and the movement value required by the robot through certain data to obtain the adjustment model, which only needs to be run once. The adjustment part is to use the trained adjustment model to install the door cover during daily production. The retest part is to perform a functional check on the gap face difference at various locations after the door cover is installed and before the car is released, so as to feedback the final installation effect. The following will explain these three parts in detail:

[0157] The first is the model training part.

[0158] The machines required for model training include a controllable industrial assembly robot, a camera installed on its mechanical structure to capture gap and flushness conditions, and a connected industrial computer. There is also a set of sample vehicles and sample door covers for users to train the model.

[0159] During actual training, the technician can manually select a position where the gap and face difference has been adjusted as the teaching point X0, also known as the initial adjustment position. At this teaching point X0, the gap and face difference between the sample door cover and the sample car at each location reaches the ideal value Y0.

[0160] Based on the teaching point X0 determined by the technician, the control platform generates a certain number of random training points near the teaching point according to the preset mixed Gaussian distribution scheme. Among them, the so-called Gaussian mixed distribution scheme refers to dividing a six-dimensional sample generation space with the teaching point X0 as the center according to the preset sample extension range, and then setting an anchor point at one-third and two-thirds of each spatial axis in the space, a total of 2×6=32 anchor points, plus the central teaching point X0, a total of 33 sample generation reference points. The control platform can randomly generate certain training points with a sample generation reference point as the center each time, that is, multiple sample adjustment positions, using a Gaussian distribution.

[0161] After obtaining multiple training points, the control platform controls the robot to grab the sample door cover and move it to each training point {X1, X2, ..., X i}, and photograph the point cloud of each training point, and calculate the corresponding gap surface difference values ​​{Y1, Y2, ..., Y i}. It should be noted that X i is a six-dimensional coordinate value that can be mapped to the robot coordinates, and Y i is an N-dimensional vector, where N is equal to the number of all required gap face differences, i is an arbitrary natural number, and the specific value of N is related to the number of cameras mounted on the robot. For example, when a camera is mounted on the robot, the corresponding Y i It is a two-dimensional vector that can represent a gap value and a surface difference value. When two cameras are mounted on the robot, the corresponding Y i It is a four-dimensional vector that can represent two gap values ​​and two face difference values.

[0162] The control platform can generate training samples based on the obtained gap surface difference values ​​and training points, train the servo model F according to the training samples and the random gradient descent method, obtain the conversion relationship between the gap surface difference value at each location and the gap surface difference at the teaching position and the robot displacement, that is, X-X0=F(Y-Y0), and obtain the assembly and adjustment model corresponding to the sample door cover.

[0163] Then comes the installation part.

[0164] The control platform responds to the installation instruction for the door cover to be installed, determines the door cover information of the door cover to be installed, and determines the target adjustment model adapted to the door cover information from the trained multiple candidate adjustment models. Then the control platform can control the door cover to be installed to move to the teaching point X0, collect point cloud information at the teaching point X0, and determine the first gap face difference Y of the door cover to be installed based on the point cloud information. k Since the body and door cover change with the process, the original teaching position X0 may not be the best matching position, but it is not far from the best matching position.

[0165] The control platform sets the first gap flushness Y k Input to the target adjustment model, the target adjustment model can be based on the first gap flushness Y k Calculate the robot adjustment displacement and get the updated adjustment position X k+1 , that is, X k+1 =X k -F(Y k -Y0). The control platform controls the robot to move the door cover to be installed to the updated installation position X k+1 Then collect the point cloud information to determine the second gap surface difference Y of the door cover to be installed k+1 .

[0166] The control platform sets the second gap difference Y k+1 Compared with the ideal gap face difference Y0, if the difference between the two is within the preset difference range, for example, within ±0.5mm, the second gap face difference Y of the door cover to be installed at the updated adjustment position can be determined. k+1 Meet the installation conditions, control platform in the update installation position X k+1 Install the door cover to be installed.

[0167] If the difference between the two is not within the preset difference range, the second gap surface difference Y of the door cover to be installed at the updated adjustment position can be determined. k+1 If the installation conditions are not met, the control platform can use the second gap difference Y k+1 For the first gap flushness Y k Update to get the updated first gap flushness Y k , at this time Y k =Y k+1 The control platform will be based on the updated first gap difference Y k Return to execute the first gap flushness Y k Steps to input to the target assembly model until the second gap flush Y k+1 Meet the installation conditions.

[0168] The last part is the retest.

[0169] The control platform can respond to the installation completion event for the door cover to be installed, determine the installation surface difference of the installed door cover, and the suction leveling point surface difference of the installed door cover in the suction leveling state. According to the difference information between the installation surface difference and the suction leveling point surface difference, and the corrected rotation dimension of the installed door cover, determine the corrected adjustment position of the installed door cover, and perform adjustment correction on the installed door cover based on the corrected adjustment position. Specifically, the control platform can run the adjustment iteration process once X k+1 =X K -F(Y k-Y0), but only one rotational dimension is moved to simulate the normal opening and closing of the door. Finally, the installation and adjustment retest results of the installed door cover are determined according to the third gap face difference of the installed door cover after the installation and adjustment correction is completed.

[0170] The automobile door cover installation control method in this embodiment has the following advantages. First, it can balance each measuring point to approximate the theoretical value at the same time, preventing the loss of one thing while focusing on another caused by the order relationship of the three-two-one method. Second, the whole process does not require camera calibration, and the servo-trained model automatically learns the coefficient relationship between the gap face difference value and the installation and adjustment movement value, without manual intervention of the parameter value. Third, the process flow has a high degree of freedom, and the user can freely choose the installation and adjustment process (installation and adjustment first, then embedding, embedding first, then adjustment of posture, then embedding, then adjustment of translation, etc.), and then the modular algorithm will execute the corresponding steps. Fourth, it can be used directly on different door covers while only modifying the number of required gap face difference measuring points, and it also allows users to freely select the measuring points and movement freedoms that need to be opened, which is convenient for use in multiple scenarios. Fifth, it has high toughness. Multiple cameras can be carried on the robot. If there is a failure of an individual camera, the parameters corresponding to the camera will be automatically adjusted to zero, so that the production line can still be installed and adjusted, so that maintenance can be carried out during the idle time of the production line instead of having to stop the line for maintenance immediately, but the effect will be slightly affected. Sixth, the cycle is fast, and the door cover can be installed and adjusted in 2-3 moves at most.

[0171] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0172] Based on the same inventive concept, the embodiment of the present application also provides a vehicle door cover installation control device for implementing the vehicle door cover installation control method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the vehicle door cover installation control device provided below can refer to the limitations of the vehicle door cover installation control method above, and will not be repeated here.

[0173] In one embodiment, Figure 7As shown, a vehicle door cover installation control device 700 is provided, comprising: a command response module 701, a model determination module 702, a gap and flushness determination module 703, an adjustment position update module 704 and a door cover installation module 705, wherein:

[0174] The instruction response module 701 is used to determine the door cover information of the door cover to be installed in response to the installation instruction for the door cover to be installed.

[0175] The model determination module 702 is used to determine a target assembly model that is adapted to the door cover information from a plurality of trained candidate assembly models.

[0176] The gap and face difference determination module 703 is used to determine the first gap and face difference of the door cover to be installed at the initial adjustment position when controlling the door cover to be installed to move to the initial adjustment position; the training samples of the target adjustment model include sample gap and face differences corresponding to the sample door cover represented by the door cover information at multiple sample adjustment positions; the sample adjustment position is determined based on the initial adjustment position.

[0177] The assembly position updating module 704 is used to input the first gap flushness into the target assembly model, and determine the updated assembly position of the door cover to be installed based on the output of the target assembly model.

[0178] The door cover installation module 705 is used to install the door cover to be installed at the update adjustment position if the second gap face difference of the door cover to be installed at the update adjustment position meets the installation condition.

[0179] In some embodiments, the door cover information includes the vehicle type to which the door cover to be installed belongs and the door cover position identifier of the door cover to be installed in the vehicle type. The model determination module 702 is used to: determine multiple selected adjustment models adapted to the vehicle type from multiple candidate adjustment models according to the vehicle type to which the door cover to be installed belongs. Based on the door cover position identifier of the door cover to be installed in the vehicle type, determine the target adjustment model adapted to the door cover position identifier from multiple selected adjustment models.

[0180] In some embodiments, the automobile door cover installation control device further includes:

[0181] The initial information acquisition module is used to acquire the initial adjustment position and initial adjustment model of each sample door cover; the initial adjustment position is the adjustment position of the sample door cover when it is in an ideal adjustment state.

[0182] The sample expansion module is used to determine multiple sample installation positions according to the initial installation position and the preset training sample expansion method.

[0183] The sample gap face difference determination module is used to control the sample door cover to move to each sample adjustment position respectively, and determine the sample gap face difference of the sample door cover at each sample adjustment position.

[0184] The sample generation module is used to determine the training samples of the sample door cover based on the gap and flushness of each sample and the adjustment position of each sample.

[0185] The model training module is used to train the initial assembly model based on the training samples to obtain the target assembly model after the training is completed.

[0186] In some embodiments, the initial information acquisition module is used to: divide the sample generation space into sample door covers according to a preset sample extension range and with the initial adjustment position as the center position; establish multiple anchor points on each spatial axis of the sample generation space based on a preset anchor point establishment method; determine each anchor point and the initial adjustment position as the sample generation reference point of the sample door cover; for each sample generation reference point, with the sample generation reference point as the center point, generate multiple sample adjustment positions for the sample generation reference point based on a Gaussian distribution scheme.

[0187] In some embodiments, the automobile door cover installation control device further includes:

[0188] The gap flushness updating module is used to update the first gap flushness using the second gap flushness if the second gap flushness of the door cover to be installed does not meet the installation conditions, so as to obtain the updated first gap flushness.

[0189] The loop module is used to return to the adjustment position update module based on the updated first gap and flushness to execute the step of inputting the first gap and flushness into the target adjustment model.

[0190] In some embodiments, the automobile door cover installation control device further includes:

[0191] The event response module is used to determine the installation surface difference of the installed door cover and the surface difference of the suction leveling point of the installed door cover in the suction leveling state in response to the installation completion event of the door cover to be installed.

[0192] The correction adjustment position determination module is used to determine the correction adjustment position of the installed door cover according to the difference information between the installation surface difference and the suction level point surface difference, and the correction rotation dimension of the installed door cover.

[0193] The adjustment correction module is used to adjust the installed door cover based on the correction adjustment position.

[0194] The re-measurement module is used to determine the re-measurement result of the installed door cover according to the third gap face difference of the installed door cover after the installation and correction is completed.

[0195] Each module in the above-mentioned vehicle door cover installation control device can be fully or partially implemented by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0196] In one embodiment, a computer device is provided. The computer device may be a terminal or a server integrated with a control platform. The internal structure diagram thereof may be as follows: Figure 8 As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data such as door cover information, candidate adjustment models, initial adjustment position, first gap face difference, updated adjustment position, second gap face difference, etc. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for controlling the installation of an automobile door cover is implemented.

[0197] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0198] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, specific steps of the above-mentioned automobile door cover installation control method embodiment are implemented.

[0199] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the specific steps of the above-mentioned automobile door cover installation control method embodiment are implemented.

[0200] In one embodiment, a computer program product is provided, including a computer program, which implements the specific steps of the above-mentioned automobile door cover installation control method embodiment when executed by a processor.

[0201] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. In addition, the acquisition, storage, processing, and transmission of data are in compliance with relevant laws and regulations.

[0202] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0203] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0204] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for controlling the installation of a car door cover, characterized in that: The method comprises: In response to an installation instruction for a door cover to be installed, determining door cover information of the door cover to be installed; Determine a target assembly model adapted to the door cover information from a plurality of trained candidate assembly models; Under the condition that the door cover to be installed is controlled to move to the initial adjustment position, a first gap face difference of the door cover to be installed at the initial adjustment position is determined; the training sample of the target adjustment model includes sample gap face differences corresponding to the sample door cover represented by the door cover information at multiple sample adjustment positions; the sample adjustment position is determined based on the initial adjustment position; Inputting the first gap flushness into the target assembly model, and determining an updated assembly position of the door cover to be installed based on an output of the target assembly model; If the second gap face difference of the door cover to be installed at the update adjustment position meets the installation condition, the door cover to be installed is installed at the update adjustment position.

2. The method according to claim 1, characterized in that The door cover information includes the vehicle type to which the door cover to be installed belongs and the door cover position identifier of the door cover to be installed in the vehicle type; The step of determining a target assembly model adapted to the door cover information from among the trained multiple candidate assembly models includes: According to the vehicle type to which the door cover to be installed belongs, determining a plurality of selected installation and adjustment models adapted to the vehicle type from a plurality of candidate installation and adjustment models; Based on the door cover position identifier of the door cover to be installed in the vehicle model, a target installation and adjustment model adapted to the door cover position identifier is determined from a plurality of the selected installation and adjustment models.

3. The method according to claim 1, characterized in that The method further comprises: For each sample door cover, an initial adjustment position and an initial adjustment model of the sample door cover are obtained; the initial adjustment position is the adjustment position of the sample door cover when it is in an ideal adjustment state; Determine multiple sample adjustment positions according to the initial adjustment position and a preset training sample expansion method; Controlling the sample door cover to move to each of the sample adjustment positions respectively, and determining the sample gap face difference of the sample door cover at each of the sample adjustment positions; Determining a training sample of the sample door cover based on each of the sample gap face differences and each of the sample installation and adjustment positions; Model training is performed on the initial assembly model based on the training samples to obtain a trained target assembly model.

4. The method according to claim 3, characterized in that Determining a plurality of sample adjustment positions according to the initial adjustment position and a preset training sample expansion method includes: Taking the initial adjustment position as the center position, dividing the sample generation space for the sample door cover according to the preset sample extension range; Based on a preset anchor point establishment method, multiple anchor points are established on each spatial axis of the sample generation space; Determine each of the anchor points and the initial adjustment position as the sample generation reference points of the sample door cover; For each of the sample generation reference points, a plurality of sample adjustment positions are generated for the sample generation reference point based on a Gaussian distribution scheme, taking the sample generation reference point as a center point.

5. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: If the second gap flushness of the door cover to be installed does not meet the installation condition, the first gap flushness is updated using the second gap flushness to obtain an updated first gap flushness; Based on the updated first gap and flush difference, the step of inputting the first gap and flush difference into the target adjustment model is returned.

6. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: In response to an installation completion event for the door cover to be installed, determining an installation surface difference of the installed door cover and a surface difference of a suction leveling point of the installed door cover in a suction leveling state; Determine the corrected adjustment position of the installed door cover according to the difference information between the installation surface difference and the suction level point surface difference and the corrected rotation dimension of the installed door cover; Performing adjustment correction on the installed door cover based on the corrected adjustment position; The installation adjustment re-measurement result of the installed door cover is determined according to the third gap face difference of the installed door cover after the installation adjustment correction is completed.

7. A vehicle door cover installation control device, characterized in that: The device comprises: An instruction response module, configured to determine door cover information of the door cover to be installed in response to an installation instruction for the door cover to be installed; A model determination module, used to determine a target assembly model adapted to the door cover information from a plurality of trained candidate assembly models; A gap and flushness determination module is used to determine a first gap and flushness of the door cover to be installed at the initial adjustment position when controlling the door cover to be installed to move to the initial adjustment position; the training samples of the target adjustment model include sample gap and flushness corresponding to the sample door cover represented by the door cover information at multiple sample adjustment positions; the sample adjustment position is determined based on the initial adjustment position; An installation position updating module, used for inputting the first gap flushness into the target installation model, and determining an updated installation position of the door cover to be installed based on an output of the target installation model; The door cover installation module is used to install the door cover to be installed at the update adjustment position if the second gap surface difference of the door cover to be installed at the update adjustment position meets the installation condition.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images