End force estimation model training method and device, end force estimation method and device, equipment and medium

Abstract

The invention provides a terminal force estimation model training method and device, an estimation method and device, equipment and a medium, and relates to the technical field of mechanical arms. Multiple sets of sample state data of the mechanical arm in multiple motion states are obtained, and the sample state data comprise sample state data of multiple joints in the mechanical arm and sample force parameters of the tail end of the mechanical arm in multiple directions; adopting a preset initial tail end force estimation model, and according to the sample detection data of the plurality of joints in the sample state data and the sample force parameter of the tail end in the first preset direction, predicting to obtain an estimated force parameter of the tail end corresponding to the sample state data in a second preset direction; and according to a sample force parameter in a second preset direction in the sample state data and an estimated force parameter in the second preset direction corresponding to the sample state data, adjusting a model parameter of the initial end force estimation model to obtain a target end force estimation model. Therefore, the cost is reduced, and the tail end force is accurately obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of robotic arms, and in particular to a terminal force estimation model training method, estimation method, device, equipment and medium. Background Art

[0002] Automated industrial robotic arms have a wide range of application scenarios in production and living environments, such as industrial welding robotic arms, sorting robotic arms, etc. In the use of the robotic arm, it is usually necessary to accurately obtain the force information at the end of the robotic arm to achieve high-precision and high-efficiency operation and assembly requirements. For example, for a welding robotic arm, the end force information can help the robotic arm control the strength and pressure of the welding gun to ensure the quality and stability of the welding. For a sorting robotic arm, the end force information can help the robotic arm control the strength and posture of the clamped object so as to accurately sort the items. In addition, in some special environments, such as high temperature and high pressure, the end force information of the robotic arm can also help the robotic arm adapt to environmental changes and ensure the safety and stability of operation. Therefore, accurately obtaining the end force information of the robotic arm is crucial for the efficient operation and safety of the robotic arm.

[0003] Various sensors and measurement methods are widely used in the acquisition of the end force of the robot arm to meet the needs of different application scenarios. In general, the end force of the robot arm can be obtained by deploying a force sensor at its end and reading the value of the sensor. For example, installing a six-axis force sensor at the end of the robot arm can obtain the force information of the end in six dimensions of xyz and roll, pitch, and yaw. This information can be used to control the movement of the robot arm and achieve high-precision operation and assembly tasks. However, the price of force sensors is generally expensive. For example, an ordinary six-axis force sensor costs more than 20,000 yuan, which may increase the cost and complexity of the robot arm. In addition, the installation and calibration of force sensors also require professional technology and operation, which may also increase the difficulty and cost of use. Summary of the invention

[0004] In view of the deficiencies in the above-mentioned prior art, the present invention provides a terminal force estimation model training method, estimation method, device, equipment and medium to solve the problems of low accuracy and high cost in the prior art.

[0005] The technical solution adopted in the embodiment of the present application is as follows:

[0006] In a first aspect, an embodiment of the present application provides a method for training a terminal force estimation model, the method comprising:

[0007] Acquire multiple groups of sample state data of the robot arm in multiple motion states, the sample state data including: sample state data of multiple joints in the robot arm and sample force parameters of the end of the robot arm in multiple preset directions; the multiple preset directions include a first preset direction and a second preset direction;

[0008] Using a preset initial terminal force estimation model, based on the sample state data of the multiple joints in the sample state data and the sample force parameters of the terminal in the first preset direction, an estimated force parameter of the terminal in the second preset direction corresponding to the sample state data is predicted;

[0009] According to the sample force parameters of the second preset direction in the sample state data and the estimated force parameters of the terminal in the second preset direction corresponding to the sample state data, the model parameters of the initial terminal force estimation model are iteratively adjusted until the initial terminal force estimation model after parameter adjustment meets the iteration stop condition, thereby obtaining the target terminal force estimation model.

[0010] In a second aspect, an embodiment of the present application provides a terminal force estimation method, the method comprising:

[0011] Acquire state data of a plurality of joints in the to-be-estimated mechanical arm, and force parameters of the end of the to-be-estimated mechanical arm in a first preset direction;

[0012] Based on the state data of multiple joints in the robotic arm to be estimated, and the force parameters of the end of the robotic arm to be estimated in a first preset direction, and using an end force estimation model, the force parameters of the end of the robotic arm to be estimated in a second preset direction are estimated; wherein the end force estimation model is the target end force estimation model described in any one of the first aspects above.

[0013] In a third aspect, an embodiment of the present application provides a terminal force estimation model training device, the device comprising:

[0014] A first acquisition module is used to acquire multiple groups of sample state data of the robot arm in multiple motion states, the sample state data including: sample state data of multiple joints in the robot arm and sample force parameters of the end of the robot arm in multiple preset directions; the multiple preset directions include a first preset direction and a second preset direction;

[0015] A prediction module, configured to use a preset initial terminal force estimation model to predict, based on the sample state data of the plurality of joints in the sample state data and the sample force parameters of the terminal in the first preset direction, an estimated force parameter of the terminal in the second preset direction corresponding to the sample state data;

[0016] An adjustment module is used to iteratively adjust the model parameters of the initial terminal force estimation model according to the sample force parameters of the second preset direction in the sample state data and the estimated force parameters of the terminal in the second preset direction corresponding to the sample state data, until the initial terminal force estimation model after parameter adjustment meets the iteration stop condition, thereby obtaining the target terminal force estimation model.

[0017] In a fourth aspect, an embodiment of the present application provides a terminal force estimation device, the device comprising:

[0018] A second acquisition module is used to acquire state data of multiple joints in the mechanical arm to be estimated, and a force parameter of the end of the mechanical arm to be estimated in a first preset direction;

[0019] An estimation module is used to estimate the force parameters of the end of the robotic arm to be estimated in a second preset direction based on state data of multiple joints in the robotic arm to be estimated and force parameters of the end of the robotic arm to be estimated in a first preset direction, and adopt an end force estimation model; wherein the end force estimation model is the target end force estimation model described in any one of the first aspects above.

[0020] In a fifth aspect, an embodiment of the present application provides an electronic device, comprising: a processor and a storage medium, wherein the processor and the storage medium are communicatively connected via a bus, the storage medium stores program instructions executable by the processor, and the processor calls the program stored in the storage medium to execute the steps of the terminal force estimation model training method as described in any one of the first aspects above, or the steps of the terminal force estimation method as described in any one of the second aspects above.

[0021] In a sixth aspect, an embodiment of the present application provides a storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the terminal force estimation model training method as described in any one of the first aspects above, or the steps of the terminal force estimation method as described in any one of the second aspects above are executed.

[0022] Compared with the prior art, this application has the following beneficial effects:

[0023] The present application provides a terminal force estimation model training method, estimation method, device, equipment and medium. By obtaining multiple groups of sample state data of a manipulator in multiple motion states, the sample state data include: sample state data of multiple joints in the manipulator and sample force parameters of the end of the manipulator in multiple preset directions; the multiple preset directions include a first preset direction and a second preset direction; using a preset initial terminal force estimation model, according to the sample state data of multiple joints in the sample state data and the sample force parameters of the end in the first preset direction, predict the estimated force parameters of the end in the second preset direction corresponding to the sample state data; according to the sample force parameters in the second preset direction in the sample state data and the estimated force parameters of the end in the second preset direction corresponding to the sample state data, iteratively adjust the model parameters of the initial terminal force estimation model until the parameter-adjusted initial terminal force estimation model meets the iteration stop condition, and obtain the target terminal force estimation model. Thus, the sample data is more realistic, and the end force of the manipulator can be accurately obtained while reducing costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0025] Figure 1 A flowchart of a terminal force estimation model training method provided in this application;

[0026] Figure 2 A schematic flow chart of a method for obtaining sample force parameters of the end of a robotic arm in multiple directions provided in an embodiment of the present application;

[0027] Figure 3 A schematic diagram of a flow chart of a method for calibrating a sample force parameter provided in an embodiment of the present application;

[0028] Figure 4 A flowchart of a data preprocessing method provided in an embodiment of the present application;

[0029] Figure 5 A flowchart of another terminal force estimation model training method provided in an embodiment of the present application;

[0030] Figure 6 A schematic diagram of a process flow of a terminal force estimation method provided in this application;

[0031] Figure 7A schematic flow chart of another method for obtaining force parameters in a first preset direction provided in an embodiment of the present application;

[0032] Figure 8 A schematic flow chart of a method for calibrating initial force parameters in a first preset direction provided in an embodiment of the present application;

[0033] Fig. 9 A schematic flow chart of a data processing method for estimating terminal force provided in an embodiment of the present application;

[0034] Fig.10 A schematic diagram of a terminal force estimation model training device provided in an embodiment of the present application;

[0035] Fig.11 A schematic diagram of a terminal force estimation device provided in an embodiment of the present application;

[0036] Fig.12 A schematic diagram of an electronic device provided in an embodiment of the present application.

[0037] Icons: 1001 - first acquisition module, 1002 - prediction module, 1003 - adjustment module, 1101 - second acquisition module, 1102 - estimation module, 1201 - processor, 1202 - storage medium. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0039] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0040] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0041] In addition, the terms “first”, “second”, etc., if used, are merely used to distinguish between the descriptions and should not be understood as indicating or implying relative importance.

[0042] It should be noted that, in the absence of conflict, the features in the embodiments of the present invention may be combined with each other.

[0043] In order to accurately estimate the end force of a robotic arm, the present application provides an end force estimation model training method, estimation method, device, equipment and medium.

[0044] Before explaining a terminal force estimation model training method and an estimation method provided in the present application, the terminal force estimation system to which the terminal force estimation model training method and the estimation method are applied is first explained.

[0045] The end force estimation system includes: a processor, a torque sensor, a force sensor, and a robotic arm.

[0046] The processor is connected to the torque sensor, the force sensor and the mechanical arm respectively. The torque sensor is arranged at each joint of the mechanical arm to obtain the torque state data of each joint, and the force sensor is arranged at the end of the mechanical arm.

[0047] The robot arm is equipped with joint torque sensors and end force sensors. The processor can obtain the torque of each joint through the torque sensor, and the end force of the robot arm through the end force sensor. The processor can obtain the state data of the robot arm such as joint angle and angular velocity in different states through the interface of the robot arm itself.

[0048] For example, the robotic arm may be a six-axis robotic arm, and torque sensors are respectively installed on the six joints of the six-axis robotic arm, and a force sensor is installed at the end of the six-axis robotic arm.

[0049] For example, the processor may be an electronic device with computing processing capability disposed at the robotic arm site, or may be an electronic device or server with computing processing capability disposed outside the robotic arm site.

[0050] The following is an explanation of a terminal force estimation model training method provided in this application through a specific example. Figure 1 The present invention provides a flow chart of a method for training a terminal force estimation model. The method may be performed by an electronic device, which may be a device with computing and processing functions, such as a desktop computer, a laptop computer, etc. Figure 1 As shown, the method includes:

[0051] S101, obtaining multiple groups of sample state data of a robotic arm in multiple motion states.

[0052] The sample state data includes: sample state data of multiple joints in the robot arm and sample force parameters of the end of the robot arm in multiple directions; the multiple preset directions include a first preset direction and a second preset direction. The sample state data includes: joint angle, angular velocity, and joint torque.

[0053] For example, the sample force parameters in multiple directions are the end forces in three directions of X (transverse), Y (longitudinal), and Z (vertical), and may also be the end forces in multiple directions under other direction division rules.

[0054] It should be noted that when training the end force estimation model, the force sensor is a multi-axis (dimensional) force sensor. For example, the multi-axis force sensor is a three-axis sensor, which can simultaneously obtain the end force of the end of the robot arm in three directions. Although the three-axis sensor is expensive, it is only used when obtaining sample data for training the end force estimation model.

[0055] For example, when collecting sample data for training the model, it is only necessary to install three-axis sensors on a few robotic arms that collect data. If multiple three-axis sensors are needed, multiple robotic arms can work in parallel to collect data. Alternatively, if time permits, only one three-axis sensor can be used, and multiple robotic arms take turns using the three-axis sensor to collect data. The three-axis sensor is only used in the sample data collection stage, so the number of sensors used is small. After the training is completed, the three-axis sensor is no longer needed, which reduces the cost of end force estimation.

[0056] When the robot arm is working normally, it is in motion and will make multiple sets of actions. Multiple sets of sample state data of the robot arm in multiple motion states are obtained to enrich the sample data. For example, data can be obtained at a fixed frequency (for example, 1000Hz). Therefore, there is no need for complex calibration, the process is simple, and the data is collected from the real robot arm working scene. The data has a certain robustness to the influence of environmental noise such as temperature and friction.

[0057] S102, using a preset initial terminal force estimation model, based on the sample state data of multiple joints in the sample state data and the sample force parameters of the terminal in the first preset direction, predict the estimated force parameters of the terminal in the second preset direction corresponding to the sample state data.

[0058] Among them, the multiple directions are three directions of XYZ, then the first preset direction is any direction of the three directions of XYZ, and the second preset direction is the remaining two directions of the three directions of XYZ.

[0059] The input data of the terminal force estimation model is the detection data of multiple joints and the force of the terminal in the first preset direction, and the output data of the terminal force estimation model is the force of the terminal in the second preset direction. In this way, the force of the terminal in the second preset direction can be estimated based on the detection data of multiple joints and the force of the terminal in the first preset direction.

[0060] For example, the initial end force estimation model may be a fully connected neural network model.

[0061] S103. According to the sample force parameters of the second preset direction in the sample state data and the estimated force parameters of the terminal in the second preset direction corresponding to the sample state data, iteratively adjust the model parameters of the initial terminal force estimation model until the parameter-adjusted initial terminal force estimation model meets the iteration stop condition, thereby obtaining the target terminal force estimation model.

[0062] According to the sample force parameters of the second preset direction in the sample state data and the estimated force parameters of the terminal in the second preset direction corresponding to the sample state data, the loss function of the force parameters in the second preset direction is calculated. If the loss function value is greater than or equal to the preset threshold, the model parameters of the initial terminal force estimation model are adjusted, and the model training model is continued using the sample data. The iteration stop condition may include a preset threshold or a preset number of times. If the loss function value is less than the preset threshold, the model obtained by the current training is used as the target terminal force estimation model. Alternatively, if the number of model training times reaches the preset number, the model obtained by the current training is used as the target terminal force estimation model. So that the model training accuracy meets the preset requirements.

[0063] The loss function may be a MSE (mean square error) function.

[0064] By training the target end force estimation model that meets the accuracy requirements, it is only necessary to collect the detection data of multiple joints and the force parameters of the end in the first preset direction to estimate the accurate force parameters in the second preset direction. The acquisition of the force parameters of the end in the first preset direction only requires a single-axis force sensor, which reduces the acquisition cost compared to a multi-axis force sensor.

[0065] In summary, in this embodiment, by obtaining multiple groups of sample state data of the robot arm in multiple motion states, the sample state data include: sample state data of multiple joints in the robot arm and sample force parameters of the end of the robot arm in multiple preset directions; the multiple preset directions include a first preset direction and a second preset direction; using a preset initial end force estimation model, according to the sample state data of multiple joints in the sample state data and the sample force parameters of the end in the first preset direction, predict the estimated force parameters of the end in the second preset direction corresponding to the sample state data; according to the sample force parameters in the second preset direction in the sample state data and the estimated force parameters of the end in the second preset direction corresponding to the sample state data, iteratively adjust the model parameters of the initial end force estimation model until the parameter-adjusted initial end force estimation model meets the iteration stop condition, and obtain the target end force estimation model. Therefore, the sample data is more realistic, and the end force of the robot arm can be accurately obtained while reducing costs.

[0066] exist Figure 1On the basis of the corresponding embodiments, the embodiments of the present application further provide a method for obtaining sample force parameters of the end of a robotic arm in multiple directions. Figure 2 A flow chart of a method for obtaining sample force parameters of the end of a robotic arm in multiple directions provided in an embodiment of the present application. Figure 2 As shown, in S101, sample force parameters of the end of the robot arm in multiple directions are obtained, including:

[0067] S201. Under multiple motion states of the robot arm, obtain initial sample force parameters of the end of the robot arm in multiple directions.

[0068] This is similar to step S101 in the above embodiment and will not be described again here.

[0069] S202: Calibrate initial sample force parameters of the end of the robotic arm in multiple directions to generate sample force parameters of the end of the robotic arm in multiple directions.

[0070] The force sensor installed at the end of the robot arm will have an offset, which will affect the final model training results. Therefore, after obtaining the initial sample force parameters of the end in multiple directions, the initial sample force parameters are calibrated first to minimize the impact of the offset on the model training results. The calibrated sample force parameters are then used for further training.

[0071] For example, a preset calibration method may be used for calibration, or an empirical value may be used for calibration.

[0072] In summary, in this embodiment, in multiple motion states of the robotic arm, initial sample force parameters of the end of the robotic arm in multiple directions are obtained; the initial sample force parameters of the end of the robotic arm in multiple directions are calibrated to generate sample force parameters of the end of the robotic arm in multiple directions. Thus, accurate sample force parameters of the end in multiple directions are obtained.

[0073] exist Figure 2 On the basis of the corresponding embodiments, the embodiments of the present application also provide a method for calibrating sample force parameters. Figure 3 A flow chart of a method for calibrating a sample force parameter provided in an embodiment of the present application is shown in FIG. Figure 3 As shown, before calibrating the initial sample force parameters of the end of the mechanical arm in multiple directions in S202 to generate the sample force parameters of the end of the mechanical arm in multiple directions, the method further includes:

[0074] S301 , obtaining sample force parameters in multiple directions when the robot arm is stationary and the end is suspended in the air.

[0075] Since the offset of the force sensor is caused by the installation relationship between the force sensor and the robotic arm, and when the robotic arm is stationary and the end is suspended in the air, the end force is 0, and the offset can be measured. Therefore, in order to reduce the impact of the offset on the model training results, the end force of the robotic arm obtained by the force sensor when it is stationary and the end is suspended in the air is used as a reference for calibration.

[0076] For example, a posture is randomly selected, the angle values ​​of each joint in the posture are determined, the angular velocity is 0, and the end of the robot arm is not in contact with external objects. At this time, the force parameters measured by the force sensor are read. Theoretically, the end force in multiple directions should be 0 at this time, but in reality it is not 0 due to the offset; therefore, the force parameter measured by the force sensor at this time is the offset value.

[0077] Furthermore, in S202, the initial sample force parameters of the end of the robot arm in multiple directions are calibrated to generate sample force parameters of the end of the robot arm in multiple directions, including:

[0078] S302 , calibrating initial sample force parameters of the end of the robotic arm in multiple directions according to sample force parameters of the end of the robotic arm in multiple directions when the robotic arm is stationary and the end is suspended in the air, and generating sample force parameters of the end of the robotic arm in multiple directions.

[0079] Based on the initial sample force parameters of the end of the robotic arm in multiple directions, the sample force parameters of the end of the robotic arm in multiple directions when the robotic arm is stationary and the end is suspended are subtracted to generate the sample force parameters of the end of the robotic arm in multiple directions, thereby eliminating the offset and realizing the calibration of the sample force parameters.

[0080] In summary, in this embodiment, sample force parameters in multiple directions of the robot arm are obtained when the robot arm is stationary and the end is suspended in the air; according to the sample force parameters in multiple directions of the robot arm when the robot arm is stationary and the end is suspended in the air, the initial sample force parameters of the end of the robot arm in multiple directions are calibrated to generate sample force parameters of the end of the robot arm in multiple directions, thereby eliminating the offset and achieving calibration of the sample force parameters.

[0081] exist Figure 1 On the basis of the corresponding embodiments, the embodiments of the present application also provide a data preprocessing method. Figure 4 A flow chart of a data preprocessing method provided in an embodiment of the present application. Figure 4 As shown, after obtaining multiple groups of sample state data of the robot arm in multiple motion states in S101, the method further includes:

[0082] S401, performing feature analysis on multiple groups of sample state data to obtain feature analysis results.

[0083] When obtaining multiple sets of sample status data, the data may be abnormal or missing due to hardware or software reasons. In order to make the training results more accurate, the data needs to be preprocessed to make the data more reasonable.

[0084] The feature analysis result may represent normal data features. For example, the feature analysis result may be a visualized angle change curve or the like.

[0085] S402: Preprocess multiple groups of sample status data according to the feature analysis results.

[0086] According to the feature analysis results, the abnormal data in the multiple groups of sample status data are eliminated and the missing data are supplemented. For example, the averaging method can be used to supplement the missing data.

[0087] In summary, in this embodiment, feature analysis is performed on multiple groups of sample state data to obtain feature analysis results, and preprocessing is performed on multiple groups of sample state data according to the feature analysis results, thereby obtaining more accurate sample state data.

[0088] In another embodiment of the present application, after acquiring multiple groups of sample state data of the robot arm in multiple motion states in S101, the method further includes:

[0089] Normalize multiple groups of sample status data.

[0090] In order to make the model calculation more accurate, multiple groups of sample state data are normalized according to the data range of each feature, and the data are dispersed to meet the independent and identically distributed requirements.

[0091] In summary, in this embodiment, multiple groups of sample state data are normalized, thereby improving the model calculation accuracy.

[0092] exist Figure 1 Based on the corresponding embodiments, the embodiments of the present application also provide another method for training a terminal force estimation model. Figure 5 A flowchart of another terminal force estimation model training method provided in an embodiment of the present application.

[0093] The multiple groups of sample state data include: a training set and a test set. For example, 70% of the sample state data are used as the training set, and 30% of the sample state data are used as the test set.

[0094] like Figure 5 As shown, in S102, a preset initial terminal force estimation model is used to predict the estimated force parameters of the terminal in the second preset direction corresponding to the sample state data according to the sample state data of multiple joints and the sample force parameters of the terminal in the first preset direction, including:

[0095] S501, using a preset initial terminal force estimation model, based on sample state data of multiple joints in a training set and sample force parameters of the terminal in a first preset direction, predict the estimated force parameters of the terminal in a second preset direction corresponding to the training set.

[0096] The contents are similar to those in the above step S102 and will not be repeated here.

[0097] Further, in S103, according to the sample force parameter of the second preset direction in the sample state data and the estimated force parameter of the second preset direction corresponding to the sample state data, the model parameters of the initial terminal force estimation model are adjusted to obtain the target terminal force estimation model, including:

[0098] S502 . According to the sample force parameters of the second preset direction in the training set and the estimated force parameters of the second preset direction corresponding to the training set, the model parameters of the initial terminal force estimation model are adjusted to obtain the terminal force estimation model to be tested.

[0099] The content is similar to that in the above step S103 and will not be repeated here.

[0100] S503: Use the test set to test the end force estimation model to be tested to obtain a test result.

[0101] During the model training, the test set is used to test the end force estimation model to be tested, and the test results are obtained. The test results characterize the estimation accuracy of the model to test the degree of model training.

[0102] S504 , adjusting the model parameters of the terminal force estimation model to be tested according to the test results to obtain a target terminal force estimation model.

[0103] If the test result indicates that the estimation accuracy is greater than or equal to the preset accuracy value, the model parameters of the terminal force estimation model to be tested are adjusted, and the training of the terminal force estimation model to be tested continues.

[0104] If the test result indicates that the estimation accuracy is less than the preset accuracy value, the terminal force estimation model to be tested is determined to be the target terminal force estimation model.

[0105] In summary, in this embodiment, the multiple groups of sample state data include: a training set and a test set; using a preset initial terminal force estimation model, according to the sample state data of multiple joints in the training set and the sample force parameters of the terminal in the first preset direction, predict the estimated force parameters of the terminal corresponding to the training set in the second preset direction; according to the sample force parameters in the second preset direction in the training set and the estimated force parameters in the second preset direction corresponding to the training set, adjust the model parameters of the initial terminal force estimation model to obtain the terminal force estimation model to be tested; using the test set to test the terminal force estimation model to obtain the test results; adjusting the model parameters of the terminal force estimation model to be tested according to the test results to obtain the target terminal force estimation model. Thus, the target terminal force estimation model is accurately trained and obtained.

[0106] A terminal force estimation method provided in the present application is explained below through a specific example. Figure 6 This is a flow chart of a method for estimating the end force provided by the present application. The execution subject of the method may be an electronic device, which may be a device with computing and processing functions, such as a desktop computer, a laptop computer, etc. Figure 6 As shown, the method includes:

[0107] S601, obtaining state data of a plurality of joints in a to-be-estimated robotic arm and force parameters of a terminal end of the to-be-estimated robotic arm in a first preset direction.

[0108] It should be noted that when estimating the end force, the force sensor is a single-axis sensor, which can obtain the end force of the end of the robotic arm in the first preset direction. The cost of a single-axis sensor is lower than that of a multi-axis sensor, which reduces the cost of end force estimation.

[0109] The state data includes: joint angle, angular velocity, and joint torque. The joint angle and angular velocity transmitted by the manipulator are obtained, the joint torque transmitted by the torque sensor is obtained, and the force parameter of the first preset direction transmitted by the uniaxial force sensor is obtained.

[0110] S602. According to the state data of multiple joints in the robotic arm to be estimated and the force parameters of the end of the robotic arm to be estimated in a first preset direction, and using an end force estimation model, estimate the force parameters of the end of the robotic arm to be estimated in a second preset direction.

[0111] The terminal force estimation model is any target terminal force estimation model in the above embodiments.

[0112] By using the trained end force estimation model, we only need to substitute the state data of multiple joints in the robotic arm to be estimated and the force parameters of the end of the robotic arm to be estimated in the first preset direction, and we can accurately obtain the force parameters of the end of the robotic arm to be estimated in the second preset direction.

[0113] The force parameter of the end of the to-be-estimated mechanical arm in the first preset direction is combined with the force parameter of the second preset direction to obtain the end force of the to-be-estimated mechanical arm, thereby accurately obtaining the end force of the mechanical arm while reducing the cost.

[0114] In summary, in this embodiment, the state data of multiple joints in the mechanical arm to be estimated and the force parameters of the end of the mechanical arm to be estimated in the first preset direction are obtained; according to the state data of multiple joints in the mechanical arm to be estimated and the force parameters of the end of the mechanical arm to be estimated in the first preset direction, the force parameters of the end of the mechanical arm to be estimated in the second preset direction are estimated by using the end force estimation model; wherein the end force estimation model is any target end force estimation model in the above embodiments. Thus, the end force of the mechanical arm can be accurately obtained while reducing the cost.

[0115] exist Figure 6 On the basis of the corresponding embodiment, the embodiment of the present application also provides another method for obtaining the force parameters in the first preset direction. Figure 7 A flow chart of another method for obtaining force parameters in a first preset direction provided in an embodiment of the present application. Figure 7 As shown, obtaining the force parameter of the end of the to-be-estimated mechanical arm in the first preset direction in S601 includes:

[0116] S701, obtaining initial force parameters of the end of the robot arm to be estimated in a first preset direction.

[0117] This is similar to step S601 in the above embodiment and will not be described again here.

[0118] S702: Calibrate the initial force parameters of the first preset direction to generate the force parameters of the first preset direction.

[0119] The force sensor installed at the end of the robotic arm will have an offset, and the model obtained by the above training uses calibrated data. Therefore, after obtaining the force parameters of the end in the first preset direction, the force parameters of the first preset direction are first calibrated. For example, the calibration can be performed using a preset calibration method or using empirical values.

[0120] In summary, in this embodiment, the initial force parameters of the end of the mechanical arm to be estimated in the first preset direction are obtained, and the initial force parameters in the first preset direction are calibrated to generate the force parameters in the first preset direction. Thus, accurate force parameters in the first preset direction are obtained.

[0121] exist Figure 7 On the basis of the corresponding embodiment, the embodiment of the present application also provides a method for calibrating the initial force parameters in a first preset direction. Figure 8A schematic diagram of a flow chart of a method for calibrating initial force parameters in a first preset direction provided in an embodiment of the present application. Figure 8 As shown, before calibrating the initial force parameter of the first preset direction and generating the force parameter of the first preset direction in S702, the method further includes:

[0122] S801. Obtain force parameters in a first preset direction of a robot arm to be estimated when the robot arm is stationary and the end is suspended in the air.

[0123] Since the offset of the force sensor is caused by the installation relationship between the force sensor and the robotic arm, and when the robotic arm is stationary and the end is suspended in the air, the end force is 0, and the offset can be measured. Therefore, in order to reduce the impact of the offset on the model training results, the end force of the robotic arm obtained by the force sensor when it is stationary and the end is suspended in the air is used as a reference for calibration.

[0124] For example, a posture is randomly selected, the angle values ​​of each joint in the posture are determined, the angular velocity is 0, and the end of the robot arm is not in contact with external objects. At this time, the force parameter measured by the force sensor is read. Theoretically, the end force in the first preset direction should be 0 at this time, but in reality it is not 0 due to the offset; therefore, the force parameter measured by the force sensor at this time is the offset value.

[0125] Furthermore, the initial force parameter of the first preset direction is calibrated in S702 to generate the force parameter of the first preset direction, including:

[0126] S802. According to the force parameters of the first preset direction to be estimated when the robot arm is stationary and the end is suspended in the air, calibrate the initial force parameters of the first preset direction to generate the force parameters of the first preset direction.

[0127] Based on the initial force parameters in the first preset direction, the force parameters in the first preset direction when the mechanical arm to be estimated is in a stationary state with the end suspended in the air are subtracted to generate the force parameters in the first preset direction, thereby eliminating the offset and achieving the force parameter calibration in the first preset direction.

[0128] In summary, in this embodiment, the force parameters of the first preset direction of the mechanical arm to be estimated are obtained when the mechanical arm is stationary and the end is suspended in the air; according to the force parameters of the first preset direction of the mechanical arm to be estimated when the mechanical arm is stationary and the end is suspended in the air, the initial force parameters of the first preset direction are calibrated to generate the force parameters of the first preset direction. Thus, the offset is eliminated and the force parameters of the first preset direction are calibrated.

[0129] exist Figure 6 On the basis of the corresponding embodiments, the embodiments of the present application also provide a data processing method for estimating end force. Fig. 9 A flow chart of a data processing method for estimating terminal force provided in an embodiment of the present application. Fig. 9 As shown, after obtaining the state data of multiple joints in the to-be-estimated mechanical arm and the force parameters of the end of the to-be-estimated mechanical arm in the first preset direction in S601, the method further includes:

[0130] S901 , normalizing state data of a plurality of joints in a robot arm to be estimated and force parameters of a terminal of the robot arm to be estimated in a first preset direction.

[0131] In order to make the model calculation more accurate, multiple groups of sample state data are normalized according to the data range of each feature. In addition, the above training model uses normalized data, so when using the model, the data also needs to be normalized.

[0132] Further, in S602, according to the state data of multiple joints in the mechanical arm to be estimated, and the force parameters of the end of the mechanical arm to be estimated in the first preset direction, and using the end force estimation model, after estimating the force parameters of the end of the mechanical arm to be estimated in the second preset direction, the method further includes:

[0133] S902: Perform a denormalization process on the force parameters of the end of the robot arm to be estimated in the second preset direction.

[0134] After the force parameters of the end of the to-be-estimated mechanical arm in the second preset direction are obtained, an inverse normalization process is performed to make the force parameters of the end in the second preset direction more visualized.

[0135] In summary, in this embodiment, the state data of multiple joints in the estimated mechanical arm and the force parameters of the end of the estimated mechanical arm in the first preset direction are normalized, and the force parameters of the end of the estimated mechanical arm in the second preset direction are denormalized. Thus, the end force is accurately estimated.

[0136] The following is a description of an end force estimation model training device, equipment, storage medium, etc. provided in the present application for execution. The specific implementation process and technical effects are described above and will not be repeated below.

[0137] Fig.10 A schematic diagram of a terminal force estimation model training device provided in an embodiment of the present application is shown in FIG. Fig.10 As shown, the device comprises:

[0138] The first acquisition module 1001 is used to obtain multiple groups of sample state data of the robotic arm in multiple motion states, and the sample state data include: sample state data of multiple joints in the robotic arm and sample force parameters of the end of the robotic arm in multiple directions; multiple preset directions include a first preset direction and a second preset direction.

[0139] Prediction module 1002 is used to use a preset initial terminal force estimation model to predict the estimated force parameters of the terminal in the second preset direction corresponding to the sample state data based on the sample state data of multiple joints and the sample force parameters of the terminal in the first preset direction.

[0140] The adjustment module 1003 is used to iteratively adjust the model parameters of the initial terminal force estimation model according to the sample force parameters of the second preset direction in the sample state data and the estimated force parameters of the terminal in the second preset direction corresponding to the sample state data, until the initial terminal force estimation model after parameter adjustment meets the iteration stop condition, thereby obtaining the target terminal force estimation model.

[0141] Furthermore, the first acquisition module 1001 is specifically used to obtain initial sample force parameters of the end of the robotic arm in multiple directions under multiple motion states of the robotic arm; calibrate the initial sample force parameters of the end of the robotic arm in multiple directions to generate sample force parameters of the end of the robotic arm in multiple directions.

[0142] Furthermore, the first acquisition module 1001 is specifically used to obtain sample force parameters in multiple directions when the robot arm is stationary with the end suspended in the air; based on the sample force parameters in multiple directions when the robot arm is stationary with the end suspended in the air, the initial sample force parameters of the end of the robot arm in multiple directions are calibrated to generate sample force parameters of the end of the robot arm in multiple directions.

[0143] Furthermore, the first acquisition module 1001 is specifically used to perform feature analysis on the multiple groups of sample state data to obtain feature analysis results; and pre-process the multiple groups of sample state data according to the feature analysis results.

[0144] Furthermore, the first acquisition module 1001 is specifically used to perform normalization processing on multiple groups of sample state data.

[0145] Further, the prediction module 1002 is specifically used for multiple groups of sample state data including: a training set and a test set; using a preset initial terminal force estimation model, according to the sample state data of multiple joints in the training set and the sample force parameters of the terminal in the first preset direction, predict the estimated force parameters of the terminal in the second preset direction corresponding to the training set;

[0146] Furthermore, the adjustment module 1003 is specifically used to adjust the model parameters of the initial terminal force estimation model according to the sample force parameters of the second preset direction in the training set and the estimated force parameters of the second preset direction corresponding to the training set, so as to obtain the terminal force estimation model to be tested; use the test set to test the terminal force estimation model to be tested to obtain the test results; adjust the model parameters of the terminal force estimation model to be tested according to the test results to obtain the target terminal force estimation model.

[0147] The following is a description of an end force estimation device, equipment, storage medium, etc. provided by the present application for execution. The specific implementation process and technical effects are described above and will not be repeated below.

[0148] Fig.11 A schematic diagram of a terminal force estimation device provided in an embodiment of the present application is shown in FIG. Fig.11 As shown, the device comprises:

[0149] The second acquisition module 1101 is used to acquire state data of multiple joints in the mechanical arm to be estimated, and force parameters of the end of the mechanical arm to be estimated in a first preset direction;

[0150] The estimation module 1102 is used to estimate the force parameters of the end of the robotic arm to be estimated in a second preset direction based on the state data of multiple joints in the robotic arm to be estimated and the force parameters of the end of the robotic arm to be estimated in a first preset direction, and adopt an end force estimation model; wherein the end force estimation model is the target end force estimation model of any of the above-mentioned embodiments.

[0151] Furthermore, the second acquisition module 1101 is used to acquire initial force parameters of the end of the robot arm to be estimated in a first preset direction; calibrate the initial force parameters in the first preset direction to generate force parameters in the first preset direction.

[0152] Furthermore, the second acquisition module 1101 is specifically used to obtain the force parameters of the first preset direction of the robotic arm to be estimated when it is in a stationary state with its end suspended in the air; based on the force parameters of the first preset direction of the robotic arm to be estimated when it is in a stationary state with its end suspended in the air, the initial force parameters of the first preset direction are calibrated to generate the force parameters of the first preset direction.

[0153] Furthermore, the second acquisition module 1101 is specifically used to normalize the state data of multiple joints in the robot arm to be estimated, and the force parameters of the end of the robot arm to be estimated in the first preset direction.

[0154] Furthermore, the estimation module 1102 is specifically configured to perform a denormalization process on the force parameter of the end of the manipulator to be estimated in the second preset direction.

[0155] Fig.12 A schematic diagram of an electronic device provided in an embodiment of the present application, which electronic device may be a device with computing and processing functions.

[0156] The electronic device includes: a processor 1201 and a storage medium 1202. The processor 1201 and the storage medium 1202 are connected via a bus.

[0157] The storage medium 1202 is used to store programs, and the processor 1201 calls the programs stored in the storage medium 1202 to execute the above method embodiment. The specific implementation method and technical effect are similar and will not be repeated here.

[0158] Optionally, the present invention also provides a storage medium, including a program, which is used to execute the above-mentioned method embodiment when executed by a processor. In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0159] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0160] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0161] The above-mentioned integrated unit implemented in the form of a software functional unit can be stored in a storage medium. The above-mentioned software functional unit is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (English: processor) to perform some steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (English: Read-Only Memory, abbreviated: ROM), random access memory (English: Random Access Memory, abbreviated: RAM), disk or optical disk and other media that can store program codes.

Claims

1. A terminal force estimation model training method, characterized in that: The method comprises: Acquire multiple groups of sample state data of the robot arm in multiple motion states, the sample state data including: sample state data of multiple joints in the robot arm and sample force parameters of the end of the robot arm in multiple preset directions; the multiple preset directions include a first preset direction and a second preset direction; Using a preset initial terminal force estimation model, based on the sample state data of the multiple joints in the sample state data and the sample force parameters of the terminal in the first preset direction, predict an estimated force parameter of the terminal in the second preset direction corresponding to the sample state data; According to the sample force parameters of the second preset direction in the sample state data and the estimated force parameters of the terminal in the second preset direction corresponding to the sample state data, the model parameters of the initial terminal force estimation model are iteratively adjusted until the initial terminal force estimation model after parameter adjustment meets the iteration stop condition, thereby obtaining the target terminal force estimation model.

2. The method according to claim 1, characterized in that The obtaining of sample force parameters of the end of the robotic arm in multiple directions includes: Under multiple motion states of the robotic arm, obtaining initial sample force parameters of the end of the robotic arm in multiple directions; Initial sample force parameters of the end of the mechanical arm in multiple directions are calibrated to generate sample force parameters of the end of the mechanical arm in multiple directions.

3. The method according to claim 2, characterized in that Before calibrating the initial sample force parameters of the end of the mechanical arm in multiple directions to generate the sample force parameters of the end of the mechanical arm in multiple directions, the method further includes: Obtaining sample force parameters in multiple directions of the robotic arm when the robotic arm is stationary and the end is suspended in the air; The calibrating the initial sample force parameters of the end of the mechanical arm in multiple directions to generate the sample force parameters of the end of the mechanical arm in multiple directions includes: According to the sample force parameters in multiple directions when the robot arm is stationary and the end is suspended, the initial sample force parameters in multiple directions of the end of the robot arm are calibrated to generate the sample force parameters in multiple directions of the end of the robot arm.

4. The method according to claim 1, characterized in that: After acquiring multiple groups of sample state data of the robot arm in multiple motion states, the method further includes: Performing feature analysis on the multiple groups of sample state data to obtain feature analysis results; The multiple groups of sample state data are preprocessed according to the feature analysis results.

5. The method according to claim 1, characterized in that: After acquiring multiple groups of sample state data of the robot arm in multiple motion states, the method further includes: The multiple groups of sample state data are normalized.

6. The method according to claim 1, characterized in that The multiple groups of sample state data include: a training set and a test set; The method adopts a preset initial terminal force estimation model, predicts the estimated force parameter of the terminal in the second preset direction corresponding to the sample state data according to the sample state data of the multiple joints and the sample force parameter of the terminal in the first preset direction, including: Using the preset initial terminal force estimation model, based on the sample state data of the multiple joints in the training set and the sample force parameters of the terminal in the first preset direction, the estimated force parameters of the terminal in the second preset direction corresponding to the training set are predicted; The method of adjusting the model parameters of the initial terminal force estimation model according to the sample force parameters of the second preset direction in the sample state data and the estimated force parameters of the second preset direction corresponding to the sample state data to obtain the target terminal force estimation model includes: According to the sample force parameters of the second preset direction in the training set and the estimated force parameters of the second preset direction corresponding to the training set, the model parameters of the initial terminal force estimation model are adjusted to obtain the terminal force estimation model to be tested; Using the test set to test the end force estimation model to be tested to obtain a test result; The model parameters of the terminal force estimation model to be tested are adjusted according to the test results to obtain the target terminal force estimation model.

7. A terminal force estimation method, characterized in that: The method comprises: Acquire state data of a plurality of joints in the to-be-estimated mechanical arm, and force parameters of the end of the to-be-estimated mechanical arm in a first preset direction; Based on the state data of multiple joints in the robotic arm to be estimated, and the force parameters of the end of the robotic arm to be estimated in a first preset direction, and using an end force estimation model, the force parameters of the end of the robotic arm to be estimated in a second preset direction are estimated; wherein the end force estimation model is the target end force estimation model described in any one of claims 1-6.

8. The method according to claim 7, characterized in that The step of obtaining the force parameter of the end of the to-be-estimated mechanical arm in the first preset direction includes: Obtaining initial force parameters of the end of the to-be-estimated robotic arm in a first preset direction; The initial force parameter of the first preset direction is calibrated to generate the force parameter of the first preset direction.

9. The method according to claim 8, characterized in that Before calibrating the initial force parameter of the first preset direction to generate the force parameter of the first preset direction, the method further includes: Obtaining force parameters in a first preset direction of the to-be-estimated robotic arm when the robotic arm is stationary and the end is suspended in the air; The step of calibrating the initial force parameter of the first preset direction to generate the force parameter of the first preset direction includes: According to the force parameters of the first preset direction of the to-be-estimated robotic arm in a stationary state with its end suspended in the air, the initial force parameters of the first preset direction are calibrated to generate the force parameters of the first preset direction.

10. The method according to claim 7, characterized in that After acquiring the state data of a plurality of joints in the to-be-estimated mechanical arm and the force parameter of the end of the to-be-estimated mechanical arm in the first preset direction, the method further comprises: Normalizing state data of multiple joints in the to-be-estimated mechanical arm and force parameters of the end of the to-be-estimated mechanical arm in a first preset direction; After estimating the force parameter of the end of the to-be-estimated mechanical arm in a second preset direction based on the state data of the plurality of joints in the to-be-estimated mechanical arm and the force parameter of the end of the to-be-estimated mechanical arm in a first preset direction and using the end force estimation model, the method further comprises: The force parameters of the end of the to-be-estimated mechanical arm in the second preset direction are subjected to inverse normalization processing.

11. A terminal force estimation model training device, characterized in that: The device comprises: A first acquisition module is used to acquire multiple groups of sample state data of the robot arm in multiple motion states, the sample state data including: sample state data of multiple joints in the robot arm and sample force parameters of the end of the robot arm in multiple preset directions; the multiple preset directions include a first preset direction and a second preset direction; A prediction module, configured to use a preset initial terminal force estimation model to predict, based on the sample state data of the plurality of joints in the sample state data and the sample force parameters of the terminal in the first preset direction, an estimated force parameter of the terminal in the second preset direction corresponding to the sample state data; An adjustment module is used to iteratively adjust the model parameters of the initial terminal force estimation model according to the sample force parameters of the second preset direction in the sample state data and the estimated force parameters of the terminal in the second preset direction corresponding to the sample state data, until the initial terminal force estimation model after parameter adjustment meets the iteration stop condition, thereby obtaining the target terminal force estimation model.

12. A terminal force estimation device, characterized in that: The device comprises: A second acquisition module is used to acquire state data of multiple joints in the mechanical arm to be estimated, and a force parameter of the end of the mechanical arm to be estimated in a first preset direction; An estimation module is used to estimate the force parameters of the end of the robotic arm to be estimated in a second preset direction based on state data of multiple joints in the robotic arm to be estimated and force parameters of the end of the robotic arm to be estimated in a first preset direction, and using an end force estimation model; wherein the end force estimation model is the target end force estimation model described in any one of claims 1-6.

13. An electronic device, characterized in that: include: A processor and a storage medium, wherein the processor and the storage medium are communicatively connected via a bus, the storage medium stores program instructions executable by the processor, and the processor calls the program stored in the storage medium to execute the steps of the terminal force estimation model training method as described in any one of claims 1 to 6, or the steps of the terminal force estimation method as described in any one of claims 7 to 10.

14. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the terminal force estimation model training method as described in any one of claims 1 to 6, or the steps of the terminal force estimation method as described in any one of claims 7 to 10 are executed.

Images