Force calibration method for integrated parallel platform

Abstract

The present invention relates to the field of force measurement platform calibration, and in particular to a force calibration method for an integrated parallel platform, including S1: setting the full scale of a force sensor when the integrated parallel platform is in different positions; S2: performing a calibration test of six-dimensional force loading and unloading for each position of the integrated parallel platform to obtain test data; wherein the test data includes the theoretical calculated stiffness of the integrated parallel platform in each position and the measured force of the force sensor; S3: constructing and training a neural network model based on the test data; S4: adjusting the integrated parallel platform to the position to be calibrated, inputting the stiffness of the integrated parallel platform in this position and the measured force of the force sensor into the neural network model, and obtaining the loading force of the integrated parallel platform in this position. The present invention utilizes the self-learning function of the neural network, introduces the position stiffness for data training, and realizes the calibration of the force sensor under the variable configuration.

Description

Technical Field

[0001] The invention belongs to the technical field of force measurement platform calibration, and in particular relates to a force calibration method for an integrated parallel platform. Background Art

[0002] The on-orbit service of a space manipulator is a high-investment and high-risk task. The most important way to reduce the risk is to conduct sufficient ground simulation tests before the mission is carried out. Currently, the most widely used method of ground simulation is semi-physical test simulation.

[0003] As my country's space station enters the on-orbit operation period, the large robotic arm on the Tianhe experimental module and the small robotic arm on the Wentian experimental module are important tools for on-orbit service during the operation period. As the size of on-orbit maintenance products increases, the weight and volume of their maintenance components also increase, and the corresponding maintenance interfaces become larger and more complex. At present, the 2m-aperture space telescope (space station optical test module) being developed by my country has entered the prototype stage, and its back-end module as an integral maintenance unit has reached a weight of 500kg and a volume envelope of 925×880×1300mm. In the future, my country's 10m-aperture on-orbit assembly telescope will have a further increase in the volume and weight of its on-orbit assembly module unit, which puts forward greater load-bearing capacity requirements for the ground semi-physical simulation system of the space robotic arm. An integrated parallel platform with an ultra-large force measurement platform and large load-bearing capacity has emerged. However, the traditional force calibration method can only realize the force sensor calibration of the parallel platform under a certain configuration. When the force sensor configuration changes, the stiffness of the force sensor changes, and the decoupling formula under the original configuration is no longer applicable to the new configuration, resulting in the inability to realize the force sensor calibration of the parallel platform under the variable configuration. Summary of the invention

[0004] In view of this, the present invention aims to provide a force calibration method for an integrated parallel platform to solve the technical problem that the traditional force calibration method cannot realize the force sensor calibration of the parallel platform under variable configuration.

[0005] To achieve the above object, the technical solution created by the present invention is implemented as follows:

[0006] A force calibration method for an integrated parallel platform, the integrated parallel platform comprises a force measuring platform, a base and a leg connected between the force measuring platform and the base, the leg comprises a first hinge, a second hinge, a force sensor and a telescopic assembly, the fixed end of the telescopic assembly is connected to the base through the first hinge, the output end of the telescopic assembly is connected to the force measuring platform through the second hinge, the force sensor is connected in series between the second hinge and the telescopic assembly, and the force calibration method comprises the following steps:

[0007] S1: Set the full scale of the force sensor in different positions of the integrated parallel platform;

[0008] S2: Perform six-dimensional force loading and unloading calibration tests on each posture of the integrated parallel platform to obtain test data; wherein the test data includes the stiffness of the integrated parallel platform in each posture and the measured force of the force sensor;

[0009] S3: Build and train a neural network model based on experimental data;

[0010] S4: Adjust the integrated parallel platform to the posture to be calibrated, input the stiffness of the integrated parallel platform in this posture and the measured force of the force sensor into the neural network model, and obtain the loading force of the integrated parallel platform in this posture.

[0011] Further, the number of the legs is six, and the six legs are arranged in a Stewart configuration.

[0012] Furthermore, the number of the legs is seven, six of which are arranged in a Stewart configuration, and the other leg is connected at the center of the force measuring platform and the base.

[0013] Furthermore, the process of building and training the neural network model is as follows:

[0014] S31: constructing a sample set based on the experimental data, and dividing the sample set into a training set, a validation set, and a test set according to a predetermined ratio;

[0015] S32: construct a neural network model, initialize all parameters in the neural network model, and train the neural network model using a training set;

[0016] S33: Use the validation set to validate the trained neural network model;

[0017] S34: Use the test set to test the verified neural network model.

[0018] Furthermore, the neural network model adopts a BP neural network model or a RBF neural network model.

[0019] Furthermore, the stiffness calculation method of the integrated parallel platform in each posture is as follows:

[0020] The mapping relationship between the force and deformation of the force measuring platform in six-dimensional directions is:

[0021] (1);

[0022] in, represents the generalized force acting on the force measuring platform, represents the virtual displacement of the force measuring platform, represents the stiffness of the integrated parallel platform;

[0023] According to Hooke's law, the mapping relationship between the force and deformation of the outrigger is:

[0024] (2);

[0025] in, represents the matrix composed of the stiffness of each leg, is the virtual displacement of each leg;

[0026] Derive the relationship between the deformation of the force measuring platform and the deformation of the legs:

[0027] (3);

[0028] in, Represents the velocity Jacobian matrix between the force platform and the outrigger.

[0029] Based on the principle of virtual work, the following relationship is obtained:

[0030] (4);

[0031] Substituting formulas (1) to (3) into (4), we can obtain the relationship between the stiffness of the integrated parallel platform, the stiffness of the legs, and the current posture of the integrated parallel platform:

[0032] (5);

[0033] According to formula (5), the stiffness matrix of the integrated parallel platform in the current position is obtained.

[0034] Compared with the prior art, the invention can achieve the following beneficial effects:

[0035] 1. By using the self-learning function of the neural network and introducing the posture stiffness information for data training, the problem that the traditional force calibration method cannot realize the calibration of the variable configuration force sensor is solved;

[0036] 2. Based on the six legs of the stewart configuration, a redundant leg is added between the force measuring platform and the center of the base. The redundant leg is used to provide preload force, eliminate the gap between hinges and other mechanisms, and improve the accuracy of the integrated parallel platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The drawings constituting part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments and descriptions of the present invention are used to explain the present invention and do not constitute an improper limitation on the present invention. In the drawings:

[0038] Figure 1It is a structural schematic diagram of an integrated parallel platform according to an embodiment of the invention;

[0039] Figure 2 is a schematic structural diagram of a support leg according to an embodiment of the present invention;

[0040] Figure 3 It is a flow chart of the force calibration method according to an embodiment of the invention.

[0041] Explanation of the reference numerals: force measuring platform 1, leg 2, base 3, first hinge 2-1, second hinge 2-2, force sensor 2-3, telescopic assembly 2-4. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the invention more clear, the invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described here are only used to explain the invention and do not constitute a limitation of the invention.

[0043] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0044] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0045] In the description of the invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the invention can be understood according to specific circumstances.

[0046] The following will refer to Figure 1-Figure 3 The invention is described in detail with reference to the embodiments.

[0047] The present invention provides a force calibration method for calibrating an integrated parallel platform, such as Figure 1 As shown, the structure of the integrated parallel platform includes a force measuring platform 1, a base 3, and a leg 2 connected between the force measuring platform 1 and the base 3. Figure 2 As shown, the structure of the leg 2 includes a first hinge 2-1, a second hinge 2-2, a force sensor 2-3 and a telescopic assembly 2-4. The fixed end of the telescopic assembly 2-4 is connected to the base 3 through the first hinge 2-1, and the output end of the telescopic assembly 2-4 is connected to the force measuring platform 1 through the second hinge 2-2. The telescopic assembly 2-4 is used to adjust the length of the leg 2, change the posture of the force measuring platform 1, and thus change the posture of the integrated parallel platform. Each posture of the integrated parallel platform is called a configuration; the force sensor 2-3 is connected in series between the second hinge 2-2 and the telescopic assembly 2-4, and is used to collect force information of the force measuring platform 1.

[0048] The integrated parallel platform provided by the present invention can adopt six legs arranged in a Stewart configuration, and a force sensor 2-3 is provided on each leg. The six force sensors 2-3 realize the six-dimensional force sensing function of the integrated parallel platform.

[0049] In order to eliminate the gap of mechanisms such as hinges, based on the six legs of the stewart configuration, a redundant leg is added between the center position of the force measuring platform 1 and the center position of the base 3. The preload force provided by the redundant leg is used to eliminate the gap of mechanisms such as hinges, thereby improving the force measurement accuracy of the integrated parallel platform.

[0050] The above content details the structure of the integrated parallel platform provided by the present invention. The force calibration method of the integrated parallel platform will be described in detail below.

[0051] like Figure 3 As shown, the force calibration method of the integrated parallel platform includes the following steps:

[0052] S1: Set the full scale of the force sensor in different positions on the integrated parallel platform.

[0053] Each posture of the integrated parallel platform is determined according to the calibration accuracy, and the full scale of the force sensor is set for each posture.

[0054] S2: Carry out six-dimensional force loading and unloading calibration tests on each posture of the integrated parallel platform to obtain test data.

[0055] The test data include the stiffness of the integrated parallel platform in each position and the measured force of the force sensor.

[0056] The traditional force calibration method calibrates the sensor for a certain configuration through the decoupling formula under the configuration. When the configuration of the force sensor changes, the stiffness of the force sensor changes, and the decoupling formula under the original configuration is no longer applicable to the new configuration, resulting in the inability of the traditional force calibration method to realize the force sensor calibration of the parallel platform under the variable configuration. In order to solve this problem, the present invention introduces the posture stiffness information and trains the neural network model to realize the calibration of the force sensor under different configurations.

[0057] S3: Build and train a neural network model based on experimental data.

[0058] The process of building and training a neural network model is:

[0059] S31: construct a sample set based on the experimental data, and divide the sample set into a training set, a validation set, and a test set according to a predetermined ratio.

[0060] S32: Build a neural network model, initialize all parameters in the neural network model, and train the neural network model using a training set.

[0061] The stiffness of the integrated parallel platform in each position and the measured force of the force sensor are used as the input of the neural network model, and the loading force of the integrated parallel platform in each position is used as the output of the neural network model.

[0062] The stiffness calculation method of the integrated parallel platform in each posture is as follows:

[0063] The stiffness characteristics of the force sensor 2-3 can be determined by the force or torque required to produce a unit deformation of the end. Therefore, the mapping relationship between the force and deformation of the force measuring platform 1 in the six-dimensional direction is described as follows:

[0064] (1);

[0065] in, represents the generalized force acting on the force measuring platform 1, represents the virtual displacement of the force measuring platform 1, Represents the stiffness of the integrated parallel platform.

[0066] According to Hooke's law, the mapping relationship between the force and deformation of leg 2 is:

[0067] (2);

[0068] in, represents the matrix composed of the stiffness of each leg 2, is the virtual displacement of each leg 2.

[0069] Using the knowledge related to the kinematics of the parallel platform, the relationship between the deformation of the force measuring platform 1 and the deformation of the leg 2 can be derived:

[0070] (3);

[0071] in, Represents the velocity Jacobian matrix between force platform 1 and leg 2.

[0072] Based on the principle of virtual work, the following relationship can be obtained:

[0073] (4);

[0074] Substituting formulas (1) to (3) into (4), the relationship between the stiffness of the integrated parallel platform, the stiffness of the legs and the current posture of the integrated parallel platform is obtained as follows:

[0075] (5);

[0076] Through the above derivation, when the stiffness of the outrigger 2 and the posture of the integrated parallel platform are known, the stiffness matrix of the integrated parallel platform in the current posture can be obtained.

[0077] In practical applications, the stiffness of each leg 2 is measured in the laboratory, and then the stiffness matrix of the leg 2 is obtained. , using the principle of virtual work to obtain the velocity Jacobian matrix under different positions , so the stiffness matrix of the integrated parallel platform in different positions can be obtained through theoretical calculation.

[0078] S33: Use the validation set to validate the trained neural network model.

[0079] The expected error E is used to verify the trained neural network model.

[0080] S34: Use the test set to test the verified neural network model.

[0081] The linearity error of the force sensor is used as an evaluation index of the neural network model. The smaller the linearity error, the better the model training.

[0082] The neural network model used in the present invention can be a BP, RBF or other neural network model.

[0083] The BP neural network model adopts a three-layer structure, a single hidden layer, and 10 nodes to model the neural network. The hidden layer uses the logarithmic S-type as the conversion function, and the linear function is used as the conversion function of the output layer. The network training uses the LM function to train the neural network. Before the neural network training, set the number of iterations to 100 and the iteration error to 10 -5 .

[0084] The RBF neural network model is divided into regularized RBF neural network and generalized RBF neural network. For sample sets with large sample sizes, generalized RBF neural network is selected, and for sample sets with large sample sizes, regularized RBF neural network is selected. A Gaussian function with good attenuation is selected as the conversion function on the hidden layer nodes, and the central vector, expansion constant and weight of the network are determined through supervised training.

[0085] S4: Adjust the integrated parallel platform to the posture to be calibrated, input the stiffness of the integrated parallel platform in this posture and the measured force of the force sensor into the trained neural network model, and obtain the loading force of the integrated parallel platform in this posture.

[0086] By utilizing the self-learning function of the neural network model, the posture stiffness information is introduced for data training, thereby solving the problem that the traditional force calibration method cannot realize the calibration of variable configuration force sensors.

[0087] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the disclosure of the present invention can be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and this document does not limit this.

[0088] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A force calibration method for an integrated parallel platform, the integrated parallel platform comprises a force measuring platform, a base and legs connected between the force measuring platform and the base, the number of the legs is seven, six of which are arranged in a Stewart configuration, and the other leg is connected to the center of the force measuring platform and the base, the legs comprise a first hinge, a second hinge, a force sensor and a telescopic assembly, the fixed end of the telescopic assembly is connected to the base through the first hinge, the output end of the telescopic assembly is connected to the force measuring platform through the second hinge, the force sensor is connected in series between the second hinge and the telescopic assembly, the force measuring platform is adjusted by adjusting the telescopic assembly, so that the integrated parallel platform is in different postures, characterized in that, The force calibration method includes the following steps: S1: Set the full scale of the force sensor in different positions of the integrated parallel platform; S2: Perform six-dimensional force loading and unloading calibration tests on each posture of the integrated parallel platform to obtain test data; wherein the test data includes the stiffness of the integrated parallel platform in each posture and the measured force of the force sensor; S3: Build and train a neural network model based on experimental data; S4: adjusting the integrated parallel platform to the posture to be calibrated, inputting the stiffness of the integrated parallel platform in this posture and the measured force of the force sensor into the neural network model, and obtaining the loading force of the integrated parallel platform in this posture; The stiffness calculation method of the integrated parallel platform in each posture is as follows: The mapping relationship between the force and deformation of the force measuring platform in six-dimensional directions is: （1）； in, represents the generalized force acting on the force measuring platform, represents the virtual displacement of the force measuring platform, represents the stiffness of the integrated parallel platform; According to Hooke's law, the mapping relationship between the force and deformation of the outrigger is: （2）； in, represents the matrix composed of the stiffness of each leg, is the virtual displacement of each leg; Derive the relationship between the deformation of the force measuring platform and the deformation of the legs: （3）； in, represents the velocity Jacobian matrix between the force platform and the outrigger; Based on the principle of virtual work, the following relationship is obtained: （4）； Substituting formulas (1) to (3) into (4), we can obtain the relationship between the stiffness of the integrated parallel platform, the stiffness of the legs, and the current posture of the integrated parallel platform: （5）； According to formula (5), the stiffness matrix of the integrated parallel platform in the current position is obtained.

2. The force calibration method of the integrated parallel platform according to claim 1, characterized in that: The process of building and training a neural network model is: S31: constructing a sample set based on the experimental data, and dividing the sample set into a training set, a validation set, and a test set according to a predetermined ratio; S32: construct a neural network model, initialize all parameters in the neural network model, and train the neural network model using a training set; S33: Use the validation set to validate the trained neural network model; S34: Use the test set to test the verified neural network model.

3. The force calibration method of the integrated parallel platform according to claim 2, characterized in that: The neural network model adopts BP neural network model or RBF neural network model.

Images