Joint offset torque calibration method and gravity balance method

Abstract

The invention provides a joint bias torque calibration method and a gravity balance method.The joint bias torque calibration method comprises the steps that firstly, an output torque error model of a driver is established, and then theoretical output force and actual output force of the driver loaded at multiple displacement positions are obtained; the parameters of the output torque error model can be obtained through fitting of the theoretical output force and the actual output force, and finally the joint offset torque can be obtained through calculation according to the torque error model. When the joint is started, the bias torque serves as the output torque of the driver, the position following error of the load can be reduced, and the position keeping function of the mechanical arm is kept. The joint bias torque calibration method is simple and practical, a large amount of prior data or neural network models are not needed, the precision of the gravity balance which can be achieved is high, and the applicability is high.

Description

Technical Field

[0001] The present invention belongs to the technical field of gravity balance, and more specifically, relates to a method for calibrating joint offset torque and a gravity balance method. Background Art

[0002] In robotic arms such as vertical joints and rotary joints composed of motors, loads, and gravity balancers, due to factors such as the non-linearity of the torque output by the gravity balancer, the slow response of the motor closed-loop feedback control, and the delay in the response of the brake, when the joint starts, the load will have a large position offset, and thus the function of maintaining the established position of the robotic arm cannot be achieved. Summary of the Invention

[0003] The purpose of the embodiments of the present invention is to provide a method for calibrating joint offset torque and a gravity balance method to solve the technical problem that the load has a large position offset when the joint starts in the prior art.

[0004] To achieve the above purpose, the technical solution adopted by the present invention is: to provide a method for calibrating joint offset torque, the joint includes a driver, a gravity balancer, and a load connected in sequence, and the method for calibrating joint offset torque includes:

[0005] Establish an output torque error model of the driver;

[0006] Obtain the theoretical output force and the actual output force of the driver at multiple displacements of the load;

[0007] Fit the parameters of the output torque error model according to the theoretical output force and the actual output force;

[0008] Calculate the joint offset torque according to the output torque error model.

[0009] In the above solution, first, an output torque error model of the driver is established, then the theoretical output force and the actual output force of the driver at multiple displacements of the load are obtained. The parameters of the output torque error model can be obtained by fitting the theoretical output force and the actual output force. Finally, the joint offset torque can be calculated according to the torque error model. When the joint starts, this offset torque is used as the output torque of the driver, which can reduce the position following error of the load and maintain the position holding function of the robotic arm. This method for calibrating joint offset torque is simple and practical, does not require a large amount of prior data or neural network models, and can achieve a relatively high gravity balance accuracy and strong applicability.

[0010] Optionally, the output torque error model is a linear function, the independent variable of the output torque error model is the displacement of the load, and the dependent variable of the output torque error model is the difference between the actual output force and the theoretical output force.

[0011] In the above solution, the difference between the actual output force and the theoretical output force can be defined as the output force error. Then, the independent variable of the output torque error model is the displacement of the load, and the dependent variable is the output force error. The product of the output force error of the driver and the force arm of the driver is the output torque error of the driver. In this output torque error model, the output force error can be used to replace the output torque error, which can simplify the calculation.

[0012] Optionally, the step of obtaining the actual output force of the driver at multiple displacements of the load includes:

[0013] Keep the load stationary at a preset displacement;

[0014] Obtain the position following error of the load that is less than a preset threshold at the moment when the driver starts;

[0015] Obtain the actual output force of the driver when the position following error is less than the preset threshold;

[0016] Change the preset displacement of the load and repeat the above steps.

[0017] In the above solution, by obtaining the actual output force of the driver through actual testing, the error influence caused by the individual differences of the joints can be eliminated. Through the first three steps, the actual output force corresponding to the driver when the load is at a preset displacement can be obtained. Through the last step, the actual output forces corresponding to the driver when the load is at multiple preset displacements can be obtained.

[0018] Optionally, the step of obtaining the position following error of the load that is less than the preset threshold at the moment when the driver starts includes:

[0019] Set the output force of the driver multiple times, and correspondingly obtain the position following error of the load at the moment when the driver starts;

[0020] Judge whether each of the position following errors is less than the preset threshold;

[0021] If at least one of the position following errors is less than the preset threshold, then obtain the actual output force of the driver;

[0022] If each of the position following errors is greater than or equal to the preset threshold, then repeat the above steps.

[0023] In the above solution, since the output force of the drive is not known before testing and it is not clear how much position following error will be generated accordingly, it is necessary to try changing the output force of the drive multiple times to obtain a position following error less than the preset threshold, and then inversely obtain the corresponding output force at this time. That is, it is necessary to obtain the actual output force of the drive through a trial-and-error method. As long as the number of attempts is large enough, a suitable output force can always be found to make the position following error at the moment of starting when the load is at the predetermined displacement small enough (less than the preset threshold).

[0024] Optionally, the step of obtaining the actual output force of the drive when the position following error is less than the preset threshold includes:

[0025] If one of the position following errors is less than the preset threshold, the output force of the drive corresponding to this position following error is the actual output force of the drive;

[0026] If multiple position following errors are less than the preset threshold, the average value of the output forces of the drive corresponding to each position following error is the actual output force of the drive, or the output force of the drive corresponding to the minimum value of each position following error is the actual output force of the drive.

[0027] In the above solution, the two steps can be carried out simultaneously, or either step can be implemented first.

[0028] Optionally, the step of obtaining the theoretical output force of the drive at multiple displacements of the load includes:

[0029] Obtain the output force of the gravity balancer at multiple displacements of the load;

[0030] Establish a theoretical output force model of the drive;

[0031] Calculate the theoretical output force of the drive according to the theoretical output force model of the drive.

[0032] In the above solution, the theoretical output force of the drive can be calculated through the theoretical output force model of the drive. The theoretical output force model of the drive has two variables, which are the theoretical output force of the drive and the output force of the gravity balancer respectively. In this way, when the output force of the gravity balancer is known, the theoretical output force of the drive can be obtained through the output force model of the theoretical output force of the drive.

[0033] Optionally, the difference between the output force of the gravity balancer and the gravity of the load is the theoretical output force of the drive.

[0034] In the above solution, the difference between the output force f(S x ) of the gravity balancer and the gravity of the load is the theoretical output force F(Sx )。Specifically, F(S x ) = -mg + f(S x ), where m is the mass of the load, g is the acceleration due to gravity, and f is the functional relationship between the position Sx of the load and the force f(S x ) output by the gravity balancer.

[0035] Optionally, the step of obtaining the output force of the gravity balancer at multiple displacements of the load includes: performing polynomial fitting on the displacements of each load and the corresponding output force of the gravity balancer to obtain the output force of the gravity balancer.

[0036] In the above solution, the output torque error model is ΔF = pS x + q, where S x is the independent variable, ΔF is the dependent variable, and p and q are the parameters of the output torque error model. ΔF is the difference between the actual output force F i ' of the driver and the theoretical output force F(S x ). At each displacement S x , there is a corresponding actual output force F i '. Since F i ' is actually a scatter function, ΔF is also a scatter function. Since the output torque error model is a linear function, a linear fit can be performed on this scatter function to obtain the Figure 4 fitting curve as shown. Through this fitting curve, the slope p and intercept q of the fitting curve can be calculated. p and q are the two parameters of the output torque error model. When p, q, and the load displacement are known, the corresponding error force can be calculated, and multiplying this error force by the force arm of the driver can obtain the error torque that needs to be compensated.

[0037] Optionally, the driver is a motor.

[0038] In the above solution, the motor can be a mechanism that provides power in the joint. Specifically, the motor has a rotating shaft, which is the power output structure of the motor, that is, the rotating shaft outputs rotational motion. The motor also includes a housing, and a part of the rotating shaft is located inside the housing, and one end of the rotating shaft extends outside the housing.

[0039] Optionally, the joint is a vertical joint or a rotary joint.

[0040] In the above solution, as long as the relationship between the torque output by the gravity balancer and the displacement of the joint is determined, the joint offset torque calibration method provided by the present invention is applicable to any form of joint.

[0041] Optionally, the gravity balancer includes a spring, and two ends of the spring are respectively connected to the driver and the load; alternatively, the gravity balancer includes a scroll spring and a sheave, the sheave is fixedly connected to an output shaft of the driver, one end of the scroll spring is fixedly arranged, and the other end is fixed to the output shaft of the driver, and the load is connected to the sheave through a traction rope wound around the sheave.

[0042] In the above solution, the joint offset torque calibration method is applicable to various gravity balancers, as long as the relationship between the torque output by the gravity balancer and the displacement of the joint is determined.

[0043] The present invention further provides a gravity balancing method, including:

[0044] Obtaining a joint offset torque through the above joint offset torque calibration method;

[0045] Taking the joint offset torque as the output torque value of the driver.

[0046] In the above solution, the torque feedforward control is adopted to realize the position holding function at the moment when the robotic arm joint starts. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0048] Figure 1 It is a schematic flow chart of the joint offset torque calibration method provided by the embodiment of the present invention;

[0049] Figure 2 It is a schematic structural diagram of a vertical joint provided by the embodiment of the present invention;

[0050] Figure 3 It is a linear fitting curve of the load displacement and the torque error provided by the embodiment of the present invention;

[0051] Figure 4 It is a load position change curve when the load is at a displacement of 180 mm and the output torque value of the driver is 0 at the moment when the driver starts provided by the embodiment of the present invention;

[0052] Figure 5 It is a load position change curve when the load is at a displacement of 180 mm and the output torque value of the driver is 10 N (offset torque) at the moment when the driver starts provided by the embodiment of the present invention.

[0053] Among them, the reference numerals in the drawings:

[0054] 1 - Driver; 2 - Gravity balancer; 3 - Load. Detailed implementation manner

[0055] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0056] It should be noted that when an element is referred to as being "fixed to" or "disposed on" another element, it can be directly on the other element or indirectly on the other element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or indirectly connected to the other element.

[0057] It should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings. It is 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 to the present invention.

[0058] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality of" means two or more unless otherwise specifically defined.

[0059] The gravity balancer 2 is a common mechanical device that uses the action of gravity to keep the load 3 in a balanced state at different positions. The gravity balancer 2 is widely used in various mechanical equipment, such as machine tools, cranes, elevators, medical devices, etc., to ensure the stability and safety of the mechanical equipment.

[0060] The main components of the gravity balancer 2 include springs. By using the gravity balancer 2 in the joints of the robot, the load 3 on the motor can be reduced, the control accuracy of the motor can be improved, and the movement of the load 3 is also smoother.

[0061] In a robotic arm with vertical joints, rotary joints, etc. composed of a motor, a load 3, and a gravity balancer 2, due to factors such as the non-linearity of the torque output by the gravity balancer 2, the slow response of the motor closed-loop feedback control, and the delay in the brake response, when the joint starts, the load 3 will have a large position offset, and thus the function of maintaining the position of the robotic arm as intended cannot be achieved. To overcome the above technical problems, the present invention proposes a method for calibrating the joint offset torque and a gravity balancing method.

[0062] The method for calibrating the joint offset torque provided in the embodiments of the present invention will now be described. The method for calibrating the joint offset torque is applicable to the joint control of a robot. The joint includes a driver 1, a gravity balancer 2, and a load 3 connected in sequence. The driver 1 is the power mechanism of the joint and is used to drive the joint to perform corresponding actions. The driver 1 can be a motor, a motor, etc. The gravity balancer 2 includes at least balance devices such as springs, and the load 3 is connected to one end of the gravity balancer 2.

[0063] By establishing an output torque error model, obtaining the parameters of the torque error model to obtain the torque error model, and calculating the joint offset torque through the torque error model, and using the joint offset torque as the output torque value of the driver 1, when the joint starts, the position following error of the load 3 can be reduced, and the function of maintaining the position of the robotic arm can be maintained.

[0064] Please refer to Figure 1 , the method for calibrating the joint offset torque includes the following steps:

[0065] S10: Establish an output torque error model of the driver 1;

[0066] S20: Obtain the theoretical output force and the actual output force of the driver 1 at multiple displacements of the load 3;

[0067] S30: Fit the parameters of the output torque error model according to the theoretical output force and the actual output force;

[0068] S40: Calculate the joint offset torque according to the output torque error model.

[0069] In step S10, the joint needs to output corresponding torques when remaining stationary, preparing for startup, and during operation to keep the load 3 balanced or perform work for movement. At the moment of joint startup, the actuator 1 also needs to output a corresponding torque to start the load 3 and make it move. The theoretical torque required by the actuator 1 to start the load 3 is called the theoretical torque of the actuator 1. However, in reality, at the moment of joint startup, due to factors such as the non-linearity of the torque output by the gravity balancer 2, the slow response of the motor closed-loop feedback control, and the delay in the brake response, there will be a large position offset of the load 3 at the moment of joint startup. The position offset of the load 3 during startup can also be called the position following error. To reduce the position following error of the load 3 during startup, the actual output torque can be adjusted accordingly. The adjusted output torque is called the actual output torque, and the difference between the actual output torque and the theoretical output torque is the output torque error. The output torque error model of the actuator 1 is a function model related to the output torque error of the actuator 1. For the sake of convenience in description, in the following text, the output torque error, output force error, actual output force, actual output force, theoretical output force, theoretical output torque, etc. of the actuator 1 all refer to the corresponding values when the load 3 remains balanced at the moment of startup of the actuator 1.

[0070] In step S20, the theoretical output force and actual output force of the actuator 1 at multiple displacements of the load 3 are obtained. When the load 3 is at different displacements, the gravity balancer 2 needs to output corresponding forces to keep the load 3 balanced. That is to say, at each displacement of the load 3, the gravity balancer 2 outputs a corresponding force to keep the load 3 balanced. And the output force of the gravity balancer 2 is related to the output force of the actuator 1. For the sake of calculation convenience, without introducing the force arm, it is directly possible to obtain the theoretical output force and actual output force of the actuator 1. When the load 3 remains balanced at each displacement, there is a corresponding theoretical output force and an actual output force.

[0071] In step S30, the parameters of the output torque error model are obtained by fitting based on the theoretical output force and actual output force. The theoretical output force and actual output force of the actuator 1 can both be obtained according to step S20. The function corresponding to the output torque error model consists of independent variables, dependent variables, and parameters, and the parameters can be constants. After the parameters are determined, the output torque error model is also determined accordingly.

[0072] In step S40, the joint offset torque is calculated according to the output torque error model. According to the output torque error model, the output force error of the actuator 1 can be obtained when the load 3 is at a predetermined position, and the joint offset torque is the product of the output force error of the actuator 1 and the force arm of the actuator 1.

[0073] In the joint offset torque calibration method in the above embodiments, first, an output torque error model of the actuator 1 is established. Then, the theoretical output force and the actual output force of the actuator 1 at multiple displacements of the load 3 are obtained. The parameters of the output torque error model can be obtained by fitting the theoretical output force and the actual output force. Finally, the joint offset torque can be calculated according to the torque error model. When the joint starts, taking this offset torque as the output torque of the actuator 1 can reduce the position following error of the load 3 and maintain the position holding function of the robotic arm. This joint offset torque calibration method is simple and practical, without the need for a large amount of prior data or neural network models, and can achieve a relatively high gravity balance accuracy and strong applicability.

[0074] In some embodiments of the present invention, in step S10, the output torque error model is a linear function. The independent variable of the output torque error model is the displacement of the load 3, and the dependent variable of the output torque error model is the difference between the actual output force and the theoretical output force. At each displacement of the load 3, there is a corresponding output torque error. A linear function generally has the form y = kx + b, where x is the independent variable, y is the dependent variable, k and b are the parameters of the linear function, and b can be 0, and the linear function is then transformed into y = kx. In this embodiment, the difference between the actual output force and the theoretical output force can be defined as the output force error. Then, the independent variable of the output torque error model is the displacement of the load 3, and the dependent variable is the output force error. The product of the output force error of the actuator 1 and the force arm of the actuator 1 is the output torque error of the actuator 1. In this output torque error model, the output force error can be used to replace the output torque error, which can simplify the calculation. Therefore, the offset torque calibration method proposed by the present invention is actually the calibration of the output force error of the actuator 1, rather than the calibration of the output torque of the actuator 1. The advantage of doing this is that when the mass of the load 3 and the joint position change, there is no need to calibrate the parameters again, which improves the applicability of this method.

[0075] In some embodiments, the actual output force of the actuator 1 is F i ', the theoretical output force of the actuator 1 is F i , and the output force error ΔF of the actuator 1 is the difference between the actual output force F i ' and the theoretical output force F i , that is, ΔF = F i ' - F i . The output torque error model is ΔF = pS x + q, where S x is the independent variable, ΔF is the dependent variable, and p and q are the parameters of the output torque error model. Specifically, S x is the displacement of the load 3, ΔF is the output force error when the displacement of the load 3 is S x , F i ' is the actual output force when the displacement of the load 3 is S x , and Fi The displacement of the load 3 is S x The theoretical output force at this time. After obtaining the parameters p and q, the output torque error model can be determined, so that the output force error of the load 3 at the preset displacement can be obtained, and then the bias torque of the load 3 at the preset displacement can be obtained.

[0076] In some embodiments of the present invention, in step S20, the steps of obtaining the actual output force of the driver 1 at multiple displacements of the load 3 include:

[0077] S21: Keep the load 3 stationary at the preset displacement;

[0078] S22: Obtain the position following error of the load 3 less than the preset threshold at the moment when the driver 1 starts;

[0079] S23: Obtain the actual output force of the driver 1 when the position following error is less than the preset threshold;

[0080] S24: Change the preset displacement of the load 3 and repeat the above steps.

[0081] In the above steps, the actual output force of the driver 1 is obtained by the method of actual test, and the error influence caused by the individual differences of the joints can be eliminated. From step S21 to step S23, the actual output force corresponding to the driver 1 when the load 3 is at a preset displacement is obtained. Through step S24, the actual output force corresponding to the driver 1 when the load 3 is at multiple preset displacements can be obtained.

[0082] When obtaining the actual output force of the driver 1, first, the load 3 needs to be kept stationary at the preset displacement, and then the output force of the driver 1 is set. The position following error of the load 3 less than the preset threshold at the moment when the driver 1 starts is obtained. The preset threshold can be set according to the precision requirements of the movement of the load 3, and it is not limited here. In actual operation, the output torque of the driver 1 can be set, or the output force of the driver 1 can be set. When the position following error is less than the preset threshold, the output force of the driver 1 at this time is the above-mentioned actual output force. In this way, it can be ensured that when the driver 1 is set to output torque with the actual output force, the position following error of the load 3 is less than the preset threshold, meeting the actual working precision requirements.

[0083] In some embodiments, in the state where the load 3 is kept stationary, when the displacement of the load 3 is S i the actual output force F i ' of the driver 1 is obtained. The actual output force F i ' satisfies that at the moment when the driver 1 starts, the position following error of the load 3 is less than the preset threshold, where i = 2, 3, 4... In this way, the corresponding actual output forces F i ' at multiple displacements can be obtained.

[0084] In some embodiments of the present invention, step S22 includes:

[0085] S221: Set the output force of the driver 1 multiple times, and correspondingly obtain the position following error of the load 3 at the moment of starting the driver 1;

[0086] S222: Determine whether each position following error is less than a preset threshold;

[0087] S223: If at least one position following error is less than the preset threshold, obtain the actual output force of the driver 1;

[0088] S224: If each position following error is greater than or equal to the preset threshold, repeat the above steps until.

[0089] Since it is not clear before the test how much position following error will be generated corresponding to the output force of the driver 1, it is necessary to try to change the output force of the driver 1 multiple times, obtain the position following error less than the preset threshold, and then inversely obtain the corresponding output force at this time, that is, it is necessary to obtain the actual output force of the driver 1 by the trial-and-error method. As long as the number of attempts is large enough, a suitable output force can always be found to make the position following error of the load 3 at the moment of starting at the predetermined displacement small enough (less than the preset threshold).

[0090] After setting the output force of the driver 1 N times (after N attempts), N position following errors can be correspondingly obtained. If all the position following errors are greater than or equal to the preset threshold, it is necessary to retest. Only if one position following error is less than the preset threshold, can step S23 be performed. In other embodiments, it can also be set that if M position following errors are less than the preset threshold, step S23 can be performed, where M is less than or equal to N.

[0091] In some embodiments of the present invention, step S23 includes:

[0092] S231: If one position following error is less than the preset threshold, the output force of the driver 1 corresponding to this position following error is the actual output force of the driver 1;

[0093] S232: If multiple position following errors are less than the preset threshold, the average value of the output forces of the driver 1 corresponding to each position following error is the actual output force of the driver 1, or the output force of the driver 1 corresponding to the minimum value of each position following error is the actual output force of the driver 1.

[0094] Step S231 and step S232 can be carried out simultaneously, and the serial numbers of these two steps do not affect the order of their implementation.

[0095] In some embodiments, when performing step S22, after implementing S221 and S221 once, and then implementing S221 and S221 again, that is, following the test sequence one after another. When the position following error is less than the predetermined threshold for the first time, the test is stopped, as in step S231. Only one position following error is less than the preset threshold, and the output force of drive 1 corresponding to this position following error is the actual output force of drive 1.

[0096] In some embodiments, when performing step S22, it can be carried out in the test sequence one after another. When the number of tests reaches the predetermined number N times, the test is stopped. If among the N position following errors, only one following error is less than the preset threshold, then the output force of drive 1 corresponding to this position following error is the actual output force of drive 1; if among the N position following errors, there are multiple position following errors that are all less than the preset threshold, then the output force of drive 1 corresponding to the position following error with the smallest value is the actual output force of drive 1.

[0097] It should be noted that the actual output force corresponding to each displacement of load 3 can be obtained through any of the above embodiments. Taking the displacement of load 3 as the abscissa and the actual output force of drive 1 as the ordinate to establish a coordinate system, multiple scatter points can be obtained, and each scatter point corresponds to a displacement of load 3.

[0098] In some embodiments of the present invention, in step S20, the steps of obtaining the theoretical output force of drive 1 at multiple displacements of load 3 include:

[0099] S25: Obtain the output force of gravity balancer 2 at multiple displacements of load 3;

[0100] S26: Establish a theoretical output force model of drive 1;

[0101] S27: Calculate the theoretical output force of drive 1 according to the theoretical output force model of drive 1.

[0102] In the above steps, the theoretical output force of drive 1 can be calculated through the theoretical output force model of drive 1. The theoretical output force model of drive 1 has two variables, which are the theoretical output force of drive 1 and the output force of gravity balancer 2 respectively. In this way, when the output force of gravity balancer 2 is known, the theoretical output force of drive 1 can be obtained through the output force model of the theoretical output force of drive 1.

[0103] In some embodiments, step S25 includes the following steps: Obtain the displacement S of load 3 x And the functional relationship f(S) between the output force of gravity balancer 2 x)). After the specific structure of the joint is determined, to maintain the balance of the load 3, the relationship between the displacement of the load 3 and the output force of the gravity balancer 2 is determined, which can be obtained by testing methods or by looking up the relevant technical parameters of the gravity balancer 2. Specifically, to obtain f(S x ) through testing, f(S x ) can be directly tested, or the characteristic curve of the gravity balancer 2 can be tested. For example, when testing the characteristic curve of the gravity balancer 2, a large number of tests can be carried out on the gravity balancers 2 of the same type and with the same parameters, and the average value of the output forces of the tested gravity balancers 2 is taken. It should be noted that although the independent variable of f(S x ) is the displacement Sx of the load 3, in fact, the relationship between the displacement Sx of the load 3 and the deformation degree of the gravity balancer 2 is one-to-one determined. Therefore, testing f(S x ) can actually be equivalent to testing the characteristic curve of the gravity balancer 2 (the abscissa is the deformation of the gravity balancer 2, and the ordinate is the output force of the gravity balancer 2).

[0104] In some embodiments, the gravity balancer 2 includes a volute spring. The corresponding f(S x ) of the volute spring can be obtained from the characteristic curve of the volute spring. The abscissa of this characteristic curve is the rotation angle of the volute spring, and the ordinate is the output torque of the volute spring. When the gravity balancer 2 is connected to the load 3, the displacement of the load 3 can be calculated through the rotation angle of the volute spring, and the output force of the volute spring can be calculated through the output torque of the volute spring. In this way, the corresponding f(S x ) can be obtained from the characteristic curve of the volute spring. The characteristic curve of the volute spring can be obtained by looking up its factory instructions or by testing. Among them, f(S x ) can be a linear function or a non-linear function, which is related to the characteristic curve of the gravity balancer 2.

[0105] Therefore, in this embodiment, as long as the relationship between the output torque of the gravity balancer 2 and the displacement of the load 3 is determined (whether linear or non-linear), the bias torque calibration method in this embodiment is applicable.

[0106] In some embodiments, step S25 includes: performing polynomial fitting on the displacements of each load 3 and the corresponding output forces of the gravity balancer 2 to obtain the output force of the gravity balancer 2. In this embodiment, the functional relationship f(S x) Specifically, the gravity balancer 2 is hung with the load 3. At each displacement of the load 3, the output force of the gravity balancer 2 is correspondingly measured. Taking the displacement of the load 3 as the abscissa and the output force of the gravity balancer 2 as the ordinate to establish a coordinate system, a scatter plot distribution in this coordinate system can be obtained. By performing polynomial fitting on this scatter plot distribution, the displacement S of the load 3 can be obtained. x The functional relationship f(S x ) between the output force of the gravity balancer 2 and

[0107] In step S26, a theoretical output force model of the driver 1 is established. The theoretical output force model of the driver 1 is actually the functional relationship between the output force of the gravity balancer 2 and the theoretical output force of the driver 1. Step S26 can be understood as establishing the functional relationship between the output force f(S x ) of the gravity balancer 2 and the theoretical output force F(S x ) of the driver 1.

[0108] In some embodiments, the difference between the output force f(S x ) of the gravity balancer 2 and the gravity of the load 3 is the theoretical output force F(S x ) of the driver 1. Specifically, F(S x ) = -mg + f(S x ), where m is the mass of the load 3, g is the acceleration due to gravity, and f is the functional relationship between the position Sx of the load 3 and the force f(S x ) output by the gravity balancer 2.

[0109] Please refer to Figure 2 . When the load 3 is connected to the gravity balancer 2, a force analysis is performed on the load 3. The load 3 is subjected to the action of gravity and the pulling force of the gravity balancer 2 ( Figure 2 shows the action of the pulling force of the towing rope), mg - F c = ma, where m is the mass of the load 3, g is the acceleration due to gravity, a is the acceleration of the load 3, and F c is the pulling force of the towing rope. The bias torque calibration method in the present invention is all performed when the driver 1 starts instantaneously on the premise that the load 3 remains stationary. Therefore, the acceleration a of the load 3 is 0. That is, mg = F c can be obtained. The pulling force of the towing rope comes from the sum of the output forces of the driver 1 and the gravity balancer 2. The output forces of the driver 1 and the gravity balancer 2 are in opposite directions. Therefore, there is F c = f(S x ) - F(S x ), f(S x ) is the output force of the gravity balancer 2, and F(S x ) is the theoretical output force of the driver 1. Substituting mg = F cSubstituting it in, we can get \(mg = f(S x ) - F(S x ), that is, \(F(S x ) = -mg + f(S x ).

[0110] In step S27, the theoretical output force of the actuator 1 is calculated according to the theoretical output force model of the actuator 1. After the output force \(f(S x ) of the gravity balancer 2 is obtained through step S25, and then combined with \(F(S x ) = -mg + f(S x ), the theoretical output force of the actuator 1 can be calculated.

[0111] In some embodiments of the present invention, in step S30, the parameters of the output torque error model are obtained by fitting the theoretical output force and the actual output force. In some embodiments, the output torque error model is \(\Delta F = pS x +q\), where \(S x is the independent variable, \(\Delta F\) is the dependent variable, and \(p\) and \(q\) are the parameters of the output torque error model. \(\Delta F\) is the difference between the actual output force \(F i '\) of the actuator 1 and the theoretical output force \(F(S x ). There is a corresponding actual output force \(F i '\) at each displacement \(S_x\). Since \(F i '\) is actually a scatter function, \(\Delta F\) is also a scatter function. Since the output torque error model is a linear function, a linear fit can be performed on this scatter function to obtain the fitting curve as shown in Figure 4 . Through this fitting curve, the slope \(p\) and the intercept \(q\) of the fitting curve can be calculated. \(p\) and \(q\) are the two parameters of the output torque error model. When \(p\), \(q\), and the displacement of the load 3 are known, the corresponding error force can be calculated. Multiplying this error force by the force arm of the actuator 1, the error torque to be compensated can be obtained.

[0112] In some embodiments of the present invention, in step S40, the joint bias torque is calculated according to the output torque error model. The output torque error model is \(\Delta F = pS x +q\). The theoretical output torque of the actuator 1 is the product of its theoretical output force \(F(S x ) and the force arm \(L\). The output torque error is the product of \(\Delta F\) and the force arm \(L\). Then the bias torque of the joint is the sum of the theoretical output torque of the actuator 1 and the output torque error of the actuator 1, that is, \((\Delta F + F(S x ))L = [pS x +q - mg + f(S x )]L.

[0113] In some embodiments of the present invention, please refer to Figure 3, for the parameters p and q of the output torque error model fitted according to the theoretical output force and the actual output force, p = 0.0583 and q = -14.367.

[0114] To verify the effectiveness of the biasing torque calibration method in the present invention, the vertical joint of the robotic arm is tested, and the position following error at the moment of startup is tested respectively with and without compensating for the output torque error. Figure 4 and Figure 5 are respectively the position change curves of the vertical joint of the robotic arm at a displacement of 180 mm when the output torque values of the driver 1 are set to 0 and 10 N (biasing torque) respectively at the moment of joint startup. From Figure 4 it can be seen that when no biasing torque is applied, the maximum displacement of the vertical joint of the robotic arm is 0.37695 mm. From Figure 5 it can be seen that after applying a biasing torque of 10 N, the maximum displacement at the moment of startup of the vertical joint is 0.015625 mm, which is 23 times smaller than the maximum displacement without applying the biasing torque. This proves both the accuracy of the biasing torque calibration method and the effectiveness of the torque feedforward control method for the position holding function at the moment of startup of the vertical joint.

[0115] In some embodiments of the present invention, the driver 1 is a motor capable of outputting rotational motion. The motor can be a mechanism that provides power to the joint. Specifically, the motor has a rotating shaft, and the rotating shaft is the power output structure of the motor, that is, the rotating shaft outputs rotational motion. The motor further includes a housing, a part of the rotating shaft is located inside the housing, and one end of the rotating shaft extends outside the housing.

[0116] In some embodiments of the present invention, the gravity balancer 2 includes a spring, and both ends of the spring are respectively connected to the driver 1 and the load 3. Optionally, a traction rope is connected to the rotating shaft of the motor, the end of the traction rope away from the rotating shaft is connected to the spring, and the end of the spring away from the traction rope is connected to the load 3.

[0117] In some embodiments of the present invention, the gravity balancer 2 includes a volute spring and a sheave. The sheave is fixedly connected to the output shaft of the driver 1. One end of the volute spring is fixedly arranged, which can be fixed to the inner wall of the housing of the driver 1. The other end of the volute spring is fixed to the output shaft of the driver 1. The load 3 is connected to the sheave through a traction rope wound around the sheave.

[0118] It can be understood that the joint biasing torque calibration method provided by the present invention is applicable to various gravity balancers 2 as long as the relationship between the torque output by the gravity balancer 2 and the displacement of the joint is determined.

[0119] In some embodiments of the present invention, the joint is a vertical joint or a rotary joint. As long as the relationship between the torque output by the gravity balancer 2 and the displacement of the joint is determined, the joint bias torque calibration method provided by the present invention is applicable to joints of any form.

[0120] The present invention also proposes a gravity balancing method, which obtains the joint bias torque by using the joint bias torque calibration method in any of the above embodiments; then, taking this joint bias torque as the output torque value of the actuator 1, the position holding function at the moment of starting the robotic arm joint is realized by using torque feedforward control.

[0121] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for calibrating the joint offset torque, wherein the joint includes a driver (1), a gravity balancer (2), and a load (3) connected in sequence, and is characterized in that, The joint offset torque calibration method includes: Establishing an output torque error model of the driver (1); Obtaining the theoretical output force and the actual output force of the driver (1) when the load (3) is at multiple displacements; Fitting the parameters of the output torque error model according to the theoretical output force and the actual output force; Calculating the joint offset torque according to the output torque error model.

2. The method for calibrating the joint offset torque according to claim 1, and is characterized in that, The output torque error model is a linear function. The independent variable of the output torque error model is the displacement of the load (3), and the dependent variable of the output torque error model is the difference between the actual output force and the theoretical output force.

3. The method for calibrating the joint offset torque according to claim 1, and is characterized in that, The step of obtaining the actual output force of the driver (1) when the load (3) is at multiple displacements includes: Keeping the load (3) stationary at a preset displacement; Obtaining the position following error of the load (3) less than a preset threshold value at the instant when the driver (1) starts; Obtaining the actual output force of the driver (1) when the position following error is less than the preset threshold value; Changing the preset displacement of the load (3) and repeating the above steps.

4. The method for calibrating the joint offset torque according to claim 3, and is characterized in that, The step of obtaining the position following error of the load (3) less than the preset threshold value at the instant when the driver (1) starts includes: Setting the output force of the driver (1) multiple times and correspondingly obtaining the position following error of the load (3) at the instant when the driver (1) starts; Judging whether each of the position following errors is less than the preset threshold value; If at least one of the position following errors is less than the preset threshold value, obtaining the actual output force of the driver (1); If each of the position following errors is greater than or equal to the preset threshold value, repeating the above steps.

5. The method for calibrating the joint offset torque according to claim 4, and is characterized in that, The step of obtaining the actual output force of the driver (1) when the position following error is less than the preset threshold value includes: If one of the position following errors is less than the preset threshold value, the output force of the driver (1) corresponding to this position following error is the actual output force of the driver (1); If multiple position following errors are less than the preset threshold value, the average value of the output forces of the driver (1) corresponding to each position following error is the actual output force of the driver (1), or the output force of the driver (1) corresponding to the minimum value of each position following error is the actual output force of the driver (1).

6. The method for calibrating the joint offset torque according to claim 1, and is characterized in that, The step of obtaining the theoretical output force of the driver (1) when the load (3) is at multiple displacements includes: Obtaining the output force of the gravity balancer (2) when the load (3) is at multiple displacements; Establishing a theoretical output force model of the driver (1); Calculating the theoretical output force of the driver (1) according to the theoretical output force model of the driver (1).

7. The method for calibrating the joint offset torque according to claim 6, and is characterized in that, The difference between the output force of the gravity balancer (2) and the gravity of the load (3) is the theoretical output force of the driver (1).

8. The method for calibrating the joint offset torque according to claim 6, and is characterized in that, The step of obtaining the output force of the gravity balancer (2) at multiple displacements of the load (3) includes: performing polynomial fitting on the displacements of the respective loads (3) and the output forces of the corresponding gravity balancers (2) to obtain the output force of the gravity balancer (2).

9. The method for calibrating the joint offset torque according to any one of claims 1-8, and is characterized in that, The gravity balancer (2) includes a spring, and two ends of the spring are respectively connected to the driver (1) and the load (3); alternatively, the gravity balancer (2) includes a scroll spring and a sheave, the sheave is fixedly connected to the output shaft of the driver (1), one end of the scroll spring is fixedly arranged, the other end is fixed to the output shaft of the driver (1), and the load (3) is connected to the sheave through a towing rope wound around the sheave.

10. A gravity balancing method, and is characterized in that,including: obtaining the joint offset torque by the joint offset torque calibration method according to any one of claims 1-9; using the joint offset torque as the output torque value of the driver (1).

Images