Two-degree-of-freedom rotary joint angle non-contact measurement method based on three-dimensional Hall sensor

Abstract

The invention discloses a two-degree-of-freedom rotary joint angle non-contact measurement method based on a three-dimensional Hall sensor, and relates to the field of robots. The invention aims to solve the problems of low integration level, large occupied space and low measurement dimension of the existing robot joint angle sensor. According to the two-degree-of-freedom rotary joint angle non-contact measurement method based on the three-dimensional Hall sensor, the mapping M-J from a motor space M to a joint space J is constructed, and the mapping S-M from a sensor space S to the motor space M is calibrated through motor closed-loop control based on the mapping M-J; the motor space M is used as a medium to realize mapping S-J from the sensor space S to the joint space J, and non-contact measurement of the two-degree-of-freedom rotary joint angle is completed. And the method has a wide application prospect in a scene requiring high integration level and high reliability.

Description

Technical Field

[0001] The present invention belongs to the field of robots, and particularly relates to the measurement of joint angles of robots. Background Art

[0002] In order to achieve high integration and high reliability in the design of two-degree-of-freedom joints, while reducing the complexity of the structural design, it is urgently necessary to break through the limitations of traditional technologies. Generally speaking, two-degree-of-freedom joints need to be measured by two joint angle sensors. However, when designing complex joint structures, it is usually difficult to meet the design requirements of high integration. At present, common absolute angle encoders have a large volume, which limits the application of micro motors in the field of highly integrated robots. In addition, potentiometer-type angle sensors have frictional losses, and their service life is limited, making it difficult to meet the requirements for high reliability and high stability. The single-joint angle sensor solution based on Hall sensors and radially magnetized permanent magnets can only measure a single joint angle and needs to be installed at the shaft end of the rotating joint, restricting the flexibility and innovation of joint structure design. Therefore, in order to achieve higher integration in the robot system while reducing the complexity of the design space, it is urgently necessary to develop a highly integrated, highly reliable, and multi-dimensional joint angle sensor. Summary of the Invention

[0003] The present invention is to solve the problems of low integration, large occupied space, and low measurement dimension existing in the existing robot joint angle sensors, and provides a non-contact measurement method for the angles of two-degree-of-freedom rotating joints based on three-dimensional Hall sensors.

[0004] The non-contact measurement method for the angles of two-degree-of-freedom rotating joints based on three-dimensional Hall sensors includes:

[0005] Construct a mapping M→J from the motor space M to the joint space J, and based on this mapping M→J, calibrate the mapping S→M from the sensor space S to the motor space M for motor closed-loop control, and use the motor space M as an intermediary to realize the mapping S→J from the sensor space S to the joint space J, so as to complete the non-contact measurement of the angles of two-degree-of-freedom rotating joints.

[0006] Further, the above realization of the mapping S→J from the sensor space S to the joint space J using the motor space M as an intermediary to complete the non-contact measurement of the angles of two-degree-of-freedom rotating joints includes:

[0007] In the static state of the two-degree-of-freedom rotating joint, use a three-dimensional Hall sensor to collect the three-dimensional magnetic induction intensity components of the two-degree-of-freedom rotating joint for multiple times, and a permanent magnet is embedded inside the two-degree-of-freedom rotating joint;

[0008] Perform mean filtering on all the three-dimensional Hall sensor data collected to obtain sensor filtered data;

[0009] Look up the cumulative count value of the motor position encoder corresponding to the sensor filtering data in the calibration mapping table, and use it as the initial cumulative count value of the two drive motor position encoders [[H 10 ,H 20 ;

[0010] Drive the two-degree-of-freedom rotating joint to move, and obtain the cumulative count values H of the two drive motor position encoders in real time 1 =H 10 +ΔH 1 and H 2 =H 20 +ΔH 2 , where ΔH 1 and ΔH 2 are the cumulative count increments of the two drive motor position encoders respectively;

[0011] According to the cumulative count values H of the two drive motor position encoders 1 and H 2 , look up the corresponding joint angle in the target model table to achieve real-time output of the absolute joint angle;

[0012] The calibration mapping table is used to represent the correspondence between three-dimensional Hall sensor data and the cumulative count values of the two drive motor position encoders, and the target model table is used to represent the correspondence between joint angles and the cumulative count values of the two drive motor position encoders.

[0013] Furthermore, the above construction of the mapping M→J from the motor space M to the joint space J includes:

[0014] Let δ be the simulation step size, let the initial values of the two rotation angles of the two-degree-of-freedom rotating joint be both 0, and the intervals of the two rotation angles of the two-degree-of-freedom rotating joint are θ∈[0,θ m , θ m is the upper limit of θ, and are respectively the lower limit and the upper limit, then the total number of simulation steps of the two rotation angles of the two-degree-of-freedom rotating joint are respectively

[0015] S101: Judge whether i≤N 1 holds, where i = 1, 2,..., N 1 , if yes, set j = 1 and then execute S102, otherwise execute S104;

[0016] S102: Judge whether j≤N 2 holds, where j = 1, 2,..., N 2, if so, execute S103; otherwise, set i = i + 1, then return to S101;

[0017] S103: Calculate the current number of model table rows k = (i - 1)N 2 +j, and calculate the two rotation angles θ of the two-degree-of-freedom rotary joint in the k-th row respectively k =(i - 1)·δ and Collect the current cumulative count values [H 1k ,H 2k of the two drive motor position encoders, and generate the k-th row vector of the target model table Then set j = j + 1 and return to S102;

[0018] S104: Generate a target model table with N 1 ×N 2 rows.

[0019] Furthermore, the above-mentioned calibration of the mapping S→M from the sensor space S to the motor space M based on the mapping M→J for motor closed-loop control includes:

[0020] Let λ be a fixed calibration step size, and λ = n·δ, where n is a positive integer. Then the total number of calibration steps for the two rotation angles of the two-degree-of-freedom rotary joint are respectively

[0021] S201: Judge whether r ≤ N 3 holds, where r = 1, 2,..., N 3 , if so, set g = 1, then execute S202; otherwise, execute S206;

[0022] S202: Judge whether g ≤ N 4 holds, where g = 1, 2,..., N 4 , if so, execute S203; otherwise, set r = r + 1, then return to S201;

[0023] S203: Calculate the current model table row number k = (r - 1)·n·N 4 +(g - 1)·n + 1 and the current calibration step number q = (r - 1)·N 4 +g, the expected joint angle θ q =(r - 1)·λ, Take the motor position encoder cumulative count value [H 1k ,H 2k in the k-th row vector of the target model table as the expected motor position encoder cumulative count value [H 1q ,H 2q , then execute S204;

[0024] S204: During the set trajectory planning time period, the drive motor performs closed-loop motion based on until |H 1 - H 1q | < e and |H 2 - H 2q | < e, where e is the set error, the motor stops rotating, enters the steady-state sampling time period, and then S205 is executed;

[0025] S205: During the steady-state sampling time period, the three-dimensional Hall sensor data and the cumulative count value of the motor position encoder are sampled multiple times, and mean filtering is performed. The filtered result is used as the q-th row vector [B xq , B yq , B zq , H′ 1q , H′ 2q of the calibration mapping table. Then, make g = g + 1 and return to S202, where [B xq , B yq , B zq represents the three-dimensional magnetic induction intensity component after mean filtering, and [H′ 1q , H′ 2q represents the cumulative count value of the motor position encoder after mean filtering;

[0026] S206: Generate a calibration mapping table with N 3 N 4 rows.

[0027] Furthermore, the above-mentioned mean filtering is performed on all the collected three-dimensional Hall sensor data to obtain the sensor filtered data, which is expressed as:

[0028] [∑B x / M, ∑B y / M, ∑B z / M],

[0029] where M is the number of sampling times, and [B x , B y , B z represents the three-dimensional magnetic induction intensity component of the two-degree-of-freedom rotating joint.

[0030] Furthermore, after obtaining the initial cumulative count values [H 10 , H 20 of the two drive motor position encoders,

[0031] According to the initial cumulative count values [H 10 , H 20 of the two drive motor position encoders, look up the corresponding joint angle in the target model table to obtain the initial absolute joint angle.

[0032] The non-contact measurement method for the angles of a two-degree-of-freedom rotating joint based on a three-dimensional Hall sensor according to the present invention has the following beneficial effects:

[0033] The present invention measures the magnetic induction intensity of a permanent magnet embedded in a motion mechanism through a three-dimensional Hall sensor, and uses an indirect space mapping method to achieve decoupled measurement of the two-degree-of-freedom joint angles. The measurement device occupies extremely little space, which can greatly save the installation space. The present invention greatly reduces the number of required sensors, reduces the complexity of the sensor system and streamlines the installation volume, and improves the integration and flexibility of the overall design. The present invention utilizes the intermediate relationship between the motor space M in the joint space J and the sensor space S, establishes a mapping M→J from the motor space M to the joint space J based on the model analysis method, realizes the closed-loop control of the motor based on this mapping using a position encoder, and then collects the sensor and motor position signal datasets, calibrates the mapping S→M from the sensor space S to the motor space M, and finally realizes the mapping S→J from the sensor space S to the joint space J, achieving the decoupled calculation of the joint angles of the two-degree-of-freedom rotating joint. The non-contact measurement method effectively eliminates the service life problem caused by contact wear of traditional sensors, and improves the reliability and stability of the system. In summary, the present invention can not only meet the requirements of miniaturization and light weight, but also has the characteristics of high precision and high stability, greatly promoting the development and application of robots and measurement systems, especially in scenarios requiring high integration and high reliability, and has broad application prospects. Description of the Drawings

[0034] Figure 1 It is a schematic diagram of a two-degree-of-freedom rotating joint mechanism;

[0035] Figure 2 It is the overall flowchart of the non-contact measurement method for the angles of a two-degree-of-freedom rotating joint based on a three-dimensional Hall sensor;

[0036] Figure 3 It is the flowchart of the mapping method from the motor space M to the joint space J;

[0037] Figure 4 It is the flowchart of the mapping method from the sensor space S to the motor space M;

[0038] Figure 5 It is the flowchart of the mapping method from the sensor space S to the joint space J. Detailed Embodiments

[0039] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention. It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments may be combined with each other.

[0040] Referring to Figures 1 to 5 The present embodiment will be specifically described. The non-contact measurement method for the angles of a two-degree-of-freedom rotating joint based on a three-dimensional Hall sensor includes:

[0041] For the two-degree-of-freedom rotating joint mechanism, the two-degree-of-freedom motion of the joint is realized by a coupled driving mechanism, and the mechanism design scheme has a certain replaceability. The two degrees of freedom are specifically manifested as the rotational degree of freedom about the X-axis, and the rotation angle is the rotational degree of freedom about the Y-axis, and the rotation angle is θ. The power elements of the driving mechanism are motors with position encoders. The two motors jointly drive the two-degree-of-freedom joint. The cumulative count values of the motor position encoders are output to the controller in real time. The schematic diagrams of the two joints are as shown in Figure 1 A and B in; Figure 1 In, C is a permanent magnet embedded in the two-degree-of-freedom joint solid mechanism as the excitation source; Figure 1 In, D is a three-dimensional Hall sensor chip for measuring the three-dimensional magnetic induction intensity components outside the mechanism.

[0042] The two-degree-of-freedom motion mechanism has certain coupling characteristics. The joint angles are coupled, and it is impossible to calculate the relative position of the joint through the cumulative count values of a single motor position encoder. The mapping from the motor space M to the joint space J can be represented by a binary function, that is, θ = f(H 1 , H 2 ), where H 1 and H 2 respectively represent the cumulative count values of the two driving motor position encoders. The form of this binary function is related to the motion characteristics of the mechanism. The form of this function is obtained by fitting. The independent variables of the function are H 1 and H 2 , and the dependent variables are θ and A better fitting effect can be achieved through a quartic polynomial, and the fitting error is 0.1 degree. It is applicable to the decoupling of joint angles of similar coupled mechanisms. The fitting method and the function form depend on the characteristics of the mechanism.

[0043] Such as Figure 2As shown in the figure, this embodiment uses the motor space M as the intermediary between the joint space J and the sensor space S. The mapping from the motor space M to the joint space J is M→J, the mapping from the sensor space S to the motor space M is S→M, and finally the mapping from the sensor space S to the joint space J is realized as S→J.

[0044] As Figure 3 shown, the mapping M→J from the motor space M to the joint space J is a simulation method based on model analysis. The simulation model and the physical mechanism need to design corresponding mechanical limits according to the zero-finding requirements for the motion zero position. Identify the zero position. In this embodiment, the zero position satisfies θ 0 = 0, H 1 = 0, H 2 = 0. In the mechanical design process before the simulation, the motion range of the mechanism should have been given. Therefore, in the simulation, set the corresponding interval: θ∈[0,θ m , The M→J mapping is related to the motion accuracy of the mechanism. Therefore, the simulation step size δ of the simulation model should be one order of magnitude lower than the estimated joint angle measurement accuracy. In this embodiment, the simulation step size δ = 0.05 degrees is designed. Therefore, for this step size, the corresponding total number of simulation steps can be calculated Use a double loop to achieve incremental scanning of the joint angle. The inner loop is counted by j, and the outer loop is counted by i. Initialize the counting variable before the start of each loop. When the outer loop i is counted at any time, the inner loop j counts from 1 to N 2 The count is performed, and j = j + 1 is incremented each time the loop is executed. The following operations are performed for each loop:

[0045] 1. Calculate the row number k of the model table corresponding to this loop = (i - 1)N 2 + j;

[0046] 2. Calculate the joint angle θ k = (i - 1)·δ and

[0047] 3. Collect the current cumulative count value [H 1k , H 2k ;

[0048] Based on the above operation results, generate the k-th row vector of the target model table

[0049] The outer loop i counts from 1 to N 1 The count is performed, and i = i + 1 is incremented each time the loop is executed. After the loop ends, construct a target model table with a starting step size of δ and a number of rows of N 1 ×N 2 using this row vector.

[0050] The specific process can be written as:

[0051] Initialization: θ 0 = 0, H 1 = 0, H 2 = 0, i = 1, j = 1;

[0052] S101: Determine whether i ≤ N 1 holds. If so, set j = 1 and then execute S102; otherwise, execute S104;

[0053] S102: Determine whether j ≤ N 2 holds. If so, execute S103; otherwise, set i = i + 1, and then return to S101;

[0054] S103: Calculate the current model table row number k = (i - 1)N 2 + j, calculate the two - degree - of - freedom joint angle θ k = (i - 1)·δ and collect the current cumulative count value of the simulation position encoder [H 1k ,H 2k , and then generate the k - th row vector of the target model table Then set j = j + 1 and return to S102;

[0055] S104: Generate a target model table with N 1 ×N 2 rows.

[0056] The target model table realizes the corresponding relationship between the joint angle and the cumulative count value of the drive motor position encoder. This corresponding relationship can realize the mapping M→J from the motor space M to the joint space J based on the motion zero position. Through the mapping M→J, the actual joint angle in the joint space can be calculated by the cumulative count value of the motor position encoder for motor closed - loop control.

[0057] The mapping M→J is established on the basis of returning to the motion zero position at the initial power - on, while the mapping S→M from the sensor space S to the motor space M calibrates the motor space through the three - dimensional Hall sensor data, realizing the detection of the offset of any state before power - on relative to the mechanism motion zero position. Therefore, it can effectively avoid the limitation of the motion zero position on motor control and realize the initial position detection of the joint in any state within the motion range when powered on.

[0058] As Figure 4 shown, the mapping S→M from the sensor space S to the motor space M is realized by the method of full - space closed - loop motion calibration on the basis of the mapping M→J. The full - space scanning calibration first needs to ensure the high consistency between the physical device and the simulation and the consistency of the motion zero position of the model and the physical device. In this embodiment, the mapping S→M is realized as:

[0059] At the initial power-on, the motor is made to move in open loop, and it is ensured to return to the zero position of the mechanism's movement through mechanical limit, i.e., θ = 0, and then the cumulative count value of the current motor position encoder is cleared, i.e., H 1 = 0, H 2 = 0. The discrete point interval λ of the joint angle is set as the calibration step size. In order to query the target model table, this calibration step size λ should be an integer multiple of the simulation step size δ, i.e., λ = n·δ (n is a positive integer representing the multiple), and in this embodiment, λ = 10. Then the total number of calibration steps for the two rotation angles of the two-degree-of-freedom rotating joint are respectively The calibration process consists of two nested loops. The outer loop is counted by r, the inner loop is counted by g, and the total number of calibration steps is counted by q, q = (r - 1)·N 4 + g, and the data of the q-th step of calibration corresponds to the row number k of the target model table as k = (r - 1)·n·N 4 + (g - 1)·n + 1. Based on this corresponding relationship, the k-th row in the target model table corresponding to the total number of calibration steps q can be found to obtain the desired joint angle θ q , and the desired cumulative count value H of the motor position encoder 1q , H 2q , and then closed-loop control is implemented based on this controlled variable. At each calibration step size λ, the single-step period is T, which includes the trajectory planning time t 1 and the steady-state sampling time T - t 1 . During the set trajectory planning time t 1 , the motor moves in closed loop, and the cumulative count value H of the motor position encoder 1 = ΔH 1 , H 2 = ΔH 2 , ΔH 1 and ΔH 2 are respectively the cumulative count increments of the two drive motor position encoders. Through closed-loop movement, based on the set calibration step size λ, a full-space scan of the desired joint angle is achieved. The scanning method is not restricted, and all combinations of the discrete joint angle positions need to be traversed. The cumulative count value [H 1 , H 2 of the motor position encoder during movement and the desired cumulative count value [H 1q , H 2qWhen the error reaches within the set error e, it indicates that the target position of the movement is reached. At this time, the motor stops rotating and enters the steady-state sampling time. During the steady-state sampling time, the cumulative count value of the motor position encoder and the magnetic induction intensity component data of the three-dimensional Hall sensor are collected multiple times. After the timing time t of the timer reaches the calibration step period T, the sampling stops, and then mean filtering is performed to obtain the calibration data row vector [B xq ,B yq ,B zq ,H 1q ,H 2q of each calibration step, and this row vector is output to the q-th row of the calibration mapping table. Then, it enters the next calibration step, and the count g is incremented by g = g + 1. When g > N 4 , it enters the outer loop, and r is incremented by r = r + 1. When r > N 3 , the outer loop stops, and the closed-loop calibration process is completed, generating a calibration mapping table with N 3 N 4 rows, completing the mapping S→M from the sensor space S to the motor space M. This calibration mapping table is incorporated into the controller program and needs to be retrieved once before each power-on startup to determine the absolute joint angle at the initial power-on, and it will not be used subsequently.

[0060] The specific process can be written as:

[0061] Initialization: Under the open-loop motion condition of the motor, adjust θ 0 = 0, H 1 = 0, H 2 = 0, g = 1, r = 1;

[0062] S201: Determine whether r ≤ N 3 holds. If so, set g = 1 and then execute S202; otherwise, execute S206.

[0063] S202: Determine whether g ≤ N 4 holds. If so, execute S203; otherwise, set r = r + 1 and then return to S201.

[0064] S203: Calculate the current model table row number k = (r - 1)·n·N 4 + (g - 1)·n + 1 and the current calibration step number q = (r - 1)·N 4 + g, the expected joint angle θ q = (r - 1)·λ, Take the cumulative count value [H 1k ,H 2k of the motor position encoder in the k-th row vector of the target model table as the expected cumulative count value [H 1q ,H 2q, and then execute S204;

[0065] S204: During the set trajectory planning time period, the drive motor is based on to perform closed-loop motion until |H 1 - H 1q | < e and |H 2 - H 2q | < e, the motor stops rotating, enters the steady-state sampling time period, and then executes S205;

[0066] S205: During the steady-state sampling time period, sample the three-dimensional Hall sensor data and the cumulative count value of the motor position encoder multiple times, and perform mean filtering to obtain the q-th row vector [B xq , B yq , B zq , H′ 1q , H′ 2q of the calibration mapping table, then make g = g + 1, and return to S202;

[0067] S206: Generate a calibration mapping table with N 3 N 4 rows.

[0068] The calibration mapping table gives the corresponding relationship between the three-dimensional Hall sensor data and the actual cumulative count value of the corresponding motor position encoder.

[0069] As Figure 5 shown, when the system is powered on, the joint may be in any state within the joint range. At this time, the mechanism is stationary. After power-on, the three-dimensional Hall sensor data is sampled in this state. Due to the data fluctuation of the sensor, multiple samplings are required. In this embodiment, the set number of samplings M = 50, and M samplings are realized through a loop. After the sampling is completed, the corresponding mean filtering is performed to obtain the sensor filtered data [ΣB x / M, ∑B y / M, ∑B z / M]. According to this data, find the corresponding row in the calibration mapping table. In this embodiment, the search method finds the nearest neighbor row vector through the combined data of the above three-dimensional magnetic induction intensity components, or other methods can also be used to obtain the corresponding cumulative count value of the motor position encoder in this row. Take the offset value of the cumulative count of the motor position encoder as the initial cumulative count value of the motor position encoder [H 10 , H 20 . At this time, the initial absolute joint angle detection is completed. This process should be independent of the real-time control program of the motor. After the absolute joint angle detection is completed, the motor drives the joint to move. Due to the movement of the motor, the cumulative count value of the motor position encoder changes, and H 1 = H 10 + ΔH 1 , H2 = H 20 + ΔH 2 , this count represents the relative joint angle change during the movement of the motor, and realizes the real-time output of the absolute joint angle through the mapping M→J, that is, realizes the mapping S→J from the sensor space to the joint space.

[0070] The specific process can be written as:

[0071] S301: Sample the three-dimensional Hall sensor data M times, and then perform mean filtering on all the collected three-dimensional Hall sensor data to obtain the sensor filtered data [∑B x / M, ∑B y / M, ∑B z / M];

[0072] S302: Search for the row where [ΣB x / M, ΣB y / M, ΣB z / M] is located in the calibration mapping table, and use the accumulated count value of the motor position encoder in this row as the initial value of the accumulated count of the motor position encoder [H 10 , H 20 , search for the corresponding joint angle of [H 10 , H 20 in the target model table to complete the detection of the initial absolute joint angle;

[0073] S303: The motor drives the joint to move, and obtains the real-time accumulated count value H of the motor position encoder 1 = H 10 + ΔH 1 , H 2 = H 20 + ΔH 2 ;

[0074] S304: According to the accumulated count value H of the motor position encoder 1 , H 2 , search for the corresponding joint angle of the H 1 , H 2 in the target model table to realize the real-time output of the absolute joint angle.

[0075] The measurement method proposed in this embodiment is applicable to various two-degree-of-freedom and one-degree-of-freedom rotational motion mechanisms. The measurement device has no special requirements for the design of the mechanism, and only needs to ensure the installation space for embedding the permanent magnet and the installation space for the sensor below the permanent magnet. An axially polarized cylindrical permanent magnet is used, and the size, polarization intensity, etc. of the magnet are determined according to the design configuration of the mechanism. In this embodiment, a permanent magnet with a diameter of 3-6 mm can ensure that the sensor operates within the measurable range of the sensor, ±160 mT. A 4-mm permanent magnet is selected in this embodiment. Since the magnetic induction intensity decays rapidly outside the permanent magnet, the size of the air gap is generally in the order of millimeters, and the air gap is 2 mm in this embodiment. The above dimensions are the physical dimensions required for the measurement device. It should be noted that the distribution shape of the magnetic induction intensity of the permanent magnet is affected by multiple factors such as the designed motion range of the mechanism, the installation direction of the three-dimensional Hall sensor, the air gap between the permanent magnet and the three-dimensional Hall sensor, the installation position, the installation deviation, the polarization direction of the permanent magnet, the diameter of the permanent magnet, and the height of the permanent magnet. Before specific implementation, reasonable design dimensions can be obtained through theoretical analysis.

[0076] In summary, the present invention proposes a non-contact measurement method for the angles of a two-degree-of-freedom rotating joint based on a single three-dimensional Hall sensor. This method measures the three-dimensional magnetic induction intensity components of the moving joint with the embedded permanent magnet through the three-dimensional Hall sensor, and uses the indirect space mapping method to achieve the decoupled measurement of the two-degree-of-freedom joint angles. First, using the model analysis method, the motion mapping model from the motor space M to the joint space J is obtained; then, based on the above motion mapping model, the motor closed-loop control is carried out to obtain the data set of the sensor and motor position signals, and the mapping model from the sensor space S to the motor space M is calibrated; finally, using the motor space M as a medium, the mapping from the sensor space S to the joint space J is realized, and the angle detection of the two-degree-of-freedom joint is completed. This method innovatively utilizes the characteristics of the three-dimensional Hall sensor, applies it to the two-degree-of-freedom motion joint, realizes the calculation of the two motion degrees of freedom through the data of a single sensor, can calibrate the initial joint absolute angle for the joint powered at any joint angle, and uses a drive motor with an incremental encoder to calculate the relative motion angle, thereby realizing the angle measurement of the two-degree-of-freedom joint by a single sensor.

[0077] The magnetic induction intensity of the permanent magnet embedded in the motion mechanism is measured by a three-dimensional Hall sensor, and the decoupled measurement of the two-degree-of-freedom joint angle is realized by using the indirect space mapping method. The size of the measuring device is affected by the size of the permanent magnet, the size of the sensor, and the gap between the sensor and the permanent magnet, occupying extremely small space and being able to greatly save the installation space. This method utilizes the intermediate relationship of the motor space M in the joint space J and the sensor space S, establishes the mapping M→J from the motor space M to the joint space J based on the model analysis method, realizes the closed-loop control of the motor by using a position encoder based on this mapping, and then collects the data sets of the sensor and motor position signals, calibrates the mapping S→M from the sensor space S to the motor space M, and finally realizes the mapping S→J from the sensor space S to the joint space J, achieving the decoupled calculation of the joint angle of the two-degree-of-freedom rotary joint. The measurement method proposed by the present invention greatly reduces the number of required sensors, reduces the complexity of the sensor system and streamlines the installation volume, and improves the integration and flexibility of the overall design. The non-contact design of this method also effectively eliminates the service life problem caused by contact wear of traditional sensors, and improves the reliability and stability of the system. This new type of rotary joint angle measurement method can not only meet the requirements of miniaturization and light weight, but also has the characteristics of high precision and high stability, greatly promoting the development and application of robots and measurement systems, especially in scenarios requiring high integration and high reliability, and having broad application prospects.

[0078] Although the present invention has been described herein with reference to particular embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. It should therefore be understood that numerous modifications may be made to the exemplary embodiments, and that other arrangements may be designed, provided they do not depart from the spirit and scope of the invention as defined by the appended claims. It should be understood that the different dependent claims and the features described herein may be combined in a manner different from that described in the original claims. It should also be understood that the features described in connection with a separate embodiment may be used in other described embodiments.

Claims

1. A non-contact measurement method for a two-degree-of-freedom rotation joint angle based on a three-dimensional Hall sensor, characterized in that: include: A mapping M→J from motor space M to joint space J is constructed, and the motor closed-loop control calibration is performed based on the mapping M→J. The mapping S→M from sensor space S to motor space M is implemented using motor space M as an intermediary, thereby completing the contactless measurement of the two-degree-of-freedom rotational joint angle.

2. The contactless measurement method of a two-degree-of-freedom rotation joint angle based on a three-dimensional Hall sensor according to claim 1 is characterized in that: The method uses the motor space M as an intermediary to realize the mapping S→J from the sensor space S to the joint space J, and completes the contactless measurement of the two-degree-of-freedom rotation joint angle, including: When the two-degree-of-freedom rotating joint is in a stationary state, the three-dimensional magnetic induction intensity components of the two-degree-of-freedom rotating joint are collected multiple times by using a three-dimensional Hall sensor, and a permanent magnet is embedded in the two-degree-of-freedom rotating joint; Perform mean filtering on all collected three-dimensional Hall sensor data to obtain sensor filtering data; The motor position encoder cumulative count value corresponding to the sensor filter data is searched in the calibration mapping table and used as the initial value of the cumulative count of the two drive motor position encoders [H 10 ,H 20 ]; Drive the two-degree-of-freedom rotation joint to move, and obtain the accumulated count value H1=H of the two drive motor position encoders in real time 10 +ΔH1 and H2 = H 20 +ΔH2, where ΔH1 and ΔH2 are the accumulated count increments of the position encoders of the two drive motors respectively; According to the accumulated count values ​​H1 and H2 of the two drive motor position encoders, the corresponding joint angles are searched in the target model table to achieve real-time output of the absolute joint angles; The calibration mapping table is used to represent the correspondence between the three-dimensional Hall sensor data and the accumulated count values ​​of the two drive motor position encoders, and the target model table is used to represent the correspondence between the joint angle and the accumulated count values ​​of the two drive motor position encoders.

3. The contactless measurement method of a two-degree-of-freedom rotation joint angle based on a three-dimensional Hall sensor according to claim 1 or 2, characterized in that: The construction of the mapping M→J from the motor space M to the joint space J includes: Let δ be the simulation step length, let the initial values ​​of the two rotation angles of the two-degree-of-freedom rotation joint be 0, and the intervals of the two rotation angles of the two-degree-of-freedom rotation joint are θ m is the upper limit of θ, and They are The lower and upper limits of the two rotation angles of the two-degree-of-freedom rotational joint are S101: Determine whether i≤N1 holds, where i=1,2,...,N1. If yes, set j=1 and execute S102. Otherwise, execute S104. S102: Determine whether j≤N2 holds, where j=1,2,...,N2. If yes, execute S103; otherwise, set i=i+1 and return to S101; S103: Calculate the number of rows in the current model table k=(i-1)N2+j, and calculate the two rotation angles θ of the two-degree-of-freedom rotation joint in the kth row respectively. k =(i-1)·δsum Collect the current accumulated count values ​​of the two drive motor position encoders [H 1k ,H 2k ], generate the kth row vector [H 1k ,H 2k ,θ k , ], then set j=j+1 and return to S102; S104: Generate a target model table with N1×N2 rows.

4. The contactless measurement method of two-degree-of-freedom rotation joint angle based on three-dimensional Hall sensor according to claim 3 is characterized in that: The method of performing a motor closed-loop control calibration mapping S→M from a sensor space S to a motor space M based on the mapping M→J includes: Assume λ is the calibration step length, and λ = n·δ, n is a positive integer, then the total number of calibration steps for the two rotation angles of the two-degree-of-freedom rotation joint is S201: Determine whether r≤N3 holds, where r=1,2,...,N3. If so, set g=1 and then execute S202. Otherwise, execute S206. S202: Determine whether g≤N4 holds, where g=1,2,...,N4. If so, execute S203. Otherwise, set r=r+1, and then return to S201. S203: Calculate the current model table row number k = (r-1)·n·N4+(g-1·n+1) and the current calibration step number q = (r-1·N4+g), and the expected joint angle θ q =(r-1)·λ, The motor position encoder of the k-th row vector in the target model table accumulates the count value [H 1k ,H 2k ] as the expected motor position encoder accumulated count value [H 1q ,H 2q ], then execute S204; S204: During the set trajectory planning time period, the driving motor is based on [θ q , ,H 1q ,H 2q ] to perform closed-loop motion until |H1-H 1q |<e and |H2-H 2q |<e, e is the set error, the motor stops, enters the steady-state sampling period, and then executes S205; S205: In the steady-state sampling time period, the three-dimensional Hall sensor data and the motor position encoder accumulated count value are sampled multiple times, and mean filtering is performed, and the filtering result is used as the qth row vector [B xq ,B yq ,B zq ,H′ 1q ,H′ 2q ], then set g=g+1, and return to S202, where [B xq ,B yq ,B zq ] represents the three-dimensional magnetic induction intensity component after mean filtering, [H′ 1q ,H′ 2q ] represents the accumulated count value of the motor position encoder after mean filtering; S206: Generate a calibration mapping table with N3N4 rows.

5. The contactless measurement method of two-degree-of-freedom rotation joint angle based on three-dimensional Hall sensor according to claim 1, 2, 3 or 4, characterized in that: The sensor filtering data is obtained by performing mean filtering on all the collected three-dimensional Hall sensor data. The sensor filtering data is expressed as: [∑B x / M,∑B y / M,∑B z / M], Where M is the number of sampling times, [B x ,B y ,B z ] represents the three-dimensional magnetic induction intensity component of the two-degree-of-freedom rotation joint.

6. The contactless measurement method of two-degree-of-freedom rotation joint angle based on three-dimensional Hall sensor according to claim 5 is characterized in that: After obtaining the initial value of the accumulated count of the two drive motor position encoders [H 10 ,H 20 ]after, According to the initial value of the accumulated count of the two drive motor position encoders [H 10 ,H 20 ], look up the corresponding joint angle in the target model table and obtain the initial absolute joint angle.

Images