Kinematic calibration method of hybrid robot based on internal sensor information
By dividing motion chains and designing internal sensing systems in a five-degree of freedom hybrid robot, building an error model, real-time pose error compensation is achieved, and the problems of insufficient error detection and unreal-time compensation in the existing technology are solved, improving The robot position accuracy is achieved.
Patent Information
- Application Number
- CN202211337039.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-10-28
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2042-10-28
AI Technical Summary
The kinematic calibration method of the existing five-degree-of-freedom hybrid robot has insufficient attitude error detection of end components, cannot obtain position and attitude error information in real time, can only perform offline compensation, the compensation program is complex and takes up a large storage space, and cannot effectively compensate the driving joints. Reverse gap and passive joint motion uncertainty errors lead to limited improvement in posture accuracy.
Using a hybrid robot kinematic calibration method based on internal sensing information, the internal sensing system is designed and installed by dividing active and driven motion chains, and the error model is constructed using regular regression algorithm and least squares method to realize the zero-point compensation and Structural error identification, establish a real-time pose error feedback system, and perform online real-time error compensation.
Real-time detection and precise compensation of end-position pose errors of hybrid robots are realized, compensation procedures are simplified, storage requirements of CNC systems are reduced, and reverse gaps of driving joints and passive joint errors are effectively compensated, greatly improving posture accuracy.
Smart Images

Figure CN115533876B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a robot kinematic calibration technology, and belongs to the field of robots. More specifically, the invention relates to a hybrid robot kinematic calibration method based on internal sensor information. Background Art
[0002] The five-degree-of-freedom hybrid robot, which is built by a three-degree-of-freedom parallel mechanism and a two-degree-of-freedom serial rotary head, is a new type of five-axis linkage processing equipment. It combines the characteristics of a large working space of a serial mechanism and a large acceleration and high rigidity of a parallel mechanism. It aims at high-speed and high-precision processing and has broad application prospects in the field of processing and manufacturing in large-scale non-structural environments. Five-degree-of-freedom hybrid robots such as Tricept and Exechon have been initially applied in aerospace, automobile manufacturing and other fields, and have achieved good results. Accuracy is an important performance indicator of this type of five-degree-of-freedom hybrid robot. On the premise of having good basic manufacturing accuracy, kinematic calibration is an effective means to further improve the accuracy of this type of equipment. Usually, kinematic calibration is divided into the following four steps: error modeling, error detection, error identification and compensation implementation. However, the limitations of the above methods are mainly manifested in the following aspects: first, the means for detecting the posture errors of the robot's end components are insufficient, the detection accuracy is limited, and the position and posture error information of the end components cannot be obtained in real time; second, only offline measurement and offline compensation can be used, and online real-time error compensation cannot be achieved; third, the posture error compensation solution procedure is complex, the real-time performance of the compensation is poor, and a large storage space of the CNC system is occupied; fourth, the reverse clearance of the hybrid robot's driving joints and the motion uncertainty errors of the passive joints cannot be effectively compensated, and the effect of improving the robot's posture accuracy is limited. Summary of the invention
[0003] The purpose of the present invention is to overcome the deficiencies in the prior art and to propose a high-precision, high-efficiency, and industrial field-friendly hybrid robot kinematic calibration method based on internal sensor information, which can directly complete posture error compensation based on the robot's internal error detection data.
[0004] The objectives of the present invention are achieved through the following technical solutions.
[0005] The present invention discloses a kinematic calibration method for a hybrid robot based on internal sensor information. The hybrid robot is composed of a parallel mechanism and a serial rotary head. The parallel mechanism is composed of a first active telescopic device, a second active telescopic device, a third active telescopic device, a driven telescopic device, a moving platform, a first rotating support structure, a second rotating support structure, a first frame bearing seat, and a second frame bearing seat. The serial rotary head is composed of a first rotating component, a second rotating component, a spindle, and a tool. The calibration method comprises the following steps:
[0006] Step 1: Design, installation and calibration of the internal sensor system of the hybrid robot
[0007] 1) Division of hybrid robot motion system
[0008] According to the kinematic equivalence principle, the hybrid robot is divided into two independent motion systems, an active motion chain and a driven motion chain; the active motion chain is composed of a first active telescopic device, a second active telescopic device, a third active telescopic device, a moving platform, a first rotating support structure, a second rotating support structure, and a serial rotating head; the driven motion chain is composed of a first rotating support structure, a driven telescopic device, a moving platform, and a serial rotating head;
[0009] 2) Design and installation of internal sensor system
[0010] According to the principle that the motion generating element corresponding to the internal sensor can form a controllable motion space of the terminal motion output component, an internal sensing system is designed and installed on the driven motion chain; the internal sensing system is composed of a first circular grating, a second circular grating, a third linear grating, a fourth circular grating, and a fifth circular grating, wherein the first circular grating is installed on the rotating pair between the first rotating support structure and the first frame bearing seat, the second circular grating is installed on the rotating pair between the first rotating support structure and the driven telescopic device, the third linear grating is installed on the moving pair in the driven telescopic device, the fourth circular grating is installed inside the rotating pair between the first rotating component and the moving platform, and the fifth circular grating is installed inside the rotating pair between the first rotating component and the second rotating component;
[0011] 3) Calibration of internal sensor system
[0012] An external measuring instrument is used to obtain the six-dimensional posture error of the tool when the hybrid robot moves to different positions in the working space, and the error transfer equation group is constructed using the drive joint error transfer model, and the zero-point error of the five drive joints of the robot is solved using the regular regression algorithm; the inverse of the obtained zero-point error is stored in the zero-point bias variable corresponding to each drive joint in the numerical control system to achieve the zero-point error compensation of the drive joint of the hybrid robot; on this basis, each drive joint of the hybrid robot is controlled to move to the zero position, and the readings of each grating in the current position are cleared, and the position of each grating at this time is defined as the zero position of each grating, thereby achieving the zero-position calibration of the internal sensor system;
[0013] Step 2: Identification of structural error parameters of the driven kinematic chain of the hybrid robot
[0014] 1) Construction of the ideal kinematic relationship between the active kinematic chain and the passive kinematic chain
[0015] The mapping model between the active kinematic chain driving joint variables and the tool posture under error-free condition is established by using the mechanism method.主 (q 主 ), where p is the six-dimensional position vector of the tool, q 主 is the active kinematic chain driving joint variable, f 主 Q 主 The mapping function between p and p; the mapping model between the reading of the internal sensor system of the driven motion chain and the tool posture is established by using the mechanism method under the error-free state. Among them, q 从 is the generalized driving joint variable of the driven motion chain, and its value is obtained from the reading of the internal sensor system of the driven motion chain. For p and q 从 The mapping function between
[0016] Combining the above two functional relationships, we can solve the active kinematic chain driving joint variable q under error-free state: 主 The generalized driving joint variable q of the driven kinematic chain 从 The mapping function between:
[0017]
[0018] 2) Establishment of comprehensive error model of driven motion chain
[0019] According to whether the error source of the driven motion chain can be directly detected by the internal sensor system, the error source in the driven motion chain is decomposed into the generalized driving joint motion error Δq 从 and structural error Δs; the generalized drive joint motion error Δq 从 It refers to the joint motion error corresponding to the measured value of the internal sensor system, which is generalized to the measured value q of the joint variable driven by the driven motion chain 从,实测 With the theoretical value q 从,理论 Take the difference and get, that is, Δq 从 =q 从,实测 -q 从,理论 ,q 从,理论 Calculated by formula (1); the structural error Δs includes the processing error, assembly error and axis deviation of the components of the driven motion chain;
[0020] The integrated error model of the driven motion chain with linear expression is established by using the screw method, POE method or DH method as follows:
[0021] Δp=AΔq 从 +BΔs (2)
[0022] Where Δp is the tool posture error vector, A is the generalized drive joint motion error mapping matrix, and B is the structural error mapping matrix;
[0023] 3) Measurement and solution of tool posture error
[0024] When the robot moves to any position, the laser tracker is used to measure the spatial position errors of the three non-collinear points P1, P2, and P3 on the main axis, which are recorded as Δr P,1 , Δr P,2 , Δr P,3 ; Point P m Position error Δr (m=1,2,3) P,m Expressed as
[0025] Δr P,m =Δr+Δθ×Rr P,m,0 ,m=1,2,3 (3)
[0026] Where Δr and Δθ are the position and attitude error vectors of the tool, respectively. P,m,0 The tool tip points to point P m The vector of , R is the posture matrix of the tool relative to the hybrid robot frame;
[0027] Combining the three equations in equation (3) and rewriting them into a matrix format, we have
[0028] Δr P =KΔp (4)
[0029]
[0030] In the formula, Δr P is the vector formed by the position errors of the three points, Δp is the tool posture error vector, and K is Δr P and Δp, I3 is the third-order unit matrix; according to formula (4) and the least squares method, the tool posture error vector Δp of the robot at the current configuration is solved
[0031] Δp=(K T K) -1 K T Δr P (5)
[0032] In the formula, K T is the transposed matrix of matrix K;
[0033] 4) Identification of structural errors of driven kinematic chains
[0034] The robot is controlled to move to n positions in its workspace. At each position, the tool posture error vector Δp is obtained by using the tool posture error measurement and solution method, and the robot tool posture error vector at the i-th position is recorded as Δp i (i=1~n); using the integrated error model of the n driven kinematic chains corresponding to the n configurations (i.e., formula (2)), the identification equations of the structural error of the integrated driven kinematic chain are formed:
[0035]
[0036] Among them, Δp i is the robot tool pose error vector at the i-th configuration, A i is the generalized driving joint motion error mapping matrix of the driven kinematic chain at the i-th configuration, B i is the structural error mapping matrix of the driven kinematic chain at the i-th configuration, Δq 从,i is the generalized driving joint motion error of the driven kinematic chain at the i-th configuration, which can be calculated from the reading of the internal sensor system of the driven kinematic chain;
[0037] The selected n configurations must make the driven kinematic chain structural error Δs identifiable. Δs is identified using equation (6) and the least squares method. The solution is:
[0038]
[0039] In the formula, the letter T in the upper right corner of the matrix represents the transpose of the matrix;
[0040] Step 3: Compensation of robot posture error based on real-time measurement information from the internal sensor system of the driven motion chain
[0041] 1) Real-time prediction of the position and posture error of the end component of the hybrid robot
[0042] When the hybrid robot moves to any position in its workspace, the theoretical value q of the generalized driving joint variable of the driven kinematic chain is first calculated using equation (1): 从,理论 At the same time, the internal sensor system of the driven motion chain is used to read the measured value q of the generalized driving joint variable of the driven motion chain 从,实测 , and then solve the generalized drive joint motion error Δq 从 =q 从,实测 -q 从,理论 ;
[0043] Using the kinematic model of the hybrid robot and the comprehensive error model of the driven motion chain, the generalized drive joint motion error mapping matrix A of the driven motion chain and the structural error mapping matrix B of the driven motion chain at the current configuration are solved, and the driven motion chain structural error Δs obtained in step 2 and the generalized drive joint motion error Δq of the driven motion chain obtained in real time are converted into 从 Substitute the comprehensive error model of the driven motion chain into equation (2), and the tool posture error vector Δp of the hybrid robot is obtained in real time;
[0044] 2) Real-time solution of hybrid robot drive joint compensation
[0045] The solved tool posture error vector Δp and the ideal tool posture vector p 理想 Sum and get the tool's true pose vector p 实际 ; 实际 Substitute the mapping model between the active kinematic chain driving joint variables and the tool posture in the error-free state described in step 2 to obtain the active kinematic chain driving joint variables q′ under the posture error state. 主 , the calculation method is as follows:
[0046]
[0047] In the formula, is the ideal kinematic inverse function of the active kinematic chain of the hybrid robot;
[0048] The active kinematic chain ideally drives the joint variables q 主 The active kinematic chain driving joint variable q′ under the state of considering posture error 主 The real-time compensation value Δq of the hybrid robot driving joint is obtained by subtracting 主,补偿 , the calculation method is as follows:
[0049] Δq 主,补偿 =q 主 -q′ 主 (9)
[0050] 3) Real-time compensation of hybrid robot posture error based on internal sensor information
[0051] Control the active kinematic chain of the hybrid robot to drive the joint to move to q 主 +Δq 主,补偿 , use the internal sensor system of the driven motion chain to read the measured value of the generalized driving joint variable of the driven motion chain in the current state, and calculate the generalized driving joint motion error Δq in the current state 从 ; If Δq 从 The 2-norm of ||Δq 从 || is less than or equal to the given threshold ε, the real-time compensation process under the current configuration ends; if Δq 从 2-norm of ||Δq 从 If || is greater than the given threshold ε, the process described in step 3 is repeated until the accuracy reaches the target.
[0052] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:
[0053] (1) The method proposed in the present invention creatively sets an internal sensing system in the driven branch chain of the hybrid robot, uses the driven motion chain to achieve precision feedback, and uses the active motion chain to achieve posture error compensation, cleverly constructs a real-time feedback system for the terminal posture error of the hybrid robot, solves the problem that the terminal posture accuracy of the hybrid robot cannot be detected in real time and accurately, and realizes the online detection and precise real-time compensation of the engineering hybrid robot posture error.
[0054] (2) The method proposed in the present invention can conveniently embed the linearized posture error prediction model into a commercial CNC system, and then use the internal sensor information to realize posture error compensation, thereby solving the problems of the traditional kinematic calibration method, such as complex solution procedures, poor real-time compensation, and large storage space occupied by the CNC system.
[0055] (3) The method proposed in the present invention can effectively compensate for the backlash of the hybrid robot's driving joints and the motion uncertainty error of the passive joints, and can greatly improve the posture accuracy of the hybrid robot. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] Figure 1 It is a schematic diagram of a five-degree-of-freedom hybrid robot and its internal sensing system.
[0057] Figure markings: 1-first active telescopic device, 2-second active telescopic device, 3-third active telescopic device, 4-driven telescopic device, 5-moving platform, 6-first rotating support structure, 7-second rotating support structure, 8-first frame bearing seat, 9-second frame bearing seat, 11-first rotating component, 12-second rotating component, 13-spindle, 14-tool, 21-first circular grating, 22-second circular grating, 23-third linear grating. DETAILED DESCRIPTION
[0058] The present invention will be further described below in conjunction with the accompanying drawings.
[0059] like Figure 1As shown, the hybrid robot is composed of a parallel mechanism and a serial rotary head. The parallel mechanism is composed of a first active telescopic device 1, a second active telescopic device 2, a third active telescopic device 3, a driven telescopic device 4, a moving platform 5, a first rotating support structure 6, a second rotating support structure 7, a first frame bearing seat 8, and a second frame bearing seat 9. Among them, two first frame bearing seats 8 are provided, which are respectively connected to the two ends of the first rotating support structure 6 through a rotating pair, and the first active telescopic device 1, the second active telescopic device 2, and the driven telescopic device 4 are respectively connected to the first rotating support structure 6 through a rotating pair. Two second frame bearing seats 9 are provided, which are respectively connected to the two ends of the second rotating support structure 7 through a rotating pair, and the third active telescopic device 3 is connected to the second rotating support structure 7 through a rotating pair. The moving platform 5 is fixedly connected to the driven telescopic device 4, and the first active telescopic device 1, the second active telescopic device 2, and the third active telescopic device 3 are respectively connected to the moving platform 5 through a ball pair. The serial rotary head is composed of a first rotating component 11, a second rotating component 12, a spindle 13, and a tool 14. One end of the first rotating component 11 is connected to the moving platform 5 through a rotating pair, and the other end is connected to the second rotating component 12 through a rotating pair. The spindle 13 is fixedly connected to the second rotating component 12 , and the tool 14 is fixedly connected to the spindle 13 .
[0060] The kinematic calibration method of a hybrid robot based on internal sensor information of the present invention specifically comprises the following steps:
[0061] Step 1: Design, installation and calibration of the internal sensor system of the hybrid robot
[0062] 1) Division of hybrid robot motion system
[0063] According to the kinematic equivalence principle, the hybrid robot is divided into two independent motion systems: an active motion chain and a driven motion chain. The active motion chain is composed of a first active telescopic device 1, a second active telescopic device 2, a third active telescopic device 3, a moving platform 5, a first rotating support structure 6, a second rotating support structure 7, and a tandem rotating head. The driven motion chain is composed of a first rotating support structure 6, a driven telescopic device 4, a moving platform 5, and a tandem rotating head.
[0064] 2) Design and installation of internal sensor system
[0065] According to the principle that the motion generating element corresponding to the internal sensor can form a controllable motion space of the terminal motion output component, an internal sensing system is designed and installed on the driven motion chain. The internal sensing system is composed of a first circular grating 21, a second circular grating 22, a third linear grating 23, a fourth circular grating, and a fifth circular grating. Among them, the first circular grating 21 is installed on the rotating pair between the first rotating support structure 6 and the first frame bearing seat 8, the second circular grating 22 is installed on the rotating pair between the first rotating support structure 6 and the driven telescopic device 4, the third linear grating 23 is installed on the moving pair in the driven telescopic device 4, the fourth circular grating is installed inside the rotating pair between the first rotating component 11 and the moving platform 5, and the fifth circular grating is installed inside the rotating pair between the first rotating component 11 and the second rotating component 12.
[0066] 3) Calibration of internal sensor system
[0067] External measuring instruments such as laser trackers are used to obtain the six-dimensional position error of the tool when the hybrid robot moves to multiple different positions in the working space. The error transfer equation group is constructed using the drive joint error transfer model, and the zero-point error of the five drive joints of the robot is solved using the regular regression algorithm. The inverse of the obtained zero-point error is stored in the zero-point bias variable corresponding to each drive joint in the CNC system to achieve zero-point error compensation of the drive joint of the hybrid robot. On this basis, each drive joint of the hybrid robot is controlled to move to the zero position, and the readings of each grating in the current position are cleared, and the position of each grating at this time is defined as the zero position of each grating, thereby achieving zero-position calibration of the internal sensor system.
[0068] Step 2: Identification of structural error parameters of the driven kinematic chain of the hybrid robot
[0069] 1) Construction of the ideal kinematic relationship between the active kinematic chain and the passive kinematic chain
[0070] The mapping model between the active kinematic chain driving joint variables and the tool posture under error-free condition is established by using the mechanism method. 主 (q 主 ). Among them, p is the six-dimensional position vector of the tool, q 主 is the active kinematic chain driving joint variable, f 主 Q 主 The mapping function between and p. The mapping model between the reading of the internal sensor system of the driven motion chain and the tool posture is established by using the mechanism method under the error-free state. Among them, q 从 is the generalized driving joint variable of the driven motion chain, and its value is obtained from the reading of the internal sensor system of the driven motion chain. For p and q 从 The mapping function between .
[0071] Combining the above two functional relationships, we can solve the active kinematic chain driving joint variable q under error-free state: 主 The generalized driving joint variable q of the driven kinematic chain 从 The mapping function between:
[0072]
[0073] 2) Establishment of comprehensive error model of driven motion chain
[0074] According to whether the error source of the driven motion chain can be directly detected by the internal sensor system, the error source in the driven motion chain is decomposed into the generalized driving joint motion error Δq 从 The generalized drive joint motion error Δq 从 It refers to the joint motion error corresponding to the measured value of the internal sensor system, which can be generalized to the measured value q of the joint variable driven by the driven kinematic chain 从,实测 With the theoretical value q 从,理论 Take the difference and get, that is, Δq 从 =q 从,实测 -q 从,理论 , where q 从,理论 It can be calculated by formula (1). The structural error Δs includes error elements such as component processing error, assembly error, and axis deviation of the driven motion chain.
[0075] Using the screw method, POE method or DH method, the linear expression of the driven motion chain comprehensive error model can be established as follows:
[0076] Δp=AΔq 从 +BΔs (2)
[0077] Where Δp is the tool posture error vector, A is the generalized drive joint motion error mapping matrix, and B is the structural error mapping matrix.
[0078] 3) Measurement and solution of tool posture error
[0079] When the robot moves to any position, the spatial position errors of the three non-collinear points P1, P2, and P3 on the main axis are measured using instruments such as laser trackers, which are recorded as Δr P,1 , Δr P,2 , Δr P,3 . Set point P m Position error Δr (m=1,2,3) P,m Expressed as
[0080] Δr P,m =Δr+Δθ×Rr P,m,0 ,m=1,2,3 (3)
[0081] Where Δr and Δθ are the position and attitude error vectors of the tool, respectively. P,m,0 The tool tip points to point P m R is the posture matrix of the tool relative to the hybrid robot frame.
[0082] Combining the three equations in equation (3) and rewriting them into a matrix format, we have
[0083] Δr P =KΔp (4)
[0084]
[0085] In the formula, Δr P is the vector formed by the position errors of the three points, Δp is the tool posture error vector, and K is Δr P The mapping matrix between and Δp, I3 is the third-order unit matrix. According to equation (4) and the least squares method, the tool posture error vector Δp of the robot at the current configuration is solved:
[0086] Δp=(K T K) -1 K T Δr P (5)
[0087] In the formula, K T is the transposed matrix of matrix K.
[0088] 4) Identification of structural errors of driven kinematic chains
[0089] The robot is controlled to move to n positions in its workspace. At each position, the tool posture error vector Δp is obtained by using the tool posture error measurement and solution method, and the robot tool posture error vector at the i-th position is recorded as Δp i (i=1~n); using the integrated error model of the n driven kinematic chains corresponding to the n configurations (i.e., formula (2)), the identification equations of the structural error of the integrated driven kinematic chain are formed:
[0090]
[0091] Among them, Δp i is the robot tool pose error vector at the i-th configuration, A i is the generalized driving joint motion error mapping matrix of the driven kinematic chain at the i-th configuration, B i is the structural error mapping matrix of the driven kinematic chain at the i-th configuration, Δq 从,i is the generalized driving joint motion error of the driven motion chain at the i-th configuration, which can be calculated from the readings of the internal sensor system of the driven motion chain.
[0092] The selected n configurations must make the driven kinematic chain structural error Δs identifiable. In this way, Δs can be identified using equation (6) and the least squares method. The solution is:
[0093]
[0094] In the formula, the letter T in the upper right corner of the matrix represents the transpose of the matrix.
[0095] Step 3: Compensation of robot posture error based on real-time measurement information from the internal sensor system of the driven motion chain
[0096] 1) Real-time prediction of the position and posture error of the end component of the hybrid robot
[0097] When the hybrid robot moves to any position in its workspace, the theoretical value q of the generalized driving joint variable of the driven kinematic chain is first calculated using equation (1): 从,理论 At the same time, the internal sensor system of the driven motion chain is used to read the measured value q of the generalized driving joint variable of the driven motion chain 从,实测 , and then solve the generalized drive joint motion error Δq 从 =q 从,实测 -q 从,理论 .
[0098] Using the kinematic model of the hybrid robot and the comprehensive error model of the driven motion chain, the generalized drive joint motion error mapping matrix A of the driven motion chain and the structural error mapping matrix B of the driven motion chain at the current configuration are solved, and the driven motion chain structural error Δs obtained in step 2 and the generalized drive joint motion error Δq of the driven motion chain obtained in real time are converted into 从 Substituting the comprehensive error model of the driven motion chain described in equation (2) into the error vector Δp of the tool position of the hybrid robot can be obtained in real time.
[0099] 2) Real-time solution of hybrid robot drive joint compensation
[0100] The solved tool posture error vector Δp and the ideal tool posture vector p 理想 Sum and get the tool's true pose vector p 实际 ; 实际 Substitute the mapping model between the active kinematic chain driving joint variables and the tool posture in the error-free state described in step 2 to obtain the active kinematic chain driving joint variables q′ under the posture error state. 主 , the calculation method is as follows:
[0101]
[0102] In the formula, It is the ideal inverse kinematics function of the active kinematic chain of the hybrid robot.
[0103] The active kinematic chain ideally drives the joint variables q 主 The active kinematic chain driving joint variable q′ under the state of considering posture error 主 The real-time compensation value Δq of the hybrid robot driving joint is obtained by subtracting 主,补偿 , the calculation method is as follows:
[0104] Δq 主,补偿 =q 主 -q′ 主 (9)
[0105] 3) Real-time compensation of hybrid robot posture error based on internal sensor information
[0106] Control the active kinematic chain of the hybrid robot to drive the joint to move to q 主 +Δq 主,补偿 , use the internal sensor system of the driven motion chain to read the measured value of the generalized driving joint variable of the driven motion chain in the current state, and calculate the generalized driving joint motion error Δq in the current state 从 ; If Δq 从 2-norm of ||Δq 从 || is less than or equal to the given threshold ε, the real-time compensation process under the current configuration ends; if Δq 从 2-norm of ||Δq 从 || is greater than a given threshold ε, the process described in step 3 is repeated until the accuracy reaches the standard. The aforementioned threshold ε is determined according to the actual engineering requirements for robot accuracy. The higher the required accuracy, the smaller the value of the threshold ε (ε>0).
[0107] Although the functions and working processes of the present invention have been described above in conjunction with the accompanying drawings, the present invention is not limited to the above-mentioned specific functions and working processes. The above-mentioned specific implementation methods are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which are all within the protection of the present invention.
Claims
1. A kinematic calibration method for a hybrid robot based on internal sensor information, wherein the hybrid robot is composed of a parallel mechanism and a serial rotary head, wherein the parallel mechanism is composed of a first active telescopic device (1), a second active telescopic device (2), a third active telescopic device (3), a driven telescopic device (4), a moving platform (5), a first rotating support structure (6), a second rotating support structure (7), a first frame bearing seat (8), and a second frame bearing seat (9), and the serial rotary head is composed of a first rotating component (11), a second rotating component (12), a spindle (13), and a tool (14), characterized in that: The calibration method includes the following steps: Step 1: Design, installation and calibration of the internal sensor system of the hybrid robot 1) Division of hybrid robot motion system According to the kinematic equivalence principle, the hybrid robot is divided into two independent motion systems, an active motion chain and a driven motion chain; the active motion chain is composed of a first active telescopic device (1), a second active telescopic device (2), a third active telescopic device (3), a moving platform (5), a first rotating support structure (6), a second rotating support structure (7), and a tandem rotating head; the driven motion chain is composed of a first rotating support structure (6), a driven telescopic device (4), a moving platform (5), and a tandem rotating head; 2) Design and installation of internal sensor system According to the principle that the motion generating element corresponding to the internal sensor can form a controllable motion space of the terminal motion output component, an internal sensing system is designed and installed on the driven motion chain; the internal sensing system is composed of a first circular grating (21), a second circular grating (22), a third linear grating (23), a fourth circular grating, and a fifth circular grating, wherein the first circular grating (21) is installed on the rotating pair between the first rotating support structure (6) and the first frame bearing seat (8), the second circular grating (22) is installed on the rotating pair between the first rotating support structure (6) and the driven telescopic device (4), the third linear grating (23) is installed on the moving pair in the driven telescopic device (4), the fourth circular grating is installed inside the rotating pair between the first rotating component (11) and the moving platform (5), and the fifth circular grating is installed inside the rotating pair between the first rotating component (11) and the second rotating component (12); 3) Calibration of internal sensor system An external measuring instrument is used to obtain the six-dimensional posture error of the tool when the hybrid robot moves to different positions in the working space, and the error transfer equation group is constructed using the drive joint error transfer model, and the zero-point error of the five drive joints of the robot is solved using the regular regression algorithm; the inverse of the obtained zero-point error is stored in the zero-point bias variable corresponding to each drive joint in the numerical control system to achieve the zero-point error compensation of the drive joint of the hybrid robot; on this basis, each drive joint of the hybrid robot is controlled to move to the zero position, and the readings of each grating in the current position are cleared, and the position of each grating at this time is defined as the zero position of each grating, thereby achieving the zero-position calibration of the internal sensor system; Step 2: Identification of structural error parameters of the driven kinematic chain of the hybrid robot 1) Construction of the ideal kinematic relationship between the active kinematic chain and the passive kinematic chain The mapping model between the active kinematic chain driving joint variables and the tool posture under error-free condition is established by using the mechanism method. 主 (q 主 ), where p is the six-dimensional position vector of the tool, q 主 is the active kinematic chain driving joint variable, f 主 for q 主 The mapping function between p and p; the mapping model between the reading of the internal sensor system of the driven motion chain and the tool posture is established by using the mechanism method under the error-free state. in, is the generalized driving joint variable of the driven motion chain, and its value is obtained from the reading of the internal sensor system of the driven motion chain. For p and q 从 The mapping function between Combining the above two functional relationships, we can solve the active kinematic chain driving joint variable q under error-free state: 主 The generalized driving joint variable q of the driven kinematic chain 从 The mapping function between: 2) Establishment of comprehensive error model of driven motion chain According to whether the error source of the driven motion chain can be directly detected by the internal sensor system, the error source in the driven motion chain is decomposed into the generalized driving joint motion error Δq 从 and structural error Δs; the generalized drive joint motion error Δq 从 It refers to the joint motion error corresponding to the measured value of the internal sensor system, which is generalized to the measured value q of the joint variable driven by the driven motion chain 从,实测 With the theoretical value q 从,理论 Take the difference and get, that is, Δq 从 =q 从,实测 -q 从,理论 ,q 从,理论 Calculated by formula (1); the structural error Δs includes the processing error, assembly error and axis deviation of the components of the driven motion chain; The integrated error model of the driven motion chain with linear expression is established by using the screw method, POE method or DH method as follows: Δp=AΔq 从 +BΔs (2) Where Δp is the tool posture error vector, A is the generalized drive joint motion error mapping matrix, and B is the structural error mapping matrix; 3) Measurement and solution of tool posture error When the robot moves to any position, the laser tracker is used to measure the spatial position errors of the three non-collinear points P1, P2, and P3 on the main axis, which are recorded as Δr P,1 , Δr P,2 , Δr P,3 ; Point P m Position error Δr (m=1,2,3) P,m Expressed as Δr P,m =Δr+Δθ×Rr P,m,0 ,m=1,2,3 (3) Where Δr and Δθ are the position and attitude error vectors of the tool, respectively. P,m,0 The tool tip points to point P m The vector of , R is the posture matrix of the tool relative to the hybrid robot frame; Combining the three equations in equation (3) and rewriting them into a matrix format, we have Δr P =KΔp (4) In the formula, Δr P is the vector formed by the position errors of the three points, Δp is the tool posture error vector, and K is Δr P and Δp, I3 is the third-order unit matrix; according to formula (4) and the least squares method, the tool posture error vector Δp of the robot at the current configuration is solved Δp=(K T K) -1 K T Δr P (5) In the formula, K T is the transposed matrix of matrix K; 4) Identification of structural errors of driven kinematic chains The robot is controlled to move to n positions in its workspace. At each position, the tool posture error vector Δp is obtained by using the tool posture error measurement and solution method, and the robot tool posture error vector at the i-th position is recorded as Δp i (i=1~n); using the n driven motion chain comprehensive error model group corresponding to the n configurations to integrate the identification equation group of the driven motion chain structure error: Among them, Δp i is the robot tool pose error vector at the i-th configuration, A i is the generalized driving joint motion error mapping matrix of the driven kinematic chain at the i-th configuration, B i is the structural error mapping matrix of the driven kinematic chain at the i-th configuration, Δq 从,i is the generalized driving joint motion error of the driven kinematic chain at the i-th configuration, which can be calculated from the reading of the internal sensor system of the driven kinematic chain; The selected n configurations must make the driven kinematic chain structural error Δs identifiable. Δs is identified using equation (6) and the least squares method. The solution is: In the formula, the letter T in the upper right corner of the matrix represents the transpose of the matrix; Step 3: Compensation of robot posture error based on real-time measurement information from the internal sensor system of the driven motion chain 1) Real-time prediction of the position and posture error of the end component of the hybrid robot When the hybrid robot moves to any position in its workspace, the theoretical value q of the generalized driving joint variable of the driven kinematic chain is first calculated using equation (1): 从,理论 At the same time, the internal sensor system of the driven motion chain is used to read the measured value q of the generalized driving joint variable of the driven motion chain 从,实测 , and then solve the generalized drive joint motion error Δq 从 =q 从,实测 -q 从,理论 ; Using the kinematic model of the hybrid robot and the comprehensive error model of the driven motion chain, the generalized drive joint motion error mapping matrix A of the driven motion chain and the structural error mapping matrix B of the driven motion chain at the current configuration are solved, and the driven motion chain structural error Δs obtained in step 2 and the generalized drive joint motion error Δq of the driven motion chain obtained in real time are converted into 从 Substitute the comprehensive error model of the driven motion chain into equation (2), and the tool posture error vector Δp of the hybrid robot is obtained in real time; 2) Real-time solution of hybrid robot drive joint compensation The solved tool posture error vector Δp and the ideal tool posture vector p 理想 Sum and get the tool's true pose vector p 实际 ; 实际 Substitute the mapping model between the active kinematic chain driving joint variables and the tool posture in the error-free state described in step 2 to obtain the active kinematic chain driving joint variables q′ under the posture error state. 主 , the calculation method is as follows: In the formula, is the ideal kinematic inverse function of the active kinematic chain of the hybrid robot; The active kinematic chain ideally drives the joint variables q 主 The active kinematic chain driving joint variable q′ under the state of considering posture error 主 The real-time compensation value Δq of the hybrid robot driving joint is obtained by subtracting 主,补偿 , the calculation method is as follows: Δq 主,补偿 =q 主 -q′ 主 (9) 3) Real-time compensation of hybrid robot posture error based on internal sensor information Control the active kinematic chain of the hybrid robot to drive the joint to move to q 主 +Δq 主,补偿 , use the internal sensor system of the driven motion chain to read the measured value of the generalized driving joint variable of the driven motion chain in the current state, and calculate the generalized driving joint motion error Δq in the current state 从 ; If Δq 从 The 2-norm of ||Δq 从 || is less than or equal to the given threshold ε, the real-time compensation process under the current configuration ends; if Δq 从 The 2-norm of ||Δq 从 If || is greater than the given threshold ε, the process described in step 3 is repeated until the accuracy reaches the target.
Citation Information
Patent Citations
Kinematics calibration method for series-parallel robot
CN108015808A
Plane constraint error model and robot self-calibration method
CN108406771A
Robot calibration method based on FIS theory
CN115026819A
Robot device, method of controlling the same, and program
JP2004283957A
Device and Method for Kinematic Calibration of Robots
KR1020050018545A
