Six-degree-of-freedom hybrid robot system based on error identification and compensation

Abstract

The invention discloses a six-degree-of-freedom hybrid robot system based on error identification and compensation, and belongs to the technical field of robots and automatic equipment. The system comprises a static platform, a movable platform, three parallel branch chain assemblies, a parallel robot support, a three-degree-of-freedom series rotary joint assembly, an end effector assembly and a controller module. The static platform is of an equilateral triangle structure, the parallel robot support is installed above the static platform, and the three parallel branch chain assemblies are arranged above the static platform at equal intervals through the parallel robot support. The movable platform is mounted on the three parallel branch chain assemblies to form a 3-PTT parallel mechanism; the parallel mechanism provides three-degree-of-freedom translational motion; the three-degree-of-freedom series-connection rotary joint assembly is hinged to the movable platform, and the three-degree-of-freedom series-connection rotary joint assembly is a 3R series-connection mechanism composed of three direct drive motors, a base connecting base, a middle support and a terminal connecting base and used for achieving three-dimensional posture adjustment of the end effector assembly; the end effector assembly is used for executing tasks; and the controller module is used for realizing coordinated control of a parallel mechanism and a series mechanism. Through the modular 3-PTT + 3R series-parallel topological design, decoupling and accurate connection of translation and rotation freedom degrees are achieved, and the three-degree-of-freedom parallel connection mechanism has high rigidity, high maintainability and excellent geometric consistency.

Description

Technical Field

[0001] This invention belongs to the field of robotics technology, specifically relating to a six-degree-of-freedom hybrid robot system based on error identification and compensation, falling under the category of robot structure calibration and electromechanical system control technology. This system integrates the high rigidity of parallel mechanisms with the flexibility of serial mechanisms. Through error identification and parameter compensation, it achieves comprehensive correction of structural and installation errors in the hybrid robot, enabling precise calibration of the robot's actual geometric parameters, thereby improving end-effector pose accuracy and overall system assembly consistency. Background Technology

[0002] With the rapid development of high-end manufacturing and intelligent equipment technologies, the demand for robots in precision assembly, complex processing, automatic inspection, and intelligent manufacturing is constantly increasing. Existing industrial robots are mainly divided into two categories: serial and parallel. Serial robots have a larger workspace and more flexible posture control capabilities, but their joint errors accumulate step-by-step along the kinematic chain, resulting in lower end-effector positioning accuracy. Parallel robots, on the other hand, are known for their high rigidity, high dynamic response performance, and high positioning accuracy, but their limited workspace and insufficient posture adjustment capabilities restrict their application in complex work scenarios.

[0003] To combine the advantages of both, hybrid robots have emerged. Hybrid robots typically use parallel mechanisms for translational motion and serial mechanisms for rotational motion, ensuring both high rigidity and load capacity while also possessing flexible attitude adjustment capabilities. This has become an important development direction for high-precision robot systems. However, due to the complex structure and strong motion coupling of hybrid robots, their overall accuracy is highly dependent on the geometric manufacturing and assembly accuracy of each component. Manufacturing errors, assembly deviations, joint clearances, link length errors, and installation position offsets can all lead to geometric deviations between the actual mechanism and the theoretical model. This causes the forward and inverse kinematic equations established based on the ideal geometric model to fail to accurately reflect the actual end effector position and attitude, resulting in positioning errors, attitude drift, and trajectory deviations, severely affecting the accuracy and reliability of the robot system.

[0004] Common methods for improving robot accuracy include parameter identification based on least squares, nonlinear modeling based on neural networks or fuzzy logic, and geometric calibration methods based on direct measurement. Parameter identification methods based on least squares typically rely on linearization assumptions, making them difficult to handle nonlinear coupling problems in hybrid structures. While neural network-based learning compensation methods possess some nonlinear fitting capabilities, they are highly sensitive to the number of training samples and the model structure, and lack physical interpretability. Geometric calibration methods based on high-precision measurement equipment (such as laser trackers, 3D probes, or photogrammetry systems) can directly acquire high-precision end-effector pose data, but the equipment is expensive and requires stringent experimental conditions, limiting their widespread application in general engineering environments.

[0005] To address the aforementioned problems, this invention proposes a six-degree-of-freedom hybrid robot system based on error identification and compensation. The system employs a parallel-series hybrid structure, with three parallel branches providing translational motion and a three-stage series rotational structure achieving rotational posture. The slider guide rail in the parallel section is driven by a servo motor, which converts the motor's rotational motion into linear motion via a ball screw mechanism, thereby driving the platform to achieve three-dimensional translation. The series section uses a direct drive motor (DD motor) structure, improving system compactness and transmission accuracy. The system is internally equipped with a grating ruler and photoelectric limit sensors for real-time detection of the position and limit status of the ball screw and slider, ensuring safe and high-precision control of the parallel drive section.

[0006] In the error identification stage, this invention uses external high-precision measuring equipment (such as a laser tracker) to collect the actual position of the robot's end effector under multiple known postures and the calculation results of the theoretical model, comparing them to establish structural error and installation error models. The parallel part uses a genetic algorithm for global optimization identification to determine the geometric error parameters of the branches and the dynamic and static platforms; the serial part is based on DH parameter modeling and uses a differential evolution algorithm to identify the geometric parameters of the installation error. This algorithm, by introducing an adaptive damping adjustment mechanism, achieves rapid convergence and stable identification under highly nonlinear conditions, and can simultaneously obtain the optimal estimates of parameters such as link length and link offset. In the control stage, the controller module introduces the identified error parameters into the corrected kinematic model to achieve parameter compensation and motion accuracy correction, enabling the robot to perform trajectory planning and attitude control based on the corrected model after offline identification, thereby significantly improving the system's positioning and attitude accuracy.

[0007] Compared with existing technologies, this invention achieves offline identification and model compensation of structural and installation errors in hybrid robots without relying on real-time sensor feedback. The solution features a rational structure and efficient algorithm, significantly improving the performance of hybrid robots in high-precision operation scenarios such as assembly, grinding, and inspection. It provides an engineering-based and scalable implementation path for hybrid robot calibration and accuracy compensation, demonstrating significant structural and algorithmic innovation. Summary of the Invention

[0008] This invention relates to a six-degree-of-freedom hybrid robot system based on error identification and compensation, with a hybrid structure consisting of a 3-PTT parallel mechanism and a 3R series mechanism as the core. Furthermore, genetic algorithms and differential evolution algorithms are introduced into the controller of the system to realize the phased identification and correction of geometric error parameters, thereby effectively improving the accuracy and consistency of the end-effector pose calculation of the hybrid robot.

[0009] To achieve the above objectives, this invention proposes a six-degree-of-freedom hybrid robot system based on error identification and compensation, comprising a static platform 1, a moving platform 2, three parallel branch assembly 3, a parallel robot support 4, a three-degree-of-freedom serial rotary joint assembly 5, an end effector assembly 6, and a controller module 7. The static platform 1 has an equilateral triangular structure, with the parallel robot support 4 mounted above it. The three parallel branch assembly 3 are arranged at equal intervals above the static platform 1 via the parallel robot support 4. The moving platform 2 is mounted on the three parallel branch assembly 3, forming a 3-PTT parallel mechanism. The parallel mechanism provides three-degree-of-freedom translational motion. The three-degree-of-freedom serial rotary joint assembly 5 is hinged to the moving platform 2, and the three-degree-of-freedom serial rotary joint assembly 6 provides three-degree-of-freedom translational motion. The rotary joint assembly 5 consists of three direct drive motors 5-1, a base connecting seat 5-2, an intermediate support 5-3, and a terminal connecting seat 5-4, forming a 3R series mechanism. The first rotary joint rotates around the Z-axis using the direct drive motors 5-1 and the base connecting seat 5-2; the second rotary joint rotates around the Y-axis using the direct drive motors 5-1 and the intermediate support 5-3; and the third rotary joint rotates around the X-axis using the direct drive motors 5-1 and the terminal connecting seat 5-4. This mechanism enables three-dimensional attitude adjustment of the end effector assembly 6. The end effector assembly 6 is mounted on the rotor of the direct drive motor 5-1 of the third rotary joint and performs tasks. The controller module 7 is used to coordinate the control of the parallel and series mechanisms.

[0010] Furthermore, the static platform 1 is an equilateral triangle structure machined from an integral aluminum alloy, serving as the mounting base for the parallel robot support 4; the moving platform 2 is an integral high-strength aluminum alloy machined structure, with three bearing seats evenly distributed on the moving platform 2 for hinged connection with the three parallel branch assembly 3 to achieve smooth three-way translational motion.

[0011] Furthermore, the parallel branch assembly 3 consists of a lower lead screw guide linear drive component and an upper connecting rod component. The parallel branch assembly 3 is mounted on the parallel robot support 4. One end of the connecting rod component is hinged to the lead screw guide linear drive component, and the other end of the connecting rod component is hinged to the moving platform 2. The three parallel branch assemblies 3 are distributed at equal angles above the static platform 1, respectively driving the moving platform 2 to achieve translational motion in three spatial directions.

[0012] Furthermore, the parallel branch assembly 3 is composed of a lead screw guide rail linear drive component and a connecting rod component. The lead screw guide rail linear drive component includes a mounting column 3-1, a linear slider 3-2, a ball screw 3-3, a guide rail 3-4, a servo motor 3-5, a grating ruler 3-6, a motor support 3-7, a U-groove photoelectric switch 3-8, a plum blossom coupling 3-9, a bearing seat 3-10, a lead screw support seat 3-11, and an angular contact ball bearing 3-12. The connecting rod component includes a pin 3-13, a connecting rod 3-14, a connecting rod joint 3-15, a locking nut 3-16, and a locking washer. Plate 3-17; Ball screw 3-3 and servo motor 3-5 are coaxially connected via a plum blossom coupling 3-9 to transmit torque; linear slider 3-2 is mounted on guide rail 3-4 to form linear reciprocating motion; guide rail 3-4, servo motor 3-5, motor support 3-7, and screw support 3-11 constitute a modular linear drive unit. Guide rail 3-4, motor support 3-7, and screw support 3-11 are all mounted on mounting column 3-1. Servo motor 3-5 is mounted on motor support 3-7, and the servo motor 3-5 drives motor support 3-7 and screw support 3-11. The rotational motion of the ball screw 3-3 drives the linear slider 3-2, which is threadedly engaged with the ball screw 3-3, to move along the guide rail 3-4. A grating ruler 3-6 is mounted on the side of the mounting column 3-1, and its reading head is connected to the linear slider 3-2 to detect the position of the linear slider 3-2 on the ball screw 3-3. U-shaped photoelectric switches 3-8 are installed at both ends of the mounting column 3-1 for limit protection. Two parallel connecting rod connectors 3-15 are installed at both ends of the two parallel connecting rods 3-14. Parallel pins 3-13 are used to connect and form a parallelogram mechanism to ensure the stability of motion transmission and the posture of the moving platform 2. Pins 3-13 are hinged to bearing housings 3-10 and 2 respectively. Angular contact ball bearings 3-12 are installed on both sides of bearing housings 3-10 and 2. Each pin 3-13 is locked to the angular contact ball bearing 3-12 on one side by a shoulder, and a locking washer 3-17 and a locking nut 3-16 are installed on the other side to lock the angular contact ball bearing 3-12, thereby realizing the translation drive of the moving platform 2.

[0013] Furthermore, the parallel robot support 4 is composed of columns 4-1, side reinforcing plates 4-2, and reinforcing plate mounting blocks 4-3; the parallel robot support 4 is installed above the static platform 1 to support the three parallel branch chain components 3 and maintain their parallelism and structural rigidity; the three columns 4-1 are vertically fixed at equal intervals on the static platform 1, and any two adjacent columns 4-1 are connected by side reinforcing plates 4-2 and reinforcing plate mounting blocks 4-3 to form an overall frame structure.

[0014] Furthermore, the direct drive motor 5-1 is an integrated DD motor structure with an internal high-resolution encoder; the direct drive motor 5-1 in the first rotary joint is mounted on the base connecting seat 5-2, the direct drive motor 5-1 in the second rotary joint is mounted on the intermediate support 5-3, the direct drive motor 5-1 in the third rotary joint is mounted on the terminal connecting seat 5-4, the terminal connecting seat 5-4 is mounted on the rotor of the direct drive motor 5-1 in the second rotary joint, the intermediate support 5-3 is mounted on the rotor of the direct drive motor 5-1 in the first rotary joint, and the base connecting seat 5-2 is mounted on the moving platform 2. The attitude adjustment is achieved by driving the end effector assembly 6 through the three rotary joints.

[0015] Furthermore, the end effector assembly 6 is installed at the output end of the series rotary joint assembly 5 and is fixedly connected to the third rotary joint through a standard flange interface. The end effector assembly 6 can be replaced with different functional modules according to the operation requirements. The functional modules include, but are not limited to, grinding heads and pin devices.

[0016] Furthermore, the controller module 7 employs an error identification and compensation control strategy to model and correct the kinematic errors of the parallel and series mechanisms, thereby achieving coordinated control of the parallel and series mechanisms.

[0017] Furthermore, the error identification and compensation control strategy includes the following steps:

[0018] 1) Offline data acquisition stage: Under various typical robot poses, record the actual measurement data of the position of linear slider 3-2 in the parallel mechanism, the rotation angle of the serial mechanism, and the end pose; at the same time, calculate the theoretical end pose based on the theoretical geometric parameters to form a theoretical-actual corresponding dataset.

[0019] 2) Independent error identification stage: For parallel mechanisms, a genetic algorithm is used to optimize and identify the geometric parameters to obtain the error model of the parallel mechanism; for series mechanisms, a differential evolution algorithm is used to identify the length deviation of each rotary joint link to obtain the error model.

[0020] 3) Error compensation and coordinated control stage: In the controller module 7, the corrected actual geometric parameters are replaced with theoretical parameters, the forward and inverse kinematics calculations are updated in real time, and the target displacement and attitude commands of each drive axis are generated; and closed-loop correction is performed in each control cycle based on the feedback information of servo motor 3-5, grating ruler 3-6 and direct drive motor 5-1 to realize the coordinated compensation control of parallel and series mechanisms.

[0021] The beneficial effects of this invention are as follows: Compared with existing technologies, the technical solution of this invention has significant advantages in three aspects: structural innovation, identification strategy, and compensation mechanism. Structurally, through a modular 3-PTT+3R hybrid topology design, the decoupling and precise connection of translational and rotational degrees of freedom are achieved, resulting in high rigidity, high maintainability, and excellent geometric consistency. Algorithm-wise, the proposed hybrid identification system of "genetic algorithm-differential evolution algorithm" exhibits stronger convergence stability and global search capability in nonlinear multi-parameter coupled models. In terms of compensation, by combining offline geometric identification with online error correction, high-precision error compensation and pose control are achieved under conditions where high-precision external measurement equipment is unavailable. The overall system significantly improves the robot's positioning accuracy and posture stability in complex workspaces while ensuring structural simplicity and controllable cost.

[0022] In summary, this invention, through collaborative innovation in structure, algorithm, and control, constructs a high-precision, low-cost, and scalable six-degree-of-freedom hybrid robot system, providing a new technical path for parameter calibration, structural optimization, and high-precision control of hybrid mechanisms, and has broad engineering application prospects. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall structure of the six-degree-of-freedom hybrid robot system of the present invention.

[0024] Figure 2 This is a three-dimensional isometric view of the six-degree-of-freedom hybrid robot of the present invention.

[0025] Figure 3 This is the front view of the six-degree-of-freedom hybrid robot of the present invention.

[0026] Figure 4 This is a left view of the six-degree-of-freedom hybrid robot of the present invention.

[0027] Figure 5 This is a top view of the six-degree-of-freedom hybrid robot of the present invention.

[0028] Figure 6 This is a schematic diagram of the 3-PTT parallel mechanism of the present invention.

[0029] Figure 7 This is a schematic diagram of the three-degree-of-freedom series rotary joint assembly of the present invention.

[0030] Figure 8 This is a schematic diagram of the static platform structure of the present invention.

[0031] Figure 9 This is a schematic diagram of the moving platform structure of the present invention.

[0032] Figure 10 This is a schematic diagram of the single-branch structure of the present invention.

[0033] Figure 11 This is a schematic diagram of the linear drive assembly for the lead screw guide rail of the present invention.

[0034] Figure 12 This is a schematic diagram of the linkage assembly structure of the present invention.

[0035] Figure 13 This is a schematic diagram of the parallel robot support structure of the present invention.

[0036] Figure 14 This is a schematic diagram of the end effector pin device of the present invention.

[0037] The following components are labeled in the diagram: 1. Static platform; 2. Moving platform; 3. Parallel support chain assembly; 3-1. Mounting column; 3-2. Linear slider; 3-3. Ball screw; 3-4. Guide rail; 3-5. Servo motor; 3-6. Grating ruler; 3-7. Motor support; 3-8. U-slot photoelectric switch; 3-9. Plum blossom coupling; 3-10. Bearing housing; 3-11. Screw support seat; 3-12. Angular contact ball bearing; 3-13. Pin; 3-14. Connecting rod; 3-15. Connecting rod joint; 3-16. Locking nut; 3-17. Locking washer; 4. Parallel robot bracket; 4-1. Column; 4-2. Side reinforcement plate; 4-3. Reinforcement plate mounting block; 5. Series rotary joint assembly; 5-1. Direct drive motor; 5-2. Base connecting seat; 5-3. Intermediate bracket; 5-4. Terminal connecting seat; 6. End effector assembly; 7. Controller module. Detailed Implementation

[0038] The invention will be further described below with reference to the accompanying drawings and embodiments, but the scope of the invention is not limited to the description.

[0039] Example 1: Design of a six-degree-of-freedom hybrid robot system.

[0040] like Figure 1 As shown, this invention discloses a six-degree-of-freedom hybrid robot system based on error identification and compensation. Its overall structure consists of a three-degree-of-freedom parallel mechanism (3-PTT) and a three-degree-of-freedom serial joint assembly 5. This hybrid system achieves decoupled control of spatial position and attitude through mechanism integration, combining the high rigidity and high precision of the parallel mechanism with the large attitude range and flexibility of the serial mechanism, making it suitable for various complex task scenarios.

[0041] Furthermore, the parallel mechanism adopts a 3-PTT structure, consisting of a static platform 1, a moving platform 2, parallel branch assembly 3, and a parallel robot support 4. Both the static platform 1 and the moving platform 2 are arranged in an equilateral triangle configuration. The moving platform is located above the static platform and is connected to the parallel robot support 4 via three parallel branches 3. The parallel robot support 4 is mounted above the static platform 1, and the three parallel branch assembly 3 are equally spaced above the static platform 1 via the parallel robot support 4 to maintain structural symmetry and force balance. The moving platform 2 is connected via pins 3-13 of the parallel branch assembly 3, forming a 3-PTT parallel mechanism. The parallel portion provides three degrees of freedom for spatial translational motion.

[0042] Furthermore, the parallel branch assembly 3 is the core execution unit of the parallel mechanism, which includes a mounting column 3-1, a linear slider 3-2, a ball screw 3-3, a guide rail 3-4, a servo motor 3-5, a grating ruler 3-6, a motor support 3-7, a U-groove photoelectric switch 3-8, a plum blossom coupling 3-9, a bearing seat 3-10, a screw support seat 3-11, an angular contact ball bearing 3-12, a pin 3-13, a connecting rod 3-14, a connecting rod joint 3-15, a locking nut 3-16, and a locking washer 3-17; the ball screw 3-3 and the servo motor 3-4 are coaxially connected through the plum blossom coupling 3-9 to transmit torque, and the linear slider 3-2 is mounted on the guide rail 3-4 to form a linear reciprocating motion; the guide rail 3-4, the servo motor, and the linear slider 3-2 are mounted on the guide rail 3-4 to form a linear reciprocating motion; the guide rail 3-4, the servo motor, and the linear slider 3-2 are coaxially connected through the plum blossom coupling 3-9 to transmit torque, and the linear slider 3-2 is mounted on the guide rail 3-4 to form a linear reciprocating motion; the guide rail 3-4, the servo motor, and the linear slider 3-2 are coaxially connected through the guide rail 3-9 to transmit torque, and the linear slider 3-2 is mounted on the guide rail 3-4 to form a linear reciprocating motion; the linear slider 3-2, the linear slider 3-2, the linear slider 3-2, the linear slider 3-3, the linear slider 3-4, the linear slider 3-5, the linear slider 3-6, the linear slider 3-7, the linear slider 3-8, the linear slider 3-9, the linear slider 3-1, the The motor 3-5, motor support 3-7, and lead screw support 3-11 constitute a modular linear drive unit. The guide rail 3-4, motor support 3-7, and lead screw support 3-11 are all mounted on the mounting column 3-1. A servo motor 3-5 is mounted on the motor support 3-7, driving the rotation of the ball screw 3-3 between the motor support 3-7 and the lead screw support 3-11. This rotation drives the linear slider 3-2, which is threaded into the ball screw 3-3, to move along the guide rail 3-4. A grating ruler 3-6 is mounted on the side of the mounting column 3-1, with a reading head connected to the linear slider 3-2 for detecting its position on the ball screw 3-3. U-shaped photoelectric switches 3-8 are mounted at both ends of the side of the mounting column 3-1 for limit protection. Pin 3-13 is hinged to bearing housing 3-10 and the bearing housing of moving platform 2 respectively. Angular contact ball bearings 3-12 are installed on both sides of the bearing housings 3-10 and moving platform 2. Pin 3-13 is locked to the angular contact ball bearing 3-12 on one side via a shoulder, and a locking washer 3-17 and a locking nut 3-16 are installed on the other side to lock the angular contact ball bearing 3-12. Linear slider 3-2 moves linearly along guide rail 3-4, thereby driving the connecting rod to move and adjusting the position of moving platform 2. This component adopts a modular design, which is convenient for maintenance and replacement, and also has high repeatability and rigidity.

[0043] Furthermore, the three-degree-of-freedom serial rotary joint assembly 5 consists of a direct drive motor 5-1, a base connector 5-2, an intermediate support 5-3, and a terminal connector 5-4; wherein: the first rotary joint rotates around the Z-axis, the second rotary joint rotates around the Y-axis, and the third rotary joint rotates around the X-axis; the direct drive motor assembly 5-1 is an integrated DD motor structure with an integrated high-resolution encoder; the direct drive motor 5-1 in the first rotary joint is mounted on the base connector 5-2, the direct drive motor 5-1 in the second rotary joint is mounted on the intermediate support 5-3, the direct drive motor 5-1 in the third rotary joint is mounted on the terminal connector 5-4, the terminal connector 5-4 is mounted on the rotor of the direct drive motor 5-1 in the second rotary joint, the intermediate support 5-3 is mounted on the rotor of the direct drive motor 5-1 in the first rotary joint, and the base connector 5-2 is mounted on the moving platform 2. The three rotary joints are used to drive the end effector assembly 6 to achieve attitude adjustment.

[0044] Furthermore, the end effector assembly 6 is connected to the output end of the series rotary joint assembly 5 via a standard flange structure, providing replaceability and versatility. Different functional modules can be replaced according to operational requirements, including but not limited to grinding heads and pin devices. The controller module 7 is the core of the system, responsible for data acquisition, kinematic calculation, error compensation, and closed-loop control. It achieves coordinated driving of the parallel and series components, and, combined with error identification and compensation control strategies, models and corrects the kinematic errors of the parallel and series mechanisms.

[0045] Specifically, the structural innovation of this invention lies in the following: the parallel section adopts a symmetrically distributed 3-PTT configuration, with each branch being an actively driven linear servo unit. The static and moving platforms are arranged in equilateral triangles, with the center point of the moving platform serving as the end-effector position reference point. The serial section adopts a three-degree-of-freedom RRR configuration, enabling the overall robot to achieve a hybrid system layout of three translations and three rotations. This structure, while maintaining high rigidity and lightweight, significantly expands the end-effector workspace and improves posture flexibility.

[0046] Example 2: Kinematic modeling and intelligent optimization algorithm parameter identification process of hybrid system.

[0047] Furthermore, this invention establishes a clear hierarchical analytical model for kinematic modeling. For parallel mechanisms, the forward and inverse solution models of the center point of the moving platform are derived based on the geometric constraint equations, with the center of the static platform as the origin of the coordinate system. Taking the installation position of the first parallel branch on the static platform and the straight line containing the center of the static platform as... The axis is perpendicular to the stationary platform and passes through the center of the circle. Upward Establish a base coordinate system based on the axes. Let the coordinates of the center of the moving platform be... The radii of the static and dynamic platforms are respectively and The hinged mounting points of the bearing housings and pins in the three branches are as follows: Installation points on the moving platform The length of the center hole of the parallel pin mounting holes at both ends of each parallel branch is denoted as . Then the constraint equations of the parallel mechanism can be expressed as:

[0048] .

[0049] Where X, Y, and Z are the spatial coordinates of the center point of the moving platform in the base coordinate system; , These are the radii of the static platform and the moving platform, respectively. For the first The length of the center hole of the pin mounting holes at both ends of a parallel branch chain; , for the first The angle of the distribution of the parallel branches on the circumference; For the first The vertical height of the parallel branched linear slider from the static platform. ,in For the first The amount of movement of a parallel branched linear slider on the guide rail Let be the height of the linear slider on the parallel chain from the stationary platform when it is at its extreme position below the guide rail. Using the above three constraint equations, the displacement of each driving unit can be solved based on the position of the end-effector moving platform, thus achieving the inverse kinematics solution of the parallel mechanism. This part of the model achieves rapid solution of geometric constraints and improved numerical stability; the calculated inverse kinematic equations are as follows:

[0050] .

[0051] ℎ 2 =Z− L 2 − [X+ 1 2 R−r ] 2 − [Y+ 3 2 r−R ] 2 .

[0052] ℎ 3 =Z− L 2 − [X+ 1 2 R−r ] 2 − [Y+ 3 2 R− r ] 2 .

[0053] Similarly, the coordinates of the end-effector moving platform can be calculated from the driving quantities of the three parallel-branch linear sliders using the constraint equations. The forward kinematic equations can then be obtained:

[0054] .

[0055] Where the coefficient of Z is:

[0056] .

[0057] .

[0058] c= [ ℎ 1 2 − ℎ 2 2 + ℎ 3 2 2 ] 2 9(R−r) 2 + ( ℎ 2 2 -ℎ 3 2 ) 2 12(R−r) 2 + (R−r) 2 − L 2 − 2[ ℎ 1 2 − (ℎ 2 2 + ℎ 3 2 ) 2 ] 3 + ℎ 1 2 .

[0059] Furthermore, the kinematics of the series mechanism are modeled using a modified DH parameter method. Let the rotation angles of the three rotary joints be... , , The DH parameters for each link are defined as follows: Link length: Linkage torsion angle: Linkage offset: Joint variables: ,in The homogeneous transformation matrix between the joints can then be expressed as:

[0060] .

[0061] The overall transformation matrix for the end-effector attitude is: Considering that the main source of error in the series section is the link length deviation, this invention only addresses the link length parameter in the DH parameters. Offset with connecting rod Identification corrections are made without considering joint angle errors. This identification method, which combines parameter simplification with differential evolution (DE), can effectively improve identification accuracy and reduce model dimensionality.

[0062] Furthermore, through the above modeling process, the motion relationship of the series mechanism can be accurately described, and it can be coupled with the end position of the parallel mechanism to realize a unified kinematic model of the entire hybrid system. By synthesizing the constraint equations of the hybrid system and superimposing the end displacement of the parallel part with the attitude transformation of the series part, the overall position and attitude expression of the end effector in the base coordinate system can be obtained: ;in, Let be the pose transformation matrix of the moving platform of the parallel mechanism relative to the base coordinate system. This is the attitude transformation matrix of the end of the serial mechanism relative to the moving platform.

[0063] Furthermore, at the algorithm level, this invention proposes a hierarchical modeling and joint optimization strategy: first, independent forward and inverse kinematic models are established for parallel and series mechanisms respectively; second, joint coordinate transformation is implemented in controller module 7 to achieve consistency of the coordinate systems of the two mechanisms; finally, in the error compensation stage, the overall kinematic model of the hybrid system is accurately calibrated by offline identification parameter correction.

[0064] Furthermore, this embodiment provides an error identification and compensation method for a six-degree-of-freedom hybrid robot based on an intelligent optimization algorithm, used for geometric parameter calibration and error correction of the hybrid robot structure described in the aforementioned embodiment. This method performs offline identification of structural geometric errors generated during manufacturing, assembly, and installation, without considering kinematic dynamic errors, focusing on correcting static structural deviations in parallel and series components. A two-stage identification strategy combining a parallel mechanism genetic algorithm (GA) and a series mechanism differential evolution algorithm (DE) is employed to achieve offline identification and error compensation of overall structural parameters.

[0065] Furthermore, in the error identification stage of the parallel mechanism, an ideal geometric constraint model of the 3-PTT parallel mechanism is first established. Under ideal conditions, the position vector of the end-effector and the lengths of the three branches satisfy the following constraint equations:

[0066] .

[0067] in, For the first The coordinate vector of the hinged mounting point between the bearing housing and the pin shaft in the parallel branch chain. For the first on the dynamic platform The coordinate vectors of the hinge points; For the first The center distance between the mounting holes of the parallel branch chains with parallel pins at both ends. This is the position vector of the end platform.

[0068] Furthermore, considering the machining errors of the moving platform and static platform, as well as the length errors of the three connecting rods in actual manufacturing and assembly, the main geometric error items of the parallel mechanism can be summarized into five: These are the length errors of the three connecting rods; Static platform installation and machining radius error; This refers to the machining radius error of the moving platform.

[0069] Furthermore, by collecting measured positions of the end effector platform in different poses... and with theoretical position By comparison, construct the geometric error objective function for the parallel part:

[0070] .

[0071] in, The measured positions of the end effector platform under different poses. For theoretical position; Indicates the end sampling point number; error parameter set To minimize Given the objective function, a genetic algorithm (GA) is used to perform a global optimization search in the multidimensional error parameter space. The genetic algorithm first optimizes the parameter set... Encode the data to generate an initial population; the fitness function of each individual is defined as the error function. The reciprocal of the value. By selecting operators to retain individuals with high fitness, and using crossover and mutation operations to generate new individuals, the population continuously evolves in each generation. The algorithm gradually converges to the globally optimal error parameters during the iteration process, thereby obtaining the geometric parameter correction amount that minimizes the end position error, realizing offline error identification and structural correction of the 3-PTT parallel mechanism.

[0072] Furthermore, in the error identification stage of the series mechanism, only the link length parameter error is considered. Based on the improved DH parameter method, the ideal parameter is set as follows: , The corrected parameters are , The attitude error function of the cascaded section is:

[0073] .

[0074] in This is the set of error parameters for the cascaded mechanism. The differential evolution algorithm (DE) is used to... Minimize the parameter vector by using differential mutation, crossover and selection operations to achieve the globally optimal identification of attitude accuracy.

[0075] Furthermore, by employing error identification and compensation strategies, the parallel and series components of the system achieve joint optimization of translation and attitude errors. The compensated correction parameters are written into controller module 7 to replace the original theoretical geometric parameters, realizing model correction and error compensation at the control level. After compensation, the end-effector positioning and attitude errors are significantly reduced, fully demonstrating the significant advantages of the six-DOF hybrid robot system based on error identification and compensation in terms of accuracy improvement and stability enhancement.

[0076] Example 3: Error Compensation and Performance Verification.

[0077] Furthermore, to verify the effectiveness of the hybrid six-DOF robot error identification and compensation algorithm described in Example 2, this example provides a simulation verification method for the error compensation algorithm.

[0078] The method described is based on the hybrid six-DOF robot structural and kinematic model established in Example 1. A complete forward and inverse kinematics solution module is constructed using the MATLAB platform, and predefined structural error parameters are introduced to simulate actual manufacturing and assembly errors. The introduced errors include the length errors of the three connecting rods of the parallel mechanism. Manufacturing errors of dynamic and static platforms , There are a total of five error terms; the series mechanism introduces link length error. and connecting rod offset error There are a total of six error terms.

[0079] During the simulation, the theoretical pose data of the parallel mechanism and the end effector under ideal parameter conditions are first calculated. Then, the actual pose data of the parallel mechanism and the end effector are recalculated under an error-containing model, and the difference between the two is used as the error dataset. The error data of the parallel mechanism is input into a genetic algorithm, and the error data of the end effector is input into a differential evolutionary algorithm optimization module. Through selection, crossover, and mutation operations, the algorithm iterates continuously to gradually approach the optimal solution of the true error parameters. When the error cost function converges, the parameter set output by the algorithm is the identified structural error result.

[0080] Furthermore, the identified error parameters are substituted into the kinematic model for correction, forming a compensated kinematic model. By re-simulating and calculating the change trend of the end-effector pose before and after compensation, the effectiveness of the proposed error compensation algorithm in improving pose consistency can be verified. Simulation results show that the method can converge stably and accurately reflect the main structural error characteristics of the system, thus verifying the feasibility and applicability of the algorithm.

[0081] Furthermore, this simulation verification method can verify the algorithm performance without physical experiments, providing a reliable basis for the subsequent calibration and error compensation of the robot system, and also providing algorithmic support for improving the control accuracy of hybrid robots.

[0082] Furthermore, to verify the comprehensive performance and practical application effect of the six-degree-of-freedom hybrid robot system described in this invention, this embodiment takes the automatic pin-locking task in the power inspection scenario as an example to explain the working principle and application advantages of the system.

[0083] In power line inspection operations, cotter pins are often installed on bolts and connecting holes of transmission equipment to ensure the reliability of mechanical connections. Traditional manual pin insertion methods are not only labor-intensive and inefficient, but also pose significant safety risks and positioning errors in high-altitude or confined spaces. To address these issues, this invention proposes a six-degree-of-freedom hybrid robot system based on error identification and compensation. Combining visual recognition with error identification and compensation technologies, it achieves automatic identification, posture adjustment, and precise pin insertion for the pin insertion task.

[0084] Before operation, the robot controller module 7 uses error identification and compensation technology to correct the forward and inverse kinematics models, thereby compensating for geometric errors in parallel and series components. The system first acquires images of the work area through a vision sensor module mounted on the end effector assembly 6, identifying the spatial position and orientation of the target bolt holes. Based on the identification results, the control system calculates the target pose of the end effector and uses the corrected inverse kinematics model to generate corresponding drive commands.

[0085] During the operation execution phase, the parallel branch assembly 3 provides three-dimensional translational positioning, ensuring that the axis of the end pin clamping mechanism reaches the center of the bolt hole. The series rotary joint assembly 5 enables flexible adjustment of the end pin's posture, ensuring that the axis of the end pin clamping mechanism is coaxially aligned with the center of the bolt hole and that the pin direction is consistent with the axis of the target hole. Subsequently, the end effector assembly 6 activates the pin pushing mechanism, pushing a single cotter pin into the bolt mounting hole, achieving automatic pin insertion. Under the action of the prying mechanism, the cotter pin is pried open to meet the installation requirements. After insertion, the three-degree-of-freedom series rotary joint assembly 5 drives the end effector to move and rotate slightly, allowing the end pin clamping mechanism to be removed from the bolt.

[0086] Furthermore, the offline error compensation algorithm of this invention plays a crucial role throughout the entire operation. By identifying and correcting the robot's geometric parameters, the system can automatically compensate for structural deviations caused by manufacturing, assembly, and installation when performing pin-locking tasks, thereby significantly improving the accuracy of end-effector pose calculation and the stability of control. The introduction of the compensation algorithm enables the robot to maintain accurate spatial positioning and reliable attitude control even in complex spatial environments, ensuring smooth pin-locking movements and assembly precision.

[0087] This hybrid robot system achieves a synergy between rigidity and flexibility in its structure: the parallel section provides high stability and load-bearing capacity, while the serial section provides a large range of posture adjustment. Together, they form a high-precision, highly flexible working platform. Through algorithm-level error compensation and model correction, the robot can achieve long-term stable operation without external calibration.

[0088] Furthermore, the control system in this embodiment can switch parameters according to different cotter pin models and installation locations, automatically adjusting the end effector attitude path and propulsion force control to achieve adaptive insertion of bolt holes of various specifications. The system has good modularity and scalability, and can be widely applied in various scenarios such as transmission line inspection, substation maintenance, and power equipment maintenance.

[0089] As demonstrated in this embodiment, the six-degree-of-freedom hybrid robot system of the present invention achieves automated and highly reliable control of complex pin-operating tasks through the deep integration of structural innovation and error identification and compensation algorithms. This system not only effectively improves the intelligence and safety of power line inspection operations but also provides a new technological path for the automation upgrade of assembly manufacturing and automated grinding.

[0090] The hybrid robot consists of a 3-PTT parallel mechanism and a 3R serial mechanism. The parallel mechanism provides the translational motion of the moving platform in three-dimensional space, while the serial mechanism enables the three-dimensional rotational attitude of the end effector. The two are connected by a standard flange to form a detachable and assembleable modular hybrid structure. The slider guide rail of the parallel section is driven by a servo motor, which converts the motor's rotational motion into the slider's linear motion through a ball screw mechanism, achieving high-precision translation of the moving platform. The serial section uses a direct drive motor (DD motor) structure with an integrated high-resolution encoder to improve system compactness, rigidity, and transmission accuracy. This modular structure not only facilitates segmented calibration and error compensation but also benefits future system expansion and maintenance. The system is equipped with a grating ruler and photoelectric switch sensors to monitor the slider position, limit status, and feedback calibration, ensuring smooth and precise control of the parallel section.

[0091] In terms of kinematic modeling, this invention establishes a unified coordinate system and kinematic equations for parallel and serial mechanisms. The parallel mechanism uses geometric constraint equations to describe the spatial constraints between the moving and stationary platforms, and solves the forward and inverse kinematic equations numerically to calculate the three-dimensional position of the moving platform. The serial mechanism uses an improved DH parameter model to describe the mapping relationship from joint space to the end effector coordinate system, constructing a unified framework for the robot's overall forward and inverse kinematics. Only linear dimensional parameters in the DH parameter model are considered. and Geometric errors are ignored, angular parameter errors are disregarded to simplify the model and focus on identifying geometric length deviations. This method allows for the separate modeling of translational errors in the parallel section and attitude errors in the series section.

[0092] In the error modeling and parameter identification process, this invention employs an offline geometric calibration method. Theoretical and measured pose data are collected at multiple end-point attitude points to construct an error identification objective function. The main error sources in the parallel section include deviations in the length of the connecting rods and deviations in the radii of the moving and stationary platforms. The main error source in the series section is the link length error, including the errors in the lengths of each link in the DH modeling. and To reduce parameter coupling, this invention employs a two-stage identification strategy: The first stage uses a genetic algorithm to optimize global parameters for the parallel section, obtaining geometric correction values ​​for the branches and the platform; the second stage, based on the corrected parallel structure, uses a differential evolution algorithm for local fine-grained identification of the link length parameters in the series section, achieving precise compensation for the link lengths of the series mechanism. This combined optimization approach complements global search and local fitting, avoiding getting trapped in local extrema while maintaining high identification accuracy and convergence speed.

[0093] In terms of error compensation and control implementation, this invention updates the model parameters based on the identification results to correct the kinematic equations and form an updated control mapping relationship. During the operation phase, the control system performs trajectory planning and attitude calculation based on the updated geometric model, thereby achieving real-time error compensation at the motion control layer. This compensation process does not rely on additional sensor closed-loop control, but achieves geometric consistency compensation through offline correction and online mapping, effectively reducing the impact of complex calculations on real-time control performance.

[0094] The application embodiments of the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to these embodiments. Those skilled in the art can make equivalent substitutions and extensions to the mechanism structure, algorithm model, or control logic without departing from the spirit and essence of the present invention, and all such substitutions and extensions should be considered to fall within the protection scope of the present invention. Taking the automatic latching task in a power line inspection scenario as an example, the working principle and application effects of the hybrid robot system of the present invention are explained.

Claims

1. A six-degree-of-freedom hybrid robot system based on error identification and compensation, characterized in that, It includes a static platform (1), a dynamic platform (2), three parallel branch assembly (3), a parallel robot support (4), a three-degree-of-freedom serial rotary joint assembly (5), an end effector assembly (6), and a controller module (7). The static platform (1) is an equilateral triangle structure. The parallel robot support (4) is installed above the static platform (1). Three parallel branch components (3) are arranged at equal intervals above the static platform (1) through the parallel robot support (4). The moving platform (2) is installed on the three parallel branch components (3) to form a 3-PTT parallel mechanism. The parallel mechanism provides three degrees of freedom translational motion. The three-degree-of-freedom serial rotary joint assembly (5) is hingedly mounted on the moving platform (2). The three-degree-of-freedom serial rotary joint assembly (5) consists of three direct drive motors (5-1), a base connecting seat (5-2), an intermediate support (5-3), and a terminal connecting seat (5-4) forming a 3R serial mechanism. The first rotary joint is formed by the direct drive motor (5-1) and the base connecting seat (5-2) rotating around the Z-axis. The second rotary joint is formed by the direct drive motor (5-1) and the intermediate support (5-3) rotating around the Y-axis. The third rotary joint is formed by the direct drive motor (5-1) and the terminal connecting seat (5-4) rotating around the X-axis. It is used to realize the three-dimensional attitude adjustment of the end effector assembly (6). The end effector assembly (6) is mounted on the rotor of the direct drive motor (5-1) of the third rotary joint for performing tasks; The controller module (7) is used to achieve coordinated control of parallel and series mechanisms.

2. The six-degree-of-freedom hybrid robot system based on error identification and compensation according to claim 1, characterized in that, The static platform (1) is an equilateral triangle structure machined from an integral aluminum alloy and is used as the mounting base for the parallel robot support (4). The moving platform (2) is an integrated high-strength aluminum alloy machined structure. Three bearing seats are evenly distributed on the moving platform (2) for hinged connection with three parallel branch chain components (3) to achieve smooth three-way translational movement.

3. The six-degree-of-freedom hybrid robot system based on error identification and compensation according to claim 1, characterized in that, The parallel branch assembly (3) consists of a lower lead screw guide linear drive component and an upper connecting rod component. The parallel branch assembly (3) is installed on the parallel robot bracket (4). One end of the connecting rod component is hinged to the lead screw guide linear drive component, and the other end of the connecting rod component is hinged to the moving platform (2). The three parallel branch assemblies (3) are distributed at equal angles above the static platform (1) and drive the moving platform (2) to achieve translational motion in three spatial directions.

4. The six-degree-of-freedom hybrid robot system based on error identification and compensation according to claim 3, characterized in that, The parallel branch assembly (3) consists of a lead screw guide rail linear drive component and a connecting rod component. The lead screw guide rail linear drive component includes a mounting column (3-1), a linear slider (3-2), a ball screw (3-3), a guide rail (3-4), a servo motor (3-5), a grating ruler (3-6), a motor support (3-7), a U-groove photoelectric switch (3-8), a plum blossom coupling (3-9), a bearing seat (3-10), a lead screw support seat (3-11), and an angular contact ball bearing (3-12). The connecting rod component includes a pin (3-13), a connecting rod (3-14), a connecting rod joint (3-15), a locking nut (3-16), and a locking washer (3-17). 3-17); The ball screw (3-3) and servo motor (3-5) are coaxially connected via a plum blossom coupling (3-9) to transmit torque. The linear slider (3-2) is mounted on the guide rail (3-4) to form a linear reciprocating motion. The guide rail (3-4), servo motor (3-5), motor support (3-7), and screw support (3-11) constitute a modular linear drive unit. The guide rail (3-4), motor support (3-7), and screw support (3-11) are all mounted on the mounting column (3-1). The servo motor (3-5) is mounted on the motor support (3-7). The servo motor (3-5) drives the motor support (3-7) and screw support (3-11) through the servo motor (3-5). The rotational motion of the ball screw (3-3) between 3-11 drives the linear slider (3-2), which is threadedly engaged with the ball screw (3-3), to move along the guide rail (3-4); the grating ruler (3-6) is installed on the side of the mounting column (3-1), and the reading head on the grating ruler is connected to the linear slider (3-2) to detect the position of the linear slider (3-2) on the ball screw (3-3); the U-shaped photoelectric switch (3-8) is installed at both ends of the mounting column (3-1) for limit protection; the two parallel connecting rod joints (3-1) installed at both ends of the two parallel connecting rods (3-1) 5) A parallel quadrilateral mechanism structure is formed by connecting parallel pins (3-13) to ensure the stability of motion transmission and the posture of the moving platform (2). The pins (3-13) are hinged to the bearing housing (3-10) and the bearing housing of the moving platform (2) respectively. Angular contact ball bearings (3-12) are installed on both sides of the bearing housing (3-10) and the bearing housing of the moving platform (2). Each pin (3-13) is locked to the angular contact ball bearing (3-12) on one side through the shoulder, and a locking washer (3-17) and a locking nut (3-16) are installed on the other side to lock the angular contact ball bearing (3-12) and realize the translation drive of the moving platform (2).

5. The six-degree-of-freedom hybrid robot system based on error identification and compensation according to claim 1, characterized in that, The parallel robot support (4) consists of columns (4-1), side reinforcing plates (4-2), and reinforcing plate mounting blocks (4-3). The parallel robot support (4) is installed above the static platform (1) to support the three parallel branch components (3) and maintain their parallelism and structural rigidity. The three columns (4-1) are vertically fixed at equal intervals on the static platform (1), and any two adjacent columns (4-1) are connected to each other through the side reinforcing plates (4-2) and the reinforcing plate mounting blocks (4-3) to form an overall frame structure.

6. The six-degree-of-freedom hybrid robot system based on error identification and compensation according to claim 1, characterized in that, The direct drive motor (5-1) is an integrated DD motor structure with an integrated high-resolution encoder. The direct drive motor (5-1) in the first rotary joint is mounted on the base connector (5-2), the direct drive motor (5-1) in the second rotary joint is mounted on the intermediate support (5-3), the direct drive motor (5-1) in the third rotary joint is mounted on the terminal connector (5-4), the terminal connector (5-4) is mounted on the rotor of the direct drive motor (5-1) in the second rotary joint, the intermediate support (5-3) is mounted on the rotor of the direct drive motor (5-1) in the first rotary joint, and the base connector (5-2) is mounted on the moving platform (2). The attitude adjustment is achieved by driving the end effector assembly (6) through the three rotary joints.

7. The six-degree-of-freedom hybrid robot system based on error identification and compensation according to claim 1, characterized in that, The end effector assembly (6) is installed at the output end of the series rotary joint assembly (5) and is fixed to the third rotary joint through a standard flange interface. The end effector assembly (6) can be replaced with different functional modules according to the operation requirements. The functional modules include, but are not limited to, grinding head and pin device.

8. The six-degree-of-freedom hybrid robot system based on error identification and compensation according to claim 1, characterized in that, The controller module (7) adopts an error identification and compensation control strategy to model and correct the kinematic errors of parallel and series mechanisms, thereby realizing coordinated control of parallel and series mechanisms.

9. The six-degree-of-freedom hybrid robot system based on error identification and compensation according to claim 1, characterized in that, The error identification and compensation control strategy includes the following steps: 1) Offline data acquisition stage: Under various typical robot poses, record the actual measurement data of the position of the linear slider (3-2) in the parallel mechanism, the rotation angle of the serial mechanism, and the end pose; at the same time, calculate the theoretical end pose based on the theoretical geometric parameters to form a theoretical-actual corresponding dataset. 2) Independent error identification stage: For parallel mechanisms, a genetic algorithm is used to optimize and identify the geometric parameters to obtain the error model of the parallel mechanism; for series mechanisms, a differential evolution algorithm is used to identify the length deviation of each rotary joint link to obtain the error model. 3) Error compensation and coordinated control stage: In the controller module (7), the corrected actual geometric parameters are replaced with theoretical parameters, the forward and inverse kinematics calculations are updated in real time, and the target displacement and attitude commands of each drive shaft are generated; and in each control cycle, closed-loop correction is performed based on the feedback information of the servo motor (3-5), the grating ruler (3-6) and the direct drive motor (5-1) to realize the coordinated compensation control of parallel and series mechanisms.

Images