A robotic arm motion accuracy detection system and method based on moiré fringe

Abstract

The present invention provides a system and method for detecting the motion accuracy of a robotic arm based on moiré fringes, and relates to the field of image data processing. The system comprises: a grating projection module, for determining grating parameters based on motion information of the robotic arm, and projecting a grating pattern onto a detection surface of the robotic arm based on the grating parameters, wherein the grating parameters include at least grating density, grating angle, and grating type; an image acquisition module, for acquiring continuous-frame moiré fringe images during the motion of the robotic arm; and a data processing module, for performing phase calculation on the continuous-frame moiré fringe images, and resolving the displacement error, rotation angle error, and vibration error of the robotic arm. The system has the advantage of improving the accuracy of robotic arm motion detection.

Description

Technical Field

[0001] The present invention relates to the field of image data processing, and in particular to a system and method for detecting the motion accuracy of a robotic arm based on moiré fringes. Background Art

[0002] Robotic arms are widely used in industrial manufacturing, medical equipment, and scientific research experiments. Their motion accuracy directly affects work efficiency and product quality. With the rapid development of intelligent manufacturing, the precision requirements for robotic arms are also increasing.

[0003] Existing methods for measuring the motion accuracy of robotic arms are gradually revealing their limitations in certain application scenarios. For example, contact measurement methods use encoders or displacement sensors to directly measure the motion accuracy of robotic arms. While these methods offer a certain degree of reliability, the measurement process is affected by friction, mechanical installation errors, and motion inertia, which can lead to reduced accuracy. Furthermore, contact measurement can damage the robotic arm's surface, limiting its application in high-precision scenarios. Laser interferometry, a typical non-contact measurement technique, enables high-precision measurement at the nanometer level. However, laser interferometer systems are complex and costly, and are sensitive to environmental vibrations, air turbulence, and temperature fluctuations, making them difficult to meet the stability requirements of complex industrial environments. Especially in multi-degree-of-freedom measurement, laser interferometry requires multiple sets of equipment to work together, increasing system complexity. For another example, using texture features from optical projection to analyze the displacement and deformation of robotic arms is simple to operate, but its detection accuracy is limited by the resolution of the projection system and is susceptible to interference from ambient light, limiting its application in dynamic detection scenarios.

[0004] Therefore, it is necessary to provide a robotic arm motion accuracy detection system and method based on moiré fringes to improve the accuracy of robotic arm motion detection. Summary of the Invention

[0005] The present invention provides a moiré-based robotic arm motion accuracy detection system, comprising: a grating projection module, configured to determine grating parameters based on motion information of the robotic arm, and project a grating pattern onto a detection surface of the robotic arm based on the grating parameters, wherein the grating parameters include at least grating density, grating angle, and grating type; an image acquisition module, configured to acquire continuous-frame moiré fringe images through multiple image acquisition devices during the motion of the robotic arm; and a data processing module, configured to perform phase calculation on the continuous-frame moiré fringe images, and resolve displacement error, rotation angle error, and vibration error of the robotic arm.

[0006] Furthermore, the grating projection module determines grating parameters based on the motion information of the robotic arm, including: determining grating density based on detection accuracy requirements and the motion speed of the robotic arm.

[0007] Furthermore, the grating projection module determines the grating parameters based on the motion information of the robotic arm, including: if the motion type of the robotic arm is translational motion, determining the grating type to be a linear grating; if the motion type of the robotic arm is rotational motion, determining the grating type to be a radial grating; if the motion type of the robotic arm is compound motion, dynamically switching the grating type according to the size of the motion component.

[0008] Furthermore, the grating projection module determines the grating parameters based on the motion information of the robotic arm, including: if the motion type of the robotic arm is translational motion, determining that the grating angle is perpendicular to the motion direction; if the motion type of the robotic arm is rotational motion, determining that the grating center is aligned with the rotation axis.

[0009] Furthermore, the data processing module performs phase calculation on continuous frame moiré fringe images to solve the displacement error, rotation angle error and vibration error of the robotic arm, including: when the motion type is translational motion, converting each moiré fringe image from the spatial domain to the frequency domain, extracting the phase information of the moiré fringe, generating a phase map, converting the phase map into a continuous phase through a phase unwrapping algorithm, and calculating the displacement error based on the continuous phase.

[0010] Furthermore, the data processing module performs phase calculation on continuous frame moiré fringe images to solve the displacement error, rotation angle error and vibration error of the robotic arm, including: when the motion type is rotational motion, converting each moiré fringe image from the spatial domain to the frequency domain, extracting the phase information of the moiré fringe, generating a phase map, converting the phase map into a continuous phase through a phase unwrapping algorithm, and calculating the rotation angle error based on the continuous phase.

[0011] Furthermore, the data processing module is also used to perform error compensation for multi-degree-of-freedom motion of the robotic arm through multi-channel data fusion technology.

[0012] Furthermore, the data processing module performs error compensation for multi-degree-of-freedom motion of the robotic arm through multi-channel data fusion technology, including: preprocessing continuous frame moiré fringe images to generate preprocessed continuous frame moiré fringe images, wherein the preprocessing includes at least data alignment, data normalization and filtering processing; and performing data association, state estimation and error compensation on the preprocessed continuous frame moiré fringe images.

[0013] Furthermore, the system also includes: a display module for visualizing the displacement error, rotation angle error and vibration error of the robotic arm.

[0014] The present invention provides a method for detecting the motion accuracy of a robotic arm based on moiré fringes, which is applied to the above-mentioned robotic arm motion accuracy detection system based on moiré fringes. The method determines grating parameters based on the motion information of the robotic arm, and projects a grating pattern onto the detection surface of the robotic arm based on the grating parameters, wherein the grating parameters include at least grating density, grating angle, and grating type. During the motion of the robotic arm, continuous frame moiré fringe images are acquired through multiple image acquisition devices. Phase calculation is performed on the continuous frame moiré fringe images to resolve the displacement error, rotation angle error, and vibration error of the robotic arm.

[0015] Compared with the existing technology, the system and method for detecting the motion accuracy of a robotic arm based on moiré fringes provided by the present invention have at least the following beneficial effects:

[0016] 1. By accurately determining grating parameters (such as grating density, angle, and type) based on the robot arm's motion information, it can adapt to detection needs under different motion states, so that the projected grating pattern is more in line with the actual motion situation, which helps to obtain more accurate and clearer Moiré fringe images, laying the foundation for subsequent high-precision error resolution.

[0017] 2. The data processing module can perform phase calculation on continuous frame moiré fringe images, which can not only solve the displacement error and rotation angle error, but also solve the vibration error, and comprehensively evaluate the motion accuracy of the robotic arm. Compared with the traditional single-dimensional detection method, it can more deeply understand the motion performance of the robotic arm.

[0018] 3. For the translational motion, rotational motion and combined motion of the robot arm, the appropriate grating type (linear grating, radial grating or dynamic switching) and grating angle are determined respectively, so that the system can flexibly adapt to various complex robot arm motion scenarios and has a wide range of applicability.

[0019] 4. For compound motion, the grating type can be dynamically switched according to the size of the motion component. This dynamic adjustment mechanism can better capture the various motion characteristics of the robot arm during compound motion and improve the accuracy and reliability of detection.

[0020] 5. Through multi-channel data fusion technology, error compensation of the multi-degree-of-freedom motion of the robot arm can be performed. This means that its motion accuracy information can be obtained in real time during the movement of the robot arm, potential problems can be discovered in time, and a basis can be provided for the real-time control and adjustment of the robot arm, thereby improving production efficiency and safety.

[0021] 6. The system adopts advanced data processing methods such as phase calculation, phase unwrapping algorithm and multi-channel data fusion technology, which can quickly and accurately process and analyze a large number of moiré fringe images, greatly improving detection efficiency and reducing manual intervention and errors.

[0022] 7. In the multi-channel data fusion process, the continuous frame moiré fringe images obtained by multiple image acquisition devices are pre-processed (data alignment, data normalization and filtering) and fusion processed (data association, state estimation and error compensation), which can effectively eliminate noise and interference in the image, improve the accuracy and reliability of the measurement results, and ensure the stable operation of the detection system in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] This specification will be further described in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, the same numbers represent the same structures, wherein:

[0024] Figure 1 is a module schematic diagram of a robotic arm motion accuracy detection system based on moiré fringes according to some embodiments of this specification;

[0025] Figure 2 This is a flow chart of a method for detecting the motion accuracy of a robotic arm based on moiré fringes according to some embodiments of this specification. DETAILED DESCRIPTION

[0026] To more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly describes the drawings required for describing the embodiments. Obviously, the drawings described below are merely examples or embodiments of this specification. Those skilled in the art can apply this specification to other similar scenarios based on these drawings without inventive effort. Unless otherwise apparent from the context or otherwise noted, the same reference numerals in the figures represent the same structure or operation.

[0027] Figure 1 is a module diagram of a robotic arm motion accuracy detection system based on moiré fringes according to some embodiments of this specification, such as Figure 1 As shown, a moiré-based robotic arm motion accuracy detection system can include a grating projection module, an image acquisition module, and a data processing module. This moiré-based robotic arm motion accuracy detection system can be applied in a variety of scenarios. For example, it can be used for production line inspection of industrial robotic arms, monitoring the motion accuracy of the robotic arms in real time to ensure product quality. Another example is the ability to optimize equipment performance by detecting the motion accuracy of the robotic arms during robot assembly and calibration. Another example is the dynamic measurement used in high-precision mechanical motion experiments, providing high-precision data support for research. The following provides a detailed explanation of each module.

[0028] The grating projection module uses a projected moiré structure based on the robot's motion information. This structure consists of a set of high-precision grating patterns and a DLP-projected grating pattern. The high-precision grating pattern is placed on the moving part of the robot (for example, a joint or linear guide). The density and angle of the grating pattern are designed based on the movement method. Specifically, linear gratings are used to measure linear motion, and radial gratings are used for high-precision detection of rotational motion. To further improve accuracy, the DLP projection module generates projection fringes of the same shape but different periods based on the different grating forms. When these projection fringes are projected directly onto the grating surface, moiré fringes appear.

[0029] Specifically, the grating projection module uses an adjustable grating generator and digital light processing (DLP) technology to project high-precision grating patterns, including linear and radial gratings. A linear grating consists of a series of equally spaced, parallel stripes. These stripes can be alternating light and dark, or a grid with alternating translucent and opaque areas. A radial grating's stripes radiate radially outward from a central point, forming a pattern resembling "sun rays." The stripe spacing increases with radius (equal angular spacing) or remains constant (equal linear spacing).

[0030] The grating projection module can dynamically adjust the grating pattern according to the motion information of the robotic arm and the detection accuracy requirements.

[0031] For example, the grating projection module can determine the grating density based on the detection accuracy requirement and the movement speed of the robotic arm.

[0032] Specifically, the input parameters are: the movement speed of the robotic arm and the detection accuracy requirements.

[0033] Adjustment Logic: When the robot moves at high speeds, the grating density is reduced (the grating period is increased) to avoid blurring the moiré pattern. When the robot moves at low speeds, the grating density is increased (the grating period is reduced) to improve detection sensitivity. Based on the required detection accuracy, the grating density is dynamically adjusted to ensure that the phase solution meets the required accuracy.

[0034] As an example only, the grating density can be calculated according to the following formula:

[0035]

[0036] Where d is the grating period, k is the scale factor, v is the robot arm's velocity, and P is the required detection accuracy. Detection accuracy (P) refers to the smallest dimensional change that the detection system can resolve or measure, expressed in microns (μm). It directly reflects the system's ability to discern subtle differences in target features.

[0037] As an example only, the robot arm performs precision detection under high-speed translation (v=1m / s), and the detection accuracy requirement is P=10μm. Based on the movement speed (v=1m / s) and the detection accuracy (P=10μm), the grating density is calculated as follows:

[0038]

[0039] Among them, translational motion refers to the movement of the end of the robotic arm in a single direction (such as the X, Y, or Z axis) along a straight line.

[0040] For example, the robot arm performs precision detection under low-speed rotation (angular velocity ω = 0.1 rad / s). The detection accuracy requirement is P = 0.001 μm. Calculate the grating density:

[0041]

[0042] Among them, rotational motion refers to the rotation of the robot arm joint around the axis.

[0043] Understandably, during high-speed motion, if the grating period is too small, the displacement of the robotic arm will cause the moiré fringes to change rapidly, exceeding the sampling frequency of the camera and causing image blur. By increasing the grating period, the rate of change of the moiré fringes is reduced, ensuring that the camera can clearly capture the fringe information. During low-speed motion, the displacement of the robotic arm is small, and the grating density needs to be increased to enhance the accuracy of phase resolution. Denser grating fringes can provide richer phase information, thereby improving the detection system's ability to resolve subtle deformations. By dynamically adjusting the grating density, the system can achieve optimal performance under different motion speeds and detection accuracy requirements. Increase the period at high speeds to avoid blurring, and reduce the period at low speeds to enhance sensitivity, thereby ensuring that the accuracy of phase resolution always meets the requirements. This strategy not only improves detection accuracy and system stability, but also enhances the system's adaptability and cost-effectiveness, providing strong support for dynamic detection tasks.

[0044] In some embodiments, the grating projection module determines grating parameters based on the motion information of the robotic arm, including:

[0045] If the motion type of the robot arm is translational motion, determine that the grating type is linear grating;

[0046] If the motion type of the robot arm is rotational motion, determine that the grating type is radial grating;

[0047] If the motion type of the robot arm is compound motion, the grating type is dynamically switched according to the size of the motion components.

[0048] In some embodiments, the grating projection module determines grating parameters based on the motion information of the robotic arm, including:

[0049] If the motion type of the robot arm is translational motion, the grating angle is determined to be perpendicular to the motion direction to maximize the contrast of the moiré fringes;

[0050] If the robot arm's motion type is rotational, ensure that the center of the grating is aligned with the axis of rotation to increase the sensitivity of rotation detection.

[0051] As an example, consider a robotic arm performing precision inspection during high-speed translation (v = 1m / s). The required inspection accuracy is P = 10μm. The robotic arm's motion is translational, with a direction of θ = 30°. Therefore, the grating angle α = θ + 90° = 30° + 90° = 120°. A linear grating with a 10mm period and a 120° grating angle is used. The system can clearly capture moiré patterns during high-speed translation, ensuring an inspection accuracy of 10μm.

[0052] For another example, the robotic arm performs precision detection during low-speed rotation (angular velocity ω = 0.1 rad / s). A radial grating is used, and the center of the grating is aligned with the rotation axis. The system can accurately detect angular changes in low-speed rotation with a detection accuracy of 0.001°.

[0053] For another example, the grating projection module can dynamically switch the grating type according to the size of the motion component based on the following process:

[0054] S11. Determine the size of the motion component:

[0055] Translational component: Linear velocity of the end of the robot along the X, Y, and Z axes ,in, is the X-axis speed, is the Y-axis speed, is the Z axis speed;

[0056] Rotational component: angular velocity ω of the manipulator joint around its axis (unit: rad / s);

[0057] S12. Weight calculation:

[0058] Define the normalized translation component weight and the rotation component weight ,in, and are the maximum allowed translational and angular velocities.

[0059] S13, Dynamic switching of grating type:

[0060] When the translation component weight Greater than or equal to When , the grating type is determined to be linear grating;

[0061] When the translation component weight Less than , the grating type is determined to be a radial grating.

[0062] As you can understand, in translational motion (such as robotic arm grasping and surface scanning), linear gratings can accurately capture linear displacement, avoiding fringe blur or phase errors caused by mismatches between the grating period and motion speed. In rotational motion (such as robotic arm joint rotation and curved surface scanning), radial gratings can efficiently measure angular changes, reducing detection errors caused by mismatches between the grating orientation and the rotation axis. By switching grating types in real time, the system can adapt to different motion scenarios, ensuring optimal detection accuracy. In complex motion (such as translation and rotation), traditional fixed gratings are prone to error accumulation due to coupled motion components. A dynamic switching mechanism separates the translational and rotational components, reducing the impact of interference. For example, when a robotic arm performs high-speed translation accompanied by slight rotation, the system can prioritize linear gratings and use filtering algorithms to compensate for the rotational component, avoiding distortion in detection results. The same grating projection module can be used for multiple tasks (such as assembly, inspection, and scanning), eliminating the need to redesign hardware for different tasks. For example, in industrial assembly, a robotic arm might use linear gratings for grasping tasks and switch to radial gratings for screwdriving.

[0063] The image acquisition module is used to collect the moiré fringe image facing the grating during the movement of the robotic arm.

[0064] Specifically, the image acquisition module is equipped with a high-resolution industrial camera, which, combined with a lens assembly, captures moiré patterns on the surface of the robotic arm. The camera's high acquisition frequency meets the dynamic detection requirements of the robotic arm's rapid movements.

[0065] The data processing module is used to perform phase calculation on continuous frame moiré fringe images and solve the displacement error, rotation angle error and vibration error of the robotic arm.

[0066] For example, when the motion type is translational motion, the data processing module can convert each moiré fringe image from the spatial domain to the frequency domain, extract the phase information of the moiré fringe, generate a phase map, convert the phase map into a continuous phase through a phase unwrapping algorithm, and calculate the displacement error based on the continuous phase.

[0067] Specifically, before extracting the phase information of the moiré fringe, it is necessary to preprocess the continuous frame moiré fringe images, specifically including:

[0068] 1. Data alignment

[0069] Time alignment: Synchronize the time of continuous frame moiré fringe images acquired by multiple image acquisition devices to ensure consistent timestamps of the data.

[0070] Spatial alignment: Perform spatial registration on the continuous frame moiré fringe images acquired by multiple image acquisition devices to ensure that the continuous frame moiré fringe images are in the same coordinate system.

[0071] 2. Data filtering

[0072] Perform filtering (such as Kalman filtering or low-pass filtering) on ​​the moiré fringe image with large noise to improve the signal-to-noise ratio of the moiré fringe image.

[0073] A two-dimensional fast Fourier transform (FFT) is performed on the preprocessed moiré fringe image to generate a spectrogram. The peaks in the spectrogram correspond to the dominant frequency components of the moiré fringe. From the spectrogram, a bandpass filter is used to extract the dominant frequency components of the moiré fringe, removing low-frequency background and high-frequency noise.

[0074] Perform inverse Fast Fourier Transform (IFFT) on the filtered spectrum to obtain a complex image containing phase information.

[0075] Compute the phase image from the real and imaginary parts of the complex image:

[0076]

[0077] in, For the complex image, is the imaginary part of the complex image, is the real part of the complex image, The phase diagram.

[0078] Since the phase map is a wrapped phase (ranging from -π to π), it needs to be converted into a continuous phase through a phase unwrapping algorithm to calculate the displacement and rotation errors of the robot arm. The goal of phase unwrapping is to eliminate the 2π jumps in the phase map and restore the continuous phase distribution. Phase unwrapping algorithms can include: Quality-Guided Phase Unwrapping: Determine the unwrapping path based on the phase quality map (such as the phase derivative variance). Least-Squares Phase Unwrapping: Solve the continuous phase by minimizing the phase gradient error.

[0079] Mass-guided phase unwrapping steps:

[0080] Compute Phase Quality Map: Calculate the phase quality of each pixel based on phase derivative variance or phase consistency.

[0081] Determine the unwrapping path: start from the pixels with the highest phase quality and gradually unwrap pixels with lower phase quality.

[0082] Phase unwrapping: Accumulate 2π jumps along the unwrapping path to restore the continuous phase.

[0083] For a linear grating, the relationship between the displacement d and the phase Φ is:

[0084]

[0085] Where p is the grating period.

[0086] For a radial grating, the relationship between the rotation angle θ and the phase Φ is:

[0087]

[0088] Where r is the radius of the radial grating.

[0089] According to the phase change of the linear grating, the displacement error of the robot arm in the X, Y, and Z directions is calculated.

[0090] Specifically, the displacement error is calculated as follows:

[0091] Reference phase and measurement phase:

[0092] Reference phase: When the robotic arm is in the initial position, the moiré fringe image is collected and the phase map is calculated to obtain the reference phase .

[0093] Measuring phase: When the robotic arm moves to the target position, the moiré fringe image is collected and the phase diagram is calculated to obtain the measured phase .

[0094] Phase difference calculation:

[0095] Calculate the phase difference for each pixel:

[0096]

[0097] Displacement calculation:

[0098] Calculate the displacement of each pixel based on the phase difference:

[0099]

[0100] Average displacement error:

[0101] The average displacement of all pixels is averaged to obtain the average displacement error of the robotic arm:

[0102]

[0103] Wherein, M is the number of rows of pixels in the moiré fringe image, and N is the number of columns of pixels in the moiré fringe image.

[0104] Continuous frame moiré fringe images are acquired by multiple image acquisition devices, and multi-channel data are fused through the least squares method to improve the displacement detection accuracy.

[0105] For example, when the motion type is rotational motion, the data processing module can convert each moiré fringe image from the spatial domain to the frequency domain, extract the phase information of the moiré fringe, generate a phase map, convert the phase map into a continuous phase through a phase unwrapping algorithm, and calculate the rotation angle error based on the continuous phase.

[0106] For radial moiré patterns, polar coordinate fast Fourier transform is used to convert the moiré fringe image from the spatial domain to the frequency domain, extract the phase information of the moiré fringes, and generate a phase map to improve the detection accuracy of rotational motion.

[0107] The rotation angle error of the robot arm is calculated based on the phase change of the radial grating.

[0108] For multi-DOF robotic arms, the rotation error is converted into Euler angle or quaternion representation through coordinate transformation.

[0109] Perform time-frequency analysis (such as short-time Fourier transform or wavelet transform) on the displacement and rotation errors to extract the vibration frequency and amplitude of the manipulator. Spectral analysis identifies the source of vibration and provides data support for vibration control of the manipulator.

[0110] Just as an example, displacement detection of a linear grating:

[0111] Input: Moiré fringe image of a linear grating.

[0112] Processing steps:

[0113] Image preprocessing: denoising, enhancement, and cropping.

[0114] FFT calculation: extract the main frequency component of the moiré fringes.

[0115] Phase Extraction: Computes the wrapped phase map.

[0116] Phase Unwrapping: Restoring a continuous phase.

[0117] Displacement calculation: Calculate the displacement of the robotic arm based on the phase change.

[0118] Output: Displacement of the robot arm in the X, Y, and Z directions.

[0119] Another example is the rotation detection of the radial grating:

[0120] Input: Moiré fringe image of a radial grating.

[0121] Processing steps:

[0122] Image preprocessing: denoising, enhancement, and cropping.

[0123] Polar coordinate FFT calculation: extract the main frequency component of rotational motion.

[0124] Phase Extraction: Computes the wrapped phase map.

[0125] Phase Unwrapping: Restoring a continuous phase.

[0126] Rotation calculation: Calculate the rotation angle of the robot arm based on the phase change.

[0127] Output: The rotation angle of the robotic arm.

[0128] The data processing module is also used to:

[0129] Through multi-channel data fusion technology, error compensation of multi-degree-of-freedom motion of the robotic arm is carried out.

[0130] Specifically include:

[0131] Preprocessing the continuous frame moiré fringe images to generate preprocessed continuous frame moiré fringe images, wherein the preprocessing at least includes data alignment, data normalization, and filtering processing;

[0132] Data association, state estimation and error compensation are performed on the preprocessed continuous frame moiré fringe images.

[0133] 1. Multi-channel data sources

[0134] The input data for multi-channel data fusion includes:

[0135] Raster data: Linear raster data: used to detect the translational movement of the robot arm (X, Y, Z directions) or radial raster data: used to detect the rotational movement of the robot arm (the robot arm joint rotates around the axis).

[0136] Time series data: Continuous frame moiré fringe images during the robot arm's motion, used to analyze dynamic errors and vibration characteristics.

[0137] 2. Data preprocessing

[0138] Before data fusion, multi-channel data needs to be preprocessed to ensure data consistency and reliability.

[0139] (1) Data alignment and time alignment: Time synchronization is performed on the continuous frame moiré fringe images acquired by multiple image acquisition devices to ensure that the timestamps of the data are consistent.

[0140] Spatial alignment: Perform spatial registration on multiple consecutive frames of moiré fringe images to ensure that the data are in the same coordinate system.

[0141] (2) Data normalization

[0142] Normalize data from different channels to the same dimension and range to facilitate subsequent fusion calculations.

[0143] For phase data, convert the wrapped phase (-π to π) to a continuous phase.

[0144] (3) Data filtering

[0145] Perform filtering (such as Kalman filtering or low-pass filtering) on ​​data with large noise to improve the signal-to-noise ratio of the data.

[0146] 3. Multi-channel data fusion algorithm

[0147] The core algorithms for multi-channel data fusion include data association, state estimation, and error compensation. The specific steps are as follows:

[0148] (1) Data association

[0149] Target matching: Associating data from different channels to the same target (such as the end effector of a robotic arm).

[0150] Feature extraction: Extract features (such as displacement, angle, phase change, etc.) from the data of each channel.

[0151] Similarity calculation: Calculate the similarity between data from different channels and determine the correlation between the data.

[0152] (2) State estimation

[0153] State vector definition: Define the state vector s of the robot arm, including position, instantaneous linear velocity, instantaneous angular velocity, rotation angle, etc.:

[0154]

[0155] in, is the state vector, is the absolute position of the end of the robot arm in the X-axis direction, is the absolute position of the end of the robot arm in the Y-axis direction, is the absolute position of the end of the robot arm in the Z-axis direction, is the rotation angle of the end of the robot arm around the X axis, is the rotation angle of the end of the robot arm around the Y axis, is the rotation angle of the end of the robot arm around the Z axis, is the instantaneous linear velocity of the end of the robot arm in the X-axis direction, is the instantaneous linear velocity of the end of the robot arm in the Y-axis direction, is the instantaneous linear velocity of the end of the robot arm in the Z-axis direction, is the instantaneous angular velocity of the end of the robotic arm around the X axis, is the instantaneous angular velocity of the end of the robotic arm around the Y axis, is the instantaneous angular velocity of the end of the robotic arm around the Z axis.

[0156] State prediction: Based on the motion model of the robot arm (such as the rigid body motion model), predict the state at the next moment:

[0157]

[0158] in, is the state transition matrix, is the process noise.

[0159] Status Update:

[0160] Based on multi-channel observation data, update the state estimate:

[0161]

[0162] in, is the observation vector, is the observation matrix, is the Kalman gain.

[0163] (3) Error compensation

[0164] Residual Calculation: Calculate the residual between the observed data and the predicted state:

[0165]

[0166] Error compensation: Adjust the state estimate based on the residual to compensate for system errors (such as sensor bias, installation error, etc.).

[0167] Just as an example, the synchronous detection of translation and rotation:

[0168] Input: Linear raster data: displacement of the robot in the X, Y, and Z directions. Radial raster data: rotation angle of the robot around the X, Y, and Z axes.

[0169] Processing steps:

[0170] Data alignment: Continuous frames of moiré fringe images acquired by multiple image acquisition devices are registered to the same coordinate system, and displacement and angle features are extracted from each image.

[0171] State prediction: Predict the state at the next moment based on the motion model of the robotic arm.

[0172] State update: Update the state estimate based on the observed data.

[0173] Error compensation: Compensate for system errors and output accurate displacement and angle.

[0174] Output: 6-DOF motion state of the robot arm (x, y, z, θx ,θ y ,θ z ).

[0175] In some embodiments, a robotic arm motion accuracy detection system based on moiré fringes may further include a display module for visualizing the displacement error, rotation angle error, and vibration error of the robotic arm.

[0176] Specifically, the display module can display the robot arm's displacement error, rotation angle error, and vibration error as real-time curves or 3D models. The data can be exported to a standard format for quality analysis and equipment calibration.

[0177] Figure 2 is a flow chart of a method for detecting the motion accuracy of a robotic arm based on moiré fringes according to some embodiments of this specification, such as Figure 2 As shown, in some embodiments, a method for detecting the motion accuracy of a robotic arm based on moiré fringes may include the following process:

[0178] Determining grating parameters based on motion information of the robotic arm, and projecting a grating pattern onto a detection surface of the robotic arm based on the grating parameters, wherein the grating parameters include at least grating density, grating angle, and grating type;

[0179] During the movement of the robotic arm, continuous frame moiré fringe images are acquired;

[0180] Phase calculation is performed on continuous frame moiré fringe images to solve the displacement error, rotation angle error and vibration error of the robotic arm.

[0181] A method for detecting the motion accuracy of a robotic arm based on moiré fringes can be applied to a system for detecting the motion accuracy of a robotic arm based on moiré fringes, which will not be described in detail here.

[0182] Finally, it should be understood that the embodiments described in this specification are intended only to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly described and illustrated in this specification.

Claims

1. A robotic arm motion accuracy detection system based on moiré fringes, characterized in that: include: a grating projection module, configured to determine grating parameters based on motion information of the robotic arm, and project a grating pattern onto a detection surface of the robotic arm based on the grating parameters, wherein the grating parameters include at least grating density, grating angle, and grating type; An image acquisition module is used to acquire continuous frame moiré fringe images through multiple image acquisition devices during the movement of the robotic arm; The data processing module is used to calculate the phase of the continuous frame moiré fringe image and solve the displacement error, rotation angle error and vibration error of the robotic arm; The grating projection module determines the grating parameters based on the motion information of the robotic arm, including: S11. Determine the size of the motion component: Translational component: Linear velocity of the end of the robot along the X, Y, and Z axes ,in, is the X-axis speed, is the Y-axis speed, is the Z axis speed; Rotational component: angular velocity ω of the manipulator joint around the axis; S12. Weight calculation: Define the normalized translation component weight and the rotation component weight ,in, and is the maximum allowed translational velocity and angular velocity; S13, Dynamic switching of grating type: When the translation component weight Greater than or equal to When , the grating type is determined to be linear grating; When the translation component weight Less than When , the grating type is determined to be a radial grating; When the grating type is linear grating, the grating density is calculated based on the following formula: in, is the grating period, is the proportionality coefficient, is the movement speed of the robot arm, The detection accuracy requirement is the minimum dimensional change that the detection system can distinguish or measure; When the grating type is radial grating, the grating density is calculated based on the following formula: in, is the angular velocity of the robot arm joint around the axis; The data processing module performs phase calculation on the continuous frame moiré fringe images to solve the displacement error, rotation angle error and vibration error of the robotic arm, including: When the motion type is rotational motion, each moiré fringe image is converted from the spatial domain to the frequency domain, the phase information of the moiré fringe is extracted, a phase map is generated, the phase map is converted into a continuous phase using a phase unwrapping algorithm, and the rotation angle error is calculated based on the continuous phase; The data processing module is further configured to: Through multi-channel data fusion technology, error compensation of multi-degree-of-freedom motion of the robot arm is carried out, including: Preprocessing the continuous frame moiré fringe images to generate preprocessed continuous frame moiré fringe images, wherein the preprocessing at least includes data alignment, data normalization, and filtering processing; Data association, state estimation and error compensation are performed on the preprocessed continuous frame moiré fringe images.

2. The robotic arm motion accuracy detection system based on moiré fringes according to claim 1, characterized in that: The grating projection module determines grating parameters based on the motion information of the robotic arm, including: If the motion type of the robot arm is translational motion, make sure the grating angle is perpendicular to the motion direction; If the robot arm's motion type is rotational, ensure that the center of the grating is aligned with the axis of rotation.

3. The robotic arm motion accuracy detection system based on moiré fringes according to claim 2, characterized in that: The data processing module performs phase calculation on the continuous frame moiré fringe images to solve the displacement error, rotation angle error and vibration error of the robotic arm, including: When the motion type is translational motion, each moiré fringe image is converted from the spatial domain to the frequency domain, the phase information of the moiré fringe is extracted, a phase map is generated, the phase map is converted into a continuous phase through a phase unwrapping algorithm, and the displacement error is calculated based on the continuous phase.

4. A robotic arm motion accuracy detection system based on moiré fringes according to any one of claims 1 to 3, characterized in that: Also includes: Display module, used to visualize the displacement error, rotation angle error and vibration error of the robot arm.

5. A method for detecting the motion accuracy of a robotic arm based on moiré fringes, characterized in that: A robotic arm motion accuracy detection system based on moiré fringes, as claimed in any one of claims 1 to 4, comprising: Determining grating parameters based on motion information of the robotic arm, and projecting a grating pattern onto a detection surface of the robotic arm based on the grating parameters, wherein the grating parameters include at least grating density, grating angle, and grating type; During the movement of the robotic arm, continuous frame moiré fringe images are acquired through multiple image acquisition devices; Phase calculation is performed on continuous frame moiré fringe images to solve the displacement error, rotation angle error and vibration error of the robotic arm.

Images