A livestock grabbing manipulator control method and system

Abstract

The present invention relates to the field of manipulator control technology, and in particular to a livestock grabbing manipulator control method and system. The method comprises the following steps: collecting the stress state of livestock in the manipulator grabbing state and monitoring the breaking force to obtain the monitoring data of the breaking force of the livestock in the stress state; quantifying the difference in contraction of the subcutaneous tissue gap of the whole body according to the monitoring data of the breaking force of the livestock in the stress state, and estimating the risk of breaking of livestock syringe injection, to obtain the risk data of breaking of livestock syringe injection; regulating the injection behavior of the manipulator according to the risk data of breaking of livestock syringe injection, to obtain the injection behavior regulation data of the manipulator; designing the automatic control firmware of the manipulator based on the injection behavior regulation data of the manipulator, to obtain the automatic control firmware of the manipulator. The present invention makes the manipulator control technology more accurate by optimizing the manipulator control technology.

Description

Technical Field

[0001] The invention relates to the technical field of manipulator control, and in particular to a livestock grabbing manipulator control method and system. Background Art

[0002] The research background of livestock grasping manipulator control method stems from the growing demand for automated and intelligent operations in modern agriculture and animal husbandry. With the continuous expansion of breeding scale and the increase in labor intensity, the previous manual grasping and handling of livestock has gradually exposed problems such as inefficiency, danger and unsustainability. Especially in animal care, vaccination, physical examination, slaughter and other links, how to grasp and handle livestock efficiently and safely has become a difficult problem that the industry needs to solve urgently. To this end, the application of automated grasping technology and robotic arms has become a key direction to improve production efficiency. The robotic arm has high precision, high reliability and high repeatability in these operations, which can effectively reduce the physical burden of personnel and reduce the risk of accidental injury to livestock. However, the traditional manipulator control method has low accuracy in recognizing the escape behavior of livestock under stress, and thus cannot accurately perform safe and stable grasping control of the manipulator. Summary of the invention

[0003] Based on this, it is necessary to provide a livestock grasping manipulator control method and system to solve at least one of the above technical problems.

[0004] To achieve the above object, a livestock grabbing manipulator control method is provided, the method comprising the following steps:

[0005] Step S1: The stress state of the livestock in the grasping state of the robot is collected by an electronic monitoring device to obtain a video of the stress state of the livestock; based on the video of the stress state of the livestock, the breaking force of the livestock in the grasping state of the robot is monitored by a piezoelectric sensor to obtain monitoring data of the breaking force of the livestock in the stress state;

[0006] Step S2: matching the struggling action sequence breaking force fluctuations according to the breaking force monitoring data of the livestock under stress state to obtain the action sequence breaking force fluctuation data; quantifying the whole body subcutaneous tissue gap contraction difference based on the action sequence breaking force fluctuation data to obtain the subcutaneous tissue gap contraction difference data; estimating the livestock needle tube injection breakage risk according to the subcutaneous tissue gap contraction difference data to obtain the livestock needle tube injection breakage risk data;

[0007] Step S3: matching the dynamic control force output of the manipulator according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the dynamic control force output data of the manipulator; regulating the injection behavior of the manipulator based on the dynamic control force output normalization data and the livestock needle injection breakage risk data to obtain the injection behavior regulation data of the manipulator;

[0008] Step S4: Design the manipulator automation control firmware based on the manipulator dynamic control force output data and the manipulator injection behavior regulation data, obtain the manipulator automation control firmware, and send the manipulator automation control firmware to the cloud platform to execute the livestock grasping manipulator control method.

[0009] Preferably, step S1 comprises the following steps:

[0010] Step S11: deploying electronic monitoring equipment and piezoelectric sensors on the manipulator;

[0011] Step S12: using electronic monitoring equipment to collect the stress state of the livestock in the state of being grasped by the robot arm, and obtaining a video of the stress state of the livestock;

[0012] Step S13: performing time-series frame processing on the livestock stress state video to obtain continuous frames of the livestock stress state;

[0013] Step S14: Continuously divide the frames based on the stress state of the livestock, and monitor the breaking force of the livestock in the grasping state of the manipulator through a piezoelectric sensor to obtain the breaking force monitoring data of the stress state of the livestock.

[0014] Preferably, step S2 comprises the following steps:

[0015] Step S21: extracting the struggling motion trajectory of the livestock in the stress state from the video of the livestock in the stress state to obtain the struggling motion trajectory of the stress state;

[0016] Step S22: matching the struggling action time sequence breaking free force fluctuation of the struggling action trajectory of the stress state according to the breaking free force monitoring data of the livestock in the stress state, and obtaining the breaking free force fluctuation data of the action time sequence;

[0017] Step S23: analyzing the dynamic tension of livestock muscles based on the action sequence breaking free force fluctuation data and the struggling action trajectory in the stress state to obtain the dynamic tension data of livestock muscles;

[0018] Step S24: quantifying the subcutaneous tissue gap contraction difference of the whole body based on the livestock muscle dynamic tension data to obtain subcutaneous tissue gap contraction difference data;

[0019] Step S25: Estimate the risk of livestock syringe injection breakage based on the subcutaneous tissue gap contraction difference data and the action sequence breakage force fluctuation data to obtain livestock syringe injection breakage risk data.

[0020] Preferably, step S23 includes the following steps:

[0021] Step S231: estimating the livestock volume based on the livestock stress state video to obtain livestock volume estimation data, wherein the livestock include pigs, cattle and sheep;

[0022] Step S232: deducing the force stretching amplitude intervals of different muscle groups of the livestock based on the action sequence breaking free force fluctuation data and the livestock volume estimation data, and obtaining the muscle force stretching amplitude intervals of different muscle groups of the livestock;

[0023] Step S233: performing tensile strength correlation coefficient analysis on the muscle force stretching amplitude interval according to the livestock volume estimation data to obtain the volume-related tensile strength correlation coefficient;

[0024] Step S234: performing spatial posture recognition on the struggling motion trajectory in the stress state to obtain struggling posture recognition data;

[0025] Step S235: Deducing the stress distribution of muscle fibers of the force-scheduling livestock between different struggling postures according to the struggling posture recognition data, and obtaining the stress distribution data of the force-scheduling muscles;

[0026] Step S236: Perform livestock muscle dynamic tension analysis based on the volume-related tensile strength correlation coefficient and the force scheduling muscle stress distribution data to obtain livestock muscle dynamic tension data.

[0027] Preferably, step S3 comprises the following steps:

[0028] Step S31: matching the dynamic control force output of the manipulator according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the dynamic control force output data of the manipulator;

[0029] Step S32: normalizing the dynamic control force output data of the manipulator to obtain dynamic control force output normalized data;

[0030] Step S33: Based on the dynamic control force output normalization data, the subcutaneous tissue gap contraction difference data and the livestock needle injection breakage risk data, the manipulator injection behavior regulation is performed to obtain the manipulator injection behavior regulation data.

[0031] Preferably, step S31 includes the following steps:

[0032] Step S311: Draw a strength time series fluctuation curve for the action time series breakaway force fluctuation data to obtain a breakaway force strength time series fluctuation curve;

[0033] Step S312: calculating the instantaneous strength rise smoothness of the break-away force strength time series fluctuation curve to obtain the instantaneous strength rise smoothness of the break-away force;

[0034] Step S313: performing strength multi-peak area pattern recognition on the strength time series fluctuation curve of the breaking force according to the rising smoothness of the instantaneous strength of the breaking force, and obtaining the strength multi-peak area pattern of the breaking force;

[0035] Step S314: performing dynamic matching of the manipulator pressing control force based on the multi-peak area mode of the strength of the break-away force to obtain the manipulator pressing dynamic control data;

[0036] Step S315: matching the manipulator control area according to the subcutaneous tissue gap contraction difference data, and optimizing the multi-point control force of the manipulator pressing dynamic control data to obtain multi-point area control force data;

[0037] Step S316: matching the dynamic control force output of the manipulator according to the dynamic control data of the manipulator's pressing and the multi-point area control force data to obtain the dynamic control force output data of the manipulator.

[0038] Preferably, the present invention further provides a livestock grabbing manipulator control system for executing the livestock grabbing manipulator control method as described above, the livestock grabbing manipulator control system comprising:

[0039] The breaking force monitoring module is used to collect the stress state of the livestock in the state of being grasped by the manipulator through the electronic monitoring equipment to obtain the livestock stress state video; based on the livestock stress state video and using the piezoelectric sensor to monitor the breaking force of the livestock in the state of being grasped by the manipulator, to obtain the livestock stress state breaking force monitoring data;

[0040] The injection fracture risk estimation module is used to match the struggling action sequence breaking force fluctuation according to the breaking force monitoring data of the livestock under stress state, and obtain the action sequence breaking force fluctuation data; quantify the whole body subcutaneous tissue gap contraction difference based on the action sequence breaking force fluctuation data, and obtain the subcutaneous tissue gap contraction difference data; estimate the livestock needle tube injection fracture risk according to the subcutaneous tissue gap contraction difference data, and obtain the livestock needle tube injection fracture risk data;

[0041] The manipulator injection behavior control module is used to match the manipulator dynamic control force output according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the manipulator dynamic control force output data; the manipulator injection behavior is controlled based on the dynamic control force output normalization data and the livestock needle injection breakage risk data to obtain the manipulator injection behavior control data;

[0042] The automation control firmware design module is used to design the automation control firmware of the manipulator based on the dynamic control force output data of the manipulator and the injection behavior regulation data of the manipulator, obtain the automation control firmware of the manipulator, and send the automation control firmware of the manipulator to the cloud platform to execute the livestock grasping manipulator control method.

[0043] The beneficial effect of the present invention is that the livestock in the state of being grasped by the manipulator is video-captured by the electronic monitoring equipment, and the stress response data of the livestock is obtained in real time. The key to this process is to capture the behavior of the livestock through video monitoring, so as to analyze its stress level. At the same time, the piezoelectric sensor is used to monitor the breaking force of the livestock in the state of being grasped by the manipulator. These monitoring data can effectively reflect the intensity of the struggling behavior of the livestock under stress, and provide basic data for subsequent data analysis and risk assessment. By performing time series analysis on the monitoring data of the livestock's breaking force, the law of the fluctuation of the breaking force is identified. This fluctuation data can reflect the intensity and periodic changes of the livestock's struggle, and provide a basis for the subsequent quantitative analysis of the difference in contraction of the subcutaneous tissue gap of the whole body. By quantifying the data of the difference in contraction of the subcutaneous tissue gap, the changes in the subcutaneous tissue under different breaking forces can be evaluated. This process ultimately helps to deduce the risk of breaking of the needle tube injection of the livestock in the state of being grasped by the manipulator, and ensures the safe operation of the manipulator during the injection process. By analyzing the breaking force fluctuation data and the difference in contraction of the subcutaneous tissue gap, the dynamic control force that the manipulator needs to output is accurately calculated. Based on these data, the control system of the manipulator can adjust the force output in real time to ensure that the force is applied appropriately during the grasping process, avoiding excessive force causing injury or discomfort to the livestock. At the same time, the injection behavior of the manipulator is regulated by combining the injection break risk data. This control mechanism can effectively prevent the injection needle from breaking due to excessive force during the injection process of livestock, ensuring the accuracy and safety of the operation. Based on the aforementioned dynamic control force output data and injection behavior control data, the manipulator automation control firmware is designed. The design focus of this firmware is to integrate the various dynamic control and injection control functions of the manipulator into an intelligent system, so that the manipulator can respond appropriately under different stress and physical conditions. After the design is completed, the firmware will be uploaded to the cloud platform for remote monitoring and operation. This innovation enables the manipulator to automatically perform livestock grasping and injection tasks, improving operational efficiency while ensuring the safety and health of livestock. Therefore, the present invention is an optimization treatment of the traditional manipulator control method, which solves the problem that the traditional manipulator control method has low accuracy in identifying the breaking free behavior of livestock under stress, and thus cannot accurately perform safe and stable grasping control of the manipulator, improves the accuracy of identifying the breaking free behavior of livestock under stress, and improves the stability and safety of the manipulator's grasping control. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of the steps of a livestock grasping manipulator control method;

[0045] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0046] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG. DETAILED DESCRIPTION

[0047] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0048] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0049] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0050] To achieve this, please refer to Figures 1 to 3 , a livestock grasping manipulator control method, the method comprising the following steps:

[0051] Step S1: The stress state of the livestock in the grasping state of the robot is collected by an electronic monitoring device to obtain a video of the stress state of the livestock; based on the video of the stress state of the livestock, the breaking force of the livestock in the grasping state of the robot is monitored by a piezoelectric sensor to obtain monitoring data of the breaking force of the livestock in the stress state;

[0052] Step S2: matching the struggling action sequence breaking force fluctuations according to the breaking force monitoring data of the livestock under stress state to obtain the action sequence breaking force fluctuation data; quantifying the whole body subcutaneous tissue gap contraction difference based on the action sequence breaking force fluctuation data to obtain the subcutaneous tissue gap contraction difference data; estimating the livestock needle tube injection breakage risk according to the subcutaneous tissue gap contraction difference data to obtain the livestock needle tube injection breakage risk data;

[0053] Step S3: matching the dynamic control force output of the manipulator according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the dynamic control force output data of the manipulator; regulating the injection behavior of the manipulator based on the dynamic control force output normalization data and the livestock needle injection breakage risk data to obtain the injection behavior regulation data of the manipulator;

[0054] Step S4: Design the manipulator automation control firmware based on the manipulator dynamic control force output data and the manipulator injection behavior regulation data, obtain the manipulator automation control firmware, and send the manipulator automation control firmware to the cloud platform to execute the livestock grasping manipulator control method.

[0055] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of a livestock grabbing manipulator control method of the present invention. In this example, the livestock grabbing manipulator control method includes the following steps:

[0056] Step S1: The stress state of the livestock in the grasping state of the robot is collected by an electronic monitoring device to obtain a video of the stress state of the livestock; based on the video of the stress state of the livestock, the breaking force of the livestock in the grasping state of the robot is monitored by a piezoelectric sensor to obtain monitoring data of the breaking force of the livestock in the stress state;

[0057] In the embodiment of the present invention, firstly, a high-resolution electronic monitoring device is used to collect the whole-body dynamic image of the livestock in the state of being grasped by the manipulator in real time, so as to ensure that the video data can clearly record the behavioral characteristics of the livestock, including small body movements, facial expressions and muscle tension. The collected video is transmitted to the signal processing module through the cache hardware. At the same time, the piezoelectric sensor is installed on the contact surface of the grasping part of the manipulator to monitor the change of the livestock's escape force in real time. The piezoelectric sensor adopts the stress-charge conversion principle to convert the physical and mechanical signals generated by the livestock during the grasping process of the manipulator into electrical signals, and the signal sampling rate is set to 1000Hz to capture the high-frequency details in the livestock's movements. The collected electrical signal is amplified by a low-noise amplifier and converted into a digital signal by an analog-to-digital converter. The digital signal is then synchronized with the livestock's stress state video through a data integration module, and finally generates livestock stress state escape force monitoring data including the mechanical data of the escape action and the video behavior characteristics.

[0058] Step S2: matching the struggling action sequence breaking force fluctuations according to the breaking force monitoring data of the livestock under stress state to obtain the action sequence breaking force fluctuation data; quantifying the whole body subcutaneous tissue gap contraction difference based on the action sequence breaking force fluctuation data to obtain the subcutaneous tissue gap contraction difference data; estimating the livestock needle tube injection breakage risk according to the subcutaneous tissue gap contraction difference data to obtain the livestock needle tube injection breakage risk data;

[0059] In an embodiment of the present invention, by performing a time series analysis on the monitoring data of the breaking free force of livestock in a stress state, a wavelet transform method is used to perform a multi-scale decomposition of the mechanical data, and the high-frequency and low-frequency components of the action sequence are extracted, and the components are accurately aligned with the time series of the livestock's struggling action in the video to form the action sequence breaking free force fluctuation data. Thereafter, the fluctuation data is subjected to a quantitative analysis of the difference in contraction of the subcutaneous tissue gap, and based on the finite element analysis method, the deformation and recovery of the subcutaneous tissue caused by the external force are simulated to define the dynamic contraction amount of the subcutaneous tissue gap. Combined with the division of livestock body surface areas, the gap contraction values ​​of different areas are statistically analyzed, and the principal component analysis method is used to extract the regional difference characteristics to obtain the subcutaneous tissue gap contraction difference data. Based on the above-mentioned contraction difference data, the force and breakage risk value of the needle tube during the injection process are calculated through the mechanical distribution model, and the livestock needle tube injection breakage risk data is generated.

[0060] Step S3: matching the dynamic control force output of the manipulator according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the dynamic control force output data of the manipulator; regulating the injection behavior of the manipulator based on the dynamic control force output normalization data and the livestock needle injection breakage risk data to obtain the injection behavior regulation data of the manipulator;

[0061] In an embodiment of the present invention, the dynamic control force output matching algorithm of the manipulator is designed by using the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data. The algorithm is based on minimizing the local stress of livestock as the objective function, and uses the particle swarm optimization method to dynamically adjust the grasping force and force application path of the manipulator, so that the output control force of each stage matches the dynamic behavior characteristics of the livestock. The output dynamic control force data of the manipulator is normalized to ensure the adaptability of the data of livestock of different body sizes. The normalized dynamic control force data and the needle tube injection breakage risk data are input into the behavior control module, and the angle, speed and depth of the manipulator during injection are adjusted through the interval constraint optimization method to generate the manipulator injection behavior control data.

[0062] Step S4: Design the manipulator automation control firmware based on the manipulator dynamic control force output data and the manipulator injection behavior regulation data, obtain the manipulator automation control firmware, and send the manipulator automation control firmware to the cloud platform to execute the livestock grasping manipulator control method.

[0063] In an embodiment of the present invention, a manipulator automation control firmware is designed based on the manipulator dynamic control force output data and injection behavior regulation data. The specific implementation includes converting the dynamic control force output data and the injection behavior regulation data into control signals through a discrete control algorithm, and the control signals are stored in a firmware program in a matrix form. Through embedded programming, the matrix control signal is encoded into a microinstruction set and loaded into the manipulator control chip. During the firmware design process, hardware-in-the-loop testing (HIL) is used to verify the real-time and stability of the control firmware. After the verification is completed, the manipulator automation control firmware is uploaded to the cloud platform and sent to the livestock grasping manipulator equipment through the wireless communication module to achieve precise control of the manipulator and perform livestock grasping and injection tasks.

[0064] Step S1 includes the following steps:

[0065] Step S11: deploying electronic monitoring equipment and piezoelectric sensors on the manipulator;

[0066] Step S12: using electronic monitoring equipment to collect the stress state of the livestock in the state of being grasped by the robot arm, and obtaining a video of the stress state of the livestock;

[0067] Step S13: performing time-series frame processing on the livestock stress state video to obtain continuous frames of the livestock stress state;

[0068] Step S14: Continuously divide the frames based on the stress state of the livestock, and monitor the breaking force of the livestock in the grasping state of the manipulator through a piezoelectric sensor to obtain the breaking force monitoring data of the stress state of the livestock.

[0069] In the embodiment of the present invention, the hardware foundation of data acquisition is built by deploying electronic monitoring equipment and piezoelectric sensors on the manipulator. The electronic monitoring equipment uses a high-resolution industrial camera, uses a wide-angle lens to cover the livestock grabbing area, and uses an infrared filter to adapt to low light conditions. The industrial camera is connected to the end of the manipulator through a fixed bracket. The bracket design must meet the functions of shock absorption and multi-axis adjustment to ensure the stability and accuracy of image acquisition during the movement of the manipulator. The piezoelectric sensor is installed on the inner contact surface of the grasping part of the manipulator, and converts the livestock's escape force signal into an electrical signal output through the working principle of stress-charge conversion. The sensor is connected to the amplifier circuit through a dedicated signal line. The circuit includes a low-noise amplifier module and an analog-to-digital conversion module. The quantization resolution of the output signal is set to 16 bits, and the sampling frequency is 1000Hz to capture high-frequency dynamic force changes. The electronic monitoring equipment deployed in step S11 is used to collect stress state video of the livestock in the grasping state of the manipulator. In the workflow of video acquisition, the time synchronization of the video frame and the manipulator action is ensured by a triggered synchronous acquisition controller, and the trigger signal frequency is set to 200Hz according to the operation rhythm of the manipulator. The collected video data is transmitted to the data processing module via Gigabit Ethernet. In order to avoid data packet loss or delay during transmission, cache hardware is added to the transmission link, and the buffer capacity is set to 1 second of data volume. The collected video needs to include the dynamic behavior characteristics of the whole body of the livestock and cover the force application process of the manipulator during the grasping process to form a comprehensive analysis basis. The livestock stress state video obtained in step S12 is subjected to time-sequential frame processing. The video frame processing adopts the inter-frame difference method to extract the motion key frame by calculating the change rate of the pixels of adjacent frames. The implementation process of inter-frame difference is to decompose the video signal into continuous image frames, and calculate the difference of the pixel matrix of each frame image. The difference threshold is set to the mean square error of 10 pixel values, and the frame exceeding the threshold is marked as a key frame. After the key frame extraction is completed, the extracted frame sequence is time-sorted and marked according to the timestamp to form continuous frame data of the livestock stress state. The data processing module compresses and stores the divided image frames by region to ensure data integrity and storage efficiency. On the basis of step S13, the piezoelectric sensor is used to monitor the breaking force of livestock. The continuous frame data of the livestock stress state is synchronously matched with the force signal output by the sensor. The matching algorithm adopts the dynamic time warping (DTW) method to ensure the precise alignment of the video frame and the mechanical signal on the time axis. The synchronized data is feature extracted, and the spectral characteristics of the force signal are analyzed by Fourier transform. Combined with the livestock movement characteristics in the video, the force change pattern corresponding to the key action is determined. The obtained breaking force signal and the continuous frame data are integrated into the breaking force monitoring data of the livestock stress state. The data format is a multidimensional matrix, including dimensions such as time, space, mechanical characteristics and video behavior characteristics, providing basic data support for subsequent steps.

[0070] Step S2 includes the following steps:

[0071] Step S21: extracting the struggling motion trajectory of the livestock in the stress state from the video of the livestock in the stress state to obtain the struggling motion trajectory of the stress state;

[0072] Step S22: matching the struggling action time sequence breaking free force fluctuation of the struggling action trajectory of the stress state according to the breaking free force monitoring data of the livestock in the stress state, and obtaining the breaking free force fluctuation data of the action time sequence;

[0073] Step S23: analyzing the dynamic tension of livestock muscles based on the action sequence breaking free force fluctuation data and the struggling action trajectory in the stress state to obtain the dynamic tension data of livestock muscles;

[0074] Step S24: quantifying the subcutaneous tissue gap contraction difference of the whole body based on the livestock muscle dynamic tension data to obtain subcutaneous tissue gap contraction difference data;

[0075] Step S25: Estimate the risk of livestock syringe injection breakage based on the subcutaneous tissue gap contraction difference data and the action sequence breakage force fluctuation data to obtain livestock syringe injection breakage risk data.

[0076] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0077] Step S21: extracting the struggling motion trajectory of the livestock in the stress state from the video of the livestock in the stress state to obtain the struggling motion trajectory of the stress state;

[0078] In an embodiment of the present invention, in order to extract the struggling motion trajectory of the livestock in the stress state video, it is first necessary to decompose the video data into a frame sequence and extract continuous image frame data. Through computer vision technology, the optical flow method or the posture recognition algorithm based on deep learning is adopted to accurately locate the key points of the livestock in each frame of the image, and the key points include the positions of the limb joints, the head, the center of the trunk, etc. The three-dimensional struggling motion trajectory of the livestock is generated by using the changes in the motion trajectory of the key points frame by frame. In order to ensure the accuracy of the trajectory extraction, the extracted key point coordinates need to be interpolated to eliminate the problem of key frame loss caused by insufficient sampling frequency. Subsequently, the trajectory is smoothed in combination with the Kalman filter algorithm to remove the existing noise interference. Finally, the processed trajectory data is stored in the form of a time series to form a struggling motion trajectory of the stress state for subsequent analysis.

[0079] In another embodiment, for livestock stress state videos, an optical flow field calculation method based on a combination of frame difference method and background modeling is used to extract the trajectory of struggling movements in stress states. First, the video is preprocessed, including grayscale and image noise reduction (using a median filter with a size of 3x3 pixels). Then, the difference between consecutive frames is calculated. Specifically, the grayscale value of the current frame is subtracted from the previous frame at the pixel level to obtain a differential image. The differential image is binarized, and a grayscale threshold is set (for example, 15% of the maximum grayscale value). Pixels above the threshold are marked as motion pixels, and pixels below the threshold are marked as background pixels. Next, background modeling technology is used. A background model is constructed, and the initial frame of the background model is the average grayscale image of the first N frames (N is 10) of the video. Subsequently, the current frame is compared with the background model to update the background model. The specific update method is: if the grayscale difference between a pixel in the current frame and the corresponding pixel in the background model is less than a preset threshold (for example, 10% of the grayscale value of the background pixel), the grayscale value of the background model of the pixel is close to the grayscale value of the current frame pixel at a certain learning rate (for example, 0.05); if the difference is greater than the preset threshold, the grayscale value of the background model pixel is kept unchanged. After the background model is updated, the foreground image (that is, the image containing the moving object) is obtained by calculating the difference between the current frame and the updated background model. Finally, the optical flow method is applied to the foreground image. The dense optical flow algorithm is used, specifically the Lucas-Kanade algorithm. This algorithm calculates the displacement vector of each pixel in the foreground image between adjacent frames. The core of the algorithm is to assume that the optical flow remains constant in a small neighborhood (for example, 5x5 pixels). For each pixel, a linear equation system about the displacement vector is constructed, and the coefficients of the equation system are determined by the grayscale gradient in the pixel neighborhood. The least squares method is used to solve the equation system to obtain the horizontal and vertical displacements of each pixel. These displacement vectors are connected to form a pixel-level motion trajectory. In order to obtain the overall struggling motion trajectory of livestock, the motion trajectory at the pixel level is clustered. The density-based clustering algorithm (DBSCAN) is used. The algorithm classifies pixels that are close to each other (the distance is less than a preset value, such as 5 pixels) and have a sufficient number of neighboring points (the number of neighboring points is greater than a preset value, such as 3) into one category. After clustering, the class with the largest number of pixels is selected, and the centroid of the motion trajectory of all pixels in this class is calculated, and the centroid trajectory is used as the struggling motion trajectory of livestock in stress state.

[0080] Step S22: matching the struggling action time sequence breaking free force fluctuation of the struggling action trajectory of the stress state according to the breaking free force monitoring data of the livestock in the stress state, and obtaining the breaking free force fluctuation data of the action time sequence;

[0081] In the embodiment of the present invention, the livestock stress state breaking force monitoring data is time-synchronized with the stress state struggling action trajectory extracted in step S21, and the breaking force data is mapped to the corresponding action trajectory points to obtain the action sequence breaking force fluctuation data. First, determine the video frame rate (for example, 30 frames per second) and the data sampling rate of the piezoelectric sensor (for example, 100 times per second). According to the frame rate and the sampling rate, calculate the time stamps corresponding to each video frame and the time stamps corresponding to each breaking force data point. Then, perform time alignment. Since the sampling rates of the video and the breaking force data are different, interpolation processing is required. The linear interpolation method is used. For each action trajectory point (corresponding to a video frame), find the two closest breaking force data points before and after its time stamp. According to the time stamps and values of these two breaking force data points, calculate the estimated breaking force value at the moment corresponding to this action trajectory point. The specific calculation method is: Let the time stamp of the action trajectory point be t, the time stamps of the two adjacent breaking force data points be t1 and t2 (t1 < t < t2), and the corresponding breaking force values be F1 and F2. Then the estimated breaking force value F corresponding to this action trajectory point is F = F1 + (F2 - F1) × (t - t1) / (t2 - t1). Finally, associate each action trajectory point with its corresponding estimated breaking force value to form a data sequence, and this sequence is the action sequence breaking force fluctuation data. This data sequence takes time as the horizontal axis and breaking force as the vertical axis, clearly showing the change in the magnitude of the breaking force corresponding to each action during the entire struggling process.

[0082] Step S23: Based on the action sequence breaking force fluctuation data and the stress state struggling action trajectory, perform livestock muscle dynamic tension analysis to obtain livestock muscle dynamic tension data;

[0083] In the embodiment of the present invention, when performing livestock muscle dynamic tension analysis based on the action sequence breaking force fluctuation data and the stress state struggling action trajectory, first, perform Fourier transform analysis on the breaking force fluctuation data to decompose the mechanical fluctuation characteristics of each frequency component, and extract the main frequency components and their amplitude changes from them. Combining the struggling action trajectory data, calculate the changes in the muscle stretching and contraction amplitudes of the livestock at different action stages through the kinematic inversion method. The kinematic inversion method uses the velocity and acceleration information in the trajectory data, combined with the volume estimation parameters of the livestock and the muscle group distribution model, to calculate the dynamic force generation of different muscle groups. Then, use the finite element analysis method to solve the distribution of muscle tension to obtain the dynamic tension distribution data of the whole body muscles of the livestock during the struggling process. During this process, it is necessary to perform joint analysis on the muscle dynamic tension data and the breaking force fluctuation data to ensure that the dynamic tension distribution can accurately reflect the actual mechanical situation.

[0084] Step S24: quantifying the subcutaneous tissue gap contraction difference of the whole body based on the livestock muscle dynamic tension data to obtain subcutaneous tissue gap contraction difference data;

[0085] In an embodiment of the present invention, based on the livestock muscle dynamic tension data obtained in step S23, a subcutaneous tissue deformation model is constructed to quantify the contraction difference of the subcutaneous tissue gaps throughout the body. First, the surface of the livestock body is divided into multiple regions (for example, back, abdomen, limbs, etc., each region is further divided into smaller grids, and the grid size is, for example, 2cm x 2cm). Each grid corresponds to a subcutaneous tissue gap. Then, a subcutaneous tissue deformation model is established. The model is based on the principle of finite element analysis. The subcutaneous tissue is regarded as a continuum composed of multiple interconnected elastic units. The deformation of each elastic unit is affected by the tension of the surrounding muscles. The input of the model is the muscle tension data of each muscle group at each moment obtained in step S23, and the elastic modulus of the subcutaneous tissue (obtained from the empirical database, and the elastic modulus of the subcutaneous tissue in different parts is different). The key equation of the model is: deformation of the elastic unit = the inverse matrix of the stiffness matrix of the elastic unit multiplied by the force acting on the unit. Among them, the force acting on the unit is calculated by the tension of the surrounding muscles through an interpolation function (for example, bilinear interpolation). The stiffness matrix is ​​determined by the geometry and elastic modulus of the elastic unit. By solving the model, the deformation of the subcutaneous tissue gap corresponding to each grid at each moment is obtained. Finally, the subcutaneous tissue gap contraction difference is calculated. For each moment, the standard deviation of the subcutaneous tissue gap deformation of all grids is calculated as a quantitative indicator of the subcutaneous tissue gap contraction difference at that moment. The larger the standard deviation, the greater the difference in subcutaneous tissue gap contraction. The contraction difference index at each moment is associated with time to obtain the subcutaneous tissue gap contraction difference data.

[0086] Step S25: Estimate the risk of livestock syringe injection breakage based on the subcutaneous tissue gap contraction difference data and the action sequence breakage force fluctuation data to obtain livestock syringe injection breakage risk data.

[0087] In an embodiment of the present invention, the risk of fracture of livestock needle tube injection is estimated based on the difference data of subcutaneous tissue gap contraction and the fluctuation data of action sequence breaking force. First, it is necessary to establish a needle tube force model, which includes parameters such as rigid modulus, length, and diameter of the needle tube. According to the difference data of subcutaneous tissue gap contraction, the non-uniform stress distribution of the needle tube during the injection process is calculated. Combined with the breaking force fluctuation data, the total force value of the needle tube at different time points is calculated by the principle of force superposition. In order to analyze the torsional stress of the needle tube under vibration and bending, the maximum bending stress and shear stress distribution of the needle tube are calculated by the three-point bending mechanical analysis method. By comparing the bending stress with the yield strength of the needle tube material, it is determined whether the needle tube has the risk of fracture. At the same time, the dynamic stability parameters of the needle tube are calculated by combining the dynamic shaking frequency data of the needle tube with the peak mutation point of the livestock breaking force. Finally, all the analysis data are comprehensively evaluated and the risk data of fracture of livestock needle tube injection is output.

[0088] Step S23 includes the following steps:

[0089] Step S231: estimating the livestock volume based on the livestock stress state video to obtain livestock volume estimation data, wherein the livestock include pigs, cattle and sheep;

[0090] Step S232: deducing the force stretching amplitude intervals of different muscle groups of the livestock based on the action sequence breaking free force fluctuation data and the livestock volume estimation data, and obtaining the muscle force stretching amplitude intervals of different muscle groups of the livestock;

[0091] Step S233: performing tensile strength correlation coefficient analysis on the muscle force stretching amplitude interval according to the livestock volume estimation data to obtain the volume-related tensile strength correlation coefficient;

[0092] Step S234: performing spatial posture recognition on the struggling motion trajectory in the stress state to obtain struggling posture recognition data;

[0093] Step S235: Deducing the stress distribution of muscle fibers of the force-scheduling livestock between different struggling postures according to the struggling posture recognition data, and obtaining the stress distribution data of the force-scheduling muscles;

[0094] Step S236: Perform livestock muscle dynamic tension analysis based on the volume-related tensile strength correlation coefficient and the force scheduling muscle stress distribution data to obtain livestock muscle dynamic tension data.

[0095] In an embodiment of the present invention, livestock volume estimation is performed on livestock stress state videos. First, the video data needs to be decomposed into a frame sequence, and the image area containing livestock in each frame needs to be extracted. The livestock in the image is detected using a target detection algorithm in deep learning (such as YOLO or Faster R-CNN), and the contour area of ​​the livestock is marked. The three-dimensional volume of the livestock is estimated by using a multi-view geometry method in combination with the detected contour area and three-dimensional reconstruction technology. Specifically, during the video acquisition process, the monitoring device on the manipulator needs to take images of the livestock from different perspectives, and extract the three-dimensional structural information of the livestock based on the key point matching in the image. Then, the volume of the three-dimensional contour of the livestock is calculated using a volume integration method. The volume calculation process needs to be combined with the actual body shape characteristic parameters of the livestock, such as the average trunk ratio or average volume range of pigs, cattle and sheep, and the initial estimated volume data is mapped to the actual livestock volume range by establishing a standardized calibration formula, and finally the livestock volume estimation data is obtained. Based on the action sequence breakaway force fluctuation data and livestock volume estimation data, the force and stretching amplitude interval of different muscle groups of livestock are deduced. First, the action sequence breakaway force fluctuation data and livestock volume estimation data need to be synchronized to ensure that the two sets of data are aligned on the time axis. Then, the three-dimensional structure information in the livestock volume estimation data is used, and the livestock's biomechanical characteristic parameters (such as muscle fiber length, muscle elastic modulus, etc.) are combined to divide the livestock's muscle distribution into regions. The main muscle groups of livestock include back muscles, limb muscles and neck muscles. The muscle groups in each region have different force characteristics in the breakaway action. According to the breakaway force fluctuation data, the inverse dynamics analysis method is used to deduce the tensile force values ​​​​of different muscle groups in the breakaway action. Subsequently, combined with the livestock's volume data, the geometric deformation analysis method is used to calculate the range of stretching amplitude changes of each muscle group, and the force and stretching amplitude interval data of each muscle group are obtained. According to the livestock volume estimation data, the tensile strength correlation coefficient of the muscle force stretching amplitude interval is analyzed. By jointly modeling the livestock volume estimation data and the muscle force stretching amplitude interval data, the correlation between the muscle force stretching amplitude and the livestock volume is analyzed. Specifically, the livestock volume estimation data is first normalized to convert the livestock data of different volumes into a unified volume scale. Then, the linear correlation between the volume normalized data and the muscle force stretching amplitude value is analyzed using the Pearson correlation coefficient calculation method in statistics. By clustering the muscle force patterns of livestock of different volumes, the differences in muscle force characteristics corresponding to livestock of different volume ranges are identified. Finally, based on the calculated correlation coefficient value, a volume-associated tensile strength correlation coefficient model is further established to describe the intensity relationship between livestock volume and muscle force. To perform spatial posture recognition on the struggling action trajectory in stress state, it is first necessary to perform temporal segmentation on the struggling action trajectory data and decompose the continuous trajectory data into multiple independent action segments.Then, posture recognition is performed on each action segment using a posture classification algorithm based on machine learning (such as support vector machine or random forest). The training data of the posture classification model needs to contain a large amount of livestock video data labeled with different struggling action categories, such as head raising, kicking, twisting and other action categories. By extracting key feature points (such as speed, acceleration, trajectory curvature, etc.) in the trajectory data, these features are input into the classification model for posture prediction, and the specific posture category of the livestock in each time segment is output. At the same time, combined with the spatial displacement information in the three-dimensional trajectory data, the spatial position of the identified posture category is corrected to ensure that the posture recognition result is consistent with the spatial distribution of the actual action trajectory, and finally the struggling posture recognition data is output. According to the struggling posture recognition data, the muscle fiber stress distribution of the force scheduling of livestock between different struggling postures is deduced for the muscle force stretching amplitude interval. First, the struggling posture recognition data needs to be matched with the muscle force stretching amplitude interval data, so as to associate different struggling postures with the corresponding muscle stretching amplitude. By analyzing the muscle force characteristics under different postures, for example, the head-raising action mainly involves the force of the neck muscles, and the kicking action mainly involves the force of the limb muscles, the finite element analysis method based on the mechanical equilibrium equation is used to calculate the fiber stress distribution of muscle force under each posture. In the deduction process, it is necessary to combine the biomechanical properties of livestock muscle fibers (such as fiber direction, fiber density, etc.) to accurately model the force scheduling of each muscle area. Subsequently, a step-by-step iterative numerical calculation method is used to globally optimize the force stress distribution of different postures to ensure that the force stress distribution under each posture meets the actual movement requirements of livestock, and finally generate force scheduling muscle stress distribution data. Based on the volume-related tensile strength correlation coefficient and the force scheduling muscle stress distribution data, the dynamic tension analysis of livestock muscles is carried out. First, the volume-related tensile strength correlation coefficient data and the force scheduling muscle stress distribution data need to be jointly modeled to describe the dynamic change characteristics of muscle tension. By introducing the tensile strength correlation coefficient, the influencing factor of livestock volume is incorporated into the muscle tension analysis, and the muscle tension of livestock of different volumes is normalized. Then, using the fiber stress distribution results in the force scheduling muscle stress distribution data, combined with the struggling movement trajectory of the livestock, the Lagrangian mechanics analysis method is used to calculate the muscle tension change process of the livestock at each struggling movement stage. During the analysis process, it is necessary to dynamically adjust the boundary conditions of the model, such as the initial tension value of the muscle force, the time interval of the posture conversion, etc., so as to accurately describe the dynamic change law of muscle tension. Finally, the analysis results are output as muscle dynamic tension data in the form of a time series, which is used for the subsequent optimization of the manipulator control force and injection behavior regulation.

[0096] Another embodiment of step S232 is to deduce the force stretching range of different muscle groups of livestock based on the action sequence breaking free force fluctuation data obtained in step S22 and the livestock volume estimation data obtained in step S231, combined with anatomical knowledge. First, according to the type (pig, cattle or sheep) and volume of the livestock, the main muscle group information of the livestock is obtained from the anatomical database, including the starting point, insertion point, physiological cross-sectional area and muscle fiber type (fast muscle fiber or slow muscle fiber) of each muscle. Then, the action sequence breaking free force fluctuation data is decomposed into multiple time periods, each time period corresponding to a struggling action (for example, one leg kick, one head raising, etc.). The basis for decomposition is the peak and trough of the breaking free force curve. The peak corresponds to the force generation phase of the struggling action, and the trough corresponds to the relaxation phase of the struggling action. For each time period, the average breaking free force in the time period is calculated. Next, the muscle groups mainly involved in the force generation in each time period are determined. According to the type of struggling action (determined by the action trajectory extracted in step S21), combined with anatomical knowledge, it is determined which muscle groups are involved in the action. For example, the leg-pushing action mainly involves the quadriceps and gluteus maximus, and the head-raising action mainly involves the neck muscles and latissimus dorsi. For each muscle group that mainly participates in the force generation, calculate the torque generated by its force generation. The torque is equal to the breakaway force multiplied by the lever arm. The lever arm refers to the distance from the line of action of the force to the center of the joint, which can be calculated based on the bone length of the livestock (obtained from the anatomical database and scaled according to the livestock volume) and the joint angle (determined by the movement trajectory). Knowing the torque and the physiological cross-sectional area of ​​the muscle, the average tension of the muscle can be calculated. The average tension of the muscle is equal to the torque divided by the physiological cross-sectional area of ​​the muscle and the lever arm. Finally, the stretched length of the muscle is inferred based on the average tension of the muscle and the length-tension relationship curve of the muscle. The length-tension relationship curve of the muscle describes the tension of the muscle at different stretched lengths. According to the curve, the stretched length corresponding to the calculated average tension can be found. The stretched length is divided by the resting length of the muscle (obtained from the anatomical database) to obtain the stretch amplitude of the muscle. Taking into account the unevenness of muscle force, multiply the calculated stretching amplitude by a coefficient (for example, 0.8-1.2) to obtain the muscle force stretching amplitude range.

[0097] Step S25 includes the following steps:

[0098] Step S251: performing contraction non-uniform stress interpolation data analysis on the subcutaneous tissue gap contraction difference data to obtain contraction non-uniform stress interpolation data;

[0099] Step S252: performing breakaway force peak mutation identification on the breakaway force fluctuation data of the action sequence to obtain breakaway force peak mutation data;

[0100] Step S253: performing a shaking inertia needle tube torsional stress analysis according to the breaking force peak value mutation data and the contraction non-uniform stress interpolation data to obtain shaking inertia needle tube torsional stress data;

[0101] Step S254: performing shear stress cross-section yield distribution simulation on the shaking inertia needle tube torsional stress data to obtain shear stress cross-section yield distribution data;

[0102] Step S255: Estimate the risk of livestock syringe injection breakage based on the shaking inertia syringe torsional stress data and the shear stress cross-sectional yield distribution data to obtain livestock syringe injection breakage risk data.

[0103] In an embodiment of the present invention, the subcutaneous tissue gap contraction difference data is subjected to contraction non-uniform stress interpolation data analysis. First, the subcutaneous tissue gap contraction difference data needs to be regionalized, and the livestock body is divided into several key areas, such as the neck, back, chest and abdomen, and limbs. By analyzing the contraction amount of the subcutaneous tissue gap in each area, its dynamic change value at different time points is calculated. In order to describe the non-uniform stress distribution characteristics of the subcutaneous tissue gap in each part, the Lagrange interpolation method is used to interpolate the subcutaneous tissue gap contraction data. The Lagrange interpolation method maps discrete contraction difference data into a continuous non-uniform stress distribution curve by constructing a polynomial interpolation function. The interpolation calculation needs to combine parameters such as the elastic modulus, muscle density, and fat thickness of the subcutaneous tissue in each area to improve the accuracy of the interpolation result. After the interpolation is completed, the result is stored as continuous contraction non-uniform stress distribution data and output in the form of a time series for subsequent stress analysis. To identify the peak sudden change of the breaking force intensity of the action timing breaking force fluctuation data, it is first necessary to pre-process the breaking force fluctuation data to remove noise interference. The wavelet transform method is used to decompose the breakaway force data in the frequency domain, decompose the data into different frequency components, and retain the main frequency components related to the breakaway action. Subsequently, the mutation point of the force value change is identified by calculating the first-order derivative of the breakaway force fluctuation data. The force value of each mutation point is compared to extract the peak point of the force value in the time series. In order to determine whether the peak point belongs to a mutation phenomenon, it is necessary to calculate the force value difference before and after the peak point by setting a mutation threshold. Once the force value difference exceeds the set threshold, the peak point is marked as a mutation point. Finally, the time information and corresponding force value information of all mutation points are stored as the breakaway force peak mutation data. According to the breakaway force peak mutation data and the contraction non-uniform stress interpolation data, the shaking inertia needle tube torsion stress analysis is performed. First, the needle tube force model needs to be established. The needle tube force includes axial pressure, lateral shear force and torsional moment. By analyzing the breakaway force peak mutation data, the instantaneous maximum external force applied by livestock to the needle tube during the breakaway action is calculated. Combined with the contraction non-uniform stress interpolation data, the distribution of subcutaneous tissue resistance on the needle tube at different parts is calculated. Subsequently, the torque value received by the needle tube during the shaking process is calculated using the moment balance equation. The moment balance equation is: Torque = Force × Lever Arm Length, where Force is the peak external force, and Lever Arm Length is the offset distance caused by the shaking of the needle tube under the skin. Combined with the rigidity modulus and geometric size parameters of the needle tube, the finite element analysis method is used to simulate the force distribution of the needle tube to obtain the torsional stress distribution data of the needle tube under the action of shaking inertia. To simulate the shear stress section yield distribution of the shaking inertia needle tube torsional stress data, it is first necessary to analyze the material properties of the needle tube to obtain the yield strength and shear modulus parameters of the needle tube.Combined with the torsional stress data of the swaying inertia needle tube, the shear stress formula is used to calculate the shear stress values ​​of the needle tube on different sections. The shear stress formula is: shear stress = torque × section radius / polar moment of inertia, where the section radius and polar moment of inertia are determined by the geometric structure parameters of the needle tube. For the shear stress value on each section, it is compared with the yield strength of the needle tube material to determine whether the section has reached the yield state. By analyzing the cross-sectional stress distribution of the needle tube layer by layer, the position with the maximum shear stress is identified, and the shear stress cross-sectional yield distribution diagram of the needle tube is drawn. The simulation results need to be stored as time series data for subsequent risk assessment. According to the torsional stress data of the swaying inertia needle tube and the shear stress cross-sectional yield distribution data, the risk of injection fracture of livestock needle tubes is estimated. First, the torsional stress data of the swaying inertia needle tube and the shear stress cross-sectional yield distribution data need to be jointly analyzed. By calculating the total stress value of the needle tube at different time points, it is determined whether the needle tube has the possibility of yielding or breaking. The calculation formula of the total stress value is: total stress = torsional stress + shear stress, where the torsional stress and shear stress are derived from the above data analysis results respectively. The time point when the total stress exceeds the yield strength of the needle is marked as a potential break risk point. Subsequently, the fatigue strength of the needle under repeated stress is calculated by combining the dynamic shaking frequency of the needle and the repetitive frequency of the livestock's escape action. The calculation of fatigue strength is based on material fatigue theory, taking into account the yield limit and cyclic stress amplitude of the needle. Finally, the risk assessment results are output as injection break risk data, including the probability of breakage, time period, and corresponding stress distribution information, which is used to guide the dynamic control adjustment of the manipulator.

[0104] Step S253 includes the following steps:

[0105] Obtain the rigidity modulus and length value of the needle tube; obtain the center of gravity of the needle tube based on the length value of the needle tube;

[0106] According to the peak value mutation data of the breaking force, the force statistics of the needle tube shaking frequency are performed on the needle tube gravity center to obtain the needle tube shaking frequency associated with the peak value of the force;

[0107] Based on the peak mutation data of the breaking force, the interpolation data of the non-uniform stress of the contraction, and the peak-force-related needle tube shaking frequency, the force change of the bending arc of the needle tube rigid modulus is analyzed to obtain the force change data of the bending arc;

[0108] The bending fulcrum deformation limit stress analysis is performed on the bending arc force change data to obtain the bending fulcrum deformation limit stress;

[0109] Based on the bending fulcrum deformation limit stress, the fulcrum deformation limit frequency analysis is performed to obtain the fulcrum deformation limit frequency;

[0110] The torsional stress of the swaying inertia needle tube is analyzed according to the bending fulcrum deformation limit stress and the fulcrum deformation limit frequency, and the torsional stress data of the swaying inertia needle tube is obtained.

[0111] In an embodiment of the present invention, to obtain the rigid modulus and length value of the needle tube, it is first necessary to obtain the physical properties data table of the needle tube manufacturing material from the specification data provided by the manufacturer to determine the rigid modulus and length of the needle tube. The rigid modulus of the needle tube is a parameter that describes the deformation resistance of the needle tube material within the elastic range, and is usually determined by the Young's modulus of the manufacturing material of the needle tube (such as stainless steel or polymer material). The needle tube length value is directly measured by the geometric structure parameters of the needle tube. The length and rigid modulus of the needle tube are important basic data for subsequent calculations of the force characteristics of the needle tube. The center of gravity of the needle tube is obtained based on the needle tube length value, and the center of gravity position of the needle tube is determined by the geometric structure information of the needle tube using the geometric center calculation formula. The needle tube is a slender cylindrical object, and its center of gravity is usually located at half the length of the needle tube. The specific calculation formula is: center of gravity position = needle tube length ÷ 2. The position of the center of gravity of the needle tube determines the torque distribution and shaking characteristics of the needle tube under the action of external force. According to the peak mutation data of the breaking force, the force statistics of the needle tube shaking frequency at the center of gravity of the needle tube are performed. First, the frequency domain analysis of the peak mutation data of the breaking force is required. The peak mutation data of the force is converted from the time domain to the frequency domain by the fast Fourier transform (FFT) method, and the main frequency components of the livestock breaking free action are extracted. Combined with the position of the center of gravity of the needle tube, the shaking frequency of the needle tube under the action of external forces of different frequencies is calculated. The force statistics formula of the shaking frequency of the needle tube is: shaking frequency = external force frequency × torque ÷ needle tube rigid modulus, where torque = external force × center of gravity distance. The statistical results are output in the form of time series to obtain the force peak-related needle tube shaking frequency. Based on the breaking force peak mutation data, the contraction non-uniform stress interpolation data, and the force peak-related needle tube shaking frequency, the force change analysis of the needle tube rigid modulus of the needle tube shaking bending arc is performed. First, the breaking force peak mutation data and the contraction non-uniform stress interpolation data need to be time-series aligned to ensure the consistency of the two sets of data in the time dimension. Then, the change in the bending arc generated by the needle tube during the shaking process is calculated in combination with the shaking frequency of the needle tube. The analysis of the force change of the bending arc is based on the bending mechanics formula: bending arc = external force × center of gravity distance ÷ (needle tube rigid modulus × section inertia moment), where the section inertia moment is determined by the geometric shape of the needle tube. By calculating the change value of the bending arc frame by frame, the bending arc force change data of the needle tube at different time points during the shaking process are obtained. To perform the bending fulcrum deformation limit stress analysis on the bending arc force change data, it is first necessary to determine the position of the bending fulcrum formed by the needle tube during the shaking process. The bending fulcrum is usually located in a fixed area where the needle tube contacts the subcutaneous tissue. Combined with the length and shaking amplitude of the needle tube, the fulcrum position is located by geometric analysis method. Then, the deformation limit stress at the fulcrum is calculated using the bending mechanics formula. The calculation formula for the deformation limit stress is: deformation limit stress = bending moment ÷ section inertia moment, where the bending moment is determined by the external force and the center of gravity position of the needle tube. By analyzing the stress distribution of the bending fulcrum frame by frame, the fulcrum deformation limit stress data is finally obtained.To analyze the deformation limit frequency of the fulcrum based on the deformation limit stress of the bending fulcrum, it is first necessary to count the time points when the deformation limit stress value at the fulcrum exceeds the yield strength of the needle tube. By analyzing the distribution law of these time points, the occurrence frequency of the deformation limit stress at the fulcrum is calculated. The calculation formula of the deformation limit frequency of the fulcrum is: frequency = total time exceeding the yield strength ÷ total analysis time. The frequency data is used to evaluate the fatigue strength of the needle tube under long-term repeated force and predict the risk of permanent deformation or fracture of the needle tube. To analyze the torsional stress of the shaking inertia needle tube based on the deformation limit stress of the bending fulcrum and the deformation limit frequency of the fulcrum, it is first necessary to jointly model the deformation limit stress data of the bending fulcrum and the deformation limit frequency data of the fulcrum. By superimposing the cumulative fatigue effect of the bending stress at the fulcrum and the repetition frequency on the needle tube material, the torsional stress distribution of the needle tube under the shaking inertia is calculated. The calculation formula of torsional stress is: torsional stress = torque ÷ polar moment of inertia, where the torque is determined by the bending moment and the shaking frequency. Combining the geometric parameters and material properties of the needle tube, the finite element analysis method is used to simulate the torsional stress distribution of the needle tube at different time points. Finally, the torsional stress data of the shaking inertia needle tube is obtained for subsequent risk assessment and adjustment of the manipulator control strategy.

[0112] Step S3 includes the following steps:

[0113] Step S31: matching the dynamic control force output of the manipulator according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the dynamic control force output data of the manipulator;

[0114] Step S32: normalizing the dynamic control force output data of the manipulator to obtain dynamic control force output normalized data;

[0115] Step S33: Based on the dynamic control force output normalization data, the subcutaneous tissue gap contraction difference data and the livestock needle injection breakage risk data, the manipulator injection behavior regulation is performed to obtain the manipulator injection behavior regulation data.

[0116] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0117] Step S31: matching the dynamic control force output of the manipulator according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the dynamic control force output data of the manipulator;

[0118] In an embodiment of the present invention, the dynamic control force output of the manipulator is matched according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data. First, the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data need to be synchronized on the time axis to ensure that the two sets of data are analyzed in the same time dimension. For the action sequence breakaway force fluctuation data, the mechanical peak value and fluctuation amplitude at different time points in the livestock breakaway process are extracted; for the subcutaneous tissue gap contraction difference data, the tissue deformation and dynamic distribution characteristics of different parts are calculated. By establishing a mechanical coupling model, the breakaway force fluctuation data are associated with the subcutaneous tissue contraction difference data, and the comprehensive reaction force of the livestock at each time point is calculated. Using the reaction force data and the mechanical characteristic parameters of the manipulator (such as the rigid modulus, clamping area, fulcrum position, etc. of the manipulator arm), the inverse dynamics method is used to calculate the output force requirements of the manipulator at different time points. The inverse dynamics method obtains the dynamic control force required by the manipulator to deal with livestock breakaway and subcutaneous tissue changes by solving the motion equation of the manipulator. Subsequently, based on the livestock's escape trajectory and the structural characteristics of the manipulator, a distributed optimization algorithm was used to perform multi-point matching of the control force to ensure that the manipulator's force output in different areas met the dynamic changes of the livestock, and finally the manipulator's dynamic control force output data was obtained.

[0119] Step S32: normalizing the dynamic control force output data of the manipulator to obtain dynamic control force output normalized data;

[0120] In an embodiment of the present invention, the dynamic control force output data of the manipulator is normalized. First, it is necessary to perform statistical analysis on the dynamic control force output data to extract characteristic parameters such as the maximum value, minimum value and average value of the data. According to these characteristic parameters, the control force output data is converted into a dimensionless data form using a normalization formula. The purpose of the normalization process is to limit the amplitude of the dynamic control force output data to a fixed numerical range, such as between 0 and 1, so as to facilitate subsequent analysis and calculation. During the normalization process, it is necessary to perform separate processing for different control force output dimensions, such as the vertical force output, horizontal force output and rotational torque output of the manipulator, and normalize them separately. In order to ensure the accuracy and consistency of the normalization results, it is necessary to verify the normalized data to verify whether it is consistent with the distribution characteristics of the original data. Finally, the processed dynamic control force output normalized data is stored in a standardized form for subsequent injection behavior regulation.

[0121] Step S33: Based on the dynamic control force output normalization data, the subcutaneous tissue gap contraction difference data and the livestock needle injection breakage risk data, the manipulator injection behavior regulation is performed to obtain the manipulator injection behavior regulation data.

[0122] In the embodiment of the present invention, the injection behavior of the manipulator is regulated based on the normalized data of dynamic control force output, the difference data of subcutaneous tissue gap contraction, and the risk data of injection fracture of livestock needle tube. First, the three types of data need to be jointly modeled to describe the dynamic change characteristics of mechanical conditions during the injection process of the manipulator. The normalized data of dynamic control force output provides the real-time mechanical output state of the manipulator during the injection process, the difference data of subcutaneous tissue gap contraction reflects the dynamic resistance distribution of livestock tissue to the needle tube, and the risk data of injection fracture of the needle tube is used to evaluate the safety of the needle tube under complex force conditions. By establishing a multivariable control model, the force output of the manipulator is combined with the tissue change characteristics of the livestock and the risk assessment results of the needle tube to calculate the optimal control parameters of the manipulator during the injection process. The multivariable control model adopts a linear weighted algorithm to assign different weights to each control variable. For example, the weight of the risk data of needle tube fracture is set higher to give priority to the safety of the needle tube during the injection process, while considering the non-uniform contraction of the subcutaneous tissue and the dynamic output capacity of the manipulator. Based on the calculated optimal control parameters, the injection force, injection angle and injection speed of the manipulator are adjusted in real time. Specifically, the injection force is achieved by controlling the output force amplitude of the manipulator, the injection angle is adjusted by the manipulator's rotary drive unit, and the injection speed is achieved by controlling the manipulator's needle insertion rate. All adjustment parameters are fed back and optimized in real time through a closed-loop control system to ensure that the manipulator's injection behavior meets the dynamic changes of livestock, and finally output the manipulator's injection behavior control data.

[0123] In another embodiment, the injection force of the manipulator is matched based on the obtained dynamic control force output normalized data and the obtained livestock needle tube injection fracture risk data. The higher the risk value of the livestock needle tube injection fracture risk data, the easier the needle tube is to break. The injection force F_inj of the manipulator is proportional to the dynamic control force output normalized data and inversely proportional to the needle tube fracture risk data. The manipulator injection angle is selected based on the obtained subcutaneous tissue gap contraction difference data and the obtained livestock needle tube injection fracture risk data. The more intense the subcutaneous tissue contraction area is, the more tense the muscle is, the greater the resistance to needle tube insertion is, and the higher the risk of fracture is. Therefore, it is necessary to select an area with relatively small contraction for injection. The average contraction amount S_avg of each area is calculated. The area with the smallest average contraction amount S_avg is selected as the injection area. In the injection area, the injection angle is further adjusted according to the needle tube fracture risk data. If the risk is higher, a smaller injection angle (closer to parallel with the skin surface) should be selected to reduce the bending stress on the needle tube. If the risk is lower, a larger injection angle can be selected. The injection angle θ is inversely proportional to the risk data, and the robot injection angle data is obtained. Based on the robot injection force data and the robot injection angle data, the robot injection behavior is regulated. The injection force and injection angle are used as control outputs. The injection force and injection angle data are arranged in chronological order to obtain the robot injection behavior regulation data. This data includes the changes in injection force and angle over time during the entire injection process.

[0124] Step S31 includes the following steps:

[0125] Step S311: Draw a strength time series fluctuation curve for the action time series breakaway force fluctuation data to obtain a breakaway force strength time series fluctuation curve;

[0126] Step S312: calculating the instantaneous strength rise smoothness of the break-away force strength time series fluctuation curve to obtain the instantaneous strength rise smoothness of the break-away force;

[0127] Step S313: performing strength multi-peak area pattern recognition on the strength time series fluctuation curve of the breaking force according to the rising smoothness of the instantaneous strength of the breaking force, and obtaining the strength multi-peak area pattern of the breaking force;

[0128] Step S314: performing dynamic matching of the manipulator pressing control force based on the multi-peak area mode of the strength of the break-away force to obtain the manipulator pressing dynamic control data;

[0129] Step S315: matching the manipulator control area according to the subcutaneous tissue gap contraction difference data, and optimizing the multi-point control force of the manipulator pressing dynamic control data to obtain multi-point area control force data;

[0130] Step S316: matching the dynamic control force output of the manipulator according to the dynamic control data of the manipulator's pressing and the multi-point area control force data to obtain the dynamic control force output data of the manipulator.

[0131] In an embodiment of the present invention, the strength time series fluctuation curve of the action sequence breaking free force fluctuation data is drawn. First, it is necessary to extract force value data from the breaking free force monitoring data collected by the piezoelectric sensor, and generate discrete time series force value points in chronological order. By interpolating these time series points, the discrete force value points are converted into a continuous force value curve. The interpolation algorithm can use the cubic spline interpolation method to ensure the smoothness and accuracy of the curve. During the drawing process, the force value data needs to be normalized to ensure that the curve can reflect the real relative strength change. Subsequently, the time axis is set as the horizontal axis, and the normalized breaking free force value is used as the vertical axis to generate a complete breaking free force strength time series fluctuation curve. After the drawing is completed, the key characteristic points of the curve are marked, such as peak points, valley points and turning points, for subsequent mechanical property analysis. To calculate the instantaneous strength rise smoothness of the breaking free force strength time series fluctuation curve, it is first necessary to calculate the derivative value of the fluctuation curve at each time point to obtain the instantaneous rate of change of the force value. The derivative value is calculated by numerical differentiation method, such as difference method, by dividing the difference of adjacent force value points by the corresponding time interval to obtain the instantaneous rate of rise at each time point. In order to evaluate the smoothness of the curve, it is necessary to calculate the change of the instantaneous rate of rise, that is, the second-order difference value of the derivative. The calculation formula of the smoothness of the instantaneous strength rise is: smoothness = 1-(variance of instantaneous change ÷ mean of instantaneous change), where the variance of instantaneous change reflects the unevenness of the curve change. After the calculation is completed, the smoothness result is stored as time series data for subsequent pattern recognition and dynamic force matching. According to the smoothness of the instantaneous strength rise of the breakaway force, the strength time series fluctuation curve of the breakaway force is subjected to the multi-peak area pattern recognition of the strength. First, the fluctuation curve needs to be segmented and the entire time series curve is divided into multiple continuous force value areas. The segmentation is based on the turning point of the curve, that is, the position where the derivative value is zero. For each segmented area, its peak point and the corresponding force value amplitude are calculated, and the area where the force value exceeds the set threshold is identified as the multi-peak area. Subsequently, combined with the instantaneous strength rise smoothness data, the change characteristics of each multi-peak area are analyzed, such as the distribution density of peak points and the smoothness of the rise rate. Through cluster analysis algorithms, such as K-means clustering, multiple peak areas are classified, force value areas with similar fluctuation patterns are identified, and finally a multi-peak area pattern of the strength of the escape force is generated. Based on the multi-peak area pattern of the strength of the escape force, the dynamic matching of the manipulator's pressing control force is performed. First, the multi-peak area pattern needs to be associated with the mechanical output parameters of the manipulator. The multi-peak area pattern provides the force distribution characteristics of different stages in the livestock's escape process. The mechanical output parameters of the manipulator include the output force range, reaction time, and adjustment frequency. By establishing a dynamic matching model, the control force value that the manipulator needs to apply in each multi-peak area is calculated. The matching model adopts a weighted optimization algorithm, takes the force amplitude of the multi-peak area as the weight, and adjusts the output force amplitude and frequency of the manipulator in real time to achieve a dynamic response to the livestock's escape action.After the matching is completed, the dynamic control data of the manipulator pressing is generated, including the output force value and control frequency of each time period. The manipulator control area is matched according to the difference data of subcutaneous tissue gap contraction, and the multi-point control force optimization of the dynamic control data of the manipulator pressing is performed. First, it is necessary to perform spatial distribution analysis on the difference data of subcutaneous tissue gap contraction to identify the tissue gap change areas in different parts of the livestock body. For each area, the distribution of force values ​​that the manipulator needs to apply at different positions is calculated in combination with the pressing area and clamping structure of the manipulator. The control area matching adopts an area distribution algorithm to ensure that the pressing force can be evenly distributed in the key areas of the livestock body by adjusting the contact area and clamping force of the manipulator. Subsequently, the force value of each control point is optimized in combination with the dynamic control data of the manipulator pressing. The optimization algorithm is based on the principle of minimum energy consumption. By iteratively adjusting the force value distribution of the control point, multi-point area control force data is generated for the dynamic control of the manipulator. According to the dynamic control data of the manipulator's pressing and the multi-point area control force data, the dynamic control data of the pressing and the area control force data need to be synchronized on the time axis to ensure that the two sets of data are calculated in the same time dimension. The matching process adopts a multi-objective optimization algorithm, with the dynamic characteristics of the livestock's escape action and the output capacity of the manipulator as constraints. By adjusting the amplitude and frequency of the manipulator's force output, real-time response to the dynamic changes of livestock is achieved. The specific calculation process includes: calculating the output force amplitude of the manipulator according to the dynamic control data of the pressing, and calculating the output force distribution of the manipulator according to the area control force data. By superimposing the two sets of force values, the comprehensive control force output data of the manipulator at each time point is generated. The matching results are stored in a time series form as real-time control instructions for the manipulator to perform livestock grasping and auxiliary injection operations.

[0132] Step S33 includes the following steps:

[0133] Step S331: performing manipulator injection force matching based on the dynamic control force output normalization data and the livestock needle injection breakage risk data to obtain manipulator injection force data;

[0134] Step S332: selecting the manipulator injection angle according to the subcutaneous tissue gap contraction difference data and the livestock needle injection breakage risk data to obtain the manipulator injection angle data;

[0135] Step S333: regulating the injection behavior of the manipulator based on the injection force data and the injection angle data of the manipulator to obtain injection behavior regulation data of the manipulator.

[0136] In an embodiment of the present invention, the injection force of the manipulator is matched based on the normalized data of the dynamic control force output and the risk data of the injection fracture of the livestock needle tube, and the injection force data of the manipulator is obtained. During implementation, the normalized data of the dynamic control force output is first decoded, the normalized control force output value of the manipulator at different time points is extracted, and the denormalization process is used to convert it into the force value of the actual unit. The risk data of the injection fracture of the livestock needle tube is used to set the upper limit threshold of the force value, and by comparing the actual force value with the threshold, it is judged whether there is a potential risk in the current injection force. The matching process adopts the piecewise linear interpolation method, and the slope and offset of the force output curve are adjusted according to the risk threshold, so that the output force value conforms to the change trend within the safe range. Finally, the matched force data is smoothed by cubic spline interpolation to generate the injection force data of the manipulator. The data is stored in the form of a time series, including the time point, the injection force and the corresponding safety assessment. The manipulator injection angle is selected according to the difference data of the subcutaneous tissue gap contraction and the risk data of the injection fracture of the livestock needle tube, and the manipulator injection angle data is obtained. The subcutaneous tissue gap contraction difference data provides the dynamic geometric changes of the subcutaneous tissue at the injection site, and the appropriate injection angle is determined in combination with the force characteristics in the livestock needle tube injection fracture risk data. The injection angle selection is based on geometric analysis. The three-dimensional coordinate point cloud model is first constructed using the subcutaneous tissue contraction data, and the surface mesh of the tissue gap is generated by Delaunay triangulation. The optimal injection direction is calculated based on the angle between the normal vector of each triangle in the mesh and the manipulator direction vector. At the same time, combined with the mechanical information in the fracture risk data, the stress distribution under different injection angles is analyzed, and the gradient descent method is used to optimize the angle selection to ensure that the injection direction minimizes the risk of needle tube fracture when the subcutaneous tissue is uniformly stressed. Finally, the injection angle data is stored in the form of polar angle and azimuth angle in the manipulator coordinate system to form the manipulator injection angle data. The injection behavior of the manipulator is regulated based on the injection force data and the injection angle data of the manipulator to obtain the injection behavior regulation data of the manipulator. The injection behavior regulation includes two main aspects: force output and angle adjustment. The precise operation of the manipulator in a dynamic environment is achieved through a real-time control algorithm. The force output part adopts a proportional integral differential controller (PID controller), takes the injection force data of the manipulator as the set value, monitors the deviation between the actual force and the set value during the injection process of the manipulator in real time, and adjusts the actuator current of the manipulator to achieve precise force control. The angle adjustment part drives the rotation module of the injection part of the manipulator through a stepper motor, adjusts the injection angle according to the polar angle and azimuth in the injection angle data of the manipulator, and uses the stepwise approximation method to adjust the injection angle, and uses the angle encoder to feedback and correct each adjustment to ensure that the angle error is less than the predetermined threshold. The entire control process records the force, angle and adjustment status of the manipulator at different time points to form the injection behavior control data of the manipulator, which is output in the form of a table or serialized file for subsequent injection behavior analysis and optimization.

[0137] The present invention also provides a livestock grabbing manipulator control system, which is used to execute the livestock grabbing manipulator control method as described above, and the livestock grabbing manipulator control system comprises:

[0138] The breaking force monitoring module is used to collect the stress state of the livestock in the state of being grasped by the manipulator through the electronic monitoring equipment to obtain the livestock stress state video; based on the livestock stress state video and using the piezoelectric sensor to monitor the breaking force of the livestock in the state of being grasped by the manipulator, to obtain the livestock stress state breaking force monitoring data;

[0139] The injection fracture risk estimation module is used to match the struggling action sequence breaking force fluctuation according to the breaking force monitoring data of the livestock under stress state, and obtain the action sequence breaking force fluctuation data; quantify the whole body subcutaneous tissue gap contraction difference based on the action sequence breaking force fluctuation data, and obtain the subcutaneous tissue gap contraction difference data; estimate the livestock needle tube injection fracture risk according to the subcutaneous tissue gap contraction difference data, and obtain the livestock needle tube injection fracture risk data;

[0140] The manipulator injection behavior control module is used to match the manipulator dynamic control force output according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the manipulator dynamic control force output data; the manipulator injection behavior is controlled based on the dynamic control force output normalization data and the livestock needle injection breakage risk data to obtain the manipulator injection behavior control data;

[0141] The automation control firmware design module is used to design the automation control firmware of the manipulator based on the dynamic control force output data of the manipulator and the injection behavior regulation data of the manipulator, obtain the automation control firmware of the manipulator, and send the automation control firmware of the manipulator to the cloud platform to execute the livestock grasping manipulator control method.

[0142] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0143] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A livestock grabbing manipulator control method, characterized in that: The following steps are involved: Step S1: The stress state of the livestock in the gripping state of the robot is collected by electronic monitoring equipment to obtain a video of the stress state of the livestock; Based on the livestock stress state video and using piezoelectric sensors to monitor the livestock's escape force when the robot is grasping the livestock, the livestock's escape force monitoring data under stress state is obtained; Step S2: matching the struggling action sequence breaking free force fluctuations according to the breaking free force monitoring data of the livestock under stress state to obtain the breaking free force fluctuation data of the action sequence; Based on the action sequence breakaway force fluctuation data, the subcutaneous tissue gap contraction difference of the whole body is quantified to obtain the subcutaneous tissue gap contraction difference data; The risk of broken needles for livestock injections is estimated based on the difference in subcutaneous tissue space contraction data, and the risk data of broken needles for livestock injections are obtained; Step S3: matching the dynamic control force output of the manipulator according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the dynamic control force output data of the manipulator; Based on the normalized data of dynamic control force output and the risk data of livestock needle tube breakage, the injection behavior of the manipulator is regulated to obtain the injection behavior regulation data of the manipulator; Step S4: Design the manipulator automation control firmware based on the manipulator dynamic control force output data and the manipulator injection behavior regulation data, obtain the manipulator automation control firmware, and send the manipulator automation control firmware to the cloud platform to execute the livestock grasping manipulator control method.

2. The livestock grabbing manipulator control method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: deploying electronic monitoring equipment and piezoelectric sensors on the manipulator; Step S12: using electronic monitoring equipment to collect the stress state of the livestock in the state of being grasped by the robot arm, and obtaining a video of the stress state of the livestock; Step S13: performing time-series frame processing on the livestock stress state video to obtain continuous frames of the livestock stress state; Step S14: Continuously divide the frames based on the stress state of the livestock, and monitor the breaking force of the livestock in the grasping state of the manipulator through a piezoelectric sensor to obtain the breaking force monitoring data of the stress state of the livestock.

3. The livestock grabbing manipulator control method according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: extracting the struggling motion trajectory of the livestock in the stress state from the video of the livestock in the stress state to obtain the struggling motion trajectory of the stress state; Step S22: matching the struggling action time sequence breaking free force fluctuation of the struggling action trajectory of the stress state according to the breaking free force monitoring data of the livestock in the stress state, and obtaining the breaking free force fluctuation data of the action time sequence; Step S23: analyzing the dynamic tension of livestock muscles based on the action sequence breaking free force fluctuation data and the struggling action trajectory in the stress state to obtain the dynamic tension data of livestock muscles; Step S24: quantifying the subcutaneous tissue gap contraction difference of the whole body based on the livestock muscle dynamic tension data to obtain subcutaneous tissue gap contraction difference data; Step S25: Estimate the risk of livestock syringe injection breakage based on the subcutaneous tissue gap contraction difference data and the action sequence breakage force fluctuation data to obtain livestock syringe injection breakage risk data.

4. The livestock grabbing manipulator control method according to claim 3, characterized in that: Step S23 includes the following steps: Step S231: estimating the livestock volume based on the livestock stress state video to obtain livestock volume estimation data, wherein the livestock include pigs, cattle and sheep; Step S232: deducing the force stretching amplitude intervals of different muscle groups of the livestock based on the action sequence breaking free force fluctuation data and the livestock volume estimation data, and obtaining the muscle force stretching amplitude intervals of different muscle groups of the livestock; Step S233: performing tensile strength correlation coefficient analysis on the muscle force stretching amplitude interval according to the livestock volume estimation data to obtain the volume-related tensile strength correlation coefficient; Step S234: performing spatial posture recognition on the struggling motion trajectory in the stress state to obtain struggling posture recognition data; Step S235: Deducing the stress distribution of muscle fibers of the force-scheduling livestock between different struggling postures according to the struggling posture recognition data, and obtaining the stress distribution data of the force-scheduling muscles; Step S236: Perform livestock muscle dynamic tension analysis based on the volume-related tensile strength correlation coefficient and the force scheduling muscle stress distribution data to obtain livestock muscle dynamic tension data.

5. The livestock grabbing manipulator control method according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: matching the dynamic control force output of the manipulator according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the dynamic control force output data of the manipulator; Step S32: normalizing the dynamic control force output data of the manipulator to obtain dynamic control force output normalized data; Step S33: Based on the dynamic control force output normalization data, the subcutaneous tissue gap contraction difference data and the livestock needle injection breakage risk data, the manipulator injection behavior regulation is performed to obtain the manipulator injection behavior regulation data.

6. The livestock grabbing manipulator control method according to claim 5, characterized in that: Step S31 includes the following steps: Step S311: Draw a strength time series fluctuation curve for the action time series breakaway force fluctuation data to obtain a breakaway force strength time series fluctuation curve; Step S312: calculating the instantaneous strength rise smoothness of the break-away force strength time series fluctuation curve to obtain the instantaneous strength rise smoothness of the break-away force; Step S313: performing strength multi-peak area pattern recognition on the strength time series fluctuation curve of the breaking force according to the rising smoothness of the instantaneous strength of the breaking force, and obtaining the strength multi-peak area pattern of the breaking force; Step S314: performing dynamic matching of the manipulator pressing control force based on the multi-peak area mode of the strength of the break-away force to obtain the manipulator pressing dynamic control data; Step S315: matching the manipulator control area according to the subcutaneous tissue gap contraction difference data, and optimizing the multi-point control force of the manipulator pressing dynamic control data to obtain multi-point area control force data; Step S316: matching the dynamic control force output of the manipulator according to the dynamic control data of the manipulator's pressing and the multi-point area control force data to obtain the dynamic control force output data of the manipulator.

7. A livestock grabbing manipulator control system, characterized in that: Used to execute the livestock grasping manipulator control method as claimed in claim 1, the livestock grasping manipulator control system comprises: The breaking force monitoring module is used to collect the stress state of the livestock in the state of being grasped by the manipulator through the electronic monitoring equipment to obtain the livestock stress state video; based on the livestock stress state video and using the piezoelectric sensor to monitor the breaking force of the livestock in the state of being grasped by the manipulator, to obtain the livestock stress state breaking force monitoring data; The injection fracture risk estimation module is used to match the struggling action sequence breaking force fluctuation according to the breaking force monitoring data of the livestock under stress state, and obtain the action sequence breaking force fluctuation data; quantify the whole body subcutaneous tissue gap contraction difference based on the action sequence breaking force fluctuation data, and obtain the subcutaneous tissue gap contraction difference data; estimate the livestock needle tube injection fracture risk according to the subcutaneous tissue gap contraction difference data, and obtain the livestock needle tube injection fracture risk data; The manipulator injection behavior control module is used to match the manipulator dynamic control force output according to the action sequence breakaway force fluctuation data and the subcutaneous tissue gap contraction difference data to obtain the manipulator dynamic control force output data; the manipulator injection behavior is controlled based on the dynamic control force output normalization data and the livestock needle injection breakage risk data to obtain the manipulator injection behavior control data; The automation control firmware design module is used to design the automation control firmware of the manipulator based on the dynamic control force output data of the manipulator and the injection behavior regulation data of the manipulator, obtain the automation control firmware of the manipulator, and send the automation control firmware of the manipulator to the cloud platform to execute the livestock grasping manipulator control method.

Images