Mechanical arm control method and system for livestock state monitoring

Abstract

The invention relates to the technical field of mechanical arm control, in particular to a mechanical arm control method and system for livestock state monitoring. The method comprises the following steps that an electronic monitoring device is used for obtaining livestock activity images, the livestock type activity distribution density is calculated, then mechanical arm vibration quantification is carried out based on activity density data, and the stress state probability of livestock is estimated; thirdly, the approaching speed of the mechanical arm is adjusted according to the stress state probability, and a monitoring posture control instruction is designed to ensure that the action of the mechanical arm is matched with the behavior of the livestock; and finally, an automatic control program is designed and embedded into a mechanical arm control system by encoding and processing a monitoring posture adjusting instruction, and accurate livestock state monitoring and control are achieved. The mechanical arm control technology is optimized, so that the mechanical arm control technology is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of mechanical arm control, and in particular to a mechanical arm control method and system for livestock status monitoring. Background Art

[0002] With the gradual modernization and intelligence of agricultural production methods, livestock status monitoring has become an indispensable part of modern agricultural management. Traditional livestock management relies on manual inspections and experience judgments, which is not only inefficient, but also easily affected by human factors, making it difficult to achieve accurate and efficient management. Especially in large-scale breeding environments, the lack of manual management will lead to delayed discovery of livestock health problems, which in turn affects breeding benefits and animal welfare. In order to solve this problem, scientific and technological personnel began to study livestock status monitoring systems based on advanced technologies such as sensing technology, data analysis and automatic control, and gradually realized real-time, comprehensive and accurate monitoring of livestock. As a highly flexible and precise automation equipment, robotic arms have been widely used in manufacturing, medical care, logistics and other fields. In recent years, with the rapid development of artificial intelligence technology, robotic arms have been introduced into the agricultural field for monitoring and management of livestock status. However, the traditional robotic arm control method has the problem that the robotic arm is not adaptable enough to environmental interference and cannot accurately control the robotic arm. Summary of the invention

[0003] Based on this, it is necessary to provide a method and system for controlling a robotic arm for livestock status monitoring to solve at least one of the above technical problems.

[0004] To achieve the above object, a method for controlling a robotic arm for monitoring livestock status is provided, the method comprising the following steps: Step S1: monitoring livestock activities in a livestock monitoring area through electronic monitoring equipment to obtain livestock activity monitoring images; calculating the distribution density of livestock activity monitoring images to obtain livestock activity distribution density; Step S2: quantifying the vibration of the approaching robot arm observation based on the activity distribution density of the livestock types to obtain the quantified data of the approaching robot arm vibration; estimating the probability of the livestock stress state based on the quantified data of the approaching robot arm vibration to obtain the probability data of the livestock stress state; Step S3: performing robot arm approach rate matching according to the livestock stress state probability data to obtain robot arm approach rate matching data; performing robot arm monitoring posture control instruction design based on the robot arm approach rate matching data and the livestock activity monitoring image to obtain the robot arm monitoring posture adjustment control instruction; Step S4: Encode the robot arm monitoring posture adjustment control instruction to obtain the robot arm monitoring posture adjustment coding instruction; design an automated control program based on the robot arm monitoring posture adjustment coding instruction to obtain the robot arm monitoring posture control program, and embed the robot arm monitoring posture control program into the robot arm control center to execute the robot arm control method for livestock status monitoring.

[0005] Preferably, step S1 comprises the following steps: Step S11: obtaining a livestock monitoring area; deploying a cross track in the livestock monitoring area, the cross track and the robotic arm are electrically connected via a servo motor, the servo motor controls the robotic arm to move on the cross track, and an electronic monitoring device is deployed on the robotic arm; Step S12: monitoring livestock activities in the livestock monitoring area through electronic monitoring equipment to obtain livestock activity monitoring images; Step S13: marking the livestock types on the livestock activity monitoring image to obtain a livestock type activity monitoring image; wherein the livestock types include pigs, cattle and sheep; Step S14: Calculate the distribution density of the livestock activity monitoring images to obtain the livestock activity distribution density.

[0006] Preferably, step S2 comprises the following steps: Step S21: analyzing the movable positions of livestock on the livestock activity monitoring images based on the distribution density of livestock types and activities, and obtaining a livestock movable position data set; Step S22: predicting the livestock spatial movement trajectory based on the livestock movable position data set and the livestock activity monitoring image to obtain livestock spatial movement trajectory prediction data; Step S23: quantifying the vibration of the approaching robot arm based on the predicted data of the livestock spatial movement trajectory, and obtaining the quantified data of the vibration of the approaching robot arm; Step S24: Estimating the livestock stress state probability based on the robot arm approach vibration quantified data to obtain livestock stress state probability data.

[0007] Preferably, step S23 includes the following steps: Step S231: analyzing the livestock intermittent speed change of the livestock spatial movement trajectory prediction data to obtain the livestock intermittent speed change data; Step S232: performing an inertia simulation of the instantaneous pause when the robot arm approaches for observation according to the intermittent speed change data of the livestock, and obtaining the instantaneous pause inertia data; Step S233: performing a robot arm jitter vector attenuation analysis on the instantaneous pause inertia data to obtain a robot arm jitter attenuation vector; Step S234: calculating the periodic jitter vector difference of the robot arm jitter attenuation vector to obtain the periodic jitter vector difference; Step S235: Acquire the design structure data of the robot arm; quantify the close observation vibration of the robot arm according to the design structure data of the robot arm, the robot arm jitter attenuation vector and the periodic jitter vector difference to obtain the close observation vibration quantification data of the robot arm.

[0008] Preferably, step S24 comprises the following steps: Step S241: Obtain the noise decibel acceptable to livestock, wherein the noise decibel acceptable to livestock includes: the noise decibel acceptable to pigs, the noise decibel acceptable to sheep, and the noise decibel acceptable to cattle; Step S242: performing vibration noise feedback matching on the robot arm approach vibration quantization data to obtain vibration noise feedback data; Step S243: calculating the instantaneous noise spectrum band difference on the vibration noise feedback data to obtain the instantaneous noise spectrum band difference; Step S244: performing noise band extreme value distribution analysis according to the instantaneous noise spectrum band difference to obtain noise band extreme value distribution data; Step S245: performing an excessive sound pressure instantaneous aggregation analysis on the livestock acceptable noise decibel according to the instantaneous noise spectrum band difference and the noise band extreme value distribution data, and obtaining the excessive sound pressure instantaneous aggregation data; Step S246: estimating the livestock stress state probability based on the excess sound pressure instantaneous aggregation data for the livestock acceptable noise decibels, and obtaining livestock stress state probability data.

[0009] Preferably, step S3 comprises the following steps: Step S31: performing robot arm approach rate matching according to the robot arm approach vibration quantification data and livestock stress state probability data to obtain robot arm approach rate matching data; Step S32: performing a logic error iteration test on the robot arm approach rate matching data to obtain logic error iteration test data; Step S33: performing rate error correction on the robot arm approach rate matching data according to the logical error iteration test data to obtain the robot arm approach rate optimization data; Step S34: Designing a robot arm monitoring posture control instruction based on the robot arm approach rate optimization data and the livestock activity monitoring image to obtain a robot arm monitoring posture adjustment control instruction.

[0010] Preferably, step S33 includes the following steps: Step S331: performing multi-frequency error accumulation calculation on the logic error iteration test data to obtain multi-frequency error accumulation data; Step S332: performing interval maximum error analysis on the logic error iteration test data according to the multi-frequency error accumulation data to obtain multi-frequency maximum error interval data; Step S333: performing time series linear interpolation on the multi-frequency maximum error interval data to obtain error interval time series interpolation data; Step S334: performing rate error correction on the robot arm approach rate matching data according to the error interval timing interpolation data and the multi-frequency maximum error interval data to obtain the robot arm approach rate optimization data.

[0011] Preferably, step S34 includes the following steps: Step S341: performing livestock activity posture decomposition processing on the livestock activity monitoring image to obtain livestock activity posture decomposition data; Step S342: Performing dynamic analysis of the livestock visual field blind area on the livestock activity posture decomposition data to obtain dynamic data of the livestock visual field blind area; Step S343: performing robot arm path posture control adjustment on the livestock visual field blind spot dynamic data according to the robot arm approach rate optimization data to obtain robot arm path posture control adjustment data; Step S344: Designing a robot arm monitoring posture control instruction based on the robot arm path posture control adjustment data to obtain a robot arm monitoring posture adjustment control instruction.

[0012] Preferably, step S343 includes the following steps: Performing blind zone boundary surface fitting processing on the dynamic data of the blind zone of the livestock visual field to obtain blind zone boundary surface fitting data; Perform robot arm space coverage calculation on the blind area boundary surface fitting data to obtain robot arm space coverage data; Based on the space coverage data of the robot arm, the blind area boundary surface fitting data is processed for path point distribution retrieval to obtain path point distribution data; According to the robot arm approach rate optimization data, the path point distribution data is matched with the staged motion parameters to obtain the staged motion parameters of the robot arm; The robot arm path posture control adjustment is performed according to the robot arm's staged motion parameters and path point distribution data to obtain the robot arm path posture control adjustment data.

[0013] Preferably, the present invention further provides a livestock status monitoring robot arm control system, which is used to execute the livestock status monitoring robot arm control method as described above, and the livestock status monitoring robot arm control system comprises: The distribution density analysis module is used to monitor livestock activities in the livestock monitoring area through electronic monitoring equipment to obtain livestock activity monitoring images; calculate the distribution density of livestock activity monitoring images to obtain the distribution density of livestock types; A stress state probability estimation module is used to quantify the vibration of the approaching observation of the robot arm based on the activity distribution density of the livestock species to obtain the quantified data of the approaching vibration of the robot arm; and to estimate the probability of the livestock stress state based on the quantified data of the approaching vibration of the robot arm to obtain the probability data of the livestock stress state; The posture control instruction design module is used to match the approach rate of the robot arm according to the probability data of the livestock stress state, and obtain the approach rate matching data of the robot arm; the robot arm monitoring posture control instruction is designed based on the approach rate matching data of the robot arm and the livestock activity monitoring image, and the robot arm monitoring posture adjustment control instruction is obtained; The automatic control program design module is used to encode and process the robot arm monitoring posture adjustment control instructions to obtain the robot arm monitoring posture adjustment coding instructions; based on the robot arm monitoring posture adjustment coding instructions, the automatic control program is designed to obtain the robot arm monitoring posture control program, and the robot arm monitoring posture control program is embedded into the robot arm control center to execute the robot arm control method for livestock status monitoring.

[0014] The beneficial effect of the present invention is that the livestock monitoring area is monitored in real time by electronic monitoring equipment, and the activity images of the livestock can be accurately obtained, and the distribution density of the images can be calculated. This process enables the system to understand the activities of different types of livestock in different areas, and provide accurate basic data for subsequent steps. The calculation of distribution density not only helps to identify the activity status of livestock, but also provides early warning of abnormal behavior or group behavior changes, thereby providing a scientific basis for livestock management and stress state assessment. According to the distribution density of livestock activities, the vibration signals generated during the approach of the robotic arm to the livestock can be accurately analyzed through the vibration quantification data of the robotic arm. These vibration signals are highly correlated with the stress response of livestock. Therefore, by quantifying the vibration data and combining the probability estimation of the stress state of livestock, the physiological stress state of livestock can be evaluated in real time. This step effectively improves the intelligence of the robotic arm monitoring system, so that it can automatically adjust the control strategy according to the response of the livestock to avoid excessive interference or triggering negative emotions. Based on the probability data of the livestock stress state, the system can automatically adjust the approach rate of the robotic arm, thereby achieving more accurate monitoring and operation. This step avoids unnecessary fright or stress response when the robot arm approaches livestock by optimizing the matching of the speed. At the same time, the posture control instructions are designed in combination with the livestock activity images, so that the robot arm can adjust its posture to ensure the stability and effectiveness of the monitoring process. This measure effectively improves the sensitivity and adaptability of the robot arm operation, thereby realizing the interaction with livestock and efficient execution of monitoring tasks. After encoding the robot arm monitoring posture adjustment instructions, efficient control instructions can be formed to provide technical support for the automated operation of the robot arm. By embedding the robot arm control center and executing a specially designed control program, the system can respond to the state changes of livestock in real time and automatically adjust the posture and operation mode of the robot arm. This automated control not only reduces the need for manual intervention, but also improves work efficiency and accuracy, ensuring that the robot arm can accurately and stably perform various operation tasks when monitoring livestock status, thereby achieving efficient and seamless livestock management and monitoring. Therefore, the present invention optimizes the traditional livestock status monitoring robot arm control method, solves the problems of the traditional livestock status monitoring robot arm control method, that is, the robot arm has insufficient adaptability to environmental interference and cannot accurately control the robot arm, improves the robot arm's adaptability to environmental interference and improves the robot arm's precise control capability. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic flow chart of the steps of a method for controlling a robotic arm for monitoring livestock status; Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart; Figure 3 for Figure 1Detailed implementation steps of step S3 in FIG.

[0016] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0017] The technical method of the present invention is described clearly and completely below 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.

[0018] 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.

[0019] 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.

[0020] To achieve this, please refer to Figures 1 to 3 , a method for controlling a robotic arm for monitoring livestock status, the method comprising the following steps: Step S1: monitoring livestock activities in a livestock monitoring area through electronic monitoring equipment to obtain livestock activity monitoring images; calculating the distribution density of livestock activity monitoring images to obtain livestock activity distribution density; Step S2: quantifying the vibration of the approaching robot arm observation based on the activity distribution density of the livestock types to obtain the quantified data of the approaching robot arm vibration; estimating the probability of the livestock stress state based on the quantified data of the approaching robot arm vibration to obtain the probability data of the livestock stress state; Step S3: performing robot arm approach rate matching according to the livestock stress state probability data to obtain robot arm approach rate matching data; performing robot arm monitoring posture control instruction design based on the robot arm approach rate matching data and the livestock activity monitoring image to obtain the robot arm monitoring posture adjustment control instruction; Step S4: Encode the robot arm monitoring posture adjustment control instruction to obtain the robot arm monitoring posture adjustment coding instruction; design an automated control program based on the robot arm monitoring posture adjustment coding instruction to obtain the robot arm monitoring posture control program, and embed the robot arm monitoring posture control program into the robot arm control center to execute the robot arm control method for livestock status monitoring.

[0021] In the embodiment of the present invention, reference Figure 1 FIG. 1 is a schematic flow chart of a method for controlling a robotic arm for monitoring livestock status according to the present invention. In this example, the method for controlling a robotic arm for monitoring livestock status includes the following steps: Step S1: monitoring livestock activities in a livestock monitoring area through electronic monitoring equipment to obtain livestock activity monitoring images; calculating the distribution density of livestock activity monitoring images to obtain livestock activity distribution density; In an embodiment of the present invention, the livestock monitoring area is monitored in real time by electronic monitoring equipment. These monitoring devices include a device installed on a cross track and electrically connected to a mechanical arm through a servo motor, and the mechanical arm is controlled to move on the cross track by a servo motor. The mechanical arm is equipped with an electronic monitoring device, or a multi-sensor integrated system, such as a combination of an infrared imaging camera and a conventional visible light camera. The image data obtained by the camera is used to observe the activities of the livestock in real time. After the monitoring image is obtained, it is analyzed using image processing technology. First, the livestock in the image is separated using an image segmentation algorithm to extract the outline of each livestock. In order to improve the segmentation accuracy, a threshold-based segmentation method is used to separate the livestock from the background in the image, and the edge definition is further optimized through morphological operations (such as corrosion and expansion). Next, the region growing method or the connected domain analysis algorithm is used to judge and distinguish different types of livestock, such as pigs, cattle and sheep, and classify and mark them according to characteristics such as color and texture. After the livestock type is marked, the distribution density of each livestock is calculated using a density estimation algorithm (such as kernel density estimation) to generate a livestock type activity distribution density map. This calculation can be done by calculating the frequency of each type of livestock appearing in the image and combining it with its spatial distribution characteristics to derive the distribution density of livestock in the monitored area, thus providing basic data for subsequent steps.

[0022] Step S2: quantifying the vibration of the approaching robot arm observation based on the activity distribution density of the livestock types to obtain the quantified data of the approaching robot arm vibration; estimating the probability of the livestock stress state based on the quantified data of the approaching robot arm vibration to obtain the probability data of the livestock stress state; In the embodiment of the present invention, firstly, based on the livestock type activity distribution density data obtained in step S1, the vibration quantification processing of the approach observation of the manipulator is performed. By analyzing the distribution density data, the areas with higher density and the areas with lower density are identified, and the observation position and direction of the manipulator are adjusted according to this information. At this time, the sensor data acquisition system is used to monitor the relative position change between the manipulator and the livestock in real time. Combining the motion trajectory and reaction speed of the manipulator, the vibration amplitude and frequency generated by the manipulator when approaching the livestock are calculated. Vibration quantification can read the vibration data of the manipulator during movement through the sensor, and use Fourier transform (FFT) to perform frequency domain analysis on these vibration data to obtain the amplitude and frequency distribution of the vibration, and then quantify the vibration data generated during the approach of the manipulator. Subsequently, the above vibration data is used to obtain the probability data of the stress state of the livestock. The probability estimation of the stress state adopts the Bayesian inference method, and the relationship between the vibration amplitude, frequency and the stress response of the livestock is introduced to perform inference, and finally the stress probability of the livestock in the current environment is obtained. The key to this process is to analyze the impact of the vibration of the robotic arm on different types of livestock and establish a mathematical model of vibration and stress state, thereby providing a scientific basis for subsequent adjustments to the robotic arm speed.

[0023] Step S3: performing robot arm approach rate matching according to the livestock stress state probability data to obtain robot arm approach rate matching data; performing robot arm monitoring posture control instruction design based on the robot arm approach rate matching data and the livestock activity monitoring image to obtain the robot arm monitoring posture adjustment control instruction; In an embodiment of the present invention, firstly, the approach rate of the robotic arm is matched according to the probability data of the livestock stress state obtained in step S2. In order to ensure that the movement of the robotic arm does not cause excessive stress response to the livestock, the movement rate of the robotic arm is first dynamically adjusted to ensure that the robotic arm approaches the livestock at the most appropriate speed. Rate matching uses an optimization algorithm, such as a gradient descent method, to adjust the movement speed of the robotic arm in combination with the stress probability data to adapt it to the stress response. The goal of rate matching is to minimize the probability of livestock stress while ensuring that the robotic arm can complete the monitoring task quickly and effectively. Next, the posture control instructions of the robotic arm are designed based on the optimized rate matching data and the livestock activity monitoring image. First, the spatial position of the livestock in the image is analyzed, and the optimal posture of the robotic arm is calculated using the inverse kinematics algorithm in combination with the position and operable space of the robotic arm. According to these posture data, the corresponding control instructions are designed. At this time, the control instructions will take into account the joint angles of the robotic arm, the constraints of the workspace, and the impact on the stress state of the livestock, and generate control instructions for adjusting the robotic arm's monitoring posture to ensure that the robotic arm completes the livestock monitoring task with an appropriate posture, while minimizing unnecessary vibration and noise interference during the movement of the robotic arm.

[0024] Step S4: Encode the robot arm monitoring posture adjustment control instruction to obtain the robot arm monitoring posture adjustment coding instruction; design an automated control program based on the robot arm monitoring posture adjustment coding instruction to obtain the robot arm monitoring posture control program, and embed the robot arm monitoring posture control program into the robot arm control center to execute the robot arm control method for livestock status monitoring.

[0025] In an embodiment of the present invention, first, the robot arm monitoring posture adjustment control instruction is coded. The coding process converts the designed control instruction into a command format suitable for the robot arm control system. Specifically, the posture adjustment instruction is formatted and coded using a standard control language (such as a PLC programming language or a robot control protocol) so that it can be recognized and executed by the robot arm control system. Then, based on these coded control instructions, an automated control program is designed. The design of the control program is based on a control algorithm, such as PID control, fuzzy control or adaptive control, to ensure that the robot arm can adjust its posture according to real-time feedback to meet the requirements of livestock monitoring. During the design process, by analyzing the kinematic model and dynamic model of the robot arm, it is ensured that each instruction can effectively control the movement of the robot arm, while taking into account the inertia, speed limit and workspace limit of the robot arm. After the design of the automated control program is completed, the program is embedded in the robot arm control center so that the robot arm can autonomously perform livestock status monitoring tasks. The control system can automatically adjust the position, posture and movement speed of the robot arm according to the real-time status of the livestock and environmental changes, ensuring that the monitoring task is completed without interfering with the livestock.

[0026] Step S1 includes the following steps: Step S11: obtaining a livestock monitoring area; deploying a cross track in the livestock monitoring area, the cross track and the robotic arm are electrically connected via a servo motor, the servo motor controls the robotic arm to move on the cross track, and an electronic monitoring device is deployed on the robotic arm; Step S12: monitoring livestock activities in the livestock monitoring area through electronic monitoring equipment to obtain livestock activity monitoring images; Step S13: marking the livestock types on the livestock activity monitoring image to obtain a livestock type activity monitoring image; wherein the livestock types include pigs, cattle and sheep; Step S14: Calculate the distribution density of the livestock activity monitoring images to obtain the livestock activity distribution density.

[0027] In an embodiment of the present invention, it is first necessary to obtain the physical environment information of the livestock monitoring area to ensure that there is enough space and facilities in the area for the manipulator to operate. The area is usually determined by environmental assessment, including ground flatness, spatial layout, and distance from surrounding obstacles. Next, a cross track is deployed at an appropriate location in the livestock monitoring area. The cross track consists of two intersecting tracks, which are installed in the X-axis and Y-axis directions respectively, so as to achieve precise movement of the manipulator in a two-dimensional plane. During the laying of the track, the width, length and material of the track must meet the load requirements of the manipulator movement. High-strength steel or aluminum alloy materials are usually used to ensure that the track has sufficient stability and durability. At each end of the cross track, a servo motor system is set, which controls the manipulator to move along the track. The selection of the servo motor needs to be determined according to the weight, motion accuracy and motion speed requirements of the manipulator, and the motor drive system needs to ensure that the manipulator can move smoothly and accurately along the X-axis or Y-axis on the track. The servo motor control system monitors the position of the manipulator in real time through a closed-loop feedback method, and adjusts the motor drive to ensure the precise positioning of the manipulator on the track. Electronic monitoring equipment is installed on the robot body, including high-definition cameras, infrared sensors and other necessary image acquisition devices to monitor and record the activity status of livestock. The installation location of the equipment must ensure full coverage of the monitoring area to avoid any corner being overlooked. Driven by the servo motor control system, the robot arm can be accurately positioned on the cross track, thereby realizing all-round monitoring of the livestock area. Real-time monitoring of livestock activities is carried out through the electronic monitoring equipment deployed on the robot arm. The electronic monitoring equipment mainly includes high-resolution visible light cameras, infrared imaging cameras and motion sensors. Visible light cameras are used to capture images under conventional lighting, while infrared imaging cameras are used to monitor livestock in low light or night environments. The equipment is installed on the robot arm, and the position of the robot arm is accurately controlled by the servo motor system to obtain image data at different positions in the livestock monitoring area. The acquired image data needs to be processed by the data acquisition system to ensure the quality and clarity of the image. The image is preliminarily de-noised and contrast-enhanced by the image processing algorithm, making the livestock in the image more obvious. In addition, during the monitoring process, the temperature and humidity data of the surrounding environment need to be monitored in real time, and the impact of environmental changes on image quality needs to be recorded by sensors. The image is further processed through image analysis technology (such as edge detection, contour extraction, etc.) to obtain the activity of livestock in the monitoring area. The image data provides basic visual information for subsequent steps and can support livestock species marking and distribution density calculation. The livestock activity monitoring image obtained in step S12 is marked with livestock species. To this end, an image classification algorithm is used for analysis, specifically including image feature extraction and classification based on deep learning convolutional neural network (CNN).By training a multi-category classification model, the model can identify different livestock species in the image, including pigs, cattle and sheep. During the training process, a large number of labeled livestock images are used as training data, and spatial features (such as texture, shape, color, etc.) in the image are extracted through CNN, thereby realizing automatic classification of livestock species. Specifically, the input livestock activity monitoring image is first preprocessed, such as scaling, standardization and normalization, to ensure that the size of the image input into the CNN model is consistent. Then, the image is convoluted layer by layer using the convolution layer to extract local features in the image, and then the feature dimension is compressed through the pooling layer. Finally, the classification output is performed through the fully connected layer to obtain the livestock species in each area of ​​the image. The classification results will be marked on the original monitoring image to form a livestock species activity monitoring image, in which each livestock is marked with the corresponding species label. In order to improve the marking accuracy, post-processing steps can also be used, such as manually correcting some areas suspected of being misidentified, to further ensure the accuracy of the image marking. The distribution density of the livestock species activity monitoring image obtained in step S13 is calculated. The specific steps include segmenting the marked images, identifying the distribution of each type of livestock, and calculating their spatial density in the monitoring area. To this end, the kernel density estimation (KDE) method is used to analyze the distribution of livestock species in the image. This method sets a kernel function (such as a Gaussian kernel) around the current position of each livestock in the image and calculates the density value of the kernel function around the point. The core idea of ​​kernel density estimation is to estimate the distribution density of livestock in each area of ​​the image by weighted summation. Specifically, for each type of livestock, the spatial coordinates of each livestock are first determined, and the area around the coordinates is weighted summed by the kernel density function to obtain the density value of the area. For the calculation of the entire image, kernel density estimation can be performed on each pixel point of the image to generate a density map showing the density of livestock in different areas. This process can clearly show the activity density of livestock in the monitoring area, and then provide a basis for the subsequent motion trajectory planning and monitoring adjustment of the robot arm. The distribution density map provides a basis for regional division for the monitoring work of the robot arm, so that it can effectively avoid high-density areas or adjust the monitoring method.

[0028] Step S2 includes the following steps: Step S21: analyzing the movable positions of livestock on the livestock activity monitoring images based on the distribution density of livestock types and activities, and obtaining a livestock movable position data set; Step S22: predicting the livestock spatial movement trajectory based on the livestock movable position data set and the livestock activity monitoring image to obtain livestock spatial movement trajectory prediction data; Step S23: quantifying the vibration of the approaching robot arm based on the predicted data of the livestock spatial movement trajectory, and obtaining the quantified data of the vibration of the approaching robot arm; Step S24: Estimating the livestock stress state probability based on the robot arm approach vibration quantified data to obtain livestock stress state probability data.

[0029] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes: Step S21: analyzing the movable positions of livestock on the livestock activity monitoring images based on the distribution density of livestock types and activities, and obtaining a livestock movable position data set; In the embodiment of the present invention, firstly, the livestock movable position analysis is performed on the livestock activity monitoring image based on the livestock type activity distribution density data obtained in step S14. The key to this step is to identify the activity range of different types of livestock in the monitoring area and their possible movement directions. By analyzing the distribution density map, the image segmentation technology is used to divide the different density areas. The higher density areas indicate that the livestock activities are more concentrated, while the lower density areas are open areas with less activity. On this basis, the activity boundaries and trajectories of the livestock are extracted through the image processing algorithm. Specifically, the pixel area showing the livestock in the image can be separated by the threshold segmentation method, so as to determine the outline position and activity range of each livestock. On this basis, a path search algorithm based on graph theory (such as A Algorithms) for livestock The moving direction is predicted. The algorithm combines heuristic search with path optimization, and can calculate the best movement direction and path based on the current distribution density of livestock and their activity areas. These path prediction results form a livestock movable position data set, which records the movement direction and area of ​​each type of livestock in the monitoring area. The data set contains the starting position of each livestock, the expected direction of movement, and the area it passes through, and finally forms a livestock movable position data set, which provides an important reference for the subsequent operation of the robotic arm.

[0030] Step S22: predicting the livestock spatial movement trajectory based on the livestock movable position data set and the livestock activity monitoring image to obtain livestock spatial movement trajectory prediction data; In the embodiment of the present invention, the livestock spatial movement trajectory is first predicted based on the livestock movable position data set obtained in step S21 and the livestock activity monitoring image obtained in step S12. The goal of this step is to use the known livestock movable position data set, combined with the livestock's historical movement trajectory and the visible environmental features in the monitoring image, to predict the livestock's future movement trajectory. To this end, a trajectory prediction algorithm based on Kalman filtering is used to estimate the livestock's movement trajectory. Kalman filtering is a recursive algorithm that can dynamically update the estimated value based on known measurement data and prediction models, and is suitable for accurate prediction under high uncertainty conditions. In the implementation process, the livestock position information in the monitoring image (i.e., the coordinates of each livestock at a certain moment) and the historical trajectory data are first input into the Kalman filter, and the speed, acceleration and position information of each livestock are updated through the prediction step of the filter. Then, according to the movement state of the livestock, the filter will generate a set of possible movement trajectories and output the movement path of each livestock within a certain period of time in the future. In order to further improve the prediction accuracy, the state transfer matrix of the Kalman filter can be adjusted by combining it with the current environmental information (such as terrain obstacles, the movement of other livestock, etc.) to make the prediction results more consistent with the changes in the actual environment. Ultimately, the livestock spatial movement trajectory prediction data obtained includes the path points and position coordinates that each livestock may pass through in the future time period, thus providing data support for the action planning of the robotic arm.

[0031] Step S23: quantifying the vibration of the approaching robot arm based on the predicted data of the livestock spatial movement trajectory, and obtaining the quantified data of the vibration of the approaching robot arm; In the embodiment of the present invention, based on the predicted data of the livestock spatial movement trajectory obtained in step S22, the vibration of the manipulator approaching for observation is quantified. The key to this process is to quantify the vibration impact of the manipulator when performing tasks according to the relative position relationship between the movement trajectory of the livestock and the movement path of the manipulator. First, the movement trajectory and position of the manipulator are determined according to the predicted data of the livestock spatial movement trajectory. The task of the manipulator is to observe and intervene in a predetermined area, so it is necessary to calculate the distance between the manipulator and the livestock, and adjust the movement speed of the manipulator according to this distance. In order to quantify the vibration of the manipulator, an acceleration sensor and a gyroscope are installed at key parts of the manipulator to monitor the vibration of the manipulator in real time during the movement. Specifically, the acceleration data detected by the sensor is analyzed by a digital signal processing algorithm to obtain the vibration amplitude and frequency. The process of vibration quantification includes modeling the inertial force of the manipulator in motion, and calculating the vibration amount generated by the manipulator when approaching the livestock through parameters such as the mass, movement speed and path curvature of the manipulator, combined with the vibration propagation model in physics. This process requires careful adjustment of the control system of the manipulator to ensure that the manipulator maintains stable movement during the vibration quantification process. Ultimately, the quantitative data on the robot arm's approach vibration will include the vibration amplitude and frequency generated by the robot arm at different movement phases and its relative distance from the livestock.

[0032] Step S24: Estimating the livestock stress state probability based on the robot arm approach vibration quantified data to obtain livestock stress state probability data.

[0033] In the embodiment of the present invention, firstly, the probability of the stress state of the livestock is estimated based on the quantified data of the mechanical arm approaching vibration obtained in step S23. The core task of this step is to quantify the intensity and frequency of the vibration to infer the probability of the stress response of the livestock under different vibration states. The intensity of the stress response is closely related to the amplitude, frequency and duration of the vibration. Therefore, by analyzing the amplitude change and frequency characteristics in the vibration data, the sensitivity of the livestock to the vibration and the degree of stress response can be inferred. In this process, a logistic regression model or a naive Bayes classification algorithm can be used to establish a probability model based on the mechanical arm vibration data and the historical stress data. The model outputs the probability of the stress state of the livestock under a specific vibration state by learning the relationship between the quantified vibration data and the stress response of the livestock. The specific operation is: first, a large amount of experimental data is collected, including the behavioral performance, stress response and physiological data (such as heart rate, respiratory rate, etc.) of the livestock under different vibration conditions. Then, a probability estimation model is trained based on these data, and the probability relationship between the quantified vibration data and the stress response is determined by the maximum likelihood estimation method. Finally, the probability data of livestock stress states are obtained, including the probability of each livestock's stress response under different vibration conditions. This data provides a basis for the subsequent intervention and adjustment of the robotic arm, ensuring that the movement of the robotic arm does not excessively interfere with the livestock and avoid unnecessary stress responses.

[0034] Step S23 includes the following steps: Step S231: analyzing the livestock intermittent speed change of the livestock spatial movement trajectory prediction data to obtain the livestock intermittent speed change data; Step S232: performing an inertia simulation of the instantaneous pause when the robot arm approaches for observation according to the intermittent speed change data of the livestock, and obtaining the instantaneous pause inertia data; Step S233: performing a robot arm jitter vector attenuation analysis on the instantaneous pause inertia data to obtain a robot arm jitter attenuation vector; Step S234: calculating the periodic jitter vector difference of the robot arm jitter attenuation vector to obtain the periodic jitter vector difference; Step S235: Acquire the design structure data of the robot arm; quantify the close observation vibration of the robot arm according to the design structure data of the robot arm, the robot arm jitter attenuation vector and the periodic jitter vector difference to obtain the close observation vibration quantification data of the robot arm.

[0035] In an embodiment of the present invention, the livestock spatial movement trajectory prediction data is first analyzed for intermittent speed changes of livestock. By analyzing the speed data of each livestock on its predetermined trajectory, the volatility and regularity of the speed are identified. The key to this step is to determine the speed change pattern of the livestock during movement and quantify its intermittent speed change. Specifically, by performing differential processing on the speed of each livestock in the time series, the incremental data of the speed is obtained, and then the sliding window algorithm is used to calculate the speed change rate in each window. By setting an appropriate window size, the change of livestock speed in a short period of time can be observed, and the acceleration, deceleration or pause moments in its movement can be identified. Further, the speed data can be analyzed in the frequency domain using a fast Fourier transform (FFT) to extract the frequency characteristics of the speed fluctuation and identify the periodic or non-periodic speed change pattern. Finally, the intermittent speed change data of livestock is obtained, including information such as the frequency of movement speed change, fluctuation amplitude and change period of each livestock. This data provides important input for subsequent steps, especially when the robotic arm is approaching for observation, the action strategy of the robotic arm can be adjusted according to the speed change of the livestock to avoid unnecessary interference. Based on the intermittent speed change data of livestock obtained in step S231, the inertia simulation of the instantaneous pause of the manipulator when approaching for observation is performed. The core task of this process is to simulate the inertial reaction of the manipulator when approaching these speed change points by analyzing the speed pause or change that occurs during the movement of livestock. First, according to the movement trajectory of the livestock and its speed change, the pause time and position that the manipulator needs to make when approaching are calculated. On this basis, the momentum theorem in physics is used to calculate the mass, speed and acceleration of the manipulator to estimate the inertial reaction of the manipulator when pausing. Specifically, assuming that the manipulator generates a certain speed in the process of approaching the livestock and suddenly pauses at a certain moment, the inertial effect of the manipulator will cause it to continue to move forward for a distance until it is stopped by an external force (such as braking or friction). At this time, by analyzing the speed and mass of the manipulator when it pauses, the classical dynamics formula is used to simulate the inertial effect. By calculating the inertial effect, the inertial data of the manipulator during the pause process can be obtained, including the speed, displacement and acceleration when it pauses. This data provides a theoretical basis for the subsequent vibration simulation and adjustment of the robotic arm, ensuring that the robotic arm can effectively control its inertial influence during the action process to avoid additional interference to the livestock. Based on the instantaneous pause inertial data obtained in step S232, the jitter vector of the robotic arm is attenuated and analyzed. This step is intended to simulate the jitter generated after the robotic arm pauses, and to evaluate the degree of interference of the robotic arm to the livestock by analyzing its attenuation process. First, the inertial effect generated by the robotic arm during the instantaneous pause will cause it to vibrate, and the amplitude and frequency of the vibration are related to the mass, motion state and pause time of the robotic arm.To this end, the vibration frequency and amplitude of the mechanical arm after the pause are first calculated by a mathematical model, and a simplified resonance model is used to describe the vibration mode of the mechanical arm. This model is based on the natural frequency of the system, the damping coefficient, and the mass and elastic properties of the mechanical arm, and can accurately predict the vibration response of the mechanical arm. According to the vibration data of the mechanical arm, an attenuation model (such as an exponential decay formula) is used to describe the attenuation process of the vibration over time. The attenuation coefficient is calculated based on the structure and material properties of the mechanical arm, and can reflect the decreasing rate of vibration energy over time. Through simulation calculation, the attenuation vector of the mechanical arm vibration is obtained, which characterizes the change of the vibration amplitude at different time points. Finally, the mechanical arm jitter attenuation vector is obtained, which provides data support for the subsequent periodic jitter analysis, ensuring that the impact of vibration on livestock is effectively controlled when the mechanical arm approaches livestock. Based on the mechanical arm jitter attenuation vector obtained in step S233, the periodic jitter vector difference calculation is performed. The goal of this step is to quantify the periodic changes of the mechanical arm jitter, and to evaluate the impact of different periodic jitters on the stability of the mechanical arm by calculating the periodic jitter vector difference. First, according to the attenuation vector, the periodic characteristics of the vibration of the manipulator are extracted, and the frequency of the vibration is analyzed. Through frequency domain analysis techniques such as Fourier transform, the vibration signal can be decomposed into multiple frequency components and the main vibration frequency can be identified. On this basis, the change of the vibration amplitude in each cycle is calculated, and the difference between the vibration vectors of different cycles is obtained by using the differential calculation method. Specifically, the periodic jitter vector difference evaluates the stability of the vibration of the manipulator in each cycle by calculating the difference in vibration amplitude between adjacent cycles. If the periodic jitter vector difference is large, it means that there is a large fluctuation in the vibration amplitude, which will have an adverse effect on the operating accuracy of the manipulator. Through this method, the vibration characteristics of the manipulator in different time periods can be fully understood, and data support can be provided for further optimizing the control strategy of the manipulator. First, the design structure data of the manipulator is obtained, which contains various design parameters of the manipulator, such as structural size, mass distribution, moment of inertia and other information. According to the design data of the manipulator, the jitter attenuation vector obtained in step S233, and the periodic jitter vector difference obtained in step S234, the manipulator is observed and quantified for vibration. In this process, the design structure data of the robot arm is the key input factor. By combining it with the vibration attenuation vector and periodic vibration data, the vibration response of the robot arm under different working conditions is calculated. Specifically, based on the mass distribution and moment of inertia of the robot arm, a mechanical model is used to quantify the vibration. Taking into account the external and internal forces on the robot arm during movement, the vibration amplitude and frequency are calculated. At this time, by calculating the propagation path and reflection characteristics of the vibration, the specific quantitative data of the vibration during the approach of the robot arm is obtained. The vibration quantification data will include important information such as vibration amplitude, frequency, duration, and the relative influence between the robot arm and livestock.This data can ensure that the movement of the robotic arm is smooth, avoid excessive vibration, reduce the impact on livestock, and provide precise parameters for subsequent robotic arm control.

[0036] Step S24 includes the following steps: Step S241: Obtain the noise decibel acceptable to livestock, wherein the noise decibel acceptable to livestock includes: the noise decibel acceptable to pigs, the noise decibel acceptable to sheep, and the noise decibel acceptable to cattle; Step S242: performing vibration noise feedback matching on the robot arm approach vibration quantization data to obtain vibration noise feedback data; Step S243: calculating the instantaneous noise spectrum band difference on the vibration noise feedback data to obtain the instantaneous noise spectrum band difference; Step S244: performing noise band extreme value distribution analysis according to the instantaneous noise spectrum band difference to obtain noise band extreme value distribution data; Step S245: performing an excessive sound pressure instantaneous aggregation analysis on the livestock acceptable noise decibel according to the instantaneous noise spectrum band difference and the noise band extreme value distribution data, and obtaining the excessive sound pressure instantaneous aggregation data; Step S246: estimating the livestock stress state probability based on the excess sound pressure instantaneous aggregation data for the livestock acceptable noise decibels, and obtaining livestock stress state probability data.

[0037] In the embodiment of the present invention, firstly, the acceptable noise decibel data of different types of livestock are obtained, specifically including the noise tolerance of pigs, sheep and cattle. To ensure the accuracy of the data, the basic reaction of each type of livestock in a quiet environment is first evaluated. The reaction of livestock in different environments is monitored using a noise sensor or a sound level meter. Through experiments, the acceptable noise thresholds of different types of livestock under various noise sources are gradually determined. First, each type of livestock is placed separately in an environment with adjustable noise, and the decibel value of the environmental noise is gradually increased to observe its behavioral response. Pigs react more sensitively, usually showing certain uneasy behavior when the noise reaches 85 decibels, so the acceptable noise decibel of pigs is 85 decibels. Sheep begin to show slight uneasy behavior when the noise reaches 80 decibels, so the acceptable noise decibel of sheep is 80 decibels. The noise tolerance of cattle is relatively high, and uneasy behavior usually occurs at 90 decibels, so the acceptable noise decibel of cattle is 90 decibels. Through these experimental data, the acceptable noise decibel data set of each livestock is obtained, including the noise tolerance of three types of livestock, pigs, sheep and cattle, which provides key reference data for the subsequent noise interference analysis; in another embodiment, relevant literature can also be consulted to obtain the acceptable noise decibels of pigs, sheep and cattle. First, the vibration noise feedback matching is performed on the vibration quantification data of the approaching robot arm. This step is to match the vibration signal generated by the robot arm when approaching the livestock with the actual noise data to form vibration noise feedback data. According to the vibration quantification data of the robot arm obtained in step S235, considering that the vibration generated during the movement of the robot arm will not only affect the stability of the robot arm itself, but also produce additional noise effects. Therefore, the frequency component and amplitude of the vibration signal are first extracted to establish a vibration spectrum. Then, the vibration spectrum is band filtered using a high-pass filter and a low-pass filter, and the frequency distribution of the vibration noise is divided into several main intervals of low frequency, medium frequency and high frequency. By comparing the vibration spectrum with the acceptable noise decibels of livestock, feedback matching is performed to evaluate the degree of noise interference caused by vibration to livestock. Specifically, a matching algorithm, such as the least squares method or the normalized cross-correlation method, is used to compare the matching degree between the vibration spectrum of the manipulator and the acceptable noise threshold of the livestock. In this way, vibration noise feedback data can be obtained to characterize the contribution of vibration to the environmental noise during the movement of the manipulator, thereby providing a reference for subsequent noise analysis. The instantaneous noise spectrum band difference is calculated for the vibration noise feedback data. The goal of this step is to calculate the difference in noise spectrum at different time points by analyzing the spectrum bands of the vibration noise feedback data. First, the vibration noise feedback data is divided into multiple time periods, and the noise signal in each time period is converted into a frequency domain signal using the fast Fourier transform (FFT). Then, based on the frequency domain data, the different frequency components of the noise signal are identified. The noise spectrum band difference at each time point is obtained by calculating the difference between the noise spectrum bands of adjacent time periods.The noise spectrum band difference is calculated by performing differential processing on the spectrum amplitude of each time period, that is, for any time t1 and t2, calculating the difference in the frequency components corresponding to them. If the spectrum difference is large, it indicates that the frequency component of the noise signal has changed significantly at this time point. Through this calculation method, the volatility of the noise spectrum can be quantified, and the frequency change pattern that causes the livestock stress response when the mechanical arm approaches can be identified. Based on the instantaneous noise spectrum band difference data obtained in step S243, the noise band extreme value distribution analysis is performed. The purpose of the noise band extreme value distribution analysis is to evaluate the extreme value distribution of the noise spectrum fluctuation caused by the vibration of the mechanical arm in different frequency bands, especially to identify the noise band that causes livestock stress response. First, multiple frequency ranges in the noise spectrum band are selected, and the maximum amplitude of each frequency band is calculated in groups. By counting the maximum noise amplitude and frequency of occurrence of each frequency band, the noise extreme value distribution of different frequency bands is obtained. Further, the extreme value theoretical model is used to analyze the extreme value distribution characteristics of the noise to determine whether the noise amplitude of each frequency band exceeds the acceptable range of the livestock. For each frequency band, the statistical characteristics of the noise peak in the frequency band, such as the maximum value, average value and standard deviation, can be calculated to evaluate its potential impact on livestock. At this time, the intensity and fluctuation of noise in different frequency bands can be clarified through the noise band extreme value distribution data, providing a basis for subsequent noise control and stress state analysis. Based on the instantaneous noise spectrum band difference and noise band extreme value distribution data obtained in step S243 and step S244, the instantaneous aggregation of excessive sound pressure in the acceptable noise decibel of livestock is analyzed. The goal of this analysis is to identify the aggregation effect of noise, that is, whether the noise intensity that occurs in a short period of time will produce excessive sound pressure on livestock, thereby triggering a stress response. First, according to the instantaneous band difference and extreme value distribution data of each noise frequency band, the sound pressure aggregation value at different time points is calculated. Specifically, by integrating the instantaneous noise spectrum band difference data, the total noise intensity of each time period is obtained, and combined with the noise band extreme value distribution, the excessive accumulation of sound pressure in these time periods is evaluated. By calculating the concentration of sound pressure in each time period, it is possible to identify areas of excessive sound pressure concentration in a short period of time. For each time period, it is further compared with the acceptable noise decibel of livestock to identify whether it exceeds the noise tolerance of livestock, and then predict the induced stress response. Based on the instantaneous aggregation data of excess sound pressure obtained in step S245, the probability of livestock stress state is estimated for the acceptable noise decibel of livestock. The purpose of this step is to quantify the probability of stress response of livestock when facing different noise intensities. First, combined with the noise tolerance data of different livestock, the excess sound pressure in each time period is evaluated. If the sound pressure aggregation value of a certain time period exceeds the acceptable noise decibel of livestock, it is considered that the time period causes a stress response in livestock. According to experimental data, by constructing a stress response probability model, based on the degree and duration of excessive noise aggregation, the probability of livestock producing a stress response in this time period is estimated.This probabilistic model can be fitted and trained using statistical methods such as linear regression or logistic regression to obtain a mapping function that associates excess sound pressure data with the probability of stress response. Ultimately, this model can accurately predict the probability of stress state of livestock under different noise conditions, providing decision support for subsequent monitoring and intervention.

[0038] Step S3 includes the following steps: Step S31: performing robot arm approach rate matching according to the robot arm approach vibration quantification data and livestock stress state probability data to obtain robot arm approach rate matching data; Step S32: performing a logic error iteration test on the robot arm approach rate matching data to obtain logic error iteration test data; Step S33: performing rate error correction on the robot arm approach rate matching data according to the logical error iteration test data to obtain the robot arm approach rate optimization data; Step S34: Designing a robot arm monitoring posture control instruction based on the robot arm approach rate optimization data and the livestock activity monitoring image to obtain a robot arm monitoring posture adjustment control instruction.

[0039] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes: Step S31: performing robot arm approach rate matching according to the robot arm approach vibration quantification data and livestock stress state probability data to obtain robot arm approach rate matching data; In an embodiment of the present invention, firstly, the mechanical arm approaching vibration quantitative data and livestock stress state probability data are obtained. These two data sets respectively characterize the vibration effect generated when the mechanical arm moves and the probability of stress response caused by the vibration to the livestock. According to these data, the mechanical arm approaching rate matching is performed, and the purpose is to minimize the impact of the vibration on the livestock stress state by adjusting the movement rate of the mechanical arm. Specifically, firstly, according to the vibration quantitative data of the mechanical arm at different movement rates, a relationship diagram between the vibration amplitude and the movement rate is drawn. The figure shows that there is a certain linear or nonlinear relationship between the movement rate and the vibration amplitude of the mechanical arm, and a higher movement rate will lead to stronger vibration. Then, combined with the livestock stress state probability data, the stress response probability generated by the livestock under different vibration amplitudes is identified. For example, when the vibration amplitude exceeds a certain threshold, the livestock stress state probability increases sharply, resulting in a stress response. Therefore, according to the relationship between the livestock stress state probability and the vibration quantitative data, the movement rate of the mechanical arm is adjusted so that the vibration amplitude is kept within the acceptable range of the livestock, thereby ensuring the comfort and safety of the livestock. Finally, the mechanical arm approaching rate matching data is obtained.

[0040] Step S32: performing a logic error iteration test on the robot arm approach rate matching data to obtain logic error iteration test data; In an embodiment of the present invention, a logical error iteration test is performed on the approach rate matching data of the manipulator. First, the actual motion data of the manipulator at different rates are compared with the theoretical rate matching data to analyze the deviation between the two. To this end, it is necessary to continuously monitor the actual rate of the manipulator during its motion and compare it with the optimized ideal rate. Through an error feedback control algorithm, such as a proportional-integral-differential (PID) controller or a Kalman filter algorithm, the rate of the manipulator is dynamically adjusted to make it closer to the ideal rate. When there is an error between the actual motion rate of the manipulator and the preset ideal rate, the system adjusts the rate value according to the feedback information to gradually reduce the error. Through multiple iterative tests, the actual motion of the manipulator is re-evaluated after each adjustment to ensure that the error continues to decrease. This process continues until the predetermined error tolerance range is reached. In this process, the results of each adjustment and calculation are based on the previous error feedback to achieve higher accuracy and better matching, and obtain logical error iteration test data. This data is used for further rate optimization to ensure that the motion rate of the manipulator at each stage can accurately meet the actual needs.

[0041] Step S33: performing rate error correction on the robot arm approach rate matching data according to the logical error iteration test data to obtain the robot arm approach rate optimization data; In an embodiment of the present invention, the rate error correction of the approach rate matching data of the manipulator is performed using the logical error iteration test data. First, based on the error iteration data obtained in step S32, the error between the actual motion rate of the manipulator and the target rate in each iteration is calculated. By statistically analyzing these error data, the pattern and change trend of the rate error are identified. A correction method, such as a weighted average or minimum error method, is used to correct the rate matching data of the manipulator. For example, in each rate segment, the distribution law of the error is analyzed and its average deviation is calculated, and then the error of each rate segment is corrected to obtain more accurate manipulator rate matching data. The core of rate error correction is to minimize the influence of different errors, and by adjusting the parameters of rate matching, the error between it and the ideal rate is kept within the minimum range. For each group of actual motion rates of the manipulator, a slight adjustment is made according to its error value to ensure that the corrected rate is more accurate. Finally, through this correction step, the manipulator approach rate optimization data is obtained, which provides more stable and reliable rate data for subsequent posture control and motion adjustment.

[0042] Step S34: Designing a robot arm monitoring posture control instruction based on the robot arm approach rate optimization data and the livestock activity monitoring image to obtain a robot arm monitoring posture adjustment control instruction.

[0043] In an embodiment of the present invention, the design of the manipulator monitoring posture control instruction is performed based on the manipulator approach rate optimization data and livestock activity monitoring image obtained in step S33. First, the optimized rate data of the manipulator provides a basis for designing the monitoring posture control instruction. In the design process, it is necessary to analyze the impact of the motion trajectory and posture adjustment of the manipulator on the livestock. By analyzing the livestock activity monitoring image, the position, action and state information of the livestock are extracted. Based on these image data, the current position and activity range of the livestock are identified, and the possible path and posture of the manipulator when approaching the livestock are predicted according to the optimized rate data of the manipulator. In order to ensure that the manipulator does not cause unnecessary stress response when approaching the livestock, it is necessary to adjust the monitoring posture of the manipulator in real time according to the activity of the livestock. For example, when the livestock is in a relatively tense state, the manipulator should maintain a slower approach rate and adopt a gentler posture to avoid excessive vibration or sound pressure stimulation. In the design of the control instruction, it is necessary to comprehensively consider these factors, generate a posture adjustment scheme adapted to the current livestock state, and finally obtain the manipulator monitoring posture adjustment control instruction. These instructions will ensure that the robotic arm is more precise and stable during movement, while reducing interference with livestock.

[0044] Step S33 includes the following steps: Step S331: performing multi-frequency error accumulation calculation on the logic error iteration test data to obtain multi-frequency error accumulation data; Step S332: performing interval maximum error analysis on the logic error iteration test data according to the multi-frequency error accumulation data to obtain multi-frequency maximum error interval data; Step S333: performing time series linear interpolation on the multi-frequency maximum error interval data to obtain error interval time series interpolation data; Step S334: performing rate error correction on the robot arm approach rate matching data according to the error interval timing interpolation data and the multi-frequency maximum error interval data to obtain the robot arm approach rate optimization data.

[0045] In an embodiment of the present invention, multi-frequency error accumulation calculation is performed on the logic error iteration test data. First, the test data of multiple logic error iterations are obtained and arranged in chronological order. The data of each round of error test is usually based on the time step, and a time series can be formed. In order to calculate the multi-frequency error accumulation, it is necessary to accumulate the value of each error. The specific operation is to add the error value of each round with the error value of the previous round to obtain the cumulative error. In order to achieve this process, the error value of each test cycle is first standardized so that the error data has the same dimension, which is convenient for subsequent addition calculation. By calculating the error at each time point and the error accumulation value at the previous time point, the cumulative value of all errors is gradually obtained. The error accumulation value at each time point reflects the accumulation of errors in the time period, forming a set of error accumulation data. This data set not only contains the time distribution of the error, but also reveals the error evolution trend at different rates, which helps to analyze the maximum fluctuation range of the error in the subsequent analysis. The maximum error analysis of the interval is performed according to the multi-frequency error accumulation data. First, the time interval analysis of the multi-frequency error accumulation data is required. The error accumulation value at each time point can be compared with the error accumulation value at the previous time point, so as to calculate the error change in each time interval. In order to perform the maximum error analysis, it is necessary to find the maximum error value of each time interval in the error accumulation data. This process involves calculating the error difference of each time period, that is, the error accumulation value at the current time point minus the error accumulation value at the previous time point. In this way, the error increment of each time interval can be calculated and the maximum error value in each interval can be identified. Next, the maximum error in each time interval is analyzed and the time interval in which it occurs is recorded. Through this analysis, the multi-frequency maximum error interval data obtained shows the amplitude of the error value fluctuation and the time interval in which it occurs in different time periods. This data is crucial for subsequent error correction and rate optimization because it can help accurately identify which time periods have the largest error fluctuations, thereby providing an accurate basis for optimizing rate matching. Time series linear interpolation is performed on the multi-frequency maximum error interval data. First, the multi-frequency maximum error interval data obtained in step S332 are obtained. These data contain the maximum error value and its time span in each time interval. In order to smooth these data and perform interpolation processing, it is necessary to first determine the time point of interpolation. These time points can be selected between the time intervals of the original data to ensure that the interpolated data has continuity and high accuracy. Then, the linear interpolation method is used to interpolate the error values ​​between every two adjacent time points. The specific operation of linear interpolation is that for any two time points, assuming that their error values ​​are E1 and E2 respectively and the time intervals are T1 and T2, then the interpolation error value at time point T is It can be calculated by the following formula: , through this method, the error value of each interpolation time point can be obtained, thereby forming a smooth error value curve. The main purpose of time series linear interpolation is to fill the gaps in the original data, so that the entire error data set becomes more refined and has a higher time resolution. The interpolated data can more accurately reflect the changing trend of the error at each time point and provide more detailed information for subsequent error correction. The rate error of the manipulator approaching rate matching data is corrected according to the error interval time series interpolation data and the multi-frequency maximum error interval data. First, the error interval time series interpolation data in step S333 and the multi-frequency maximum error interval data in step S332 are obtained. These two sets of data represent the smooth change trend of the error and the maximum fluctuation range of the error, respectively. In order to correct the rate error of the manipulator approaching rate matching data, it is necessary to combine these two sets of data with the original rate matching data for error correction. The specific operation is to first calculate the rate error at each time point by comparing the relationship between the original rate matching data and the error accumulation data. Then, according to the error interval time series interpolation data and the multi-frequency maximum error interval data, the rate value at each time point is adjusted. The adjustment method is to calculate the correction factor based on the amplitude and time interval of the error fluctuation. The correction factor reflects the accumulation of errors in a certain period of time, and dynamically adjusts the rate matching data according to the maximum fluctuation range of the error. For the error at each time point, a weighted average or minimum error optimization algorithm is used to gradually correct the movement rate of the robot arm to make it closer to the target rate. Finally, the robot arm approach rate optimization data is obtained, which adjusts the movement of the robot arm to a more precise state through the error-corrected rate, ensuring that the robot arm can minimize interference with livestock when approaching livestock and avoid excessive vibrations that cause unnecessary stress responses.

[0046] Step S34 includes the following steps: Step S341: performing livestock activity posture decomposition processing on the livestock activity monitoring image to obtain livestock activity posture decomposition data; Step S342: Performing dynamic analysis of the livestock visual field blind area on the livestock activity posture decomposition data to obtain dynamic data of the livestock visual field blind area; Step S343: performing robot arm path posture control adjustment on the livestock visual field blind spot dynamic data according to the robot arm approach rate optimization data to obtain robot arm path posture control adjustment data; Step S344: Designing a robot arm monitoring posture control instruction based on the robot arm path posture control adjustment data to obtain a robot arm monitoring posture adjustment control instruction.

[0047] In an embodiment of the present invention, livestock activity monitoring images are subjected to livestock activity posture decomposition processing. First, livestock activity monitoring image data from a camera or other image acquisition device is obtained, and these image data generally contain the movement trajectory and specific activity posture of the livestock. In order to effectively decompose the activity posture of the livestock, the image is first preprocessed to remove noise and enhance the contrast of the image, so that the outline of the livestock is clearer. During the processing, an edge detection algorithm (such as Canny edge detection) is used to extract the outline and joint position of the livestock in the image. Then, a posture estimation algorithm, such as the OpenPose algorithm, is used to identify the position information of various parts of the livestock (such as the head, limbs, trunk, etc.) through key point detection technology. These key points are connected according to different activity postures to form a skeleton diagram, which represents the posture of the livestock at different time points. Further, the joint position of each skeleton diagram is matched with time to obtain the time series data of the livestock activity, thereby completing the decomposition of the livestock activity posture. Finally, the livestock activity posture decomposition data obtained contains the specific posture information of the livestock at different time points, including key features such as the start, duration, and direction of movement of the activity, providing basic data for subsequent visual blind spot analysis and path control. Livestock activity posture decomposition data is subjected to dynamic analysis of livestock visual blind spots. The purpose of this step is to identify and analyze the changing rules of the visual blind spots of livestock during their activities, so as to use the blind spots for effective monitoring and intervention. According to the livestock activity posture decomposition data obtained in step S341, firstly, by analyzing the direction and movement trajectory of the livestock's head and eyes, combined with their activity state, the visual field range of the livestock at each time node is calculated. Since the direction of the livestock's head and eyes often changes during the movement, its visual blind spot changes dynamically. In order to accurately describe this change, stereoscopic vision technology and deep learning algorithms are used to analyze the three-dimensional information of the image to identify which areas are blocked and become blind spots at different time points. For example, if the livestock's head turns faster or the body posture changes significantly, the size and position of the blind spot will also change accordingly. Based on these analyses, the dynamic area division technology is used to spatialize the blind spots as dynamic area data that changes over time, and generate dynamic data of the blind spots from the livestock's perspective. Through this analysis, the position of the livestock's blind spots can be accurately grasped, providing accurate data support for subsequent path adjustments and monitoring instructions. According to the robot arm approach rate optimization data, the robot arm path posture control adjustment is performed on the dynamic data of the livestock's blind spots. The goal of this step is to adjust the path and posture of the robot arm by analyzing the relationship between the movement speed of the robot arm and the dynamic changes of the blind spot, so as to ensure that the robot arm is always in the livestock's blind spot, thereby avoiding interfering with the livestock's visual perception. First, the robot arm approach rate optimization data that has been completed in step S333 is obtained, which describes the optimal movement rate and acceleration of the robot arm in the process of approaching livestock.Then, according to the dynamic data of the blind spot of the livestock's visual field, it is determined whether the robot arm has entered the visual range of the livestock. If the robot arm enters the livestock's visual field, its motion path needs to be adjusted so that the robot arm can stay in the livestock's blind spot as much as possible during the approach process. This adjustment is achieved by replanning the path points of the robot arm and adjusting the posture. For example, a path planning method based on the A* algorithm is used to avoid the livestock's visual field while considering the speed limit of the robot arm, so that the robot arm path is as little exposed to the livestock's visual field as possible. By adjusting the movement trajectory and speed of the robot arm, it is ensured that the robot arm is always in the blind spot of the livestock's visual field, thereby avoiding interfering with the normal activities of the livestock. Through this path adjustment, the robot arm path posture control adjustment data is obtained, which reflects the specific posture adjustment instructions that the robot arm should take when approaching the livestock. The robot arm monitoring posture control instruction is designed based on the robot arm path posture control adjustment data. The purpose of this step is to generate the final robot arm control instruction to ensure that the robot arm can accurately perform path posture adjustment and maintain in the livestock's blind spot. First, according to the robot arm path posture control adjustment data obtained in step S343, combined with the current state, target position and movement rate of the robot arm, the inverse kinematics (IK) algorithm is used to accurately calculate the angle of each joint of the robot arm. Through this process, the target position and movement angle of each joint are calculated to ensure that each part of the robot arm can move according to the predetermined path and posture. In the path control process, a trajectory tracking algorithm, such as a PID control algorithm, is used to ensure that the robot arm remains stable and smooth during the movement process, avoiding instability caused by too fast or too slow speed. Finally, the robot arm posture adjustment control instructions are generated according to these control parameters, including the angle change, movement speed, acceleration and other control parameters of each joint, to ensure that the robot arm can accurately execute the predetermined path and keep monitoring operations in the blind area of ​​livestock. Through this series of steps, the generated robot arm monitoring posture adjustment control instructions ensure that the robot arm can complete the task efficiently and accurately during the operation process, while avoiding interference with livestock to the greatest extent.

[0048] Step S343 includes the following steps: Performing blind zone boundary surface fitting processing on the dynamic data of the blind zone of the livestock visual field to obtain blind zone boundary surface fitting data; Perform robot arm space coverage calculation on the blind area boundary surface fitting data to obtain robot arm space coverage data; Based on the space coverage data of the robot arm, the blind area boundary surface fitting data is processed for path point distribution retrieval to obtain path point distribution data; According to the robot arm approach rate optimization data, the path point distribution data is matched with the staged motion parameters to obtain the staged motion parameters of the robot arm; The robot arm path posture control adjustment is performed according to the robot arm's staged motion parameters and path point distribution data to obtain the robot arm path posture control adjustment data.

[0049] In the embodiment of the present invention, the dynamic data of the blind spot of the livestock visual field is firstly subjected to blind spot boundary surface fitting processing. The main purpose of this step is to accurately describe the shape and position change of the blind spot during livestock activities by fitting the boundary surface of the blind spot. First, the dynamic data of the blind spot of the livestock visual field obtained in step S342 is obtained. The data reflects the angle and position of the livestock activities at different time points, and the resulting changes in the visual field blind spot. In order to perform the fitting processing, polynomial fitting or B-spline curve (B-spline) method is used. These two methods can generate smooth surfaces through a small number of data points, which are suitable for describing irregular boundaries. By fitting the blind spot data in three-dimensional space, the surface equation of the blind spot boundary is obtained. In the specific operation, the coordinates of each data point in space (for example, X, Y, Z axis position) are used as fitting inputs, and the change law of these coordinates is modeled by the surface fitting algorithm, and finally a set of mathematical functions that can describe the blind spot boundary surface is generated. In this way, the obtained blind spot boundary surface fitting data can accurately represent the three-dimensional shape of the dynamic visual field blind spot formed by the livestock during the activity, and further provide a basis for path planning and robot arm motion adjustment. The space coverage of the manipulator is calculated by fitting the surface of the blind spot boundary. The space coverage data of the manipulator is generated by judging the space area that the manipulator can cover to ensure that its path and movement will not enter the field of vision of the livestock. In this step, first, the effective reachable space of the manipulator is determined according to the working range and motion parameters of the manipulator, which is usually calculated by the intersection of the trajectory generation of the manipulator and the target area. In order to more accurately control the behavior of the manipulator, it is necessary to combine the structural characteristics of the manipulator (such as arm length, rotation angle, etc.) and the surface fitting data of the blind spot boundary to analyze whether the manipulator can enter or approach the field of vision of the livestock under a specific path. If the working space of the manipulator coincides with the blind spot of the livestock, it means that the manipulator can move in this area, thereby ensuring that the task is performed without interfering with the field of vision of the livestock. In the calculation process of this step, the spatial mapping technology is used to compare the reachable area of ​​the manipulator with the boundary of the blind spot to obtain the space coverage data of the manipulator, and to clarify the range of motion of the manipulator and the relative position relationship with the blind spot. In the path point distribution retrieval processing step, the path point distribution retrieval is performed on the blind spot boundary surface fitting data based on the robot arm's spatial coverage data. The purpose of this step is to ensure that the robot arm can always stay in the blind spot during the process of approaching livestock and avoid being exposed to the livestock's field of vision by retrieving and distributing path points. First, the robot arm's spatial coverage data is used to clarify the relative position of the robot arm's range of motion and the blind spot boundary. Next, the path point distribution retrieval is performed based on the robot arm's working range and blind spot data through a path planning algorithm, such as the A* algorithm or the Dijkstra algorithm.The retrieval process needs to consider the target position of the manipulator, the range of the blind spot, and the constraints of the surrounding environment, and evenly distribute the path points in the working space of the manipulator to ensure that the manipulator path always stays in the blind spot during the movement process and avoids crossing the visual range of livestock. Through this path point distribution retrieval process, the path point distribution data is obtained to ensure that the path of the manipulator movement is adapted to the blind spot of the livestock visual field. The path point distribution data is matched with the staged motion parameters according to the manipulator approach rate optimization data. First, the staged parameters of the manipulator movement are determined according to the manipulator approach rate optimization data. These parameters include the speed of the manipulator, acceleration, the interval between the target points, and the movement sequence of the path points. Then, these parameters are combined with the path point distribution data to match the staged motion parameters. This process involves ensuring that the manipulator is always in an appropriate motion state when completing the task by adjusting the time interval and movement speed between the path points, thereby ensuring that it always stays in the blind spot during the process of approaching livestock. Specifically, time segmentation and parameter adjustment algorithms (such as piecewise linear interpolation) are used to accurately match the motion parameters of the manipulator between different path points, so that the manipulator moves smoothly and efficiently, and avoids instability caused by too fast or too slow speed. After this step is completed, the staged motion parameters of the manipulator are obtained, and these data provide the motion basis for the precise control of the manipulator. Finally, the manipulator path posture control adjustment is performed according to the staged motion parameters of the manipulator and the path point distribution data to obtain the manipulator path posture control adjustment data. The purpose of this step is to ensure that the manipulator can perform tasks accurately and correctly by optimizing the path and posture of the manipulator, and always stay within the blind spot of the livestock. In this step, the angle changes of each joint are first calculated according to the staged motion parameters of the manipulator, and the posture of the manipulator is adjusted by the inverse kinematics algorithm to adapt to the changes between different path points. At the same time, combined with the path point distribution data, a trajectory tracking algorithm (such as a PID control algorithm) is used to finely control the movement of the manipulator to ensure that the manipulator does not deviate from the predetermined trajectory during the movement and is always within the blind spot. Finally, the robot arm path posture control adjustment data is generated. This data provides the control instructions such as joint angle, speed, acceleration, etc. required for the robot arm to perform tasks, ensuring that the robot arm can accurately complete the monitoring task and maximize the use of the livestock's blind spots for monitoring operations.

[0050] The present invention also provides a livestock status monitoring robot arm control system, which is used to execute the livestock status monitoring robot arm control method as described above. The livestock status monitoring robot arm control system comprises: The distribution density analysis module is used to monitor livestock activities in the livestock monitoring area through electronic monitoring equipment to obtain livestock activity monitoring images; calculate the distribution density of livestock activity monitoring images to obtain the distribution density of livestock types; A stress state probability estimation module is used to quantify the vibration of the approaching observation of the robot arm based on the activity distribution density of the livestock species to obtain the quantified data of the approaching vibration of the robot arm; and to estimate the probability of the livestock stress state based on the quantified data of the approaching vibration of the robot arm to obtain the probability data of the livestock stress state; The posture control instruction design module is used to match the approach rate of the robot arm according to the probability data of the livestock stress state, and obtain the approach rate matching data of the robot arm; the robot arm monitoring posture control instruction is designed based on the approach rate matching data of the robot arm and the livestock activity monitoring image, and the robot arm monitoring posture adjustment control instruction is obtained; The automatic control program design module is used to encode and process the robot arm monitoring posture adjustment control instructions to obtain the robot arm monitoring posture adjustment coding instructions; based on the robot arm monitoring posture adjustment coding instructions, the automatic control program is designed to obtain the robot arm monitoring posture control program, and the robot arm monitoring posture control program is embedded into the robot arm control center to execute the robot arm control method for livestock status monitoring.

[0051] 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.

[0052] 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 method for controlling a robotic arm for livestock status monitoring, characterized in that: The following steps are involved: Step S1: monitoring livestock activities in a livestock monitoring area through electronic monitoring equipment to obtain livestock activity monitoring images; calculating the distribution density of livestock activity monitoring images to obtain livestock activity distribution density; Step S2: quantifying the vibration of the approaching robot arm observation based on the activity distribution density of livestock types, and obtaining the quantified data of the approaching robot arm vibration; The probability of livestock stress state is estimated based on the quantitative data of the mechanical arm approaching vibration, and the probability data of livestock stress state is obtained; Step S3: performing robot arm approach rate matching according to the livestock stress state probability data to obtain robot arm approach rate matching data; Based on the approach rate matching data of the robot arm and the livestock activity monitoring images, the robot arm monitoring posture control instructions are designed to obtain the robot arm monitoring posture adjustment control instructions; Step S4: Encode the robot arm monitoring posture adjustment control instruction to obtain the robot arm monitoring posture adjustment coding instruction; design an automated control program based on the robot arm monitoring posture adjustment coding instruction to obtain the robot arm monitoring posture control program, and embed the robot arm monitoring posture control program into the robot arm control center to execute the robot arm control method for livestock status monitoring.

2. The method for controlling a robotic arm for monitoring livestock status according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: obtaining a livestock monitoring area; deploying a cross track in the livestock monitoring area, the cross track and the robotic arm are electrically connected via a servo motor, the servo motor controls the robotic arm to move on the cross track, and an electronic monitoring device is deployed on the robotic arm; Step S12: monitoring livestock activities in the livestock monitoring area through electronic monitoring equipment to obtain livestock activity monitoring images; Step S13: marking the livestock types on the livestock activity monitoring image to obtain a livestock type activity monitoring image; wherein the livestock types include pigs, cattle and sheep; Step S14: Calculate the distribution density of the livestock activity monitoring images to obtain the livestock activity distribution density.

3. The method for controlling a robotic arm for monitoring livestock status according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: analyzing the movable positions of livestock on the livestock activity monitoring images based on the distribution density of livestock types and activities, and obtaining a livestock movable position data set; Step S22: predicting the livestock spatial movement trajectory based on the livestock movable position data set and the livestock activity monitoring image to obtain livestock spatial movement trajectory prediction data; Step S23: quantifying the vibration of the approaching robot arm based on the predicted data of the livestock spatial movement trajectory, and obtaining the quantified data of the vibration of the approaching robot arm; Step S24: Estimating the livestock stress state probability based on the robot arm approach vibration quantified data to obtain livestock stress state probability data.

4. The method for controlling a robotic arm for monitoring livestock status according to claim 3, characterized in that: Step S23 includes the following steps: Step S231: analyzing the livestock intermittent speed change of the livestock spatial movement trajectory prediction data to obtain the livestock intermittent speed change data; Step S232: performing an inertia simulation of the instantaneous pause when the robot arm approaches for observation according to the intermittent speed change data of the livestock, and obtaining the instantaneous pause inertia data; Step S233: performing a robot arm jitter vector attenuation analysis on the instantaneous pause inertia data to obtain a robot arm jitter attenuation vector; Step S234: calculating the periodic jitter vector difference of the robot arm jitter attenuation vector to obtain the periodic jitter vector difference; Step S235: Acquire the design structure data of the robot arm; quantify the close observation vibration of the robot arm according to the design structure data of the robot arm, the robot arm jitter attenuation vector and the periodic jitter vector difference to obtain the close observation vibration quantification data of the robot arm.

5. The method for controlling a robotic arm for monitoring livestock status according to claim 3, characterized in that: Step S24 includes the following steps: Step S241: Obtain the noise decibel acceptable to livestock, wherein the noise decibel acceptable to livestock includes: the noise decibel acceptable to pigs, the noise decibel acceptable to sheep, and the noise decibel acceptable to cattle; Step S242: performing vibration noise feedback matching on the robot arm approach vibration quantization data to obtain vibration noise feedback data; Step S243: calculating the instantaneous noise spectrum band difference on the vibration noise feedback data to obtain the instantaneous noise spectrum band difference; Step S244: performing noise band extreme value distribution analysis according to the instantaneous noise spectrum band difference to obtain noise band extreme value distribution data; Step S245: performing an excessive sound pressure instantaneous aggregation analysis on the livestock acceptable noise decibel according to the instantaneous noise spectrum band difference and the noise band extreme value distribution data, and obtaining the excessive sound pressure instantaneous aggregation data; Step S246: estimating the livestock stress state probability based on the excess sound pressure instantaneous aggregation data for the livestock acceptable noise decibels, and obtaining livestock stress state probability data.

6. The method for controlling a robotic arm for monitoring livestock status according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: performing robot arm approach rate matching according to the robot arm approach vibration quantification data and livestock stress state probability data to obtain robot arm approach rate matching data; Step S32: performing a logic error iteration test on the robot arm approach rate matching data to obtain logic error iteration test data; Step S33: performing rate error correction on the robot arm approach rate matching data according to the logical error iteration test data to obtain the robot arm approach rate optimization data; Step S34: Designing a robot arm monitoring posture control instruction based on the robot arm approach rate optimization data and the livestock activity monitoring image to obtain a robot arm monitoring posture adjustment control instruction.

7. The method for controlling a robotic arm for monitoring livestock status according to claim 6, characterized in that: Step S33 includes the following steps: Step S331: performing multi-frequency error accumulation calculation on the logic error iteration test data to obtain multi-frequency error accumulation data; Step S332: performing interval maximum error analysis on the logic error iteration test data according to the multi-frequency error accumulation data to obtain multi-frequency maximum error interval data; Step S333: performing time series linear interpolation on the multi-frequency maximum error interval data to obtain error interval time series interpolation data; Step S334: performing rate error correction on the robot arm approach rate matching data according to the error interval timing interpolation data and the multi-frequency maximum error interval data to obtain the robot arm approach rate optimization data.

8. The method for controlling a robotic arm for monitoring livestock status according to claim 6, characterized in that: Step S34 includes the following steps: Step S341: performing livestock activity posture decomposition processing on the livestock activity monitoring image to obtain livestock activity posture decomposition data; Step S342: Performing dynamic analysis of the livestock visual field blind area on the livestock activity posture decomposition data to obtain dynamic data of the livestock visual field blind area; Step S343: performing robot arm path posture control adjustment on the livestock visual field blind spot dynamic data according to the robot arm approach rate optimization data to obtain robot arm path posture control adjustment data; Step S344: Designing a robot arm monitoring posture control instruction based on the robot arm path posture control adjustment data to obtain a robot arm monitoring posture adjustment control instruction.

9. The method for controlling a robotic arm for monitoring livestock status according to claim 8, characterized in that: Step S343 includes the following steps: Performing blind zone boundary surface fitting processing on the dynamic data of the blind zone of the livestock visual field to obtain blind zone boundary surface fitting data; Perform robot arm space coverage calculation on the blind area boundary surface fitting data to obtain robot arm space coverage data; Based on the space coverage data of the robot arm, the blind area boundary surface fitting data is processed for path point distribution retrieval to obtain path point distribution data; According to the robot arm approach rate optimization data, the path point distribution data is matched with the staged motion parameters to obtain the staged motion parameters of the robot arm; The robot arm path posture control adjustment is performed according to the robot arm's staged motion parameters and path point distribution data to obtain the robot arm path posture control adjustment data.

10. A robotic arm control system for livestock status monitoring, characterized in that: The method for controlling a robotic arm for livestock status monitoring according to claim 1, wherein the robotic arm control system for livestock status monitoring comprises: The distribution density analysis module is used to monitor livestock activities in the livestock monitoring area through electronic monitoring equipment to obtain livestock activity monitoring images; calculate the distribution density of livestock activity monitoring images to obtain the distribution density of livestock types; A stress state probability estimation module is used to quantify the vibration of the approaching observation of the robot arm based on the activity distribution density of the livestock species to obtain the quantified data of the approaching vibration of the robot arm; and to estimate the probability of the livestock stress state based on the quantified data of the approaching vibration of the robot arm to obtain the probability data of the livestock stress state; The posture control instruction design module is used to match the approach rate of the robot arm according to the probability data of the livestock stress state, and obtain the approach rate matching data of the robot arm; the robot arm monitoring posture control instruction is designed based on the approach rate matching data of the robot arm and the livestock activity monitoring image, and the robot arm monitoring posture adjustment control instruction is obtained; The automatic control program design module is used to encode and process the robot arm monitoring posture adjustment control instructions to obtain the robot arm monitoring posture adjustment coding instructions; based on the robot arm monitoring posture adjustment coding instructions, the automatic control program is designed to obtain the robot arm monitoring posture control program, and the robot arm monitoring posture control program is embedded into the robot arm control center to execute the robot arm control method for livestock status monitoring.

Images