Full-automatic feeding control method and system for multiple areas of cattle and sheep

Abstract

The invention discloses a full-automatic feeding control method and system for multiple areas of cattle and sheep, and the method comprises the steps: obtaining basic information and ingestion behavior data of cattle and sheep in each area, carrying out the preprocessing, and carrying out the training through a support vector machine algorithm, and obtaining a cattle and sheep ingestion speed prediction model; a relation model of the feed physical parameters and the feeding speed is constructed, and the initial feeding speed of each area is obtained based on the relation model and the feed physical parameters of each area; acquiring real-time feeding behavior data and real-time satiation degree data of the cattle and the sheep in each area, and judging whether the average feeding speed of the cattle and the sheep meets the requirement or not; if the requirements are not met, based on the change condition of the feeding speed of the cattle and the sheep, a fuzzy control algorithm is adopted to generate a feeding speed regulation and control instruction; and obtaining feeding demand data of cattle and sheep in each region and current feed supply data, and calculating an optimal feed supply speed. Accurate feeding is achieved, the feeding efficiency is improved, and the problems that in traditional feeding, supply and demand are not matched, and the efficiency is low are effectively solved.

Description

Technical Field

[0001] The present invention belongs to the field of information technology, and particularly relates to a fully automatic feeding control method and system for multiple areas of cattle and sheep. Background Art

[0002] In the fully automatic feeding control system for multiple areas of cattle and sheep, there is a key technical problem, that is, how to accurately control the feeding speed of each area according to the real-time feeding speed of cattle and sheep and the feed characteristics. The difficulty of this problem lies in that the breeds, numbers, ages, and body sizes of cattle and sheep in different areas are different, and there are also significant differences in their feeding behaviors and feeding speeds. At the same time, the physical characteristics of different feeds, such as particle size, density, fluidity, etc., are also different, and these factors will affect the control of the feeding speed. If the feeding speed is not properly controlled, problems such as feed blockage, waste, or cattle and sheep scrambling for food will occur, seriously affecting the feeding efficiency.

[0003] The control of the feeding speed not only needs to consider the feeding behaviors of cattle and sheep and the feed characteristics, but also needs to adapt to the dynamic changes of the feeding behaviors of cattle and sheep. For example, during the feeding process, as the cattle and sheep become more satiated, their feeding speeds will gradually decrease, which requires real-time adjustment of the feeding speed to adapt to this change. At the same time, the coordination problem between different areas also needs to be considered, that is, how to ensure the feeding efficiency of cattle and sheep in each area while avoiding the problem of insufficient or excessive feeding in individual areas. This requires an intelligent collaborative control strategy to dynamically adjust the feeding speed of each area according to the real-time status of each area to achieve the optimal feeding effect of the entire system. Summary of the Invention

[0004] To solve the above technical problems, the present invention proposes a fully automatic feeding control method and system for multiple areas of cattle and sheep to solve the problems existing in the above prior art.

[0005] To achieve the above object, the present invention provides a fully automatic feeding control method for multiple areas of cattle and sheep, including:

[0006] Obtain the basic information and feeding behavior data of cattle and sheep in each area, and after preprocessing the feeding behavior data, use the support vector machine algorithm for training to obtain a cattle and sheep feeding speed prediction model;

[0007] Construct a relationship model between the feed physical parameters and the feeding speed, obtain the feed physical parameters of each area, and based on the relationship model and the feed physical parameters of each area, obtain the initial feeding speed of each area;

[0008] Obtain the real-time feeding behavior data and real-time satiety level data of cattle and sheep in each area, and determine whether the average feeding speed of cattle and sheep in each area meets the requirements; if the average feeding speed does not meet the requirements, then based on the change of the feeding speed of cattle and sheep in each area, use the fuzzy control algorithm to generate a feeding speed regulation instruction;

[0009] Obtain the feeding demand data and current feed supply data of cattle and sheep in each area, and calculate the optimal feeding speed of each area.

[0010] Optionally, the process of obtaining the cattle and sheep feeding speed prediction model by training with the support vector machine algorithm includes:

[0011] Obtain the basic information of cattle and sheep in each area, and construct a basic attribute database of cattle and sheep based on the basic information; collect the feeding behavior data of cattle and sheep in each area in real time and perform preprocessing, and divide the preprocessed feeding behavior data based on the basic attribute database of cattle and sheep to obtain several data subsets; use the support vector machine to train each data subset to obtain the corresponding cattle and sheep feeding speed prediction model and store the trained model parameters in the model library, where the preprocessing includes but is not limited to feeding time, feeding frequency, and feeding duration.

[0012] Optionally, input the feed physical parameters of each area into the relationship model to obtain the initial feeding speed of each area; if the feed quality changes, obtain the feed physical parameters in real time and update the initial feeding speed of each area; where the feed physical parameters include feed particle size, feed particle density, and fluidity parameter.

[0013] Optionally, the process of determining whether the average feeding speed of cattle and sheep in each area meets the requirements includes:

[0014] Obtain the real-time position information of cattle and sheep in each area, and obtain the number of cattle and sheep in each area based on the real-time position information; after obtaining the feeding behavior images of cattle and sheep in each area and performing preprocessing, extract the key features of the feeding behavior and input them into the trained convolutional neural network model to obtain the satiety level values of cattle and sheep in each area; obtain the average feeding speed of each area based on the satiety level data, the number of cattle and sheep, and the key features of the feeding behavior; obtain the preset normal feeding speed threshold based on the breed, age, and physiological state of cattle and sheep; judge whether the average feeding speed of each area meets the requirements based on the preset normal feeding speed threshold and the average feeding speed of each area.

[0015] Optionally, the process of generating a feeding speed regulation instruction includes:

[0016] Obtain the real-time feeding speed data of cattle and sheep in the area, input the real-time feeding speed data into the fuzzy control algorithm to obtain the feeding speed adjustment value; match the feeding speed adjustment value with the real-time feeding speed data, if they do not match, regenerate the feeding speed adjustment value through the fuzzy control algorithm; among them, the fuzzy control algorithm is continuously optimized based on the feeding speed control instructions and the real-time average feeding speed of each area.

[0017] Optionally, the process of calculating the optimal feeding speed of each area includes:

[0018] Obtain the feeding demand prediction result based on the historical feeding data and the current number of cattle and sheep, and calculate the optimal feeding speed of each area by using an optimization algorithm based on the feed supply data and the feeding demand prediction result; if the difference between the actual feeding amount in the area and the feeding prediction demand result exceeds the preset value, update the optimal feeding speed.

[0019] Optionally, the process of calculating the optimal feeding speed of each area by using an optimization algorithm based on the feed supply data and the feeding demand prediction result includes:

[0020] Obtain the historical data and real-time data of the feed supply and feeding demand of each area, obtain the feed supply of each area based on the historical data and real-time data of the feed supply of each area, and obtain the feeding demand of each area based on the historical data and real-time data of the feeding demand of each area; use the predicted feed supply and feeding demand of each area as constraints to establish a feed speed optimization model, and the objective function of the feed speed optimization model is to minimize the absolute value of the difference between the total feed supply and the total demand; use a heuristic algorithm to solve the feed speed optimization model to obtain the optimal feeding speed of each area.

[0021] Optionally, it further includes real-time obtaining the feeding speed and cattle and sheep feeding behavior data of each area, performing a correlation analysis on the feeding speed and cattle and sheep feeding behavior data to obtain the key factors for efficiency evaluation, evaluating the feeding efficiency of each area based on the key factors for efficiency evaluation, and if the feeding efficiency does not meet the preset requirements, optimizing the feeding management strategy according to the cattle and sheep feeding behavior and feed supply data in the area.

[0022] The present invention also provides a full-automatic feeding control system for multi-areas of cattle and sheep, including:

[0023] An eating speed prediction module for predicting the eating speed of cattle and sheep in each area;

[0024] An initial feeding speed determination module for determining the initial feeding speed of each area;

[0025] An eating speed change detection module for judging whether the eating speed of cattle and sheep in each area has changed;

[0026] A feeding speed dynamic adjustment module, which is used to adjust the feeding speed according to the change of the feeding speed of cattle and sheep.

[0027] A multi-region feeding coordination module, which is used to coordinate the feeding speeds between regions.

[0028] A feeding efficiency evaluation and optimization module, which is used to evaluate the feeding efficiency and optimize the feeding management strategy.

[0029] Compared with the prior art, the present invention has the following advantages and technical effects:

[0030] Based on the pre-established information of cattle and sheep and real-time foraging behavior data, the present invention uses the support vector machine algorithm to predict the feeding speed of cattle and sheep in each region. Combining with the physical characteristics of the feed, the initial feeding speed is determined. During the feeding process, the foraging behavior and satiety level of cattle and sheep are analyzed in real time, and the change of the feeding speed is judged through the convolutional neural network. If a change occurs, the fuzzy control algorithm is used to dynamically adjust the feeding speed. At the same time, the multi-agent cooperative control algorithm is used to coordinate the feeding in each region to avoid uneven supply. The present invention also evaluates the feeding efficiency through big data analysis, identifies the regions with low efficiency, and uses data mining technology to optimize the feeding strategy. The present invention realizes precise feeding, improves the feeding efficiency, effectively solves the problems of mismatch between supply and demand and low efficiency in traditional feeding, and provides an advanced intelligent solution for the cattle and sheep breeding industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The drawings forming a part of this application are used to provide a further understanding of this application. The schematic embodiments and descriptions thereof of this application are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0032] Figure 1 is the method flow chart of the embodiment of the present invention;

[0033] Figure 2 is the system schematic diagram of the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will refer to the drawings and combine the embodiments to detail this application.

[0035] It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0036] Embodiment 1

[0037] As Figure 1-2As shown in the figure, in this embodiment, a full-automatic feeding control method and system for multiple areas of cattle and sheep are provided, including:

[0038] Obtain the basic information and feeding behavior data of cattle and sheep in each area, and after preprocessing the feeding behavior data, use the support vector machine algorithm for training to obtain a prediction model for the feeding speed of cattle and sheep;

[0039] In some specific embodiments, the process of using the support vector machine algorithm for training to obtain a prediction model for the feeding speed of cattle and sheep includes:

[0040] Obtain the basic information of cattle and sheep in each area, and build a basic attribute database of cattle and sheep based on the basic information; collect the feeding behavior data of cattle and sheep in each area in real time and preprocess it, and divide the preprocessed feeding behavior data based on the basic attribute database of cattle and sheep to obtain several data subsets; use the support vector machine to train each data subset to obtain the corresponding prediction model for the feeding speed of cattle and sheep and store the trained model parameters in the model library, where the preprocessing includes but is not limited to feeding time, feeding frequency, and feeding duration.

[0041] Furthermore, according to the pre-established information on the breeds, quantities, ages, and body sizes of cattle and sheep in each area, build a basic attribute database of cattle and sheep for storing and managing the basic information of cattle and sheep. By wearing sensor devices on cattle and sheep, collect the feeding behavior data of cattle and sheep in each area in real time, including feeding time, frequency, duration, etc., and transmit the collected data to the data processing center. Preprocess the collected feeding behavior data of cattle and sheep, including data cleaning, normalization, etc., remove noise data, and convert the data into a format suitable for training by machine learning algorithms. According to the information such as breed, age, and body size in the basic attribute database of cattle and sheep, divide the preprocessed feeding behavior data according to different attribute dimensions to form multiple data subsets. For each data subset, use the support vector machine algorithm for training to obtain a prediction model for the feeding speed of cattle and sheep under the corresponding area and attribute combination, and store the trained model parameters in the model library. When it is necessary to predict the feeding speed of cattle and sheep in a certain area, according to the breed, quantity, age, and body size attributes of cattle and sheep in this area, obtain the corresponding prediction model for the feeding speed from the model library, and input the real-time collected feeding behavior data of cattle and sheep in this area into the model for prediction. Visualize the predicted feeding speed information of cattle and sheep in each area to generate intuitive charts and reports, which is convenient for managers to timely understand the feeding situation of cattle and sheep and dynamically adjust the feeding plan for cattle and sheep according to the prediction results.

[0042] Specifically, the basic attribute database of cattle and sheep is used to store and manage the basic information of cattle and sheep, such as breed, quantity, age, body size, etc. This database can be imagined as an Excel table, where each row represents a cow or a sheep, and each column represents an attribute. For example, it can record a Simmental cow numbered 001, 3 years old, with a large body size, located in area A. At the same time, it can also record a Hu sheep numbered 002, 1 year old, with a medium body size, also located in area A. Establishing this database is to provide basic data for subsequent analysis and prediction of the feeding speed of cattle and sheep. The sensor devices worn on cattle and sheep can collect their feeding behavior data in real time. For example, the sensor can record the timestamp when cattle and sheep start feeding, the duration of feeding, and the number of chewing times per unit time, etc. Suppose the Simmental cow numbered 001 starts feeding at 8:00 in the morning, lasts for 30 minutes, and chews 20 times per minute on average during this period. These data will be wirelessly transmitted to the data processing center. The collected data may have noise and missing values, and data preprocessing is required. For example, the sensor may occasionally malfunction, resulting in abnormal data, and these noise data need to be identified and removed. In addition, the data ranges collected by different sensors may be different, and normalization processing is required to convert the data to a unified scale, such as between 0 and 1. Suppose the sensor of the cow numbered 001 malfunctions at 8:15, and the recorded number of chewing times is 0, which is obviously abnormal data and needs to be removed. According to the basic attribute database of cattle and sheep, the preprocessed feeding behavior data can be divided into multiple data subsets according to different attribute dimensions. For example, all the feeding data of Simmental cows in area A can be divided into one subset, and all the feeding data of Hu sheep in area A can be divided into another subset. The purpose of doing this is to establish a more accurate feeding speed prediction model for cattle and sheep of different breeds, ages, and body sizes. For each data subset, the support vector machine algorithm can be used for training to obtain the feeding speed prediction model of cattle and sheep under the corresponding area and attribute combination. The support vector machine algorithm can be understood as finding an optimal hyperplane in the data to separate samples of different categories. In this example, "fast feeding" and "slow feeding" can be used as two categories, and the support vector machine algorithm is used to find a hyperplane to distinguish them. After training, the model parameters will be stored in the model library. When it is necessary to predict the feeding speed of cattle and sheep in a certain area, first, according to the attributes of cattle and sheep in this area, such as breed, age, body size, etc., the corresponding feeding speed prediction model is obtained from the model library. For example, if it is necessary to predict the feeding speed of a 3-year-old Simmental cow in area A, the feeding speed prediction model of Simmental cows in area A needs to be found from the model library. Then, the real-time collected feeding behavior data of this cow, such as feeding time, chewing frequency, etc., are input into the model for prediction. The prediction results can be visually displayed, such as generating charts or reports to show the feeding speed of cattle and sheep in each area.For example, a bar chart can be generated to show the average feeding speeds of different breeds of cattle and sheep in area A. Managers can timely understand the feeding situation of cattle and sheep based on these visualizations and dynamically adjust the feeding plans for cattle and sheep according to the prediction results. For example, if it is predicted that the feeding speed of cattle and sheep in a certain area is slow, the nutritional content of the feed can be considered to be increased or the feeding time can be adjusted. Doing so can improve the growth rate and production efficiency of cattle and sheep.

[0043] Construct a relationship model between the physical parameters of the feed and the feeding speed, obtain the physical parameters of the feed in each area, and based on the relationship model and the physical parameters of the feed in each area, obtain the initial feeding speed of each area;

[0044] In some specific embodiments, the physical parameters of the feed in each area are input into the relationship model to obtain the initial feeding speed of each area; if the feed quality changes, the physical parameters of the feed are obtained in real time to update the initial feeding speed of each area; wherein, the physical parameters of the feed include the feed particle size, feed particle density, and fluidity parameter.

[0045] Furthermore, obtain the feed samples currently used in each area, and use the feed particle image analysis technology to obtain the average particle size parameter of the feed particles. Measure the mass of the feed per unit volume using the weighing method to obtain the feed density parameter. Place a certain amount of feed in a funnel and measure the time it takes for it to completely flow out of the funnel, and calculate to obtain the fluidity parameter of the feed. According to the pre-established mathematical relationship model between the physical characteristics of the feed and the feeding speed, substitute the feed particle size, density, and fluidity parameters obtained in the above steps, and by solving the model equation, determine the initial feeding speed control parameters for each area. If the feed quality changes during the actual production process, re-obtain the physical characteristics parameters of the feed and dynamically adjust the feeding speed parameters according to the relationship model to adapt to the new feed quality. Based on the initial feeding speed, combined with the actual production feedback data of each area, through the proportional integral derivative (PID) control algorithm, dynamically optimize and adjust the feeding speed of each area to achieve precise control of the output. Record and store the optimized feeding speed parameters for each area as the basis for subsequent production scheduling, and use them for the iterative update and optimization of the feeding speed relationship model.

[0046] Specifically, the feed pellet image analysis technology can be used to obtain the average particle size parameter of feed pellets. For example, images of feed samples used in different regions can be collected, and image processing software can be used to identify and measure the diameters of all feed pellets in the image, and then the average value can be calculated to obtain the average particle size. This can help understand whether the particle size of the feed is uniform and whether it is suitable for cattle and sheep of different ages and breeds. The particle size affects the feeding speed and digestion and absorption efficiency of cattle and sheep. The weighing method can be used to measure the mass of the feed per unit volume to obtain the feed density parameter. For example, a container with a known volume can be filled with feed and weighed, and then the total mass of the feed is divided by the volume of the container to obtain the density of the feed. The feed density is an important parameter affecting the feeding speed. Too large or too small density will affect the accuracy of feeding. For example, feed with too large density is prone to clogging during transportation, while feed with too small density is prone to waste. By measuring the time for a certain amount of feed to completely flow out of the funnel, the flowability parameter of the feed can be calculated. For example, 1 kg of feed can be poured into a standard funnel, and the time required for the feed to completely flow out of the funnel is recorded. The shorter the time, the better the flowability of the feed. The flowability of the feed is crucial for the stability and efficiency of the automatic feeding system. Feed with poor flowability is prone to uneven feeding or clogging. The mathematical model of the relationship between feed physical properties and feeding speed can be used to determine the initial feeding speed control parameters. This model can calculate the appropriate initial feeding speed based on parameters such as the particle size, density, and flowability of the feed. For example, for feed with smaller particles, larger density, and better flowability, the model may recommend a higher initial feeding speed. Establishing such a model can improve the feeding efficiency and reduce feed waste. During the actual production process, if the feed quality changes, such as changing to different batches of feed, it is necessary to re-obtain the feed physical property parameters and dynamically adjust the feeding speed parameters according to the relationship model. For example, if the new batch of feed has larger particles, smaller density, and poorer flowability, then the feeding speed needs to be reduced to ensure the stability and uniformity of feeding. This can avoid feeding problems caused by changes in feed quality and ensure the normal feeding of cattle and sheep. The proportional-integral-derivative (PID) control algorithm can dynamically optimize and adjust the feeding speed of each region according to the actual production feedback data of each region to achieve precise control of production. For example, if the feeding speed of cattle and sheep in a certain region is lower than expected, resulting in a decrease in production, the PID controller will automatically increase the feeding speed of this region according to the production deviation until the production reaches the target value. The PID control algorithm can automatically adapt to various changes and keep the feeding speed in the best state at all times. The optimized feeding speed parameters of each region will be recorded and stored as the basis for subsequent production scheduling and used for iterative update and optimization of the feeding speed relationship model. For example, the feeding efficiency and production data under different parameters can be analyzed to find the best combination of feeding speed parameters and use it to guide future production scheduling.Continuous recording and analysis can help continuously improve the feeding strategy and increase production efficiency.

[0047] Obtain the real-time feeding behavior data and real-time satiety degree data of cattle and sheep in each area, and judge whether the average feeding speed of cattle and sheep in each area meets the requirements;

[0048] In some specific embodiments, the process of judging whether the average feeding speed of cattle and sheep in each area meets the requirements includes:

[0049] Obtain the real-time position information of cattle and sheep in each area, and obtain the number of cattle and sheep in each area based on the real-time position information; after obtaining and preprocessing the feeding behavior images of cattle and sheep in each area, extract the key feeding behavior features and input them into the pre-trained convolutional neural network model to obtain the satiety degree values of cattle and sheep in each area; obtain the average feeding speed of each area based on the satiety degree data, the number of cattle and sheep, and the key feeding behavior features; obtain the preset normal feeding speed threshold based on the breed, age, and physiological state of cattle and sheep; judge whether the average feeding speed of each area meets the requirements based on the preset normal feeding speed threshold and the average feeding speed of each area.

[0050] Furthermore, according to the real-time position information of cattle and sheep in each area, obtain the number and distribution of cattle and sheep in each area. For the cattle and sheep in each area, collect the feeding behavior images and video data of cattle and sheep through video monitoring equipment. Preprocess the collected feeding behavior images and video data, and extract key feature parameters, including feeding frequency, feeding duration, etc. Input the extracted feeding behavior feature parameters into the pre-trained convolutional neural network model, and obtain the satiety degree values of cattle and sheep in each area through model inference and calculation. Calculate the average feeding speed of each area according to the number of cattle and sheep in each area, the feeding behavior characteristics, and the satiety degree values. Compare the average feeding speed of each area with the preset normal feeding speed threshold to judge whether there is an obvious change in the feeding speed of cattle and sheep in each area. If the average feeding speed of a certain area is significantly lower than the normal threshold, trigger a warning message to prompt the breeding personnel to pay attention to the feeding situation of cattle and sheep in this area and adjust the feeding plan as needed.

[0051] Specifically, feed supply is a complex process that requires precise control to ensure the healthy growth and production efficiency of cattle and sheep. First, it is necessary to understand the real-time location information of cattle and sheep. A positioning system based on radio frequency identification technology (RFID) can be adopted. Electronic tags are installed on cattle and sheep, and readers are deployed in the breeding area to obtain the activity range and location information of each cattle and sheep in real time. In this way, the number of cattle and sheep in each area and their distribution can be accurately counted. For example, there are 30 cows in Area A and 25 sheep in Area B, and so on. After knowing the number of cattle and sheep, it is also necessary to master their feeding behaviors. High-definition video monitoring devices are installed in each area to capture the feeding behavior images and video data of cattle and sheep in real time. After these data are preprocessed, key feature parameters can be extracted. For example, the number of times cattle and sheep lower their heads to feed per minute, that is, the feeding frequency, can be counted; the duration of each feeding, that is, the feeding duration, can also be counted. For example, if a cow lowers its head to feed 10 times in one minute and the average duration of each feeding is 5 seconds, these data can be used as the feeding behavior feature parameters of the cow. These extracted feeding behavior feature parameters are then input into a pre-trained convolutional neural network (CNN) model. This model has previously learned a large amount of image and video data of cattle and sheep feeding behaviors and can infer the satiety level of cattle and sheep based on the input feature parameters. The satiety level can be represented as a value between 0 and 1. For example, 0.8 means that the cattle and sheep are relatively full, and 0.2 means that the cattle and sheep are still very hungry. Suppose the average satiety level of 30 cows in Area A is 0.6, and the average satiety level of 25 sheep in Area B is 0.7. Combining the number of cattle and sheep, feeding behavior characteristics, and satiety level, the average feeding speed of each area can be calculated. For example, if the total feed intake of 30 cows in Area A is X kg / hour, then the average feeding speed is X / 30 kg / hour / head. Similarly, the average feeding speed of sheep in Area B can be calculated. To judge whether the feeding situation of cattle and sheep is normal, it is necessary to compare the calculated average feeding speed with a preset normal feeding speed threshold. This threshold is determined according to factors such as the breed, age, and physiological state of cattle and sheep. Suppose the normal feeding speed threshold for beef cattle is Y kg / hour / head. If the average feeding speed of cows in Area A is significantly lower than Y, the system will trigger a warning message to notify the breeding staff to pay attention to the feeding situation of cows in Area A. The breeding staff can adjust the feeding plan according to the actual situation, such as increasing the feed supply, changing the feed type, and checking the health status of cattle and sheep. By monitoring and analyzing the real-time location, feeding behavior, satiety level, and feeding speed of cattle and sheep, abnormal situations can be detected in a timely manner, and corresponding measures can be taken to ensure the healthy growth and production efficiency of cattle and sheep. This data-based intelligent feeding management method can effectively improve the breeding efficiency and reduce the production cost. Record and store the optimized feed supply speed parameters to provide data support for subsequent production scheduling and to be used for the iterative update and optimization of the feed supply speed relationship model, forming a data-driven virtuous cycle and continuously improving the accuracy and efficiency of feeding management.

[0052] If the average feeding speed does not meet the requirements, based on the changes in the feeding speeds of cattle and sheep in each area, a fuzzy control algorithm is used to generate a feeding speed regulation instruction;

[0053] In some specific embodiments, the process of generating a feeding speed regulation instruction includes:

[0054] Obtain the real-time feeding speed data of cattle and sheep in the area, input the real-time feeding speed data into the fuzzy control algorithm to obtain a feeding speed adjustment value; match the feeding speed adjustment value with the real-time feeding speed data, if they do not match, then regenerate the feeding speed adjustment value through the fuzzy control algorithm; wherein, the fuzzy control algorithm is continuously optimized based on the feeding speed regulation instructions and the real-time average feeding speed in each area.

[0055] Furthermore, according to the area where the cattle and sheep are located, obtain the real-time feeding speed data of the cattle and sheep in that area. Use the obtained cattle and sheep feeding speed data as the input of the fuzzy control algorithm, and determine the adjustment value of the feeding speed in that area through fuzzy reasoning. Dynamically adjust the actual feeding speed in that area according to the feeding speed adjustment value output by the fuzzy control algorithm. Determine whether the adjusted feeding speed in the area matches the feeding speed of the cattle and sheep. If they do not match, return to the second step to continue the adjustment until they match. Continuously monitor the changes in the feeding speeds of cattle and sheep in each area. When the speed in a certain area changes, trigger the dynamic adjustment process of the feeding speed in that area. Real-time feedback the feeding speed adjustment situation in each area to the data analysis module, and optimize the rule base of the fuzzy control algorithm through big data analysis. Continuously and dynamically adjust the feeding speeds in each area according to the optimized fuzzy control algorithm to achieve real-time matching of the feeding speed and the feeding speed of the cattle and sheep.

[0056] Specifically, the feeding speed of cattle and sheep varies with factors such as time, individual differences, and feed types. To ensure the healthy growth of cattle and sheep and the effective utilization of feed, it is necessary to dynamically adjust the feeding speed. This is like a buffet, where dishes are replenished in real-time according to the dining speed of guests to avoid food waste or insufficient supply. First, it is necessary to obtain the feeding speed data of cattle and sheep in real-time. Suppose through cameras and image recognition technology, information such as the feeding frequency and chewing times of each cattle and sheep can be monitored, and the feed intake per minute can be calculated. For example, the average feeding speed of 10 cattle and sheep in Area 1 is 200 grams per minute. Next, the obtained feeding speed data is input into the fuzzy control algorithm. The core of the fuzzy control algorithm is to simulate human experience for control. Unlike traditional control algorithms that require precise mathematical models, it conducts reasoning through fuzzy rules. For example, the following rules can be set: If the feeding speed is very fast, then reduce the feeding speed; if the feeding speed is very slow, then increase the feeding speed; if the feeding speed is moderate, then maintain the current feeding speed. The process of fuzzy reasoning can be understood as a weighted average process. Suppose the current feeding speed in Area 1 is 300 grams per minute, and the feeding speed of cattle and sheep is 200 grams per minute. According to the fuzzy rules, the feeding speed needs to be reduced. Suppose the membership degrees of the three fuzzy sets of "very fast", "moderate", and "very slow" are 0.2, 0.5, and 0.3 respectively, and the corresponding feeding speed adjustment values are -50, 0, and +50. Then the final feeding speed adjustment value is 0.2*(-50)+0.5*0+0.3*50 = 5 grams per minute. Then, according to the adjustment value output by the fuzzy control algorithm, the actual feeding speed in this area is dynamically adjusted. In the above example, the feeding speed in Area 1 will be adjusted to 300 + 5 = 305 grams per minute. To ensure the matching of the feeding speed and the feeding speed of cattle and sheep, continuous monitoring and adjustment are required. For example, after a period of time, if the feeding speed of cattle and sheep in Area 1 becomes 350 grams per minute, while the feeding speed is still 305 grams per minute, then fuzzy reasoning and adjustment are carried out again until the two match. This is like a continuous fine-tuning process, making the feeding speed always follow the change of the feeding speed of cattle and sheep. In addition, the adjustment situation of the feeding speed in each area is fed back to the data analysis module in real-time. Through big data analysis, the deficiencies of the fuzzy rules can be discovered and optimized. For example, if it is found that the feeding speed of cattle and sheep in a certain area is always low during a specific period, then the fuzzy rules can be adjusted for this period to increase the feeding speed, so as to better meet the needs of cattle and sheep. This is like optimizing the dish supply in a restaurant based on historical data, predicting the needs of guests, and making preparations in advance. Through the above steps, the real-time matching of the feeding speed and the feeding speed of cattle and sheep can be achieved, improving the feed utilization rate and promoting the healthy growth of cattle and sheep. This is like an intelligent feeding system that can automatically adjust the feeding speed according to the needs of cattle and sheep to achieve precise feeding.

[0057] Exemplarily, the process of continuously and dynamically adjusting the feeding speed of each area according to the optimized fuzzy control algorithm to achieve real-time matching of the feeding speed and the feeding speed of cattle and sheep includes: obtaining the feeding speed data of each area according to the real-time feeding speed of cattle and sheep in different areas; inputting the obtained feeding speed data of each area into a preset fuzzy control algorithm model for processing; the fuzzy control algorithm model dynamically calculates the optimized feeding speed parameters of each area through fuzzy inference rules according to the input feeding speed data; judging whether the difference between the calculated feeding speed parameters of each area and the current actual feeding speed parameters exceeds a preset threshold; if it exceeds the preset threshold, taking the optimized feeding speed parameters as the new feeding speed setting value and adjusting the feeding speed of each area; if it does not exceed the preset threshold, maintaining the current actual feeding speed of each area unchanged; continuously and circularly executing the above steps, dynamically and optimally matching the feeding speed of each area according to the change of the feeding speed of cattle and sheep, and achieving the dynamic balance of the feeding speed and the feeding speed.

[0058] Obtain the feeding demand data of cattle and sheep in each area and the current feed supply data, and calculate the optimal feeding speed of each area.

[0059] In some specific embodiments, the process of calculating the optimal feeding speed of each area includes:

[0060] Obtain the feeding demand prediction result based on the historical feeding data and the current number of cattle and sheep, and calculate the optimal feeding speed of each area by using an optimization algorithm based on the feed supply data and the feeding demand prediction result; if the difference between the actual feeding amount in the area and the feeding prediction demand result exceeds a preset value, update the optimal feeding speed.

[0061] Furthermore, obtain the feeding demand data of cattle and sheep in each area and the current feed supply data as the input of the multi-agent collaborative control algorithm. Through the feeding demand prediction model, combined with the historical feeding data and the current number of cattle and sheep, predict the change trend of the feeding demand in each area in the next period of time. According to the feed supply data and the feeding demand prediction result, use an optimization algorithm to calculate the optimal feeding speed of each area to ensure that the total feed supply meets the total feeding demand. While optimizing the feeding speed, through the negotiation mechanism between agents, coordinate the feeding speed between each area to avoid the problem of uneven feeding. Monitor the real-time feeding data of each area. If it is found that the difference between the actual feeding amount in a certain area and the predicted demand amount is large, trigger the dynamic adjustment of the feeding speed. Transmit the adjusted feeding speed data to the feed supply system to control the actual feeding speed of each area and ensure the dynamic matching of the feed supply and the feeding demand. Continuously track the feeding status of each area and the change of the number of cattle and sheep, and regularly update the parameters of the feeding demand prediction model and the collaborative control algorithm to maintain the optimized performance of the system.

[0062] Specifically, in modern animal husbandry, it is crucial to ensure a dynamic match between the feeding requirements of cattle and sheep and the feed supply. First, by collecting real-time feeding requirement data and current feed supply data for cattle and sheep in each region, it can be effectively input into the multi-agent collaborative control algorithm. For example, assume that there are currently 500 cows in a certain region, and each cow requires an average of 2 kg of feed per day. Then the feed demand for this region in one day is 1000 kg. Next, using the feeding requirement prediction model, combined with historical feeding data and the current number of cattle and sheep, the changing trend of feeding requirements in the future period can be predicted. For example, if it is predicted that the number of cattle and sheep in this region will increase to 600 in the next week, then the predicted feed demand will increase to 1200 kg / day. This prediction helps decision-makers adjust the feeding strategy in advance to avoid feed shortages or surpluses. According to the feed supply data and the feeding requirement prediction results, an optimization algorithm is used to calculate the optimal feeding speed for each region. For example, if the current feed supply is 1100 kg / day and the demand is 1200 kg / day, the optimization algorithm may suggest increasing the feeding speed to make up for the 100 kg gap. While optimizing the feeding speed, through the negotiation mechanism among agents, the feeding speeds among different regions can be coordinated to avoid uneven feeding. For example, if the feed supply in adjacent region A is in surplus while that in region B is insufficient, the agents can coordinate to transfer some feed from region A to region B. Monitor the real-time feeding data of each region. If it is found that the actual feeding amount in a certain region varies greatly from the predicted demand, for example, the actual consumption in region C suddenly increases to 1300 kg / day, which may be due to newly added cattle and sheep or other factors. At this time, it will trigger a dynamic adjustment of the feeding speed. The adjusted feeding speed data is transmitted to the feed supply system to control the actual feeding speed of each region and ensure a dynamic match between the feed supply and the feeding requirements. Continuously track the feeding status of each region and the changes in the number of cattle and sheep, and regularly update the parameters of the feeding requirement prediction model and the collaborative control algorithm to maintain the optimized performance of the system. This continuous monitoring and updating ensure that the entire system can flexibly respond to changes in the number of cattle and sheep and fluctuations in feeding requirements, thus achieving efficient and sustainable feed management.

[0063] In some specific embodiments, the process of using an optimization algorithm to calculate the optimal feeding speed for each region based on the feed supply data and the feeding requirement prediction results includes:

[0064] Obtain the historical data and real-time data of the feed supply and feeding demand in each region, obtain the feed supply in each region based on the historical data and real-time data of the feed supply in each region, and obtain the feeding demand in each region based on the historical data and real-time data of the feeding demand in each region; use the predicted feed supply and feeding demand in each region as constraint conditions to establish a feeding speed optimization model, and the objective function of the feeding speed optimization model is to minimize the absolute value of the difference between the total feed supply and the total demand; use a heuristic algorithm to solve the feeding speed optimization model to obtain the optimal feeding speed in each region.

[0065] Further, obtain the historical data and real-time data of the feed supply in each region, establish a supply prediction model based on the historical data, and calibrate the model in combination with the real-time data to predict the feed supply in each region in the future for a period of time. Obtain the historical data and real-time data of the feeding demand in each region, establish a demand prediction model based on the historical data, and calibrate the model in combination with the real-time data to predict the feeding demand in each region in the future for a period of time. Use the predicted feed supply and feeding demand in each region as constraint conditions to establish a feeding speed optimization model, and the objective function is to minimize the absolute value of the difference between the total feed supply and the total demand. Use a heuristic optimization algorithm to solve the above optimization model to obtain the optimal feeding speed in each region, so that while meeting the feeding demand in each region, the total feed supply is as close as possible to the total demand. Set the feeding equipment parameters in each region according to the optimization results, control the actual feeding speed in each region, and monitor the feeding amount and feeding quantity in each region in real time. If it is monitored that the actual feeding amount in a certain region is significantly different from the optimization result, trigger an alarm, analyze the reason and take measures for adjustment, such as adjusting the feeding speed in that region or coordinating the feeding in other regions. Regularly evaluate the effect of the optimization model, correct and improve the model according to the actual operation data, and continuously improve the accuracy and stability of the feeding optimization to ensure long-term satisfaction of the feeding demand.

[0066] In some specific embodiments, the method further includes obtaining the feeding speed and the feeding behavior data of cattle and sheep in each region in real time, performing a correlation analysis on the feeding speed and the feeding behavior data of cattle and sheep to obtain the key factors for efficiency evaluation, evaluating the feeding efficiency in each region based on the key factors for efficiency evaluation, and if the feeding efficiency does not meet the preset requirements, optimizing the feeding management strategy according to the feeding behavior and feed supply data of cattle and sheep in the region.

[0067] Furthermore, obtain the real-time feeding speed data and the feeding behavior data of cattle and sheep in each area, and transmit the data to the big data analysis platform for processing. Adopt data cleaning and preprocessing techniques to denoise, normalize and other processes on the obtained original data to improve the data quality. Through the correlation analysis algorithm, analyze the correlation between the feeding speed and the feeding behavior of cattle and sheep to obtain the key factors affecting the feeding efficiency. According to the key factors, establish a feeding efficiency evaluation model, and use the support vector machine or random forest algorithm for model training and optimization. Input the real-time data of each area into the trained evaluation model to conduct real-time evaluation and scoring of the feeding efficiency of each area. According to the evaluation results, identify the areas with low feeding efficiency, generate a heat map of efficiency distribution through visualization technology, and intuitively display the low-efficiency areas. For the identified low-efficiency areas, combine the regional characteristics and the analysis results of key factors to formulate targeted feeding optimization strategies, and achieve precise feeding through an automated control system to improve the overall feeding efficiency.

[0068] Furthermore, the processing process for areas with low feeding efficiency includes: obtaining the historical data of the feeding behavior of cattle and sheep and the feed supply in areas with low feeding efficiency, and preprocessing the data, including operations such as data cleaning, data integration, and data transformation, to obtain a dataset suitable for data mining and analysis. According to attributes such as cattle and sheep breeds and feed quality, use the K-means clustering algorithm to conduct clustering analysis on the dataset, divide the data with similar feeding behavior and feed supply characteristics into the same category, and identify the differences in feeding efficiency among different regions and different cattle and sheep breeds. For each category, use the association rule mining algorithm to analyze the association relationship between the feeding behavior of cattle and sheep and the feed supply, and find out the key factors affecting the feeding efficiency, such as feeding time, feeding frequency, feed type, feed dosage, etc. According to the results of association rule mining, combined with expert knowledge and feeding experience, determine the specific measures to optimize the feeding management strategy, such as adjusting the feed ratio, changing the feeding time and frequency, improving the feeding environment, etc. Apply the optimized feeding management strategy to the areas with low feeding efficiency, and use the information management system to monitor the feeding behavior of cattle and sheep and the feed supply situation in real time, and dynamically adjust the strategy parameters according to the feedback data. Continuously track the optimized feeding efficiency indicators, evaluate the effect of the optimization measures by comparing the data before and after optimization, and form a replicable and popularizable feeding management optimization model. Establish a feeding efficiency warning mechanism. When the feeding efficiency indicators in a certain area show abnormal fluctuations, give a timely warning and diagnose the reasons, and take targeted optimization and adjustment measures to ensure the stable improvement of the overall feeding efficiency.

[0069] Specifically, obtaining the historical data of the areas with low feeding efficiency is to deeply analyze the reasons for the inefficiency. For example, the feeding efficiency of Area 1 is lower than the average level. It is necessary to collect the feeding behavior data of cattle and sheep in this area for a period of time, such as one month, including the feed intake, feeding time, feeding frequency of each cattle and sheep, as well as the feed supply data, such as feed type, feeding amount, feeding time, etc. These data will be used for subsequent data mining and analysis. In the data preprocessing stage, data cleaning is carried out first. For example, if it is found that the data of some days is missing or abnormal, which may be caused by equipment failure, it needs to be excluded or filled. Then data integration is carried out to integrate data from different sources. For example, the cattle and sheep feeding data and feed supply data are merged. Finally, data transformation is carried out. For example, the feeding time is converted into time periods, such as morning, noon, and evening. After these processes, a clean, complete, and consistent data set is obtained. Using the K-means clustering algorithm, the data can be divided into different categories. For example, according to the breeds of cattle and sheep and the quality of feed, the data can be divided into different categories such as beef cattle, dairy cows, and goats. The cattle and sheep in each category have similar feeding behaviors and feed supply characteristics. In this way, more precise feeding strategies can be formulated for different categories. Suppose through cluster analysis, it is found that the feed intake of beef cattle is lower at night, while the feed intake of dairy cows is higher in the early morning. Association rule mining can discover the association relationship between the feeding behavior of cattle and sheep and the feed supply. For example, it can be found that there is a strong association between "Feed Type A" and "Feeding Time in the Morning", indicating that cattle and sheep prefer to feed on Feed A in the morning. Another example is that it is found that the feed intake of beef cattle decreases and the water intake increases in hot weather, which indicates that high temperature affects the feeding behavior of beef cattle. By analyzing these association rules, the key factors affecting the feeding efficiency can be identified. According to the results of association rule mining and expert experience, optimization strategies can be formulated. For example, for the problem of low feed intake of beef cattle at night, the feed ratio can be adjusted to increase the feed with better palatability, or part of the feed feeding time can be adjusted to night. For the problem of decreased feed intake of beef cattle in hot weather, some heatstroke prevention and cooling components can be added to the feed and sufficient drinking water can be provided. Apply the optimization strategy to actual production and conduct continuous monitoring and adjustment. For example, apply the adjusted feed ratio to the feeding of beef cattle in Area 1 and monitor the feeding situation of cattle and sheep in real time through the information management system. If it is found that the feed intake still has not increased, the strategy needs to be further adjusted, such as changing the feeding frequency or improving the feeding environment. Continuously track the optimized feeding efficiency indicators. For example, after one month, the average feed intake of beef cattle in Area 1 has increased by 10%, and the feeding efficiency has also been significantly improved. This shows that the optimization strategy is effective and can be promoted to other similar areas. Establish an early warning mechanism to detect and solve problems in a timely manner. For example, when the system monitors that the feed intake of cattle and sheep in a certain area suddenly decreases, it will issue an early warning to remind the management personnel to check the reasons in time. It may be a problem with the feed or the cattle and sheep are sick.Greater losses can be avoided through timely intervention.

[0070] As Figure 2 shown, this embodiment also provides a fully automatic feeding control system for multiple areas of cattle and sheep, including:

[0071] An eating speed prediction module for predicting the eating speed of cattle and sheep in each area;

[0072] An initial feeding speed determination module for determining the initial feeding speed of each area;

[0073] An eating speed change detection module for determining whether the eating speed of cattle and sheep in each area has changed;

[0074] A feeding speed dynamic adjustment module for adjusting the feeding speed according to the change in the eating speed of cattle and sheep;

[0075] A multi-area feeding coordination module for coordinating the feeding speeds between different areas;

[0076] A feeding efficiency evaluation and optimization module for evaluating the feeding efficiency and optimizing the feeding management strategy.

[0077] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in the present application should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A fully automatic feeding control method for multiple areas of cattle and sheep, characterized in that: The following steps are involved: Obtain the basic information and feeding behavior data of cattle and sheep in each area, and after preprocessing the feeding behavior data, use the support vector machine algorithm to train and obtain the cattle and sheep feeding speed prediction model; Constructing a relationship model between feed physical parameters and feeding speed, obtaining feed physical parameters of each area, and obtaining an initial feeding speed of each area based on the relationship model and the feed physical parameters of each area; Obtain the real-time feeding behavior data and real-time satiety data of cattle and sheep in each area to determine whether the average feeding speed of cattle and sheep in each area meets the requirements; if the average feeding speed does not meet the requirements, a fuzzy control algorithm is used to generate a feeding speed control instruction based on the changes in the feeding speed of cattle and sheep in each area; Obtain the feeding demand data and current feed supply data of cattle and sheep in each area, and calculate the optimal feeding speed in each area.

2. The method according to claim 1, characterized in that: The process of using the support vector machine algorithm to train and obtain a prediction model for the feeding speed of cattle and sheep includes: The basic information of cattle and sheep in each area is obtained, and a cattle and sheep basic attribute database is constructed based on the basic information; the feeding behavior data of cattle and sheep in each area are collected in real time and preprocessed, and the preprocessed feeding behavior data is divided based on the cattle and sheep basic attribute database to obtain a number of data subsets; each data subset is trained using a support vector machine to obtain a corresponding cattle and sheep feeding speed prediction model and the trained model parameters are stored in a model library, wherein the preprocessing includes but is not limited to feeding time, feeding frequency, and feeding duration.

3. The method according to claim 1, characterized in that The physical parameters of the feed in each area are input into the relationship model to obtain the initial feeding speed of each area; if the feed quality changes, the physical parameters of the feed are obtained in real time to update the initial feeding speed of each area; wherein the physical parameters of the feed include feed particle size, feed particle density, and fluidity parameters.

4. The method according to claim 1, characterized in that: The process of judging whether the average feeding speed of cattle and sheep in each area meets the requirements includes: The real-time location information of cattle and sheep in each area is obtained, and the number of cattle and sheep in each area is obtained based on the real-time location information; the feeding behavior images of cattle in each area are obtained and preprocessed, and then the key features of the feeding behavior are extracted and input into the trained convolutional neural network model to obtain the satiety values ​​of cattle and sheep in each area; the average eating speed of each area is obtained based on the satiety data, the number of cattle and sheep, and the key features of the feeding behavior; the preset normal eating speed threshold is obtained based on the breed, age and physiological state of the cattle and sheep; based on the preset normal eating speed threshold and the average eating speed of each area, it is judged whether the average eating speed of each area meets the requirements.

5. The method according to claim 1, characterized in that The process of generating feed speed control instructions includes: The real-time feeding speed data of cattle and sheep in the area are obtained, and the real-time feeding speed data are input into the fuzzy control algorithm to obtain the feeding speed adjustment value; the feeding speed adjustment value is matched with the real-time feeding speed data. If there is a mismatch, the feeding speed adjustment value is regenerated through the fuzzy control algorithm; wherein the fuzzy control algorithm is continuously optimized based on the feeding speed control instructions of each area and the real-time average feeding speed.

6. The method according to claim 1, characterized in that The process of calculating the optimal feed rate for each zone includes: Based on historical feeding data and the current number of cattle and sheep, the feeding demand forecast results are obtained. Based on the feed supply data and the feeding demand forecast results, an optimization algorithm is used to calculate the optimal feeding speed for each area; if the difference between the actual feeding amount in the area and the feeding forecast demand result exceeds the preset value, the optimal feeding speed is updated.

7. The method according to claim 6, characterized in that Based on the feed supply data and the feeding demand prediction results, the process of using the optimization algorithm to calculate the optimal feeding rate for each area includes: The method comprises the following steps: obtaining historical data and real-time data of feed supply and feeding demand of each region, obtaining feed supply of each region based on the historical data and real-time data of feed supply of each region, and obtaining feeding demand of each region based on the historical data and real-time data of feeding demand of each region; taking the predicted feed supply and feeding demand of each region as constraint conditions, establishing a feeding speed optimization model, wherein the objective function of the feeding speed optimization model is to minimize the absolute value of the difference between the total feed supply and the total demand; and solving the feeding speed optimization model by using a heuristic algorithm to obtain the optimal feeding speed of each region.

8. The method according to claim 1, characterized in that: It also includes obtaining real-time data on the feeding speed and feeding behavior of cattle and sheep in each area, performing correlation analysis on the feeding speed and feeding behavior data of cattle and sheep, obtaining key factors for efficiency evaluation, and evaluating the feeding efficiency of each area based on the key factors for efficiency evaluation. If the feeding efficiency does not meet the preset requirements, the feeding management strategy is optimized based on the feeding behavior of cattle and sheep and feed supply data in the area.

9. A fully automatic feeding control system for cattle and sheep in multiple areas, characterized in that: include: The feeding speed prediction module is used to predict the feeding speed of cattle and sheep in each area; An initial feeding speed determination module is used to determine the initial feeding speed of each area; The eating speed change detection module is used to determine whether the eating speed of cattle and sheep in each area has changed; Feeding speed dynamic adjustment module, used to adjust the feeding speed according to the changes in the feeding speed of cattle and sheep; Multi-zone feeding coordination module, used to coordinate the feeding speed between different zones; Feeding efficiency evaluation and optimization module, used to evaluate feeding efficiency and optimize feeding management strategies.

Images