Method for adaptive regulation of feeding parameters based on growth stage of chicken flock

Abstract

The application provides a feeding parameter self-adaptive regulation method based on the growth stage of a chicken group. The application synchronously acquires growth-related parameters and multi-dimensional environmental parameters of the chicken group, combines the group average empty stomach weight and the daily feed intake per unit weight to calculate a virtual physiological age and generate a target environmental dynamic reference, fuses environmental deviation and behavior distribution deviation to construct a total deviation of comfort, and adopts dynamic damping adjustment to obtain a smooth control instruction. The application can self-adaptively regulate the feeding environment according to the real growth stage of the chicken group, improve the breeding environment control precision and animal welfare, avoid frequent shaking start and stop of the equipment, prolong the service life of the equipment, reduce energy consumption, and realize intelligent, refined, stable and efficient automatic breeding.

Description

Technical Field

[0001] This invention relates to the field of intelligent control of livestock and poultry farming environment, and in particular to an adaptive control method for feeding parameters based on the growth stage of chicken flocks. Background Technology

[0002] Currently, the existing technologies widely used in the field of livestock and poultry farming environmental control are mainly based on a fixed physical age lookup table and PID feedback control architecture. This technology divides growth stages according to age, reads the corresponding target temperature from a preset table, and adjusts actuators such as fans and heaters through a PID controller to maintain the indoor environment near the set value.

[0003] However, this technology has two inherent drawbacks: First, using physical age as the benchmark for environmental control fails to consider developmental differences in chickens due to factors such as disease and nutrition during their actual growth, resulting in a mismatch between the temperature control target and the chickens' actual physiological needs. Second, relying solely on monitoring data from fixed-point sensors cannot perceive the chickens' perceived physical condition. For example, when chickens huddle together for warmth, even if the sensors show the temperature is within the acceptable range, the actual perceived temperature may still be too low. Furthermore, control strategies based on PID or switching methods are prone to causing abrupt changes in equipment such as fans, leading to violent fluctuations in environmental parameters within the coop. This not only exacerbates the chickens' stress response but also increases unnecessary energy consumption. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides an adaptive control method for feeding parameters based on the growth stage of a chicken flock.

[0005] To achieve the above objectives, the technical solution of this invention is as follows: In a first aspect, the present invention provides an adaptive control method for feeding parameters based on the growth stage of a chicken flock, the method comprising: Within each control cycle, the average fasting body weight, daily feed intake per unit body weight, Shannon entropy of the flock, and environmental parameter vectors including current ambient temperature, current ambient humidity, ammonia concentration, and carbon dioxide concentration are acquired synchronously for the target flock. Based on the average fasting body weight and daily feed intake per unit body weight of the flock, the virtual physiological age reflecting the true developmental status of the flock is calculated, and a dynamic baseline for the target environment is generated. Using the target environment dynamic benchmark as a reference, the physical environment deviation and gas concentration deviation obtained by the calculation are integrated with the directional behavioral distribution deviation calculated based on the Shannon entropy of the population distribution to construct the total comfort deviation; Based on historical oscillation indicators, total comfort deviation, and virtual physiological age, the initial control increment generated based on the total comfort deviation is dynamically damped to obtain a smooth control command, which is then executed to adjust the feeding environment equipment.

[0006] In some embodiments, the virtual physiological age reflecting the true developmental state of the flock is calculated based on the flock's average fasting body weight and daily feed intake per unit body weight. This includes: mapping the flock's average fasting body weight to theoretical growth time by solving the inverse function of a pre-stored standard Gompertz growth curve model, using the flock's average fasting body weight as the dependent variable; interpolating and querying the theoretical feed intake corresponding to the theoretical growth time from a preset breed standard feed intake curve, and calculating the feed intake metabolic equivalent age based on the daily feed intake per unit body weight and the theoretical feed intake; dynamically configuring weight coefficients based on the growth stage determined by the current physical age, and adjusting the theoretical growth time and feed intake metabolic equivalent age accordingly. The age is weighted and fused to calculate the virtual physiological age. Between two valid weight data acquisitions, the metabolic state coefficient is determined based on the ratio of the current daily feed intake per unit weight to the theoretical feed intake. The metabolic state coefficient is used to decay and maintain the virtual physiological age calculated based on the previous control cycle to update the virtual physiological age of the current control cycle. When the change in daily feed intake per unit weight relative to the theoretical feed intake is detected to exceed the preset amplitude threshold, it is determined that the flock is in a state of stress. The decay and maintenance logic is paused, and the virtual physiological age is corrected based on the current measured feed intake data. The corrected value is used as the virtual physiological age of the current control cycle.

[0007] In some embodiments, generating a target environmental dynamic baseline includes: constructing a continuous environmental baseline mapping function with virtual physiological age as the independent variable based on pre-stored discrete data points of breed feeding standards through piecewise cubic spline interpolation; using the virtual physiological age of the current control cycle as input, calling the continuous environmental baseline mapping function to interpolate and obtain the corresponding target environmental temperature and target environmental humidity; parsing the outdoor temperature collected in real time by the weather station outside the barn from the environmental parameter vector, and combining it with a preset seasonal compensation coefficient to perform outdoor temperature feedforward compensation and seasonal trend correction on the target environmental temperature to obtain the corrected target environmental temperature; and jointly determining the corrected target environmental temperature and target environmental humidity as the target environmental dynamic baseline.

[0008] In some embodiments, constructing the total comfort deviation includes: calculating the physical environment deviation based on the target environment dynamic benchmark, the current ambient temperature, and the current ambient humidity; determining whether the ammonia and carbon dioxide concentrations in the current control cycle exceed preset safety thresholds; if they exceed them, calculating the gas concentration deviation according to a preset calculation formula; calculating the behavioral distribution deviation based on the population distribution Shannon entropy and the preset maximum theoretical entropy value; determining the conflict mode by detecting the sign consistency between the physical environment deviation and the behavioral distribution deviation to obtain the conflict detection result; and calculating the total comfort deviation using a nonlinear weighted fusion method based on the conflict detection result and the growth stage determined by virtual physiological age.

[0009] In some embodiments, conflict mode determination is performed by detecting the consistency of the signs of physical environment deviation and behavioral distribution deviation to obtain conflict detection results, including: if the signs of physical environment deviation and behavioral distribution deviation are the same, it is determined to be a somatosensory abnormal mode, a biological priority conflict resolution mechanism is activated, and the conflict marker is set to the reverse regulation mode; if the signs of physical environment deviation and behavioral distribution deviation are opposite, it is determined to be a normal mode, and the conflict marker is set to the normal mode.

[0010] In some embodiments, the total comfort deviation The calculation formula is: In the formula, Due to physical environmental deviations; This is a behavioral distribution bias; Gas concentration deviation; weighting coefficient , and All are dynamically adjusted according to the growth stage; nonlinear index and All settings are based on the growth stage; in reverse adjustment mode, the total comfort deviation... The calculation formula is: .

[0011] In some embodiments, based on historical oscillation indices, total comfort deviation, and virtual physiological age, the initial control increment generated based on the total comfort deviation is dynamically damped to obtain a smooth control command. This includes: constructing an adaptive damping coefficient based on the absolute value of the total comfort deviation and historical oscillation indices; calculating the initial control increment based on the adaptive damping coefficient, the total comfort deviation, and a preset base proportional gain; determining whether to activate the juvenile protection mechanism based on the virtual physiological age and a preset juvenile age threshold; if the virtual physiological age is less than the preset juvenile age threshold, activating the juvenile protection mechanism and forcibly limiting the initial control increment to ensure it does not exceed a preset maximum increment, thus obtaining the final adjustment increment; if the virtual physiological age is not less than the preset juvenile age threshold, directly using the initial control increment as the final adjustment increment; superimposing the final adjustment increment onto the actuator output value of the previous cycle, and after physical limiting, generating and outputting a smooth control command, which is then sent to the equipment for adjusting the feeding environment.

[0012] In some embodiments, the method further includes: calculating the growth offset rate based on the virtual physiological age and the current physical age; comparing the growth offset rate with a preset offset rate threshold; if the growth offset rate is greater than the preset offset rate threshold, marking an abnormal state identifier, temporarily shortening the control cycle to a preset high-frequency monitoring cycle, and triggering a high-frequency monitoring and early warning mechanism.

[0013] The adaptive control method for feeding parameters based on the growth stage of chicken flocks provided by this invention first achieves dynamic matching between the environmental control benchmark and the actual physiological needs of the flock by introducing virtual physiological age, fundamentally solving the problem of physiological demand mismatch caused by rigid time benchmarks in existing technologies. On this basis, by integrating physical environment deviation, gas concentration deviation, and behavioral distribution deviation based on Shannon entropy, a multi-dimensional total comfort deviation is constructed, enabling the system to perceive the physiological state of the chicken flock and overcoming the limitation that a single sensor cannot identify hidden discomfort conditions. Furthermore, an adaptive damping coefficient is constructed through historical oscillation indicators, and combined with the virtual physiological age triggering a young protection mechanism, the control increment is smoothed and the execution strategy is made safe, effectively suppressing system oscillations and animal agitation risks. Finally, the virtual physiological age forms a data closed loop that runs through the entire chain of perception-decision-execution, forming a precise, stable, and biologically-prioritized intelligent control architecture, which significantly improves the adaptability, robustness, and animal welfare level of feeding environment control. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating the adaptive control method for feeding parameters based on the growth stage of a chicken flock, provided in an embodiment of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on the present invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] In the following description, references to "some embodiments" refer to a subset of all possible embodiments; however, it is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. Unless otherwise defined, all technical and scientific terms used in the embodiments of the invention have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of the invention pertain. The terminology used in the embodiments of the invention is for the purpose of describing the embodiments of the invention only and is not intended to limit the invention.

[0017] The following describes an exemplary application of the adaptive control device for feeding parameters based on the growth stage of a chicken flock, as described in this embodiment of the invention. This device can be implemented as a terminal or a server. In one implementation, the device can be implemented as a laptop, tablet, desktop computer, mobile device, or other types of terminal. In another implementation, it can also be implemented as a server. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. The terminal and server can be directly or indirectly connected via wired or wireless communication, which is not limited in this embodiment. The following describes an exemplary application of the adaptive control device for feeding parameters based on the growth stage of a chicken flock when it is implemented as a server.

[0018] This invention provides an adaptive control method for feeding parameters based on the growth stage of a chicken flock. (See also...) Figure 1 , Figure 1 This is a flowchart illustrating an adaptive control method for feeding parameters based on the growth stage of a chicken flock, provided by an embodiment of the present invention. Figure 1 The steps shown are explained.

[0019] S110, within each control cycle, synchronously acquires the target flock's average fasting body weight, daily feed intake per unit body weight, Shannon entropy of the flock distribution, and environmental parameter vectors including current ambient temperature, current ambient humidity, ammonia concentration, and carbon dioxide concentration.

[0020] In some implementations, the control cycle refers to a fixed time interval for a complete data collection, calculation, analysis, and command execution of environmental parameters for the chicken flock. It serves as a time benchmark for adaptive adjustment of feeding parameters and can be flexibly set according to the chicken flock's growth stage. For example, during the chick stage, metabolism is rapid and environmental requirements change frequently, so the control cycle can be set to 1 hour; during the growing chicken stage, growth is stable, so the control cycle can be set to 2 hours. The control cycle ensures dynamic monitoring and timely adjustment of the flock's condition and environmental parameters, preventing the feeding environment from deviating from the suitable range due to monitoring lag.

[0021] In some implementations, the target flock refers to the collection of all chickens that are at the same growth stage and raised in the same enclosed or semi-enclosed rearing space, which is the target of the control method.

[0022] In some implementations, the average fasting weight of the flock refers to the average weight obtained by randomly sampling and weighing the target flock at the beginning of each control period, removing undigested feed and feces from the gastrointestinal tract of the sampled chickens, i.e., in a fasting state, and then calculating the average weight by arithmetic mean. This average weight is used to reflect the overall growth and development level of the flock.

[0023] In some implementations, daily feed intake per unit body weight refers to the average total feed intake per unit body weight of chickens in the target flock over 24 hours. It is used to reflect the flock's feeding efficiency and nutrient intake level and is an important indicator for judging the flock's physiological state. Its calculation is based on the total daily feed intake and the average weight of the flock.

[0024] In some implementations, population distribution Shannon entropy is based on the concept of Shannon entropy in information theory. It is an index that quantifies the uniformity of the distribution of a target flock of chickens within the rearing space, reflecting the flock's behavioral state, such as whether they gather or disperse. The lower the entropy value, the more likely the flock is to gather or huddle together, which may indicate abnormal behaviors such as environmental discomfort, gathering for warmth, or avoiding heat. The higher the entropy value, the more uniform the distribution of the flock within the rearing space, usually corresponding to a suitable environment and normal flock behavior.

[0025] In some implementations, the environmental parameter vector is used to comprehensively describe a multi-dimensional set of parameters describing the environmental state of the target chicken flock. Specifically, this invention includes four core environmental parameters: current ambient temperature, current ambient humidity, ammonia concentration, and carbon dioxide concentration. These parameters are arranged in a fixed order to form a vector, facilitating subsequent deviation calculations and data fusion. Each parameter represents an average value within a control period, synchronously collected by sensors deployed within the rearing space, such as temperature sensors, humidity sensors, ammonia sensors, and carbon dioxide sensors. This data forms the core basis for regulating the rearing environment.

[0026] In some implementations, the current ambient temperature refers to the average air temperature within the target flock's rearing space during the control period. This is a key environmental parameter affecting the flock's thermoregulation, metabolic rate, and feeding behavior; different growth stages of the flock have different suitable temperature ranges. Temperature data is collected synchronously by temperature sensors placed at different locations within the rearing space, such as the center, edges, and areas with high flock activity, and the average value is taken as the current ambient temperature.

[0027] In some implementations, the current ambient humidity refers to the average relative humidity of the air in the target flock's rearing space during the control period. It mainly affects the heat dissipation of the flock's surface, respiratory health, and the volatilization of harmful gases in the rearing environment. The suitable humidity range varies with the flock's growth stage.

[0028] S120 calculates the virtual physiological age reflecting the true developmental status of the flock based on the average fasting body weight and daily feed intake per unit body weight, and generates a dynamic baseline for the target environment.

[0029] In some implementations, virtual physiological age is an equivalent age calculated based on the target flock's average fasting body weight and daily feed intake per unit body weight. It reflects the flock's true developmental status and differs from traditional actual calendar age. This avoids the limitation of calendar age in reflecting individual differences within the flock, making environmental benchmarks and control commands more aligned with the flock's actual physiological needs. The calculation logic is based on the standard flock's weight-daily feed intake-age relationship. Using the current flock's average fasting body weight and daily feed intake per unit body weight, a standard age matching this physiological state is derived—this is the virtual physiological age. For example, a flock's actual calendar age is 20 days, but due to poor rearing conditions resulting in a low average fasting body weight and insufficient daily feed intake per unit body weight, the calculated virtual physiological age is 18 days. This indicates that the flock's actual developmental level lags behind its calendar age, requiring adjustments to the rearing environment to promote growth.

[0030] In some implementations, the target environment dynamic benchmark is a set of standard environmental parameter values ​​dynamically generated based on virtual physiological age, adapted to the current actual developmental state of the target flock. Each parameter in the environmental parameter vector has a corresponding dynamic benchmark value, serving as a reference standard for calculating environmental deviations and judging environmental comfort. Its generation logic is as follows: a model or function corresponding to the virtual physiological age and the suitable range of each environmental parameter is pre-established; based on the calculated virtual physiological age, the optimal standard value of each environmental parameter is retrieved from the model or function to form the target environment dynamic benchmark. For example, the dynamic benchmark corresponding to a virtual physiological age of 18 days is: temperature 28℃, humidity 65%RH, ammonia concentration ≤15ppm, and carbon dioxide concentration ≤2000ppm, and this benchmark is dynamically updated as the virtual physiological age changes.

[0031] S130 uses the target environment dynamic benchmark as a reference, integrates the calculated physical environment deviation and gas concentration deviation, as well as the directional behavioral distribution deviation calculated based on the Shannon entropy of the population distribution, to construct the total comfort deviation.

[0032] In some implementations, physical environment deviation refers to the difference between the current ambient temperature, current ambient humidity and the corresponding parameters in the target environment dynamic benchmark, which is used to quantify the degree of deviation between the current physical environment and the suitable environment.

[0033] In some implementations, gas concentration deviation refers to the difference between the concentration of harmful gases (ammonia concentration, carbon dioxide concentration) in the current environmental parameter vector and the corresponding standard value of the gas concentration in the target environmental dynamic benchmark. It is used to quantify the degree of deviation between the current harmful gas concentration and the suitable concentration, and is one of the core components in constructing the overall comfort deviation. A positive deviation indicates that the gas concentration exceeds the standard and needs to be adjusted promptly. The gas concentration deviation is calculated according to a preset concentration deviation calculation formula when the ammonia concentration is >20 ppm or the carbon dioxide concentration is >3000 ppm.

[0034] In some implementations, behavioral distribution deviation is a directional deviation index calculated based on the Shannon entropy of the flock distribution, used to quantify the degree of deviation between the flock's distribution state and its ideal distribution state. This deviation index reflects not only the magnitude of the deviation but also its direction, i.e., the regional characteristics of the flock's aggregation.

[0035] In this invention, the standard Shannon entropy (a preset value, corresponding to the entropy value when the chickens are evenly distributed) under the appropriate distribution state of the chicken flock is used as a benchmark to calculate the difference between the current group distribution Shannon entropy and the standard Shannon entropy. At the same time, the direction of chicken flock gathering is determined by combining the distribution ratio of chickens in each area, and finally a directional behavioral distribution deviation is formed.

[0036] It should be noted that when the behavioral distribution deviation is greater than zero, it indicates that the flock is huddled together / cold stress; when the behavioral distribution deviation is less than zero, it indicates that the flock is overly dispersed / heat stress; when the behavioral distribution deviation is equal to zero, it indicates that the flock is evenly distributed.

[0037] In some implementations, the total comfort deviation is calculated by weighting physical environment deviation, gas concentration deviation, and behavioral distribution deviation against a dynamic benchmark of the target environment using preset weighting coefficients. This comprehensive deviation index is used to quantify the comfort level of the current rearing environment for the target flock. The smaller the absolute value of the total comfort deviation, the more suitable the current environment is for the flock's growth; the larger the absolute value, the more severely the environment deviates from the suitable range. This is the core basis for generating control commands. Here, the preset weighting coefficients are set according to the flock's growth stage and the importance of environmental parameters. For example, temperature deviation has a higher weight in the chick stage, and gas concentration deviation has a higher weight in the grower stage.

[0038] S140, based on historical oscillation indicators, total comfort deviation, and virtual physiological age, dynamically dampes the initial control increment generated based on the total comfort deviation to obtain a smooth control command, and executes the smooth control command to adjust the feeding environment equipment.

[0039] In some implementations, the historical oscillation index refers to an index calculated based on the total comfort deviation and control command execution effect of historical control cycles (usually 3-5 consecutive control cycles before the current control cycle). It reflects the degree of fluctuation of environmental parameters and flock status during the control process and is used to determine whether there is oscillation in the control process, such as frequent equipment adjustments causing environmental parameters to fluctuate. It can provide a basis for damping adjustment of the initial control increment.

[0040] In some implementations, the initial control increment refers to the initial adjustment amount generated based on the total comfort deviation using a preset control algorithm, such as a proportional control algorithm or a PID control algorithm, for adjusting the feeding environment equipment. This initial value represents the magnitude and direction of the adjustment required by the equipment. Because historical oscillations are not considered, there may be issues with excessively large or small adjustment magnitudes, leading to oscillations in environmental parameters.

[0041] In some implementations, dynamic damping adjustment refers to the dynamic correction and adjustment of the initial control increment based on historical oscillation indicators, total comfort deviation, and virtual physiological age. Its purpose is to suppress oscillations during the control process, make the control commands smoother and more adapted to the physiological state of the flock, and avoid fluctuations in environmental parameters caused by frequent and large-scale adjustments, which would affect the growth of the flock.

[0042] Here, the degree of oscillation is judged based on historical oscillation indicators; the larger the indicator, the more obvious the oscillation. The control demand is judged based on the total comfort deviation; the larger the deviation, the stronger the control demand. The adaptability of the flock is judged based on virtual physiological age; the adaptability is weak in the chick stage and the damping coefficient is large; the adaptability is strong in the rearing stage and the damping coefficient is small. The damping coefficient is calculated by integrating the three factors through a preset damping formula. The damping coefficient is used to correct the initial control increment to obtain a smooth control command.

[0043] In some implementations, the smooth control command refers to the final command obtained after dynamic damping adjustment, used to control the feeding environment equipment. The control range of this command is smooth and without drastic fluctuations. It can effectively correct the total comfort deviation and make the environmental parameters tend to the target environmental dynamic benchmark, while avoiding environmental oscillations caused by excessive control range, and adapting to the physiological adaptability of the flock.

[0044] In some implementations, the term "feeding environment equipment" refers to a general term for various devices used to adjust the feeding environment parameters of the target chicken flock, whose operation is directly controlled by smooth control commands. Specifically, this includes temperature control equipment, such as heating lamps and air coolers; humidity control equipment, such as humidifiers and dehumidifiers; ventilation equipment, such as axial flow fans and ventilation windows; and harmful gas purification equipment, such as ammonia adsorption devices. Its core function is to adjust operating parameters according to smooth control commands, precisely controlling the temperature, humidity, ammonia concentration, and carbon dioxide concentration of the feeding environment to the appropriate range corresponding to the dynamic benchmark of the target environment, providing the target chicken flock with a feeding environment suitable for its growth stage, effectively ensuring the healthy growth of the flock and the stable performance of its production.

[0045] The adaptive control method for feeding parameters based on the growth stage of chicken flocks provided by this invention first achieves dynamic matching between the environmental control benchmark and the actual physiological needs of the flock by introducing virtual physiological age, fundamentally solving the problem of physiological demand mismatch caused by rigid time benchmarks in existing technologies. On this basis, by integrating physical environment deviation, gas concentration deviation, and behavioral distribution deviation based on Shannon entropy, a multi-dimensional total comfort deviation is constructed, enabling the system to perceive the physiological state of the chicken flock and overcoming the limitation that a single sensor cannot identify hidden discomfort conditions. Furthermore, an adaptive damping coefficient is constructed through historical oscillation indicators, and combined with the virtual physiological age triggering a young protection mechanism, the control increment is smoothed and the execution strategy is made safe, effectively suppressing system oscillations and animal agitation risks. Finally, the virtual physiological age forms a data closed loop that runs through the entire chain of perception-decision-execution, forming a precise, stable, and biologically-prioritized intelligent control architecture, which significantly improves the adaptability, robustness, and animal welfare level of feeding environment control.

[0046] In some embodiments, the above-described S120 can also be implemented by the following: Using the average fasting body weight of the flock as the dependent variable, the average fasting body weight of the flock is mapped to the theoretical growth time by solving the inverse function of the pre-stored standard Gompertz growth curve model. The theoretical feed intake corresponding to the theoretical growth time is interpolated from the preset breed standard feed intake curve, and the feed intake metabolic equivalent age is calculated based on the daily feed intake per unit body weight and the theoretical feed intake. The weight coefficients of the growth stage determined by the current physical age are dynamically configured to calculate the virtual physiological age by weighted fusion of the theoretical growth time and the feed intake metabolic equivalent age. Between two acquisitions of effective body weight data, the metabolic state coefficient is determined based on the ratio of the current daily feed intake per unit body weight to the theoretical feed intake. The metabolic state coefficient is used to decay and maintain the virtual physiological age calculated based on the previous control cycle to update the virtual physiological age of the current control cycle. When the change in daily feed intake per unit body weight relative to the theoretical feed intake exceeds the preset amplitude threshold, the flock is judged to be in a state of stress, the decay and maintenance logic is paused, and the virtual physiological age is corrected based on the current measured feed intake data. The corrected value is used as the virtual physiological age of the current control cycle.

[0047] Here, the standard Gompertz growth curve model refers to a pre-stored growth curve model for the target chicken flock and corresponding breed. It is constructed based on the Gompertz mathematical model and is used to describe the change of the average fasting body weight of the flock with the growth time under normal growth conditions. It is the core basis for mapping body weight to theoretical growth time.

[0048] Here, the inverse function of the standard Gompertz growth curve model refers to the function that reverses the correspondence between growth time and average fasting body weight in the standard Gompertz growth curve model. It uses average fasting body weight as the independent variable and growth time as the dependent variable. Its purpose is to deduce the corresponding theoretical growth time from the measured average fasting body weight.

[0049] Here, theoretical growth time refers to the standard growth time required for a flock to reach that weight under normal growth conditions, obtained by mapping the currently measured average fasting weight of the flock to the inverse function of the standard Gompertz growth curve model. It reflects the theoretical development level of the flock based on weight and is different from the actual physical age.

[0050] Here, the preset standard feed intake curve refers to a preset standard curve for the target flock, describing the daily feed intake per unit body weight as a function of growth time under normal growth conditions. It is obtained by fitting a large amount of experimental data and stored in the control system for querying the theoretical feed intake corresponding to different growth times.

[0051] Here, interpolation query refers to a query method that calculates the theoretical feed intake corresponding to the theoretical growth time by using mathematical methods such as linear interpolation and nonlinear interpolation when the theoretical growth time does not completely match the standard growth time node stored in the preset variety standard feeding curve. This ensures the accuracy of the theoretical feed intake and avoids errors caused by node mismatch.

[0052] Here, theoretical feed intake refers to the standard daily feed intake per unit body weight of a flock under normal growth conditions, obtained by interpolation using a preset breed standard feed intake curve based on the theoretical growth time. It serves as a reference for assessing whether the current measured daily feed intake per unit body weight is normal. The relationship between measured and actual daily feed intake per unit body weight is as follows: if the measured daily feed intake per unit body weight is greater than the theoretical feed intake, it indicates that the flock's nutrient intake efficiency is higher than the standard; if it is less than the theoretical feed intake, it indicates that the nutrient intake efficiency is lower than the standard, which may indicate abnormal feeding or environmental unsuitability.

[0053] Here, the equivalent age of feed intake metabolism refers to the equivalent age of chicken flocks, which is obtained by correcting the theoretical growth time based on the ratio of the current measured daily feed intake per unit body weight to the theoretical feed intake. It is used to quantify the impact of feed intake efficiency on the developmental status of chicken flocks and make up for the limitations of judging the developmental level solely by body weight.

[0054] Here, the current physical age refers to the actual number of days the target flock has been raised from hatching to the current control period. It is the basic basis for determining the growth stage of the flock (chick stage, growing stage, laying / marketing stage), and is different from theoretical growth time and virtual physiological age. It only reflects the actual length of time the flock has been raised. For example, if the flock hatched on March 1st and the current control period is March 30th, then the current physical age is 29 days.

[0055] Here, the growth stage refers to the stage with different growth and development characteristics and nutritional / environmental requirements, which is divided according to the current physical age of the flock. It is usually divided into the chick stage (1-7 days old), the mid-chick stage (8-21 days old), the rearing stage (22-42 days old), and the egg-laying / marketing stage (43 days old and above). There are significant differences in weight gain rate, feed intake efficiency, and physiological adaptability among flocks in different growth stages.

[0056] Here, the weighting coefficient refers to the coefficient that is dynamically configured based on the current physical age and is used to weight and integrate the theoretical growth time and the equivalent age of feed intake and metabolism. Its value is dynamically adjusted according to the needs of the growth stage, reflecting the weight of the influence of body weight development and feed intake and metabolism development on the actual development status of the flock in different growth stages.

[0057] Here, weighted fusion refers to the calculation method of multiplying the theoretical growth time and the equivalent age of food intake and metabolism by the corresponding weight coefficients, and then adding the two results to obtain the virtual physiological age.

[0058] Here, the interval between two valid weight data acquisitions refers to the control period between two consecutive successful acquisitions of the average fasting weight data of the target flock. During this period, since no new weight data is acquired, it is impossible to calculate the new theoretical growth time through the inverse function. Therefore, the virtual physiological age needs to be updated through decay maintenance logic to ensure the continuity of regulation.

[0059] Here, for example, if the control cycle is 4 hours, the weight data is successfully collected in the first control cycle, but no valid weight data is collected in the second and third control cycles due to sensor failure, and valid weight data is collected again in the fourth control cycle, then the period between the first and fourth control cycles, and the period between the second and third control cycles, is between two valid weight data acquisitions.

[0060] Here, the metabolic state coefficient refers to the coefficient determined between two valid weight data acquisitions based on the ratio of the current daily feed intake per unit body weight to the theoretical feed intake. This metabolic state coefficient is used to reflect the stability of the metabolic state of the flock when there is no new weight data, and serves as the basis for the decay and maintenance of virtual physiological age.

[0061] Here, the decay-maintaining logic refers to the use of metabolic state coefficients to moderately correct the virtual physiological age calculated in the previous control cycle when the theoretical growth time cannot be updated through weight data between two valid weight data acquisitions. This allows the virtual physiological age to be updated slowly (rather than abruptly), maintaining its continuity and stability, and avoiding abnormal fluctuations in virtual physiological age due to missing weight data, which would affect the accuracy of regulation.

[0062] Here, the preset amplitude threshold refers to the preset critical value of the daily feed intake per unit body weight used to determine whether the flock is under stress. It can be preset according to the breed and growth stage of the flock, and is usually set to ±15%~±20%. If the threshold is exceeded, the flock is determined to be under stress, such as environmental discomfort, disease, abnormal feed, etc.

[0063] Here, stress refers to the abnormal physiological state of chickens caused by abnormal external environment factors, such as excessively high / low temperature, excessive ammonia concentration, disease, feed quality problems, etc., which manifests as a drastic fluctuation in daily feed intake per unit body weight.

[0064] Here, the currently measured feed intake data refers to the actual daily feed intake per unit body weight of the target flock, calculated by the difference between the amount of feed fed and the amount of feed remaining within the current control period. This is different from the theoretical feed intake and serves as the actual measurement basis for correcting the virtual physiological age, ensuring the accuracy of the virtual physiological age under stress.

[0065] This invention breaks through the limitations of traditional methods that rely solely on body weight or physical age to determine the developmental status of chicken flocks. It combines theoretical growth time and equivalent feed intake and metabolic age, using dynamic weighting to calculate virtual physiological age. This approach more closely aligns with the actual developmental patterns of chicken flocks, resolving the issue of large biases in single-indicator assessments and significantly improving the accuracy of virtual physiological age. This provides a more reliable basis for generating dynamic benchmarks for the target environment. Simultaneously, a stress state determination mechanism is added. By monitoring the variation between daily feed intake per unit body weight and theoretical feed intake, the stress state of the flock is identified in a timely manner. The conventional decay maintenance logic is paused, and the virtual physiological age is corrected based on measured feed intake data. This avoids distortion of the virtual physiological age under stress, ensuring that even when the flock's physiological state is abnormal, the control method can still accurately adapt to the flock's actual needs, improving the control method's anti-interference capability and adaptability. Furthermore, for scenarios where body weight data is missing between two valid weight data acquisitions, a metabolic state coefficient and decay maintenance logic are designed to ensure continuous and stable updates of the virtual physiological age. This avoids abrupt changes in virtual physiological age due to data loss, thereby preventing abnormal control commands and ensuring the continuity and stability of feeding control.

[0066] In some embodiments, the above-described S120 can also be implemented by the following: Based on pre-stored discrete data points of breed feeding standards, a continuous environmental baseline mapping function with virtual physiological age as the independent variable is constructed through piecewise cubic spline interpolation. Using the virtual physiological age of the current control cycle as input, the continuous environmental baseline mapping function is called to interpolate and obtain the corresponding target environmental temperature and target environmental humidity. The outdoor temperature collected in real time by the weather station outside the barn is parsed from the environmental parameter vector, and combined with the preset seasonal compensation coefficient, the target environmental temperature is subjected to outdoor temperature feedforward compensation and seasonal trend correction to obtain the corrected target environmental temperature. The corrected target environmental temperature and target environmental humidity are jointly determined as the target environmental dynamic baseline.

[0067] Here, the pre-stored breed feeding standard discrete data points refer to the set of environmental parameter standard data corresponding to different developmental stages, which are pre-stored in the control system for the target chicken flock and the corresponding breed, according to the feeding standards. The data exists in discrete form, that is, each data point corresponds to a specific virtual physiological age and the corresponding environmental baseline parameter, which is the basic data for constructing the environmental baseline continuous mapping function.

[0068] Here, piecewise cubic spline interpolation refers to a mathematical interpolation method used to process discrete data and construct continuous functions. It can divide the virtual physiological age interval where the pre-stored breed feeding standard discrete data points are located into multiple continuous sub-intervals, construct multiple spline functions in each sub-interval, so that the function curves of adjacent sub-intervals are smoothly connected, and finally form a continuous mapping function covering the entire virtual physiological age interval, thus solving the problem that intermediate values ​​cannot be accurately obtained between discrete data points.

[0069] Here, the environmental baseline continuous mapping function refers to a continuous mathematical function constructed by piecewise cubic spline interpolation, using virtual physiological age as the independent variable and target environmental temperature and humidity as dependent variables. This mapping function can achieve accurate and continuous mapping between virtual physiological age and environmental baseline parameters, and can quickly interpolate the corresponding target environmental temperature and humidity based on any input virtual physiological age.

[0070] Here, outdoor temperature feedforward compensation refers to the process of correcting the target ambient temperature obtained by interpolation based on the outdoor temperature collected in real time by the outdoor weather station and combined with the preset seasonal compensation coefficient. This process takes into account the impact of outdoor temperature on indoor temperature control in advance, avoids lag in indoor temperature control, reduces control energy consumption, and ensures that the indoor temperature quickly approaches the corrected target value.

[0071] Here, seasonal trend correction refers to the correction of the target ambient temperature in accordance with the current seasonal climate trend by combining the preset seasonal compensation coefficient. The core is to make up for the limitations of only compensating by outdoor temperature feedforward, and take into account the overall seasonal climate characteristics, such as summer being generally hot and winter being generally cold. This makes the corrected target ambient temperature not only match the real-time outdoor temperature, but also conform to the long-term seasonal climate pattern, further improving the comfort of the flock.

[0072] It should be noted that the difference between seasonal trend correction and outdoor temperature feedforward compensation is that feedforward compensation focuses on the real-time correction of outdoor temperature, while seasonal trend correction focuses on the trend correction of the overall seasonal characteristics. The combination of the two makes the correction of the target environmental temperature more comprehensive and more in line with the actual breeding scenario.

[0073] This invention achieves a smooth and continuous mapping between virtual physiological age and environmental baseline parameters by constructing an environmental baseline continuous mapping function based on piecewise cubic spline interpolation. This mapping mechanism can accurately obtain the environmental baseline value corresponding to any virtual physiological age, making the target environmental dynamic baseline more closely match the gradual physiological needs of chickens at different growth stages, significantly improving the accuracy and adaptability of environmental baseline setting. Furthermore, this invention overcomes the limitation of setting environmental baselines solely based on the chickens' own developmental state by introducing outdoor temperature feedforward compensation and seasonal trend correction mechanisms. By integrating real-time outdoor temperature data collected from outdoor weather stations and preset seasonal compensation coefficients, the target environmental temperature is dynamically feedforward corrected, effectively avoiding stress responses in chickens caused by excessive indoor-outdoor temperature differences. Simultaneously, it optimizes the energy consumption of environmental control equipment such as fans and heaters, achieving a synergistic improvement in both the physiological comfort of the chickens and the economic benefits of farming.

[0074] In some embodiments, step S130 above can also be implemented by the following: Based on the target environment dynamic benchmark, current ambient temperature, and current ambient humidity, the physical environment deviation is calculated; it is determined whether the ammonia and carbon dioxide concentrations in the current control cycle exceed the preset safety thresholds; if they do, the gas concentration deviation is calculated according to the preset calculation formula; the behavioral distribution deviation is calculated based on the population distribution Shannon entropy and the preset maximum theoretical entropy value; conflict mode is determined by detecting the sign consistency between the physical environment deviation and the behavioral distribution deviation, and conflict detection results are obtained; based on the conflict detection results and the growth stage determined by virtual physiological age, the total comfort deviation is calculated using a nonlinear weighted fusion method.

[0075] Here, the preset safety threshold refers to the safety critical values ​​of ammonia and carbon dioxide concentrations that are pre-stored in the control system for different growth stages of the target chicken flock. It is the core basis for judging whether the concentration of harmful gases exceeds the standard and whether it is necessary to calculate the gas concentration deviation. It is preset by those skilled in the art in combination with the characteristics of the chicken flock, breeding test data and industry safety standards to ensure the healthy growth of the chicken flock.

[0076] Here, the preset calculation formula refers to a fixed mathematical formula that is pre-set and stored in the control system. It is used to calculate the gas concentration deviation when the ammonia concentration and carbon dioxide concentration exceed the preset safety threshold. It quantifies the degree of deviation between the concentration of harmful gases and the safety threshold. The formula parameters can be dynamically adjusted according to the growth stage of the flock to ensure the accuracy of the deviation calculation.

[0077] Here, the preset maximum theoretical entropy value refers to the theoretical maximum value of the Shannon entropy of the flock distribution when the target flock reaches the most uniform distribution state in the rearing space. It is the core reference for calculating the behavioral distribution deviation and is determined by factors such as the size of the rearing space, flock density, and number of regional divisions. It is preset and stored in the control system through theoretical calculation and experimental verification.

[0078] Here, sign consistency refers to whether the positive and negative signs of the physical environment deviation and the behavioral distribution deviation are the same, used to determine whether there is a conflict between the two. If the two signs are the same, that is, both are positive or both are negative, it means that the physical environment deviation and the behavioral distribution deviation reflect the same direction of environmental discomfort, both indicating that the environment is too hot and the chickens are gathering. If the two signs are opposite, that is, one is positive and the other is negative, it means that the two reflect contradictory directions of environmental discomfort. For example, a positive physical environment deviation indicates that the temperature is too high, and a negative behavioral distribution deviation indicates that the chickens are gathering, which may be due to local low temperatures causing the gathering.

[0079] Here, conflict mode determination refers to the process of determining whether there is a contradiction (conflict) between physical environment deviation and behavioral distribution deviation by detecting the consistency of the signs between the two. It is used to identify contradictory information in environmental assessment and avoid distortion of the total comfort deviation calculation due to contradictory deviations.

[0080] Here, the nonlinear weighted fusion method refers to a method that adjusts the weights of each deviation based on the conflict detection results and the growth stage of the flock through a preset nonlinear function, and then sums the three types of deviations by weight to calculate the total comfort deviation. Unlike fixed weight fusion, it dynamically adjusts the weights according to the actual situation to solve the problem of evaluation distortion caused by deviation conflicts.

[0081] This invention introduces a conflict mode determination mechanism, based on the sign consistency detection of physical environment deviation and behavioral distribution deviation, to effectively identify contradictory information in environmental assessment. This overcomes the technical shortcomings of traditional fusion methods that ignore deviation conflicts, leading to distortion of the total comfort deviation, ensuring that the total deviation truly and accurately reflects the overall discomfort level of the environment. In calculating the total comfort deviation, this invention fully considers the physiological characteristics of chickens at different growth stages, dynamically adjusting weighting coefficients and nonlinear exponents to finely adapt to the comfort needs of chickens at each developmental stage. Simultaneously, it integrates the synergistic effects of physical environment parameters, gas concentrations, and chicken behavioral characteristics to construct a multi-dimensional comfort evaluation system, making the evaluation results more comprehensive and more closely aligned with the complex environmental conditions of real-world farming scenarios.

[0082] In some embodiments, the conflict mode determination by detecting the consistency of the signs of physical environment deviation and behavioral distribution deviation in the above steps, and obtaining the conflict detection result, can be achieved as follows: if the signs of physical environment deviation and behavioral distribution deviation are the same, it is determined to be a somatosensory abnormal mode, the biological priority conflict resolution mechanism is activated, and the conflict marker is set to the reverse regulation mode; if the signs of physical environment deviation and behavioral distribution deviation are opposite, it is determined to be a normal mode, and the conflict marker is set to the normal mode.

[0083] This invention enables the system to intelligently perceive environmental-behavioral conflicts by judging the consistency of the signs of deviations in the physical environment and deviations in behavioral distribution. When the signs are the same, it is determined to be an abnormal sensory mode, and a biological priority mechanism is activated, prioritizing the sensory feedback of the chicken population to reduce the risk of mis-regulation stress. When the signs are opposite, it is determined to be a normal mode, maintaining routine regulation while balancing biosafety and control stability. By setting conflict markers, a clear pattern basis is provided for subsequent calculations, improving the robustness of the algorithm. This mechanism only requires sign comparison, has low computational load, fast response, and is easy to implement in embedded real-time execution, effectively improving the response speed and operational stability of the control system.

[0084] In some embodiments, the total comfort deviation The calculation formula is: In the formula, Due to physical environmental deviations; This is a behavioral distribution bias; Gas concentration deviation; weighting coefficient , and All are dynamically adjusted according to the growth stage; nonlinear index and All settings are based on the growth stage; in reverse adjustment mode, the total comfort deviation... The calculation formula is: .

[0085] In some embodiments, step S140 may also be implemented by the following: An adaptive damping coefficient is constructed based on the absolute value of the total comfort deviation and historical oscillation indicators. The initial control increment is calculated based on the adaptive damping coefficient, the total comfort deviation, and a preset baseline proportional gain. Whether to activate the juvenile protection mechanism is determined based on the virtual physiological age and a preset juvenile age threshold. If the virtual physiological age is less than the preset juvenile age threshold, the juvenile protection mechanism is activated, and the initial control increment is forcibly limited to not exceed the preset maximum increment, resulting in the final adjustment increment. If the virtual physiological age is not less than the preset juvenile age threshold, the initial control increment is directly used as the final adjustment increment. The final adjustment increment is superimposed on the actuator output value of the previous cycle, and after physical limiting, a smooth control command is generated and output, which is then sent to the equipment that executes the adjustment of the feeding environment.

[0086] Here, the preset basic proportional gain refers to the fixed proportional coefficient that is pre-stored in the control system and used to calculate the initial control increment in conjunction with the total comfort deviation and the adaptive damping coefficient. It is preset by those skilled in the art based on the breed of chickens, the characteristics of the feeding equipment, and the control precision requirements. Its function is to set the basic proportional relationship between the control increment and the total comfort deviation to ensure that the magnitude of the initial control increment is reasonable.

[0087] Here, the initial control increment refers to the idealized, original change in control quantity calculated based on the current deviation before any limiting and protection mechanisms are applied. It reflects the basic control requirements corresponding to the current environmental deviation. It has not undergone early protection and physical limiting, and there may be a risk of excessive control.

[0088] It's important to note that the young chick protection mechanism is a safety control logic specifically designed for very young, vulnerable chicks. Because chicks' thermoregulatory centers are not fully developed, they have extremely poor tolerance to drastic environmental changes. Therefore, this mechanism is triggered when their virtual physiological age falls below a certain threshold, forcibly and strictly limiting the range of changes in control commands to prevent cold stress, heat stress, or even death caused by sudden environmental changes. Here, the preset young age threshold is a pre-set virtual physiological age value used to define whether a flock is in the young stage. When the calculated virtual physiological age is less than this value, the system determines that the flock is in a vulnerable period and the young chick protection mechanism needs to be activated.

[0089] Here, the final adjustment increment refers to the amount of adjustment to the actuator output value that is ultimately allowed to be executed after all logical judgments and limits (especially the mandatory limit of the juvenile protection mechanism).

[0090] This invention achieves a dynamic soft landing of control by introducing historical oscillation indices and adaptive damping coefficients. When the system shows signs of oscillation, it automatically weakens the control force to avoid environmental parameter fluctuations caused by over-adjustment, allowing the environment inside the chick house to stabilize more quickly. Simultaneously, through a young chick protection mechanism and the determination of virtual physiological age, the protection of vulnerable chicks is prioritized. A mandatory preset maximum increment limit fundamentally eliminates extreme environmental changes that might result from the algorithm's pursuit of rapid response, significantly reducing stress and mortality risks for chicks. In non-young stages, the system can respond quickly based on the magnitude of deviations; when gradual changes are needed, it can automatically or forcibly smooth the output. This intelligent adjustment strategy achieves a dynamic balance between speed and stability.

[0091] In some embodiments, the method further includes: calculating the growth offset rate based on the virtual physiological age and the current physical age; comparing the growth offset rate with a preset offset rate threshold; if the growth offset rate is greater than the preset offset rate threshold, marking an abnormal state identifier, temporarily shortening the control cycle to a preset high-frequency monitoring cycle, and triggering a high-frequency monitoring and early warning mechanism.

[0092] The following will describe an exemplary application of the embodiments of the present invention in a practical application scenario.

[0093] This embodiment provides another method for adaptive adjustment of feeding parameters based on the growth stage of a chicken flock, which includes the following steps: Step 1, Data Acquisition: Multidimensional status data is collected synchronously within the poultry house using a fixed 60-minute control cycle. The average fasting weight of the flock is obtained every 6 hours via a cage weighing platform. Combined with feed tower weighing and waste estimation, the daily feed intake per unit body weight is calculated. An 8×6 grid is divided on the ground using a top-view camera. Based on a target detection model such as YOLOv5, the distribution of chickens within each grid is statistically analyzed, and the Shannon entropy of the flock distribution is calculated. Simultaneously, environmental parameters such as temperature, humidity, ammonia concentration, and carbon dioxide concentration are monitored in real-time at the height above the chickens' backs. The aforementioned biometric data and environmental parameter vectors together form the initial input basis for subsequent control measures.

[0094] For biometric data collection, 3 to 5 cage-type weighing platforms were deployed in different temperature zones within the enclosure. Each platform had an area of ​​1.2 meters × 0.8 meters. The control cycle was set to 60 minutes, and raw weight data was collected every 6 control cycles (i.e., every 6 hours). The average fasting weight of the population was then calculated. This weight value remained constant over a 6-hour data collection interval for subsequent calculations; simultaneously, the initial weight of the material at the start of each period was recorded using the weighing sensors on the silo. With end material weight The total feed intake is obtained by subtracting the waste percentage estimated by the differential weighing method (usually 5% to 8%). The average feed intake is calculated by combining the current number of chicks N (initial number of chicks minus cumulative deaths and culls, updated daily), and then further divided by... Then multiply by 1000 to standardize, and get the daily food intake per unit body weight. (Unit: grams / kilogram / day).

[0095] In terms of behavioral distribution perception, the floor of the poultry house is gridded into m×n grids (usually divided into 8×6 grids, totaling 48 grids) by a top-down camera placed in the center of the roof. During each control cycle (60 minutes), the number of chickens in each grid j is counted in real time using the YOLOv5 model. Then calculate the Shannon entropy of the population distribution. This entropy value characterizes the degree of disorder in the population distribution, and the entropy value is the largest when the distribution is uniform. When people gather together for heating, the entropy value approaches zero.

[0096] In terms of environmental gas monitoring, an electrochemical sensor array is deployed at the height of the chicken's back to monitor the current ambient temperature in real time. ammonia concentration carbon dioxide concentration and ambient humidity Ammonia, produced by the decomposition of uric acid, reflects fecal accumulation and ventilation efficiency (the threshold is usually set at 20 ppm); carbon dioxide reflects the total metabolic intensity of the population (the threshold is usually set at 3000 ppm); and humidity affects perceived temperature. These three factors together constitute an environmental parameter vector. The aforementioned environmental parameter vector is read and updated in real time for each control cycle.

[0097] Step 2, Virtual Growth Age Dynamic Benchmark Based on Physiological Progress Calculation: Based on measured body weight and feed intake data, the system dynamically calculates the virtual physiological age reflecting the true developmental state of the flock. First, the average fasting body weight of the flock is mapped to the theoretical growth time using the inverse function of the Gompertz growth curve. Then, the feed intake metabolic equivalent age is calculated by combining the preset breed feed intake curve. Subsequently, the theoretical growth time and the feed intake metabolic equivalent age are weighted and fused to obtain the virtual physiological age. Finally, using this virtual physiological age as the sole time scale, a continuous, smooth target temperature and humidity dynamic benchmark that matches the actual physiological needs of the flock is generated through piecewise cubic spline interpolation combined with seasonal trend compensation.

[0098] This step aims to break away from the traditional farming model that relies on fixed physical ages to define growth stages. By integrating measured body weight and feed intake, it dynamically calculates a virtual physiological age that reflects the true developmental state of the flock. And based on its growth offset rate relative to physical age. Mark abnormal developmental states, regardless of Whether it exceeds the limit or not, it is based on To obtain continuous, climate-compensated temperature and humidity targets based on interpolation, if and only if | When the preset offset threshold is reached, a high-frequency monitoring and early warning mechanism is triggered in parallel.

[0099] (1) Calculate the theoretical growth time The average fasting weight of the group collected in step 1 Find the inverse function f of the standard Gompertz growth curve positive model, with f as the dependent variable. -1 : ,in, For maximum weight, and These are the variety-specific inflection point time parameter and the variety-specific growth rate parameter, respectively. This calculation process is completed from the population average fasting body weight. to theoretical growth time The mapping, output This serves as a theoretical developmental timeframe based on body weight data.

[0100] (2) Calculate the metabolic equivalent age of food intake Simultaneously, the standard feeding curve F(t) of the variety is interpolated and queried. Theoretical feed intake Then the actual measurement in step 1 will be performed. The query results Calculate the metabolic equivalent age of food intake The metabolic equivalent age reflects the degree of deviation between the current metabolic intensity and the standard feeding curve. If the daily food intake per unit body weight... Higher than the standard value If the ratio is greater than 1, it indicates active metabolism. > This suggests advanced physiological maturity, or conversely, delayed physiological maturity.

[0101] (3) Calculate virtual physiological age Regarding the above results and Perform weighted fusion to calculate virtual physiological age Based on the current physical age Determine the growth stage and set and The weighted fusion coefficient β is used to calculate the virtual physiological age using a weighted fusion formula. .

[0102] Note: Since the average fasting body weight of the population is updated every 6 hours, the baseline virtual physiological age is calculated based on the measured body weight every 6 control cycles (i.e., every 6 hours). If no feeding mutation event occurs during the subsequent 5 control periods, then A linear decay maintenance strategy is adopted: Let the effective weight obtained from the kth weighing be... Based on this, in the subsequent 5 control periods (t=1,2,3,4,5), ,in, This represents the daily age increment per unit time under the standard growth curve in the feeding guidelines (v = 1 day / day, indicating that the virtual physiological age increases by 1 day per day under normal circumstances). This represents the current metabolic state coefficient; if daily food intake per unit body weight is detected... relative to theoretical feed intake If the change exceeds ±15% (this threshold has been verified by broiler feeding behavior experiments and can effectively distinguish between daily fluctuations and significant physiological changes), a temporary correction will be immediately triggered. The correction formula is as follows: Generate the effective control cycle for the current period All environmental control decisions are based on this effective... implement.

[0103] This embodiment quantifies the spatial distribution characteristics of chicken flocks into behavioral distribution deviations through gridded image acquisition and Shannon information entropy calculation, and then performs nonlinear exponential weighted fusion with physical environment deviations. In particular, a biological priority reverse regulation mechanism is established when physical environment deviations and behavioral deviations conflict. This feature overcomes the limitation of existing technologies that rely solely on a single sensor to perceive the sensory state of chicken flocks, enabling the control system to possess environmental-behavioral collaborative perception capabilities.

[0104] (4) Calculate the growth offset rate Based on this, the growth offset rate is calculated. And set a preset threshold θ (usually 10%).

[0105] When | When |≤θ: the population development is considered to be basically synchronized with physical age, and is on a normal growth trajectory. The system maintains a regular monitoring frequency (e.g., data is collected every 30 minutes), and uses... The step is to obtain a dynamic baseline of the target environment. When | When |>θ: An abnormal growth and development of the population (advanced or delayed) is determined, triggering a high-frequency monitoring mode (shortening the data acquisition interval to 5 minutes and outputting an alarm signal). The system simultaneously records this offset event to the log for reference by aquaculture personnel for manual intervention, but still uses... This serves as a benchmark for obtaining dynamic baselines of the target environment, ensuring the continuity of environmental control.

[0106] (5) Obtain the dynamic baseline of the target environment The virtual physiological age obtained from the input And the outdoor temperature collected in real time by the outdoor weather station. By using continuous curve interpolation and seasonal compensation calculations, the corrected target environmental baseline is output. And with Environmental baseline calculations are performed using a single time scale to ensure the continuity of regulation.

[0107] The environmental baseline calculation process is as follows: First, based on pre-stored discrete data points of breed feeding standards (such as standard temperature and humidity on days 10, 15, 20, and 25), a continuous mapping relationship between virtual physiological age and environmental baseline is constructed through piecewise cubic spline interpolation. This mapping transforms the discrete step-like standards into a smooth continuous curve, allowing any virtual physiological age to be precisely calculated through interpolation, rather than only being able to query fixed values ​​corresponding to integer ages as in existing technologies; subsequently, based on the calculated... Obtain the target environmental benchmark from this continuous mapping. ,in For the target temperature, The target humidity is determined; a seasonal compensation coefficient is then introduced. The target temperature is corrected using the following formula: ,in, The heat transfer coefficient of the enclosure structure was set at 0.3~0.5 for standardized chicken houses and 0.5~0.7 for open chicken houses, which were preset constants. To accommodate the differences in enclosure structures, tests were conducted on standardized chicken houses under extreme conditions of winter (-5℃) and summer (35℃). =0.3 / 0.5 / 0.7. Standardized enclosed chicken house. When the coefficient of performance is 0.4, the temperature standard deviation is 0.9℃ and energy consumption is reduced by 15% compared to 0.3; semi-open chicken house. When the coefficient of variation is 0.6, the standard deviation is 1.1℃, a significant improvement compared to 1.8℃ at 0.5. Therefore, values ​​are taken in segments according to chicken house type. The target temperature is the real-time outdoor temperature, with 20°C serving as a reference for comfortable outdoor temperatures. This correction ensures that the target temperature adapts to actual climate loads under extreme climatic conditions; the final output is the seasonally corrected target temperature. and humidity This completes the dynamic mapping from physical time to physiological time and then to environmental benchmarks.

[0108] Step 3, Method for Constructing Total Comfort Deviation: Using the target environment dynamic benchmark as a reference, the system calculates the physical environment deviation, gas concentration deviation, and behavioral distribution deviation based on Shannon entropy of the flock distribution. Conflict pattern identification is performed by detecting the consistency of the signs of the physical environment deviation and the behavioral distribution deviation: if the signs are the same (e.g., sensor readings are high while the flock huddles together for warmth), it is determined to be an abnormal sensory mode, and a biological-priority conflict resolution mechanism is activated, calculating the total comfort deviation only based on the behavioral distribution deviation; conversely, if the signs are opposite, it is determined to be a normal mode. In this case, weighting coefficients and nonlinear indices are dynamically configured according to the flock's growth stage. The brooding period emphasizes behavioral distribution deviation, and the fattening period emphasizes physical environment deviation, with the introduction of nonlinear mapping. The three types of deviations are fused according to preset rules to generate a directional total comfort deviation, thereby accurately representing the overall control needs of the current environment.

[0109] This step uses the target environmental dynamic benchmark output in step 2 as a reference to construct a multimodal comfort deviation calculation system. First, physical environment deviation, gas concentration deviation, and behavioral distribution deviation are calculated separately. Behavioral distribution deviation is characterized by Shannon entropy quantification of flock distribution dispersion to represent the actual somatic sensation. Then, a nonlinear exponential weighting strategy is used to fuse the three types of deviations, and the weighting coefficients are dynamically adjusted according to the growth stage. When there is a contradiction between physical sensor indications and behavioral perception, a biologically prioritized conflict resolution mechanism is activated, adopting the behavioral distribution deviation for reverse adjustment. Finally, the total comfort deviation with directional signs and conflict markers are output, providing accurate deviation measurement and pattern recognition for execution control.

[0110] (1) Calculate the physical environment deviation, gas concentration deviation and behavioral distribution deviation. Calculate the physical environment deviation based on the target environment dynamic benchmark output in step 2. ,in, The humidity-temperature equivalent conversion factor is 0.05~0.1℃ / %RH, or can be dynamically calculated based on the saturated water vapor pressure curve. The current ambient temperature; Current ambient humidity; Calculate gas concentration deviation: only when Concentration >20ppm or Gas deviation is only calculated when the concentration is >3000ppm. Otherwise, a mandatory order This indicates that the gas environment is acceptable and does not produce any additional negative deviations in comfort.

[0111] Based on the population distribution Shannon entropy obtained in step 1 Calculate behavioral distribution deviation ,in The maximum theoretical entropy value is given by the uniform distribution, and m×n is the total number of grid divisions described in step 1. This indicates that the chickens are huddled together / under cold stress. This indicates excessive dispersion of the flock / heat stress. This indicates that the chickens are evenly distributed.

[0112] This embodiment quantifies the spatial distribution characteristics of chicken flocks into behavioral distribution deviations through gridded image acquisition and Shannon information entropy calculation, and then performs nonlinear exponential weighted fusion with physical environment deviations. In particular, a biological priority reverse regulation mechanism is established when physical environment deviations and behavioral distribution deviations conflict. This feature overcomes the limitation of existing technologies that rely solely on a single sensor to perceive the sensory state of chicken flocks, enabling the control system to possess environmental-behavioral collaborative perception capabilities.

[0113] (2) Conflict detection and pattern determination Whether there is a contradiction between physical environment deviation and behavioral distribution deviation: If If the symbols are the same, it is determined to be a sensory abnormality pattern. At this time, there is a discrepancy between the physical sensor indications and the chickens' behavioral perceptions (for example, the temperature sensor shows that it is too hot and needs to be cooled, but the chickens are huddled together and it shows that it is too cold and needs to be heated; or the temperature sensor shows that it is too cold and needs to be heated, but the chickens are too scattered and it shows that it is too hot and needs to be cooled). The biological priority conflict resolution mechanism is activated, so the biological priority is adopted first. And generate a reverse adjustment instruction, setting the conflict flag to reverse adjustment mode; if If the signs are opposite, it is judged as the normal pattern. Physical evidence and biological evidence corroborate each other, and the conflicting positions are marked as the normal pattern.

[0114] (3) Calculate the total comfort deviation The weighting is determined based on the conflict detection results, and the total comfort deviation is calculated using a nonlinear weighted fusion algorithm. .

[0115] Normal mode: The weighting coefficient , Adjust dynamically according to the growth stage (based on) Assess growth status during the brooding period. =0.6, =0.4; fattening period =0.3, =0.7), when When >0, The value is 0.2 otherwise, representing a nonlinear exponent. and Also based on the growth stage (brooding period) =1.3, =0.9; fattening period =0.8, =1.2), when the exponent is greater than 1, it amplifies the response to larger deviations, and when the exponent is less than 1, it compresses the effect of small fluctuations.

[0116] Note that the above parameter settings were optimized through comparative experiments in multiple chicken houses. Compared with fixed parameter configurations, they reduced the daily variation coefficient of population distribution entropy by 37%, achieved a comfort prediction accuracy of over 90%, and improved the feed conversion ratio by 0.07.

[0117] Reverse adjustment mode: (make =0, ignore item).

[0118] Step 4, Smooth the output of control commands: Based on the current total comfort deviation, the variance of historical deviations over the past 24 cycles, and virtual physiological age, an adaptive damping coefficient is constructed. The adaptive damping coefficient is generated through an adaptive damping function, and then the control increment is calculated and limited. Finally, a smooth fan / wet curtain control command is output, forming a perception-decision-execution closed loop to achieve gradual regulation that matches the physiological state.

[0119] (1) Input At the start of each control cycle (e.g., every 60 minutes), immediately acquire three key input data: one is the total comfort deviation output from step three. The second is the virtual physiological age calculated in step two. The first is used to identify the developmental stage of the chicken flock; the second is the current actuator output value. This refers to the control commands actually issued to equipment such as fans, heaters, or wet curtains in the previous cycle.

[0120] (2) Maintain and update the historical deviation window The system maintains a sliding time window of length 24 to dynamically store the comfort deviation sequence over the most recent 24 control cycles (approximately 24 hours). At the start of each new cycle, the system will output the total comfort deviation from step three. Add it to the end of the window and automatically remove the oldest history when the window is full, thus always maintaining a short-term memory of recent regulatory actions.

[0121] Note that in the initial stage of system operation, when there are fewer than 24 historical deviation data points, all actually collected data will still be used. Calculate the variance; in particular, when there is only one hour, =0, ensuring that the system can generate effective control commands from the first control cycle.

[0122] (3) Calculate historical oscillation indicators Based on the deviation sequence within the sliding window, its variance is calculated in real time as a historical oscillation indicator. This indicator quantifies the system's regulatory stability over a past period: if A large value indicates that environmental parameters are frequently overshooting and correcting, and are in a state of oscillation; if A smaller value indicates that the system is running smoothly.

[0123] (4) Constructing the adaptive damping coefficient Then, using historical oscillation indicators Construct an adaptive damping coefficient This is used to dynamically adjust the strength of the control response, and its expression is: ,in, The kurtosis parameter of the Sigmoid function is dimensionless. In this scheme, we take 3 to control the sensitivity of the damping as the magnitude of the deviation changes; The inflection point offset parameter of the Sigmoid function. In this plan, we take =1.2; Design a full factorial experiment for k∈[2,4], c∈[1.0,1.5] to monitor the response time and overshoot of a step disturbance. When k=3 and c=1.2, the damping at |ΔC|=2.0℃ smoothly transitions from 0.5 to 0.8, the response time is shortened by 30% and the overshoot is <3%; when k<3, the response is sluggish, and when k>4, oscillations occur. Therefore, k=3 and c=1.2 are determined. The oscillation penalty coefficient is dimensionless. In this embodiment, we take =0.8, used to adjust the weight of the impact of historical oscillations on the current damping, set =0.5 / 0.8 / 1.0, tested under a 24-hour temperature difference of 15℃ fluctuation. When =0.8, For every 1.0 increase, the adaptive damping coefficient... Increased by 0.44, compared to =0.5 Oscillation frequency reduced by 60%, compared to =1.0 Recovery time is reduced by 40%, achieving an optimal balance between stability and response speed. Therefore, it is determined that... =0.8.

[0124] The design logic for this adaptive damping coefficient is based on the current deviation. When the value is large, the sigmoid term approaches 1, providing low damping for a fast, large-step response; when... When the value is small, the sigmoid term approaches 0, providing high damping to prevent frequent adjustments caused by small fluctuations; at the same time, if historical oscillation indicators... A larger value (indicating recent repeated overshooting in the system) will result in a larger value in the denominator. The item increases, leading to The overall reduction forces the system to adopt a more moderate adjustment strategy, effectively suppressing environmental shocks.

[0125] This embodiment constructs an adaptive damping coefficient ξ by maintaining the historical sequence of total comfort deviation and calculating its variance, thereby achieving dynamic smoothing of control increments. Simultaneously, it introduces a juvenile protection mechanism based on virtual physiological age recognition to forcibly limit the amplitude of control commands. This feature effectively solves the problems of system oscillation and animal stress caused by overly coarse execution strategies in existing technologies.

[0126] (5) Calculate and limit the increment of the control. Obtaining the adaptive damping coefficient Then, calculate the basic control increment. ,in The base proportional gain is pre-calibrated based on the response characteristics of actuators such as fans and heaters. Then, it is determined whether to activate the early-age protection mechanism; if so, the virtual physiological age... Less than the young age threshold Then, the increment amplitude of the control will be forcibly limited, that is... Less than the maximum increment (For example, the fan speed should not be adjusted by more than 10% of the full range at a time) to avoid sudden changes in environmental parameters causing stress to chicks with weak thermoregulation ability and to ensure early survival rate.

[0127] The basic proportional gain Based on the dynamic response characteristics calibration of the actuator, for variable frequency fans, the step test should select a setting that results in a system overshoot of <5% and a settling time of <30 minutes. The measured value is 8-12, so the safe range [5,15] is chosen; Early childhood threshold. =Based on a 7-day study on the development of broiler thermoregulation, it was determined that chicks under 7 days old cannot effectively maintain a constant temperature. Experiments showed that implementing ±10% temperature control during this stage could reduce the mortality rate from 3.8% to 2.1%; maximum increment The value is set to 10% of the actuator's full scale. This value has been verified through testing in multiple chicken houses and can ensure the necessary control capabilities while avoiding drastic environmental changes.

[0128] (6) Output the final instruction and close the loop. Ultimately, the system will add the control increment to the current actuator state to generate new instructions. After physical limiting (e.g., ensuring output is within the 0%–100% range), the signal is immediately sent to actuators such as fans, evaporative cooling pads, and heaters. Simultaneously, Cached for the next cycle This completes the state transmission. The actions performed change the microenvironment inside the coop, which in turn affects the state data such as temperature and flock distribution collected in the next cycle step 1, thus forming a complete closed loop of perception → evaluation → execution → feedback, achieving gradual environmental regulation that dynamically matches the physiological state of the flock.

[0129] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of the present invention are included within the scope of protection of the present invention.

[0130] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of the invention, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the invention. The sequence numbers of the above-described embodiments of the invention are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0131] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, or apparatus. Without further limitations, an element defined by the phrase "comprising a..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. In the several embodiments provided by this invention, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components may be combined, or integrated into another system, or some features may be ignored or not performed.

[0132] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for adaptive control of feeding parameters based on the growth stage of a chicken flock, characterized in that, The method includes: Within each control cycle, the average fasting body weight, daily feed intake per unit body weight, Shannon entropy of the flock, and environmental parameter vectors including current ambient temperature, current ambient humidity, ammonia concentration, and carbon dioxide concentration are acquired synchronously for the target flock. Based on the average fasting body weight and daily feed intake per unit body weight of the flock, the virtual physiological age reflecting the true developmental status of the flock is calculated, and a dynamic baseline for the target environment is generated. Using the target environment dynamic benchmark as a reference, the physical environment deviation and gas concentration deviation obtained by the calculation are integrated with the directional behavioral distribution deviation calculated based on the Shannon entropy of the population distribution to construct the total comfort deviation; Based on historical oscillation indicators, total comfort deviation, and virtual physiological age, the initial control increment generated based on the total comfort deviation is dynamically damped to obtain a smooth control command, which is then executed to adjust the feeding environment equipment.

2. The method according to claim 1, characterized in that, Based on the flock's average fasting body weight and daily feed intake per unit body weight, a virtual physiological age reflecting the true developmental state of the flock is calculated, including: Using the population's average fasting weight as the dependent variable, the population's average fasting weight is mapped to the theoretical growth time by solving the inverse function of the pre-stored standard Gompertz growth curve model. Interpolate from the preset standard feeding curves of the breed to find the theoretical feed intake corresponding to the theoretical growth time, and calculate the equivalent age of feeding metabolism based on the daily feed intake per unit body weight and the theoretical feed intake. Based on the current physical age determination, the growth stage is dynamically configured with weight coefficients, and the theoretical growth time and the equivalent age of feeding and metabolism are weighted and fused to calculate the virtual physiological age. Between two valid weight data acquisitions, the metabolic state coefficient is determined based on the ratio of the current daily food intake per unit body weight to the theoretical food intake. The metabolic state coefficient is used to decay and maintain the virtual physiological age calculated based on the previous control cycle in order to update the virtual physiological age of the current control cycle. When the change in daily feed intake per unit body weight relative to the theoretical feed intake exceeds the preset threshold, it is determined that the flock is in a state of stress. The decay maintenance logic is paused, and the virtual physiological age is corrected based on the current measured feed intake data. The corrected value is used as the virtual physiological age for the current control cycle.

3. The method according to claim 2, characterized in that, Generate a dynamic baseline for the target environment, including: Based on pre-stored discrete data points of breed feeding standards, a continuous environmental baseline mapping function with virtual physiological age as the independent variable is constructed by piecewise cubic spline interpolation. Using the virtual physiological age of the current control cycle as input, the environmental baseline continuous mapping function is called to interpolate and obtain the corresponding target environmental temperature and target environmental humidity; The outdoor temperature collected in real time by the weather station outside the building is extracted from the environmental parameter vector. Combined with the preset seasonal compensation coefficient, the target environmental temperature is subjected to outdoor temperature feedforward compensation and seasonal trend correction to obtain the corrected target environmental temperature. The corrected target ambient temperature and target ambient humidity are jointly determined as the target environment dynamic benchmark.

4. The method according to claim 1, characterized in that, Construct the total comfort deviation, including: Calculate the physical environment deviation based on the target environment dynamic benchmark, current ambient temperature, and current ambient humidity; Determine whether the ammonia and carbon dioxide concentrations in the current control cycle exceed the preset safety thresholds; if they do, calculate the gas concentration deviation according to the preset calculation formula. Based on the Shannon entropy of the population distribution and the preset maximum theoretical entropy value, the behavioral distribution deviation is calculated; Conflict detection results are obtained by determining the conflict mode by detecting the sign consistency between physical environment deviation and behavioral distribution deviation. Based on the conflict detection results and the growth stage determined by virtual physiological age, a nonlinear weighted fusion method is used to calculate the total comfort deviation.

5. The method according to claim 4, characterized in that, Conflict patterns are determined by detecting the sign consistency between physical environment deviations and behavioral distribution deviations, resulting in conflict detection results, including: If the signs of the physical environment deviation and the behavioral distribution deviation are the same, it is determined to be a somatosensory abnormality mode, and the biological priority conflict resolution mechanism is activated, setting the conflict marker to the reverse regulation mode. If the signs of the physical environment deviation and the behavioral distribution deviation are opposite, it is determined to be a normal mode, and the position conflict is marked as a normal mode.

6. The method according to claim 5, characterized in that, In normal mode, the total comfort deviation The calculation formula is: ； In the formula, Due to physical environmental deviations; This is a behavioral distribution bias; Gas concentration deviation; weighting coefficient , and All adjustments are made dynamically according to the growth stage; Nonlinear exponent and All are set according to the growth stage; In reverse adjustment mode, the total comfort deviation The calculation formula is: 。 7. The method according to claim 1, characterized in that, Based on historical oscillation indices, total comfort deviation, and virtual physiological age, dynamic damping adjustment is applied to the initial control increment generated based on the total comfort deviation to obtain smooth control commands, including: An adaptive damping coefficient is constructed based on the absolute value of the total comfort deviation and historical oscillation indicators. The initial control increment is calculated based on the adaptive damping coefficient, total comfort deviation, and preset base proportional gain. Whether to activate the child protection mechanism is determined based on virtual physiological age and a preset child age threshold. If the virtual physiological age is less than the preset young age threshold, the young age protection mechanism is activated, and the initial control increment is forcibly limited to ensure that it does not exceed the preset maximum increment, thus obtaining the final adjustment increment. If the virtual physiological age is not less than the preset young age threshold, the initial control increment will be used directly as the final adjustment increment. The final adjustment increment is superimposed on the actuator output value of the previous cycle, and after physical limiting, a smooth control command is generated and output. The smooth control command is then sent to the equipment that performs the adjustment of the feeding environment.

8. The method according to claim 1, characterized in that, The method also includes: Calculate the growth offset rate based on the virtual physiological age and the current physical age; The growth offset rate is compared with a preset offset rate threshold. If the growth offset rate is greater than the preset offset rate threshold, an abnormal state identifier is marked, and the control cycle is temporarily shortened to a preset high-frequency monitoring cycle. At the same time, the high-frequency monitoring and early warning mechanism is triggered.

Images