An automatic feeding system for adapting to the growth of a flock of chickens

Abstract

The application discloses an automatic feeding system for self-adapting chicken growth, and belongs to the technical field of livestock breeding equipment. In order to solve the problem that the height of the existing feed line cannot be adjusted automatically and cannot adapt to the dynamic change of the chicken population, the application comprises a liftable feed line, a lifting mechanism for driving the feed line to lift, ultrasonic sensors installed along the length direction of the feed line at intervals, and a controller. The controller is configured to obtain M continuous distance measurement values within a sliding time window for the ultrasonic sensors, calculate the variance, determine that there is a feeding state target when the variance is less than a preset stability threshold, take the mean value of the window as a candidate height value, eliminate the candidate values with a distance value difference from the ground reference distance value less than a preset ground threshold to obtain effective chicken back height values, statistically obtain the average back height of the current chicken population from all the effective chicken back height values, and finally calculate the target feed line height and control the lifting mechanism to adjust the feed line. The application is mainly used in chicken farms to realize self-adaptive adjustment of the height of the feed line.

Description

Technical Field

[0002] This invention belongs to the technical field of livestock breeding equipment. More specifically, this invention relates to an automatic feeding system that adapts to the growth of chicken flocks. Background Technology

[0003] In large-scale poultry farming, automated feeding systems are widely used. A typical automated feeding system includes a feed hopper, conveying pipes, feed trays, and a conveying mechanism that drives the feed along the pipes. Some systems also have a lifting mechanism to adjust the pipe height to accommodate the feeding needs of poultry at different growth stages. However, existing technologies for automatically adjusting the feed line height usually rely on manual observation or timed control, making it difficult to match the actual growth status of the flock in real time.

[0004] Using ultrasonic sensors for non-contact distance measurement is a potential technique for obtaining the back height of poultry. However, in actual poultry farming environments, the application of this technology faces multiple obstacles. First, poultry density is high and activity is frequent in poultry houses, resulting in complex echo signals received by ultrasonic sensors, including reflections from the poultry's back, ground reflections, and dynamic interference from moving poultry. When poultry move, distance measurements fluctuate drastically, making it difficult to extract stable and valid signals. Simply averaging or using fixed thresholds often leads to misinterpreting interference signals from moving birds as valid data, resulting in inaccurate height measurements.

[0005] Secondly, there are many periods in the breeding environment when there are no poultry, such as after lights out at night or during feeding intervals, when poultry leave the feed tray area. During these times, the distance values ​​measured by the ultrasonic sensors are all from the ground. If the system fails to effectively recognize this situation, it will misinterpret these ground distance values ​​as the height of the poultry, causing the feed line to be incorrectly lowered too low, preventing the poultry from properly approaching the feed tray the next day. Relying solely on the ultrasonic signal itself is insufficient to distinguish whether the feeding target is the poultry or the ground.

[0006] Furthermore, the posture of poultry while feeding also poses challenges to measurement. When poultry dip their heads into the feed tray, their backs may be obscured by the edge of the tray. In this case, the echo received by the ultrasonic sensor comes from the feed tray, not the poultry's back. If the system misjudges the height of the feed tray as the height of the poultry, the feed line height will remain unchanged or be incorrectly adjusted, especially in the young poultry stage, where the obstruction phenomenon is more common due to their small size. In addition, the frequent head movements of poultry during feeding cause slight fluctuations in the ultrasonic signal. If these fluctuations are simply attributed to movement and discarded, a large number of valid measurements will be lost during peak feeding periods, resulting in data update delays.

[0007] The aforementioned problems are intertwined, making the accurate extraction of poultry back height from complex farming environments a technical challenge. Existing ultrasonic ranging solutions are applicable in static or low-density scenarios, but in the dynamic conditions of intensive farming, the reliability and stability of the measurement results are insufficient to meet the requirements of automatic control. Therefore, a technical solution is needed that can effectively identify and filter out various interferences and accurately extract the back height of poultry while they are feeding. Summary of the Invention

[0008] One objective of this invention is to address the problem that existing ultrasonic sensors struggle to accurately extract the back height of chickens from complex echo signals in intensive farming environments. This invention provides a multi-sensor data filtering method based on sliding window variance discrimination and ground reference value elimination to effectively identify the back height of chickens during feeding.

[0009] To achieve the above objectives, the present invention provides an automatic feeding system for adaptive chicken flock growth, comprising: a liftable feeding pipeline with multiple feed trays mounted thereon; a lifting mechanism for driving the feeding pipeline to rise and fall; multiple ultrasonic sensors fixedly installed at intervals above the feeding pipeline along its length, each ultrasonic sensor continuously emitting ultrasonic waves at a sampling frequency of not less than 20Hz and receiving echoes, outputting a time-series distance signal; and a controller connected to the ultrasonic sensors and the lifting mechanism, wherein the controller has a preset ground reference distance value and is configured to perform the following steps: for each ultrasonic sensor, acquiring M consecutive distance measurement values ​​within a sliding time window. M≥20; Calculate the variance of the distance measurement values ​​within the sliding time window. When the variance is less than a preset stability threshold, it is determined that there is a feeding target below the sensor within the current time window, and the mean of the distance measurement values ​​within the window is used as a candidate height value; Compare the candidate height value with the ground reference distance value, and remove candidate values ​​whose difference is less than a preset ground threshold to obtain the effective chicken back height value; Statistically analyze all effective chicken back height values ​​obtained by all ultrasonic sensors within a set time period, and take the median or mode as the average back height of the current flock; Calculate the target feed line height based on the average back height, and control the lifting mechanism to adjust the feed pipeline to the target feed line height.

[0010] As a preferred embodiment, the system further includes a passive infrared sensor installed corresponding to each ultrasonic sensor, the detection area of ​​the passive infrared sensor spatially overlapping the beam area of ​​the ultrasonic sensor; the controller is further configured to: synchronously acquire the distance measurement value sequence of the ultrasonic sensor and the heat source signal sequence of the passive infrared sensor within the sliding time window; calculate the variance of the distance measurement value sequence and calculate the fluctuation characteristic value of the heat source signal sequence; when the variance is less than a preset stability threshold and the fluctuation characteristic value of the heat source signal sequence exceeds a preset live animal threshold, determine that there are live chickens in the current window in a feeding state, and use the average distance measurement value of the window as a candidate chicken back height value; when the variance is less than a preset... If the stability threshold is reached, but the fluctuation characteristic value of the heat source signal sequence does not exceed the preset live animal threshold, the controller further determines whether the mean value of the heat source signal sequence exceeds the ambient temperature threshold. If yes, it is determined to be a stationary non-chicken heat source interference, and the data in this window is discarded. If no, it is determined to be a period without chickens feeding, the data in this window is discarded, and a sleep mode is triggered to pause sampling until the heat source signal recovers. When the variance is greater than or equal to the preset stability threshold, but the fluctuation characteristic value of the heat source signal sequence exceeds the preset live animal threshold, it is determined to be a moving live chicken. The current distance value is not accepted, but the activity intensity of the chickens in this area is recorded. The controller calculates the average back height based on all valid candidate chicken back height values ​​and dynamically adjusts the width of the sliding time window according to the activity intensity of the chickens.

[0011] As a preferred embodiment, the controller is further configured to: acquire the current actual height of the feeding pipeline, and calculate in real time the theoretical distance value of the tray below each sensor position based on the fixed installation height of each ultrasonic sensor; when the variance within the sliding time window is less than a preset stability threshold and the fluctuation characteristic value of the heat source signal sequence exceeds a preset liveness threshold, further compare the average distance measurement value within the window with the theoretical distance value of the tray at the corresponding position; if the difference between the average value and the theoretical distance value of the tray is greater than or equal to a preset occlusion threshold, then the average value is used as a candidate chicken back height value; if the difference between the average value and the theoretical distance value of the tray is less than the preset occlusion threshold, then it is determined that the current sensor is occluded by the tray, the data of the window is discarded, and the event of the sensor being occluded is recorded.

[0012] As a preferred embodiment, the system also includes a pressure sensor installed in each feed tray to detect the load weight on the feed tray; the controller is further configured to: acquire the load weight signal within a time window corresponding to the candidate chicken back height value, and calculate its average value; determine the theoretical back height range corresponding to the average load weight based on the standard weight-back height relationship of chickens of the current age; if the candidate chicken back height value exceeds the upper limit of the theoretical back height range, and the excess exceeds a preset stacking threshold, then the candidate chicken back height value is determined to originate from chicken stacking and is discarded; for the discarded measurement points, the historical height value effectively measured by the sensor within adjacent time windows, or the weighted average of the effective height values ​​of adjacent position sensors within the same time window, is used as the representative value of the chicken back height in that area.

[0013] As a preferred embodiment, the controller is further configured to: synchronously calculate the fluctuation variance of the load weight signal detected by the pressure sensor within the sliding time window; when the variance of the ultrasonic distance measurement sequence is greater than or equal to a preset stability threshold, but the fluctuation variance of the load weight signal is less than a preset body stability threshold, and the fluctuation characteristic value of the heat source signal sequence exceeds a preset liveness threshold, determine that the chickens in the current window are in a feeding state, and use the median of the ultrasonic distance measurement sequence as a candidate chicken back height value; when the variance of the ultrasonic distance measurement sequence is greater than or equal to the preset stability threshold, and the fluctuation variance of the load weight signal is also greater than or equal to the preset body stability threshold, determine that the chickens in the current window are in a moving state, and discard the data for that window.

[0014] As a preferred embodiment, the controller is further configured to: record candidate chicken back height values ​​and their corresponding timestamps generated by multiple ultrasonic sensors sequentially distributed along the length of the feeding pipeline within multiple consecutive sliding time windows; detect whether, within a preset short time interval Δt, at least three adjacent ultrasonic sensors sequentially generate candidate chicken back height values, and the maximum difference between these candidate values ​​is less than a preset height consistency threshold; if the above situation is detected, it is determined as a chicken running past quickly, and all candidate chicken back height values ​​generated within the corresponding time window are marked as pseudo data and discarded; if the above situation is not detected, the candidate chicken back height values ​​determined according to claim 5 are retained for subsequent statistical calculation of average back height.

[0015] As a preferred embodiment, the system further includes suspension assemblies at both ends of the feed pipeline for support and vibration. The suspension assemblies include a clamp, an elastic layer, a vibration-damping rod, a vibration-isolation joint, and a vibrator. The clamp is a split-ring structure that hugs the end of the feed pipeline. The elastic layer, a ring-shaped rubber or polyurethane bushing, is disposed between its inner wall and the outer wall of the feed pipeline. This elastic layer transfers the static load of the feed pipeline to the clamp and isolates the vibration generated by the vibrator from upward transmission. The vibration-damping rod is a rigid member, its upper end connected to the suspension point of the lifting mechanism, and its lower end connected to the clamp via the vibration-isolation joint. The vibration-isolation joint is an elastic element used to further attenuate the upward transmission of vibration. The vibrator is fixedly installed on the feed pipeline to generate periodic excitation force, driving the feed pipeline to vibrate through the clamp and the elastic layer. The controller is also configured to control the start / stop, excitation frequency, and excitation amplitude of the vibrator based on the average back height of the flock, feeding activity, or a preset cleaning mode.

[0016] As a preferred embodiment, the controller is further configured to: calculate the feeding activity index of each feed tray in real time based on the load weight signal detected by the pressure sensor, wherein the feeding activity index is the high-frequency fluctuation energy value of the load weight signal within a sliding time window; statistically analyze the feeding activity index of all feed trays within the same time period to obtain the current flock's group feeding activity; compare the group feeding activity with a preset activity benchmark value to obtain the activity deviation; dynamically adjust the operating parameters of the vibrator according to the activity deviation: when the activity deviation is positive and exceeds a first threshold, it is determined that the flock is actively feeding, and the vibration frequency and / or vibration amplitude are increased to promote feed flow and feeding; when the activity deviation is negative and exceeds a second threshold, it is determined that the flock is passively feeding, and the vibration frequency and / or vibration amplitude are reduced or the vibrator is paused to avoid vibration disturbing the flock; when the activity deviation is between the first threshold and the second threshold, the current vibration parameters are maintained.

[0017] The present invention has at least the following beneficial effects: 1. This invention achieves accurate measurement of chicken back height in complex farming environments through a multi-level data filtering and fusion mechanism. Addressing interference factors in intensive farming scenarios such as chicken movement, ground reflection, feed tray obstruction, multiple chickens stacking, head swaying, and rapid running, the system employs sliding window variance discrimination, passive infrared liveness verification, theoretical feed tray distance comparison, weight-back height relationship verification, pressure fluctuation characteristic analysis, and multi-sensor spatiotemporal correlation detection to progressively eliminate invalid and interfering signals, ultimately extracting the effective measurement value that truly reflects the chicken's back height under feeding conditions. Compared to existing methods using a single sensor or simple threshold judgment, this invention's multi-dimensional fusion discrimination mechanism significantly improves the accuracy and reliability of the measurement data, enabling the system to automatically and accurately adjust the feed line height according to the actual growth status of the flock, avoiding the lag and uncertainty of manual experience-based adjustments.

[0018] 2. This invention achieves independent and controllable vibration of the feed pipeline by physically decoupling the vibration function from the lifting function. The vibration suspension mechanism, composed of a sleeve, elastic layer, vibration damping rod, and vibration isolation joint, ensures that vibration energy is effectively transferred to the feed pipeline to promote feed flow, while the two-stage vibration isolation design prevents vibration from being transmitted upwards to the lifting mechanism, eliminating the impact of vibration on lifting accuracy and mechanism lifespan. Furthermore, by calculating the feeding activity index in real time and dynamically adjusting the operating parameters of the vibrator, the vibration intensity is automatically matched to the actual feeding needs of the flock: vibration is increased to promote feed delivery when feeding is active, and vibration is reduced or stopped when feeding is passive to avoid disturbance, thus improving feeding efficiency and reducing unnecessary energy consumption and stress. Compared with existing technologies using fixed-mode vibration, the adaptive vibration control of this invention offers better energy efficiency and animal welfare.

[0019] 3. This invention actively guides the macroscopic flow of feed by identifying the spatial distribution characteristics of feeding activity and dynamically adjusting the height difference between the two ends of the feed pipeline. When feeding activity in a certain area is significantly higher than in other areas, the corresponding end is lowered while the other end is raised to create a downward slope. This allows the feed to flow naturally to the area with high demand under gravity, effectively alleviating the problem of localized feed shortages or accumulations caused by uneven distribution of chickens along long feed lines. Compared with existing technologies that rely solely on vibration to promote feed delivery, the tilt adjustment and vibration control of this invention create a synergistic effect: tilting solves the macroscopic flow problem, while vibration promotes microscopic feed delivery. Together, they improve the uniformity of feed distribution and the balance of feeding among the chickens. At the same time, the tilt amplitude is limited to a preset threshold to ensure that the overall height remains within the optimal range, balancing feeding comfort and feed guidance effectiveness.

[0020] 4. This invention achieves differentiated vibration control for different sections of a long-distance feed line by setting up multiple independently controlled vibration suspension mechanisms. Based on the comparison between the local feeding activity of each section's feed pan and the average level of the entire population, the operating parameters of the vibrator in the corresponding section are adjusted independently: vibration is strengthened in sections with insufficient feed to promote feeding, while vibration is weakened in sections with sufficient feed to prevent feed accumulation. Compared with the overall control of a single vibration source, the segmented independent control of this invention can accurately compensate for the differences in feed supply caused by long-distance conveying, making the feeding speed of the far-end feed pan more consistent with that of the near-end, further improving the uniformity of feed supply throughout the entire feed line. This zonal control, together with the aforementioned tilt adjustment and overall vibration adjustment, forms a multi-level collaborative control system, jointly realizing intelligent and precise management of feed distribution.

[0021] 5. This invention achieves a dynamic and reliable connection between the fixed hopper and the lifting feed line through a flexible connection module. The outer flexible bag ensures sealing and bending freedom, while the internal spring skeleton prevents the bag from collapsing and blocking the flow channel. The quick-change connector, composed of a ball head and an elastic retaining ring, allows for large-angle swing while maintaining constant clamping force. The redundant length design avoids tensile tearing. Compared with traditional canvas sleeves or corrugated hoses, the composite structure of this invention combines flexibility, support, and sealing, maintaining unobstructed flow and sealed interfaces during frequent lifting and lowering of the feed line. This effectively solves common on-site problems such as powder leakage, collapse and blockage, and fatigue fracture at the connection, providing a reliable hardware foundation for the long-term stable operation of the entire adaptive feeding system.

[0022] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of an automatic feeding system for adaptive chicken flock growth according to the present invention. Figure 2 This is a schematic diagram of the structure of the vibration suspension mechanism of the present invention; Figure 3 This is a schematic diagram of the flexible connection module of the present invention; Figure 4 This is a schematic diagram of the automatic feeding system for adaptive chicken flock growth according to the present invention.

[0024] In the diagram, 1-hopper, 2-hopper support, 3-flexible connection module, 4-screw conveyor motor, 5-feeding pipeline, 6-material tray, 7-lifting mechanism, 8-ultrasonic sensor, 9-passive infrared sensor, 10-hanging assembly, 11-vibrator, 12-pressure sensor, 13-drain valve, 101-vibration damping rod, 102-vibration isolation joint, 103-hoop, 104-elastic layer, 301-elastic clamp, 302-cloth bag, 303-lightweight spring frame. Detailed Implementation

[0025] The present invention will now be described in further detail with reference to the accompanying drawings, so that those skilled in the art can implement it based on the description.

[0026] like Figure 1-4 As shown, in one embodiment of the adaptive automatic feeding system for chicken flock growth of the present invention, it is applied to large-scale breeding scenarios of broilers or laying hens. The system mainly includes a liftable feeding pipeline, multiple feed trays installed on the feeding pipeline, a lifting mechanism for driving the feeding pipeline to rise and fall, multiple ultrasonic sensors, and a controller. In this embodiment, the feeding pipeline can be a circular steel pipe, and its length can be set to 10-100 meters depending on the span of the chicken house. A screw conveyor motor is installed at one end of the feeding pipeline, and this motor is connected to the screw conveyor mechanism passing through the feeding pipeline via a reducer. The screw conveyor mechanism preferably adopts a shaftless screw structure, made of spring steel or stainless steel, which rotates under the drive of the motor, pushing feed along the feeding pipeline to the far end. An emptying valve is installed at the other end of the feeding pipeline. This valve remains closed during normal feeding to seal the end of the pipeline; it can be opened in cleaning mode to discharge residual feed with vibration. The feed line's inlet is connected to the outlet of the fixed feed hopper via a flexible connection module. This module allows the feed line to maintain a dynamically sealed connection with the fixed feed hopper during lifting and lowering. Multiple receiving trays are hung on the feed line at certain intervals (e.g., 0.5-1.2 meters, adjustable according to chicken age and stocking density), each tray connected to the feed line via a hanging bracket. The lifting mechanism can use an electric winch with a steel cable suspension system. The electric winch is installed on the steel structure at the top of the chicken house, and the lower end of the steel cable is connected to the two ends of the feed line's suspension points. Lifting and lowering the feed line is achieved by rotating the electric winch in both directions. An industrial-grade ultrasonic ranging module, such as the Maxbotix MB1040, can be used as the ultrasonic sensor. It has an IP67 protection rating, a 15° beam angle, and can withstand the high humidity, dust, and ammonia environment inside the chicken house. Multiple ultrasonic sensors are fixedly installed at intervals along the length of the feed pipeline, for example, one every 1-3 meters. The sensors are fixed to independent crossbeams via dedicated brackets, ensuring the crossbeams do not contact the feed pipeline and maintain a constant sensor installation height. Each ultrasonic sensor continuously emits ultrasonic waves downwards at a sampling frequency of at least 20Hz and receives the echoes, outputting a time-series distance signal. The controller can be a programmable logic controller (PLC), such as the Siemens S7-1200 series, which has a built-in high-speed counting module for acquiring the ultrasonic sensor signals and controlling the operation of the electric winch through digital or analog output modules. The controller has a pre-stored ground reference distance value, which can be obtained by averaging measurements taken from the ground using ultrasonic sensors during system installation and commissioning, while the chicken coop is vacant.

[0027] In terms of mechanical assembly, the electric winch of the lifting mechanism is bolted to the I-beam or a special hanger on the top of the chicken house. One end of the wire rope is wound around the winch drum, and the other end is connected to the hanging points at both ends of the feed pipeline after being redirected by a fixed pulley. Each feed tray on the feed pipeline has a hanging device at its installation position. The feed tray is flexibly connected to the feed pipeline through the hanging device (e.g., the hanging device is made of elastic material, or the hanging device is made of plastic or metal with flexible gaskets at the contact points with the feed pipeline) to avoid rigid impact during lifting. The ultrasonic sensor bracket is made of stainless steel. The upper end of the bracket is fixedly connected to the crossbeam, and the lower end has mounting holes. The sensor is fixed in the mounting holes through threaded connections, with the sensor's detection surface pointing vertically downwards. All signal lines from the ultrasonic sensors are routed along the crossbeam to a waterproof junction box, and then connected to the signal input module of the controller via shielded cables. The controller is installed in an electrical control cabinet outside the chicken house and communicates with the host computer monitoring system via fieldbus or industrial Ethernet. In terms of electrical connections, the motor driver of the electric winch is connected to the digital output module of the controller. The controller controls the winch's movement by outputting forward, reverse, and stop commands. If frequency conversion speed regulation is used, the controller controls the winch's operating speed by outputting a 0-10V voltage signal or a 4-20mA current signal through the analog output module. The ultrasonic sensor is powered by the controller's DC 24V power supply module, and its output signal is an analog voltage or digital pulse, which is directly connected to the controller's analog input module or high-speed counter channel.

[0028] During the control logic and operation, after the system powers on, the controller first reads the pre-stored ground reference distance value and initializes various parameters. In operation, for each ultrasonic sensor, the controller continuously acquires distance measurements using a sliding window, with the window width M set to 20 sampling points (corresponding to 1 second at a sampling frequency of 20Hz). The controller calculates the variance of the 20 distance values ​​within the window and compares this variance with a preset stability threshold. The stability threshold can be calibrated based on the amplitude of the chicken's head swaying while feeding; for example, it can be set to 0.5 square centimeters for chicks and 2 square centimeters for adult chickens. When the variance is less than the preset stability threshold, the controller determines that there is a feeding target below the sensor within the current window and uses the average distance measurement value within that window as a candidate height value. Subsequently, the controller compares this candidate height value with the ground reference distance value. If the difference is less than the preset ground threshold (e.g., 5 centimeters), the candidate value is determined to originate from ground reflection rather than the chicken's back and is discarded; otherwise, it is retained as a valid chicken back height value. All valid chicken back height values ​​generated by ultrasonic sensors within a set time period (e.g., 5 minutes) are aggregated in the controller. The controller performs statistical analysis on these values ​​and takes the median or mode as the current average back height of the flock. Since the median is less affected by extreme values ​​when the flock moves, this embodiment preferably uses the median as the statistical measure. Finally, the controller calculates the target feed line height based on the average back height. The calculation method can use the empirical formula: Target feed line height = Average back height × 0.85. The controller compares the current feed line height with the target feed line height. If the deviation exceeds the allowable range (e.g., ±2 cm), it controls the lifting mechanism to operate, driving the feed line to rise or fall to the target height.

[0029] In a specific breeding cycle example, when the flock is in the brooding period (e.g., 10 days old), the average back height is approximately 15 cm. The controller calculates the target feed line height as 12.8 cm and controls the lifting mechanism to lower the feed line to this height. As the flock grows, when the average back height increases to 25 cm, the target feed line height is updated to 21.3 cm, and the controller again drives the lifting mechanism to raise the feed line. Throughout the process, ultrasonic sensors continuously monitor the system, updating the average back height every 6 hours (or adjustable from 30 minutes to 12 hours as needed) and determining if adjustment is necessary. To avoid erroneous adjustments due to brief flock activity, the controller can be configured with an adjustment dead zone or a delayed confirmation mechanism, such as only executing an action when three consecutive measurements indicate the need for adjustment. The deployment density of ultrasonic sensors can be adjusted according to different chicken house widths and accuracy requirements; for example, when the feed line length exceeds 50 meters, the density can be increased to one sensor every 2 meters. The preset stability threshold and ground threshold can also be fine-tuned based on actual field measurement data. Compared to conventional machine vision solutions, this system offers significant advantages in the real-world environment of a chicken farm: camera-based solutions often face challenges such as rapid dust accumulation on the lens and decreased image recognition algorithm stability due to uneven lighting, requiring frequent maintenance. In contrast, the ultrasonic sensor used in this system is insensitive to changes in ambient light, and its IP67 protection rating effectively resists dust and moisture corrosion, resulting in higher long-term reliability. Through these methods, the system achieves automatic feed line adjustment based on the actual back height of the chickens, avoiding the lag and inaccuracy of manual adjustments and ensuring that the feed line height always matches the chickens' growth status.

[0030] Furthermore, in another embodiment, a passive infrared sensor is added to the described system, corresponding to each ultrasonic sensor, to address the issues of mis-collection of ground signals and heat source interference during periods without chickens. The passive infrared sensor can be an LHI878 pyroelectric infrared sensor, whose detection area is adjusted by a Fresnel lens to spatially overlap with the beam area of ​​the ultrasonic sensor, ensuring that both sense the same spatial range. The passive infrared sensor also uses IP67 protection to adapt to the high humidity and dusty environment inside the chicken house. The controller needs to have multi-channel analog or digital input capabilities, for example, using a Siemens S7-1200 series PLC with expanded analog input modules, to simultaneously acquire distance signals from the ultrasonic sensor and heat source signals from the passive infrared sensor. The controller has several preset parameters, including a stability threshold for determining the feeding status target, a liveness threshold for determining the presence of live organisms, an ambient temperature threshold for distinguishing environmental heat sources, and an activity intensity coefficient for subsequently dynamically adjusting the width of the sliding window.

[0031] In terms of mechanical assembly, a passive infrared sensor is mounted next to each ultrasonic sensor via an integrated bracket. The orientation and angle of the sensors are calibrated to ensure that their detection area substantially coincides with the projection area of ​​the ultrasonic beam on the ground. The sensor brackets are made of stainless steel and fixed to independent crossbeams, avoiding contact with the feed lines. All sensor signal lines are routed along the crossbeams to a waterproof junction box, and then connected to different input channels of the controller via shielded cables. For electrical connections, the ultrasonic sensors output analog voltage signals (e.g., 0-10V corresponding to a distance of 0-500cm), which are connected to the PLC's analog input module. The passive infrared sensors output digital signals (0 / 1 indicating no / present heat source) or analog signals (reflecting heat source intensity), which are connected to the PLC's digital or analog input module depending on the sensor type. The controller communicates with each module via an internal bus to achieve synchronous data acquisition.

[0032] In terms of control logic, the system uses the same sliding time window (M sampling points) as a basis, synchronously acquiring ultrasonic distance measurement sequences and passive infrared heat source signal sequences within each window. The controller first calculates the variance of the ultrasonic distance values ​​and simultaneously calculates the fluctuation characteristic value of the heat source signal sequence. This fluctuation characteristic value can be taken as the standard deviation or peak-to-peak value of the heat source signal within the window, used to distinguish between live chickens (whose signals fluctuate due to breathing and slight movement) and stationary heat sources (such as a metal plate heated by the sun, whose signals are stable). When the ultrasonic variance is less than a preset stability threshold and the heat source fluctuation characteristic value exceeds a preset live animal threshold, the controller determines that there is a live chicken in a feeding state within the current window, uses the average distance measurement value of that window as the candidate chicken back height value, and proceeds to the subsequent processing flow. When the ultrasonic variance is less than the stability threshold but the heat source fluctuation characteristic value does not exceed the liveness threshold, the mean of the heat source signal sequence is further judged to see if it exceeds the ambient temperature threshold. If the mean exceeds the ambient temperature threshold (e.g., 35℃, higher than the upper limit of the chicken house ambient temperature), it is determined to be a stationary non-chicken heat source interference (such as heating equipment or heated objects), and the data in this window is discarded while maintaining normal operation. If the mean does not exceed the ambient temperature threshold, it is determined to be a period when no chickens are feeding (such as chickens resting at night), and the data in this window is discarded and a sleep mode is triggered, pausing sampling until the heat source signal recovers, in order to reduce system power consumption and extend sensor life. When the ultrasonic variance is greater than or equal to the stability threshold but the heat source fluctuation characteristic value exceeds the liveness threshold, it is determined to be a moving live chicken (such as chickens running by). At this time, the current distance value is not accepted, but the activity intensity of chickens in this area is recorded. The activity intensity can be quantified based on the effective value of the heat source signal within the window or the number of zero crossings. The controller calculates the average back height from all valid candidate chicken back height values ​​and dynamically adjusts the width of the sliding time window based on the recorded chicken activity intensity: when activity intensity is high (e.g., during peak feeding periods), the window width is appropriately reduced (e.g., from 20 points to 10 points) to improve temporal resolution and capture brief feeding moments; when activity intensity is low (e.g., at night), the window width is appropriately increased (e.g., to 30 points) to improve stability. The window width adjustment range can be limited to 10-50 sampling points, and the adjustment step size can be calibrated based on actual measurement data.

[0033] In a specific example of a breeding cycle, during peak daytime feeding periods, the chickens frequently move around the feed trays. When a chicken feeds under the sensor, the ultrasonic variance is less than 0.8 square centimeters, and the heat source fluctuation characteristic value (standard deviation) reaches 0.5V (exceeding the preset live threshold of 0.3V). The controller identifies this as a valid live chicken and records a back height of 25 cm. When a chicken runs quickly under the sensor, the ultrasonic variance reaches 3 square centimeters (exceeding the stability threshold), but the heat source fluctuation characteristic value remains at 0.5V. The controller identifies this as a moving live chicken, disregards the distance value, but records high activity intensity and dynamically adjusts the sliding window width from 20 points to 15 points to increase the probability of capturing subsequent feeding moments. After lights out at night, the chickens leave the feed trays, and there are no live chickens under the sensor. At this time, the ultrasonic variance is less than 0.5 square centimeters, the heat source has no fluctuation (characteristic value of 0), and the heat source mean is consistent with the ambient temperature (20℃). The controller identifies this as a period without chicken feeding, triggers the sleep mode, and pauses sampling. When the temperature rises in the early morning, sunlight shines through the window onto a metal object near the sensor, raising its temperature to 40°C. At this point, the average heat source temperature exceeds the ambient temperature threshold (35°C), but there is no fluctuation in the heat source temperature. The controller determines this as heat source interference, discards the data, and continues monitoring without going into sleep mode. Once the chickens begin to move around in the early morning, the heat source temperature fluctuates again, and the system automatically exits sleep mode and resumes normal sampling. Through this multi-dimensional fusion judgment and dynamic adjustment mechanism, this system effectively solves the signal recognition problem of a single ultrasonic wave in complex environments, achieving reliable measurement of the height of the chickens' backs.

[0034] Furthermore, in another embodiment, a feed tray obstruction recognition mechanism is introduced based on the described system to address the problem that ultrasonic sensors cannot measure the true chicken back height due to feed tray obstruction. To achieve this function, the controller needs to obtain the current actual height of the feed pipeline in real time. This information can be obtained by installing an encoder or displacement sensor on the lifting mechanism, such as installing a rotary encoder at the shaft of the electric winch and calculating the feed pipeline height by the release length of the wire rope, or by installing a pull-wire displacement sensor at the end of the feed pipeline to directly measure the height. The fixed installation height of each ultrasonic sensor is measured and stored in the controller during system installation and commissioning. This height refers to the vertical distance from the sensor detection surface to the ground. Since the sensor is installed on an independent crossbeam and does not contact the feed pipeline, this height value is constant. The controller has a preset obstruction threshold, which can be calibrated according to the geometry and installation position of the feed tray, for example, it can be set to 2-5 cm, to determine whether the measured value comes from the feed tray rather than the chicken back.

[0035] In terms of control logic, the system is based on the same sliding time window and fusion judgment conditions. When a sliding time window simultaneously meets the following conditions: the variance of the ultrasonic distance measurement value sequence is less than a preset stability threshold, and the fluctuation characteristic value of the heat source signal sequence exceeds a preset live animal threshold, it is determined that there are live chickens in the current feeding state within the current window. At this time, the controller uses the average distance measurement value of the window as a candidate height value. However, before including this candidate value in the effective chicken back height value, further occlusion judgment is required. The controller first calculates the theoretical distance value of the feed tray below the corresponding position of the sensor in real time based on the currently acquired actual height of the feed pipeline and the fixed installation height of the sensor. The formula for calculating the theoretical distance value of the feed tray is: sensor fixed installation height minus the current height of the feed pipeline, and then minus the vertical dimension correction value of the feed tray itself (height from the feed tray suspension point to the tray surface). For example, if the sensor is installed at a height of 150 cm, the current height of the feed line is 120 cm, and the height from the feed tray suspension point to the tray surface is 10 cm, then the theoretical distance to the feed tray is 150 - 120 - 10 = 20 cm. That is, if the distance measured by the sensor is approximately 20 cm, it corresponds to the feed tray surface. The controller compares the average distance measurement value in the current window with the theoretical distance value of the feed tray and calculates the absolute value of the difference. If the difference is greater than or equal to a preset occlusion threshold (e.g., 2 cm), it is determined that the current measurement value comes from the chicken rather than the feed tray, and the average value is used as a valid candidate chicken back height value for subsequent statistical processing. If the difference is less than the preset occlusion threshold, it is determined that the current sensor is occluded by the feed tray, and the measured value comes from the feed tray rather than the chicken. The data in this window is discarded, and the event of sensor occlusion is recorded. This event record can be used for system diagnosis and maintenance prompts. For example, if a sensor frequently experiences occlusion events over a long period, it may indicate an abnormal feed tray position or sensor installation misalignment.

[0036] In a specific farming example, the sensor installation height was uniformly set at 150 cm, the current height of the feed pipeline was 120 cm, and the theoretical distance to the feed tray was 20 cm. When a chicken was feeding near the feed tray, its back height was approximately 25 cm. The sensor measured a distance of 150 - 25 = 125 cm, with an average window value of 125 cm. This differed from the theoretical distance of 20 cm by 105 cm, significantly exceeding the preset obstruction threshold of 2 cm. Therefore, it was considered a valid chicken back height value and included in subsequent statistics. When a chicken put its head into the feed tray, and its back was just obstructed by the edge of the tray, the sensor received an echo from the surface of the tray, measuring a distance of approximately 20 cm. The average window value was 20.5 cm, differing from the theoretical distance of 20 cm by 0.5 cm, which was less than the preset obstruction threshold of 2 cm. The controller determined this as feed tray obstruction and discarded the data from that window. Within the subsequent 5-minute statistical window, the sensor may fail to provide valid data due to obstruction. In this case, the controller will calculate the average back height based on valid data from other sensors and simultaneously record the sensor obstruction event. If the same sensor is frequently obstructed within multiple consecutive statistical periods, the system can prompt maintenance personnel via the host computer to check whether the feed tray at that location is tilted or the sensor is offset. Through the above obstruction identification mechanism, this system effectively avoids interference from feed tray echoes on chicken back height measurement, ensuring that the final control data accurately reflects the flock's feeding status.

[0037] Furthermore, in another embodiment, based on the described system, a pressure sensor is introduced for each feed tray to detect the load weight on the tray, thus solving the problem of inflated measurement values ​​caused by upper-layer chickens obstructing the view when multiple chickens are stacked. The pressure sensor can be a single-point load cell, such as a cantilever beam or S-type sensor, with a range of 0-10kg, an accuracy of ±5g, and an IP67 protection rating, capable of withstanding the high humidity environment and frequent washing in the chicken house. Each feed tray is mounted on a load cell via a three-point suspension or single-point support method. The upper end of the sensor is connected to the hanging bracket of the feed pipeline, and the lower end is connected to the feed tray. Optional flexible gaskets are used to isolate the feed tray and the sensor to avoid vibration interference. The sensor signal line is centrally introduced into a waterproof junction box along the feed pipeline, and then connected to the analog input module of the controller via a shielded cable. Due to the large number of feed trays, distributed I / O modules can be used for data acquisition; for example, a remote I / O slave station can be set up for each section, communicating with the main controller via a fieldbus. The controller has pre-stored data on the relationship between standard weight and back height for chickens of different ages. This data can be obtained by fitting a breed standard curve or historical breeding data. For example, for Ross 308 broilers, the standard weight at 35 days old is approximately 2.0 kg, corresponding to a back height of approximately 20 cm; the standard weight at 42 days old is approximately 2.8 kg, corresponding to a back height of approximately 23 cm. The preset stacking threshold can be set to 2-5 cm to determine whether abnormal height is due to stacking.

[0038] In terms of control logic, the system is based on the same sliding time window and fusion judgment conditions. Once data within a certain sliding time window has passed feeding status discrimination, liveness discrimination, and occlusion discrimination in sequence, and is finally determined as a candidate chicken back height value, the controller needs to further perform layered verification on this candidate value. The controller first acquires the load weight signal within the time window corresponding to the candidate chicken back height value, i.e., the continuous weight values ​​collected by the pressure sensor within the same time window, and calculates its average value. This average value reflects the load situation on the feed tray within that time window. Subsequently, based on the current age of the flock, the controller searches for the theoretical back height range corresponding to the average load weight from the preset weight-back height relationship. For example, if the current flock is 35 days old, the standard weight-back height relationship shows that a weight of 2.0 kg corresponds to a back height of 20 cm. Considering individual differences, the theoretical back height range can be set to 20 ± 2 cm, i.e., 18-22 cm. The controller compares the candidate chicken back height values ​​with the upper limit of the theoretical back height range. If a candidate value exceeds the upper limit (e.g., if the measured value is 25cm, it exceeds the limit by 22cm) and the excess is greater than the preset stacking threshold (e.g., 3cm), it is determined that the candidate value originates from chicken stacking (the upper layer of chickens has a back height of 25cm, but the lower layer of chickens is the main feeder) and is discarded. If the candidate value is within the theoretical back height range or the excess is less than the stacking threshold, the value is retained for subsequent statistics.

[0039] For measurement points discarded due to overlap, the controller needs to use data interpolation to provide a reasonable representative value of chicken back height for the sensor within that time window, ensuring the continuity and accuracy of subsequent statistics. Data interpolation can be performed in two ways: one is to use the historical height values ​​effectively measured by the sensor within adjacent time windows, such as taking the weighted average of the first three valid values; the other is to use the weighted average of the valid height values ​​of adjacent sensors within the same time window, such as taking the average of the two valid values ​​on each side of the sensor, with closer distances having greater weights. In this embodiment, the second method is preferred because it better reflects the actual height of the chicken flock in the area at the current moment. For example, sensor S3 is discarded due to overlap. The valid values ​​of sensors S2 and S4 on its left and right sides within the same window are 24cm and 23cm respectively, while those on the outermost sides, S1 and S5, are 25cm and 22cm respectively. The controller calculates an interpolated value of 23.8cm based on distance weighting (S2 and S4 each have a weight of 0.3, S1 and S5 each have a weight of 0.2), which is used as the representative value of chicken back height for S3 within that window for subsequent statistics.

[0040] In a specific breeding example, a group of 35-day-old broilers weighed approximately 2.0 kg, with a theoretical back height range of 18-22 cm. During a peak feeding period, a stacking phenomenon occurred below sensor S3, with one broiler stepping on the back of another to feed. At this time, the ultrasonic variance within the sliding window was less than the stability threshold (broiler feeding), the heat source fluctuation exceeded the live animal threshold (live broiler present), and the difference between the average distance measurement and the theoretical distance from the feed tray was greater than the occlusion threshold (not occluded by the feed tray). Therefore, the average distance of 25 cm within this window was selected as a candidate value for the stacking verification process. Simultaneously, the average load weight detected by the pressure sensor within this window was 2.1 kg (approximately the total weight of the two broilers, but they could not be directly separated). Based on the standard weight-back height relationship for 35-day-old broilers, the controller determined that 2.1 kg corresponded to a theoretical back height of approximately 20.5 cm, with a theoretical back height range of 18.5-22.5 cm. The candidate value of 25cm exceeded the upper limit of 22.5cm by 2.5cm, which is greater than the preset stacking threshold of 2cm. Therefore, it was determined to be a stacking phenomenon and discarded. Subsequently, the controller called the valid values ​​of adjacent sensors S2, S4, S1, and S5 within the same window (24cm, 23cm, 25cm, and 22cm respectively), and interpolated them by distance weighting to obtain 23.8cm, which was used as the representative value of S3 in this window. This interpolated value is basically consistent with the measurement values ​​of surrounding sensors, reflecting the actual height of the flock in this area. Through the above weight-back height verification and data interpolation mechanism, this system effectively identified and eliminated abnormally high values ​​caused by stacking, avoiding the problem of the feed line rising too high due to the height of the upper layer of chickens, making it difficult for the lower layer of chickens to eat. At the same time, it ensured the continuity of data collection and provided a reliable data foundation for the subsequent accurate calculation of the average back height.

[0041] Furthermore, in another embodiment, based on the described system, the fluctuation characteristics of the load weight signal detected by the pressure sensor are further utilized to solve the problem of chickens' head movements during feeding being misjudged as movement due to ultrasonic variance exceeding the threshold, thus resulting in the loss of valid measurement data. To achieve this function, the controller needs to simultaneously collect and analyze three data sources: ultrasonic distance signal, passive infrared heat source signal, and pressure sensor load weight signal. A body stability discrimination based on pressure fluctuations is added to the original discrimination logic. A preset body stability threshold is added internally to the controller. This threshold is used to quantify the stability of the chicken's body (mainly the legs), and can be calibrated according to the standing stability of chickens of different ages; for example, it can be set to 0.3 kg for chicks. 2 The adult chicken stage can be set at 0.5kg. 2 This value reflects the permissible range of pressure fluctuations when a chicken is in a feeding state.

[0042] In terms of control logic, the system uses the same sliding time window as a basis, synchronously acquiring ultrasonic distance measurement sequences, passive infrared heat source signal sequences, and pressure sensor load weight signal sequences within each window. The controller first calculates the ultrasonic variance and heat source fluctuation characteristic values, then performs occlusion judgment, and then layering judgment. However, before entering these judgments, special handling is required for cases where the ultrasonic variance exceeds a threshold. Specifically, when the variance of the ultrasonic distance measurement sequence is greater than or equal to a preset stability threshold, according to the original logic, it should be judged as a moving state and the data should be discarded. However, this implementation adds a verification mechanism based on pressure fluctuation. The controller also calculates the fluctuation variance of the load weight signal within the window. This fluctuation variance reflects the stability of the load on the feed tray, i.e., the stability of the chicken's body. If the fluctuation variance of the load weight signal is less than the preset body stability threshold, it means that although the chicken's head is swinging (causing a large ultrasonic variance), its body (feet) remains almost still, and the chicken is still standing on the feed tray in a feeding state. At the same time, if the fluctuation characteristic value of the heat source signal sequence exceeds the preset live animal threshold, it further confirms the presence of a live chicken. When all three conditions are met (ultrasound variance ≥ stability threshold, pressure fluctuation variance < body stability threshold, heat source fluctuation > liveness threshold), the controller determines that the chickens in the current window are in a feeding state (body almost still, head swaying). In this case, instead of using the mean (which is easily affected by extreme values), the median of the ultrasound distance measurement sequence is used as the candidate chicken back height value. The median is not sensitive to outliers and can better represent the center value of back height during head swaying. Conversely, if the ultrasound variance ≥ stability threshold and the pressure fluctuation variance also ≥ body stability threshold, the chicken is determined to be in a moving state (such as walking or running), and the data in that window is discarded.

[0043] In a specific farming example, 49-day-old broilers (near slaughter age) were feeding in front of a feed tray, standing still but frequently raising and lowering their heads to peck at the food. At this time, the distance values ​​measured by the ultrasonic sensor fluctuated due to the head movements, with a variance of 1.5 square centimeters across 20 sampling points within the window, exceeding the preset stability threshold of 1.2 square centimeters, which should logically be discarded. However, simultaneously, the pressure sensor detected a stable load of approximately 3.2 kg on the feed tray (corresponding to the standard weight of a 49-day-old broiler), with a pressure fluctuation variance of only 0.2 kg within the window. 2 Less than the preset body stability threshold of 0.4 kg 2The passive infrared heat source signal fluctuation characteristic value is 0.6V, exceeding the liveness threshold of 0.3V. The controller determines that the chickens in the current window are in a feeding state and takes the median of the ultrasonic distance measurement value sequence as a candidate value. Assuming the distance values ​​in the window are: 121, 125, 118, 122, 119, 123, 124, 120, 122, 121, 126, 119, 120, 122, 123, 121, 120, 124, 122, 121 cm, the median is 121.5 cm, which translates to a chicken back height of 28.5 cm (sensor installation height 150 cm). This value is used as a candidate chicken back height value in the subsequent occlusion and overlay judgment processes. In another scenario, a chicken runs past the feed tray. The ultrasonic variance is also 1.5 square centimeters, but the pressure sensor detects a sudden change in load to 0 due to the chicken running away from the feed tray, and the pressure fluctuation variance in the window reaches 1.2 kg. 2 (Exceeding the body stability threshold), although the heat source fluctuation characteristic value is high (0.7V), the pressure fluctuation is large, and the controller judges it as a movement state and discards the data in this window. Through the above-mentioned verification mechanism based on pressure fluctuation variance, this system effectively distinguishes between the feeding state of "body standing still and head swinging" and the non-feeding state of "whole body moving". It retains a large amount of valid data that would otherwise be discarded during the peak feeding period, ensuring the continuity and accuracy of the measurement of the back height of the flock, and providing sufficient sample support for subsequent feed line adjustment.

[0044] Furthermore, in another embodiment, based on the described system, the spatial distribution characteristics and temporal correlation of multiple ultrasonic sensors are further utilized to solve the problem of pseudo-data contamination caused by chickens running past quickly. To achieve this function, the controller needs to record the candidate chicken back height values ​​and their corresponding timestamps generated by each ultrasonic sensor within multiple consecutive sliding time windows. Since the data of each sensor has already undergone multi-level screening (feeding status discrimination, liveness discrimination, occlusion discrimination, layering discrimination, feeding status discrimination), the data entering the candidate chicken back height value list already has high reliability, but pseudo-data generated by running past may still be mixed in. The controller internally presets two key parameters: short time interval Δt and height consistency threshold. The short time interval Δt can be set according to the typical speed of chickens running past and the sensor spacing. For example, if the sensor spacing is 2 meters and the chicken running speed is about 2 meters / second, then the time interval between adjacent sensors being triggered is about 1 second. Considering the possibility of multiple chickens running past at the same time, Δt can be set to 0.5-1.5 seconds, and in this embodiment, it is preferably set to 0.8 seconds. The height consistency threshold can be set according to the individual differences of chickens. For example, the back height of the same chicken changes very little when it runs by, so it can be set to 2-3 cm. In this embodiment, it is preferred to set it to 2.5 cm.

[0045] In terms of control logic, the system continuously generates candidate chicken back height values ​​for each sensor based on the same sliding time window. The controller maintains a circular buffer in memory to store the candidate values ​​and their timestamps generated by all sensors within a recent period (e.g., 5 seconds). Whenever a new candidate value is generated, the controller triggers a spatiotemporal correlation detection. The core logic of the detection is: within a preset short time interval Δt, are there at least three adjacent ultrasonic sensors that sequentially generate candidate chicken back height values, and the maximum difference between these candidate values ​​is less than a preset height consistency threshold? If this situation is detected, it is determined as a chicken running past quickly, and all candidate chicken back height values ​​generated within the corresponding time window are marked as pseudo data and discarded. The requirement of "at least three adjacent sensors" is to eliminate accidental single-point misjudgments and improve detection reliability; "sequential generation" means that the time sequence is consistent with the sensor arrangement order, reflecting the physical process of the chicken running along the feed line; "height consistency" reflects the biological characteristic that the back height of the same chicken is basically constant when it runs past. If the above situation is not detected, the candidate chicken back height values ​​determined according to claim 5 are retained for subsequent statistical average back height.

[0046] In practical implementation, the controller can use a sliding window scanning method for inspection. For example, whenever a new candidate value is generated, the controller expands to two adjacent sensors on each side of the original sensor, extracting all candidate values ​​from these sensors over the past time interval Δt. Then, it checks if there are three or more consecutive sensors whose candidate value generation times increase sequentially (consistent with the sensor arrangement order), and if the maximum difference between these candidate values ​​is less than a high consistency threshold. If multiple such sequences exist, the longest sequence is used as the criterion. Sensor window data determined to have passed an event are removed from the candidate value list and do not participate in subsequent statistical calculations.

[0047] In a specific breeding example, the feed pipeline is 40 meters long, with 20 ultrasonic sensors evenly installed along the line, numbered S1 to S20, with an adjacent sensor spacing of approximately 2 meters. During a feeding interval, a chicken quickly runs from one end of the feed line (near S1) to the other end (towards S20). As it passes under sensors S3, S4, and S5 in sequence, because the running direction is perpendicular to the sensor beam direction and the speed is relatively fast, the distance values ​​measured by each sensor are basically constant within a brief 0.3 seconds. The sliding window variance is less than the stability threshold (pseudo-data), and subsequent liveness detection, occlusion detection, layering detection, and feeding status detection are passed. Therefore, S3, S4, and S5 generate candidate chicken back height values ​​at t=10.0 seconds, 10.4 seconds, and 10.8 seconds, respectively, which are 24.1cm, 24.3cm, and 24.0cm. The controller detected that within a 0.8-second time interval, three adjacent sensors (S3, S4, and S5) sequentially generated candidate values, and the maximum difference between these candidate values ​​was only 0.3 cm (24.3-24.0), which is less than the preset height consistency threshold of 2.5 cm. Therefore, this was determined to be an event of a chicken running past quickly, and all candidate values ​​generated by S3, S4, and S5 at t=10.0, 10.4, and 10.8 seconds were marked as pseudo data and discarded. These data will not be included in the calculation of the average back height within the subsequent 5-minute statistical window. During the same period, a chicken was standing and feeding below sensor S7, generating a candidate value of 25.2 cm. However, since its adjacent sensors (S6 and S8) did not generate candidate values ​​with consistent height within Δt, this value was retained for statistical purposes. Through the above spatiotemporal correlation detection mechanism, this system effectively eliminated pseudo data generated by chickens running past, avoiding these short-lived and stable outliers from contaminating the statistical results, and ensuring that the average back height used for feed line control truly reflects the height of the flock during feeding. Compared with single-point discrimination, this method takes advantage of the collaborative work of multiple sensors and solves the problem of passing through pseudo data that cannot be solved by single-point variance discrimination.

[0048] Furthermore, in another embodiment, based on the described chicken flock growth status sensing system, a vibration suspension mechanism is further introduced at both ends of the feed pipeline to achieve the function of applying controllable vibration to the feed pipeline when needed. The vibration suspension mechanism is installed between both ends of the feed pipeline and the suspension points of the lifting mechanism, and mainly includes a clamp, an elastic layer, vibration damping rods, vibration isolation joints, and a vibrator. The clamp can adopt a split ring structure, made of cast aluminum or Q235 steel, and the two halves of the clamp are fastened to the ends of the feed pipeline by bolts. An elastic layer is provided between the inner wall of the clamp and the outer wall of the feed pipeline. The elastic layer is a ring-shaped rubber or polyurethane bushing with a thickness of 5-15mm and a Shore hardness of 50-70A. This elastic layer is pre-compressed when the clamp is tightened, transferring the static load of the feed pipeline to the clamp and isolating high-frequency vibrations from being transmitted upwards when the vibrator is working. The vibration damping rod is a rigid metal rod, which can be made of round steel or stainless steel pipe with a diameter of 16-20mm. Its upper end is connected to the wire rope or chain of the lifting mechanism through a lifting ring, and its lower end is connected to the sleeve through a vibration isolation joint. The vibration isolation joint is an elastic element, which can be a rubber-metal composite elastic element, such as the Rosta rubber suspension device, in which the internal rubber elastomer and metal flange are bonded together through a vulcanization process; or a ceiling vibration damper, such as a spring vibration damper or a wire rope vibration isolator, can be used. Both types of elastic elements can effectively absorb and attenuate vibration energy, achieving isolation of vibration transmission upwards. The vibrator can be a vibration motor or an electromagnetic vibrator. In this embodiment, a vibration motor is preferred, such as the MV series micro vibration motor, with a power of 0.1-0.3kW and adjustable excitation force. The speed can be steplessly regulated within the range of 10-60Hz by adjusting the power supply frequency through a frequency converter. The vibrator is fixedly installed on the supply pipeline, which can be installed on the pipeline body near the sleeve and connected to the mounting base welded to the pipeline by bolts. The controller can be the same PLC as the sensing system, and the start / stop, frequency and amplitude of the vibrator can be controlled by expanding the analog output module or relay output module.

[0049] Regarding the mechanical assembly, the two halves of the clamp are first aligned with the end of the feed pipeline, with an elastic layer pre-installed on the inner side. Bolts are then tightened to the designed torque to induce appropriate pre-compression of the elastic layer. The upper end of the vibration damping rod has a lifting eye hole, which connects to the wire rope connector of the lifting mechanism via a shackle. A metal connector is welded to or threaded onto the lower end of the vibration damping rod, and this connector is fixedly connected to the upper metal core of the vibration isolation joint. The lower metal shell of the vibration isolation joint is fixedly connected to a lifting lug welded above the clamp with bolts. The vibrator is bolted onto a mounting base on the side of the feed pipeline, positioned close to the clamp but with sufficient space for maintenance. The vibrator cable is led upwards along the vibration damping rod and connected to a frequency converter or driver in the electrical control cabinet via a cable tray. The frequency converter is controlled by the analog output module of the PLC. In terms of electrical connections, the PLC controls the start and stop of the vibrator through a digital output module and outputs a 0-10V voltage signal to the frequency converter through an analog output module, thereby adjusting the power supply frequency of the vibratory motor and thus regulating the excitation frequency. For vibratory motors with adjustable eccentric blocks, the excitation amplitude can be achieved by manually adjusting the eccentric block angle when the machine is stopped. Alternatively, a dual-axis adjustable vibratory motor can be selected during motor selection, and automatic amplitude adjustment can be achieved by controlling an additional adjustment mechanism through the PLC.

[0050] In terms of control logic, the controller comprehensively determines the vibrator's operating parameters based on the calculated average back height of the flock, feeding activity level, and preset cleaning mode. Specifically, the controller first reads the current average back height of the flock and the calculated feeding activity index from memory. In normal operation mode, the controller determines whether to start vibration and the intensity of vibration based on the feeding activity index: when feeding activity is high, vibration is started and a higher frequency (e.g., 40-50Hz) is set to promote feed flow and feeding; when feeding activity is moderate, the basic vibration frequency (e.g., 20-30Hz) is maintained; when feeding activity is low, vibration is paused to avoid disturbing the flock. In addition, the controller can indirectly determine the age of the chickens based on the average back height, and then adjust the vibration parameters accordingly. For example, the vibration frequency can be appropriately reduced for chicks and appropriately increased for adult chickens. When a cleaning command is received (which can be manually triggered or automatically triggered at a set time), the controller forcibly starts the vibrator and runs it in a preset cleaning mode. The cleaning mode usually uses high-frequency vibration (such as 50-60Hz) and / or variable frequency vibration (such as 30-50Hz reciprocating sweep frequency) for a certain period of time to assist in the shedding and emptying of residual feed.

[0051] In a specific farming example, the chickens were 35 days old and exhibited high feeding activity. The controller, based on a calculated feeding activity index of 85 (baseline value 50), determined that vibration was needed to promote feeding. The controller sent a 5V control voltage to the frequency converter via the analog output module, corresponding to a 40Hz power supply frequency for the vibration motor, which operated continuously at this frequency. Driven by the vibrators at both ends, the feed line experienced minute vibrations. These vibrations were transmitted through the pipe wall to each feed tray, promoting even distribution of feed within the trays and smooth descent to the feeding surface. During vibration, the vibration was effectively isolated within the feed line itself due to the combined effect of the elastic layer inside the hoop and the vibration isolation joint at the lower end of the vibration damping rod. The vibration transmitted upwards to the lifting mechanism's wire rope was negligible, not affecting the normal operation and lifespan of the lifting mechanism. When the peak feeding period passed and the feeding activity index dropped to 40, the controller stopped the vibrator, and the feed line returned to a static state, avoiding unnecessary energy consumption and disturbance to the chickens. In nighttime cleaning mode, the controller forcibly activates the vibrator, running it at a high frequency of 50Hz for 30 seconds. This, combined with the opening of the drain valve, shakes and discharges any residual feed from the feed line. Through this structure and control, this implementation achieves physical isolation and independent control of the vibration and lifting functions, resulting in significant vibration without affecting the reliability of the lifting mechanism. The power and number of vibrators can be adjusted according to the different lengths of the feed lines and the number of feed trays in different chicken houses. For example, for feed lines exceeding 80 meters, one vibrator can be installed at each end, or an auxiliary vibrator can be added in the middle. The material and thickness of the elastic layer can also be optimized according to the vibration frequency requirements to achieve the best vibration isolation effect and vibration transmission efficiency.

[0052] Furthermore, in another embodiment, based on the described vibration suspension mechanism, the load weight signal detected by the pressure sensor is further utilized to calculate the feeding activity index and dynamically adjust the operating parameters of the vibrator accordingly, thereby achieving automatic matching between vibration intensity and the feeding needs of the flock. To achieve this function, the controller needs to continuously collect the load weight signal from the pressure sensor of each feed tray at the same sampling frequency (e.g., 20Hz) and perform signal processing and analysis in real time. The controller has several preset parameters, including a filter cutoff frequency for extracting high-frequency fluctuation energy, a sliding time window width for calculating the activity index, an activity baseline value, and a first threshold and a second threshold for determining whether feeding is active or passive. The filter cutoff frequency can be set to 0.5-5Hz to separate the rapid impact component caused by chickens pecking from the slow change component caused by feed consumption; the sliding time window width can be consistent with the sliding window of the ultrasonic sensor (e.g., 30 seconds) or can be set independently (e.g., 1-5 minutes). In this embodiment, it is preferred to set it to 1 minute to balance response speed and stability; the activity benchmark value can be dynamically updated according to historical data, for example, taking the average activity level of the same period in the past 24 hours to eliminate the influence of circadian rhythm; the first threshold and the second threshold can be set to +20% and -20% respectively, that is, when the activity deviation exceeds 20% of the benchmark value, it is judged as active or passive feeding.

[0053] In calculating the feed intake activity index, the controller collects and preprocesses the load weight signal from each pressure sensor in real time. The raw signal contains low-frequency components of slow changes in feed weight (e.g., weight loss due to feeding, with a frequency typically below 0.1Hz) and high-frequency impact components caused by chicken feeding activities (e.g., impacts from pecking and trampling, with a frequency typically in the 0.5-5Hz range). The controller first extracts the high-frequency component in the 0.5-5Hz band using a high-speed filter or bandpass filter, which reflects the dynamic impact generated by the chickens' feeding behavior. Specific filtering methods can employ infinite impulse response (IRR) filters or finite impulse response (FIR) filters; in this embodiment, a second-order Butterworth bandpass filter is preferred to ensure minimal phase distortion. Then, the energy value of this high-frequency component is calculated within a sliding time window (e.g., 1 minute, corresponding to 1200 sampling points), which can be calculated using the root mean square (RMS) value or the absolute integral value. Taking the RMS value as an example, the calculation formula is: Activity Index = sqrt((1 / N) × Σ(x_i) 2The root mean square (RMS) value reflects the average intensity of feeding activity in a feed tray during the specified time period. Since the number of chickens in different feed trays may vary, directly comparing absolute values ​​is not meaningful; therefore, the activity index of each feed tray needs to be normalized. Normalization can be achieved by dividing by the historical average activity level of that feed tray to obtain the relative activity index. The historical average activity level can be the average of the same time period over the past 3-7 days, for example, the average activity level between 9-10 AM each day as the baseline for that time period. The controller calculates the relative activity index of all feed trays within the same time period and takes the median or weighted average as the current flock's group feeding activity level. The group feeding activity level reflects the overall intensity of the flock's feeding behavior, expressed as a relative value (dimensionless), with a baseline value typically of 1.0.

[0054] In terms of control logic, the controller first reads the current average back height of the flock (calculated according to claims 1-6) and the calculated flock feeding activity level from memory. Then, it compares the current flock feeding activity level with a preset activity benchmark value to calculate the activity deviation. The activity benchmark value can be a fixed value (e.g., set to 1.0 based on breed experience) or a dynamic benchmark (e.g., the average value of the same period over the past 7 days). When the activity deviation is positive and exceeds a first threshold (e.g., +20%), the flock is considered to be actively feeding. In this case, vibration needs to be strengthened to promote feed flow and feeding, meeting the flock's vigorous demand. The controller accordingly increases the vibration frequency and / or vibration amplitude of the vibrator, for example, increasing the frequency from the base value of 30Hz to 45Hz, or increasing the vibration force from 30% to 60%. When the activity deviation is negative and exceeds a second threshold (e.g., -20%), the flock is considered to be passively feeding. In this case, vibration should be reduced or stopped to avoid unnecessary disturbance and energy consumption. The controller reduces the excitation frequency and / or amplitude accordingly, or directly stops the exciter. When the activity level deviation is between the first and second thresholds, the feeding activity is considered normal, and the current excitation parameters are maintained unchanged. The excitation parameters can be adjusted according to a preset gradient, for example, a 5Hz frequency change corresponds to every 10% activity level deviation, to avoid sudden changes that could cause stress to the flock.

[0055] In a specific farming example, the chickens are 35 days old, and their peak feeding periods are typically between 8-10 AM and 4-6 PM. The system continuously monitors the load weight signal of each feed tray. At 9 AM, the peak feeding period arrives, and the relative activity index of each feed tray generally increases. Taking a certain feed tray as an example, its original load weight signal is filtered by a 0.5-5Hz bandpass filter to obtain a high-frequency impact component. The root mean square value of this component is calculated within a 1-minute sliding window, giving the absolute activity of this feed tray as 2.5. The historical average absolute activity of this feed tray between 9-10 AM over the past 7 days is 1.8, therefore its relative activity index is 2.5 / 1.8≈1.39. The controller calculates the relative activity index of all feed trays and takes the median to obtain the group's feeding activity as 1.35. With the activity baseline set at 1.0, the activity deviation is +35%, exceeding the first threshold of +20%. The controller, determining the flock was actively feeding, sent a control signal to the frequency converter of the vibration motor via the analog output module, gradually increasing the excitation frequency from the default 25Hz to 45Hz and the excitation amplitude from 30% to 70%. Driven by the vibrators at both ends, the feed line vibrated strongly, effectively promoting the falling and flowing of feed in the feed pans, meeting the flock's vigorous feeding needs. At noon, feeding activity decreased, with a calculated flock feeding activity value of 0.95 and a deviation of -5%, between the first and second thresholds. The controller maintained the current excitation parameters (at this point, it may have already fallen back to the baseline value from its peak). At 2 PM, the flock entered a resting period, and the flock feeding activity dropped to 0.65, with a deviation of -35%, exceeding the second threshold of -20%. The controller determined feeding was inactive, gradually reducing the excitation frequency to 10Hz and eventually pausing the vibrator. The feed line returned to stillness to avoid disturbing the resting flock. At 6 PM, activity rebounded, and the system restarted vibration. Through the aforementioned dynamic adjustment mechanism, the vibration intensity is always matched to the actual feeding needs of the flock, ensuring feed supply efficiency during peak feeding periods while avoiding ineffective vibration and energy waste during non-feeding periods. Depending on the chicken house environment and breed, the filter cutoff frequency, sliding window width, activity baseline value, and threshold can be calibrated and adjusted on-site. For example, the threshold can be appropriately lowered for breeds with milder feeding behavior, and appropriately increased for breeds with more vigorous feeding. The sampling frequency and filtering parameters of the pressure sensor can also be optimized based on the actual signal characteristics.

[0056] Furthermore, in another embodiment, based on the described vibration parameters adjusted according to feeding activity, the spatial distribution information of feeding activity in each feed tray is further utilized to dynamically adjust the height difference between the two ends of the feed pipeline to tilt the pipeline, thereby guiding feed to flow towards areas with more active feeding. To achieve this function, the lifting mechanism needs to have the ability to control both ends independently, that is, both ends of the feed pipeline are driven by independent electric winches or electric push rods, and the height of each end can be adjusted individually. For example, an electric winch can be installed at each end of the feed pipeline, with each winch equipped with an independent motor and controller, or a dual-output shaft geared motor can be used with two drums to control the steel wire ropes at both ends respectively. The controller needs to be able to send control commands to the two lifting drive units respectively and obtain height feedback from both ends in real time. Height monitoring can be achieved by installing a pull-wire displacement sensor or encoder at each end. The controller has a preset tilt angle threshold, which is used to limit the maximum tilt angle of the pipeline, ensuring that even in the tilted state, the overall height of the pipeline is still within the optimal height range determined based on the average back height of the flock. The tilt angle threshold can be set according to the pipeline length and the distribution of the trays. For example, for an 80-meter-long pipeline, the tilt angle can be limited to within 1°, corresponding to a height difference of no more than 14 centimeters between the two ends.

[0057] In terms of spatial distribution feature extraction, the controller further analyzes the spatial distribution patterns of the feeding activity index of each feed tray based on real-time calculations. Specifically, the controller divides the feed pipeline along its length into several regions, each containing several consecutive feed trays. Region division can be based on a fixed length (e.g., one region every 10 meters) or on the number of feed trays (e.g., one region every 5 trays). The controller calculates the average feeding activity index of the feed trays within each region, obtaining the spatial distribution curve of the activity level. Then, it identifies regions with significantly higher activity levels than the overall average (called high-activity zones) and regions with significantly lower activity levels than the overall average (called low-activity zones). The determination of "significance" can be achieved by setting a deviation threshold; for example, a region with activity levels more than 30% higher than the overall average is considered a high-activity zone, and a region with activity levels more than 30% lower is considered a low-activity zone. If multiple high-activity or low-activity zones exist, the most extreme zone can be used as the adjustment basis, or a weighted average method can be used to determine the direction in which the feed should flow.

[0058] In terms of tilt control logic, when the controller identifies a significant difference in the spatial distribution of activity levels—for example, the feeding activity index in the first region (e.g., the far end) is significantly higher than that in the second region (e.g., the near end)—it determines that feed needs to be guided towards the first region. The controller then calculates the required height difference between the two ends: lowering the height of the feed line end corresponding to the first region and raising the height of the feed line end corresponding to the second region, creating a downward slope from the second region to the first region. The adjustment range of the height difference between the two ends needs to meet two constraints: first, ensuring that the highest and lowest points of the tilted pipeline remain within the optimal height range determined based on the average back height of the flock (e.g., optimal height ± 5 cm); second, ensuring that the height difference between the two ends does not exceed the maximum allowable value calculated by the tilt angle threshold. The controller calculates the maximum allowable height difference based on the current average back height, pipeline length, and preset tilt angle threshold, and then limits the actual adjustment range within this range. The adjustment process can be carried out in steps, adjusting a small height difference each time (e.g., 1-2 cm), and observing the subsequent changes in activity distribution to avoid excessive adjustment at once that could cause discomfort to the flock. After adjustment, the system continuously monitors the activity distribution. If the distribution tends to be balanced, the current tilt can be maintained or the level can be gradually restored. If the distribution difference persists or worsens, further fine-tuning can be performed.

[0059] In a specific breeding example, a 10-meter-long chicken house has feed lines controlled by independent electric winches at both ends. The average back height of the flock is currently 25 cm, and the optimal feed line height is 21.3 cm (calculated using an empirical formula of 0.85). The controller calculates the feeding activity index of each feed tray using the method described in claim 8, and finds that the average feeding activity in the 5-meter area (far end) near the fan end is 90, while the average feeding activity in the 5-meter area (near end) near the feed inlet end is only 45. The overall average feeding activity of all feed trays along the entire feed line is 67.5. Calculations show that the activity level at the far end is 33% higher than the overall average, while the activity level at the near end is 33% lower, indicating that the feed demand of the flock at the far end is significantly higher than that at the near end, requiring guidance of feed flow towards the far end.

[0060] Based on a pipeline length of 10 meters, the preset tilt angle threshold is set to 0.5° (this value can be calibrated on-site based on pipeline length and flock tolerance). The allowable fluctuation range for the optimal height of 21.3 cm is set to ±5 cm, meaning the lowest point at the distal end can be lowered to 16.3 cm, and the highest point at the proximal end can be raised to 26.3 cm. The controller calculates the maximum allowable height difference between the two ends: if the proximal end is raised to the upper limit of 26.3 cm and the distal end is lowered to the lower limit of 16.3 cm, the height difference is 10 cm, corresponding to a tilt angle of arctan(10 / 1000)≈0.57°, slightly exceeding the 0.5° threshold, therefore the extreme value cannot be used. The controller selects the adjustment range within a safe range: raising the proximal end by 2 cm to 23.3 cm and lowering the distal end by 2 cm to 19.3 cm, creating a downward slope from the proximal end to the distal end with a height difference of 4 cm. The corresponding tilt angle is arctan(4 / 1000)≈0.23°, far less than the 0.5° threshold, ensuring that the overall pipeline height remains within the optimal range of 16.3-26.3 cm. During the adjustment process, the controller continuously monitors the changes in feeding activity in each feed tray. After 2 hours of observation, the distal end activity decreased from 90 to 80, while the proximal end activity increased from 45 to 55, significantly reducing the difference between the two ends. This indicates that the tilt adjustment effectively promoted feed flow to the distal end, improving the insufficient feed supply at the distal end. If the difference further widens or narrows unsatisfactorily, the controller can continue to fine-tune the height at both ends, controlling the adjustment range within 1-2 cm each time, gradually approaching the optimal tilt state. Through the coordinated control of spatial distribution sensing and tilt adjustment, this system achieves dynamic optimization of the macroscopic feed flow direction, complementing vibration intensity adjustment (microscopic feed dispensing promotion), and jointly improving the uniformity of feed distribution and the balance of feed intake within a 10-meter feed line. Based on different chicken house layouts and flock behavior characteristics, the zoning method (e.g., one zone every 2 meters) and deviation threshold (e.g., ±20%) can be calibrated on-site. For flocks with significant differences in feeding behavior, the deviation threshold can be appropriately increased; for chicken houses with high uniformity requirements, the zoning can be refined to one zone every 1 meter.

[0061] Furthermore, in another embodiment, based on the described spatial distribution sensing and tilt adjustment, an empty valve installed at the end of the feed pipeline is introduced, and a complete cleaning mode control process is defined to achieve automated internal cleaning of the feed pipeline and recovery of residual feed. The empty valve can be an electric butterfly valve or an electric ball valve, such as DN75 or DN90, matching the diameter of the feed pipeline. The valve body is made of 316L stainless steel, and the sealing material is food-grade rubber, achieving an IP67 protection rating, capable of withstanding the high humidity environment and frequent flushing in the chicken house. The empty valve is installed at the end of the feed pipeline (the end furthest from the feed hopper), connected to the pipeline end via a flange or clamp. The valve outlet can lead to a collection container or return pipeline for easy recycling. The valve actuator is a DC 24V powered switching or regulating electric actuator with position feedback function, capable of receiving the controller's switching commands and providing feedback on the current valve position status. The controller needs to be expanded with a digital output module for controlling valve opening and closing, and a digital input module for receiving valve position feedback signals. The controller has several preset parameters related to the cleaning mode, including cleaning height, vibration parameters (frequency, amplitude, duration), confidence time for determining the absence of chickens, and recovery waiting time after cleaning. The cleaning height can be set to 30-50 cm higher than the optimal height range determined based on the average back height of the flock, to ensure sufficient space under the feed tray and prevent contact with the ground during vibration. The vibration parameters for the cleaning mode can use high-frequency vibration (e.g., 50Hz) and / or variable-frequency vibration (e.g., 30-50Hz reciprocating sweep frequency), with a duration of 30-60 seconds. The determination of the absence of chickens requires confirmation through multiple consecutive sampling cycles to avoid misjudgment.

[0062] In terms of mechanical assembly, the vent valve is connected to the end of the feed pipeline via a flange or quick-release clamp, with a sealing gasket at the connection to ensure airtightness. A short pipe can be connected to the valve outlet to guide the discharged feed into the collection container. The electric actuator is fixed to the valve body via a bracket, and its cable is connected to the electrical control cabinet along the feed pipeline or an independent cable tray. Since the feed pipeline moves up and down during lifting, a flexible connection is required between the valve and the fixed pipeline. For example, a flexible hose can be connected to the valve outlet, with the other end of the hose leading to the fixed collection container. The hose length should have sufficient slack to accommodate pipeline lifting. In terms of electrical connection, the control lines (open and close command lines) and feedback lines (fully open and fully closed position signals) of the electric actuator are connected to the digital output module and digital input module of the controller, respectively. For regulating valves, an analog output module is also required to control the valve opening, but in this embodiment, the cleaning mode only requires fully open and fully closed positions, so an on / off type is sufficient.

[0063] In terms of control logic, when the controller receives a cleaning command, it first enters cleaning mode. Cleaning commands can be triggered in various ways: they can be sent remotely by a human through a host computer interface, automatically triggered by the system based on a preset time (e.g., 2 AM daily), or automatically triggered based on the detection of residual feed tray levels (e.g., when multiple feed trays are still loaded during periods without chickens). After entering cleaning mode, the controller executes the following steps: The first step is to confirm the absence of chickens. The controller uses the load weight signal detected by the pressure sensor and the heat source signal from the passive infrared sensor to determine whether all feed trays are in a chicken-free state. For each feed tray, it is determined to be chicken-free if the following conditions are met simultaneously: the load weight detected by the pressure sensor is close to the weight of the empty tray (e.g., the deviation is less than 50 grams), the passive infrared sensor has no heat source signal (or the heat source fluctuation characteristic value is below the live animal threshold), and this state lasts for a certain period of time (e.g., 30 seconds) to ensure that the chickens have actually left and are not just temporarily away. Only when all feed trays meet the chicken-free state will the system allow the cleaning mode to continue; otherwise, cleaning is paused and the reason is recorded, and the system will retry when the conditions are met or wait for manual intervention.

[0064] The second step is to raise the feed line to the clean height. After confirming there are no chickens, the controller controls the lifting mechanism to raise the entire feed line to the preset clean height. Since the lifting mechanism has independent control capabilities at both ends, it is necessary to control both ends to rise simultaneously, keeping the feed line horizontal or raised according to a preset tilt (which can be set to horizontal) to avoid twisting of the feed line due to asynchrony. During the raising process, the actual height at both ends is continuously monitored, and the process stops and locks once the clean height is reached.

[0065] The third step involves opening the valve and initiating vibration. After the feed pipeline stabilizes at the clean height, the controller first sends an opening command to the vent valve actuator and confirms that the valve is fully open via valve position feedback. Then, the controller controls the vibrator to operate in a preset cleaning mode, which can employ high-frequency vibration (e.g., 50Hz) and / or variable-frequency vibration (e.g., linear sweep from 30Hz to 50Hz, with a period of 5 seconds), for a first preset time (e.g., 30-60 seconds). Under the vibration, residual feed adhering to the inner wall of the feed pipeline, residual feed in the feed tray, and clumps of feed on the pipe wall are shaken off and flow along the inner wall of the pipeline towards the end under gravity, eventually being discharged from the opened vent valve. The discharged feed can enter a collection container for recycling.

[0066] The fourth step is to end the cleaning process and reset. After the cleaning mode has lasted for the first preset time, the controller first stops the vibrator and then sends a closing command to the vent valve to confirm that the valve is completely closed. Finally, the controller controls the lifting mechanism to slowly lower the feed line to the optimal height determined based on the average back height of the current flock. During the restoration process, both ends must be kept synchronized to avoid tilting. After restoration is complete, the system exits the cleaning mode and returns to normal operation.

[0067] In a specific cleaning example, a broiler house was set to automatically execute a cleaning mode at 10:00 PM daily. At this time, the chickens had turned off their lights and were resting, away from the feed trays. The controller first checked the pressure and infrared sensors of all feed trays: the pressure sensors showed that the load on all feed trays was close to the weight of an empty tray (deviation <20 grams), and the infrared sensors showed no heat source signal, and this state had lasted for 1 minute. After confirming that there were no chickens, the controller activated the lifting motors at both ends simultaneously, raising the feed line from a working height of 20 cm to a cleaning height of 60 cm (a 40 cm lift). Upon reaching the height, the controller sent an opening command to the electric butterfly valve at the end, and received feedback that the valve was fully open 2 seconds later. Subsequently, the controller activated the vibrators at both ends, set to a 50 Hz high-frequency vibration for 45 seconds. During the vibration, residual feed on the inner wall of the feed line was shaken off, flowed along the line to the end, and was discharged from the valve, falling into the collection bucket below. After 45 seconds, the controller stopped the vibrators, sent a valve closing command, and after confirming that the valve was closed, controlled the lifting mechanism to slowly lower the line to the 20 cm working height. The entire process took approximately 3 minutes. The following morning, the feeder found approximately 2 kilograms of residual feed in the collection bin. This feed, which would have otherwise accumulated in the pipeline and become moldy and spoiled, was now being recycled. The vibration parameters of the cleaning mode can be adjusted according to different cleaning needs. For example, for severely compacted feed, variable frequency vibration can be used to generate a resonance effect; for light cleaning, the duration can be shortened or the frequency reduced. Drain valves can also be installed at both ends of the feed pipeline to achieve bidirectional drainage; in this embodiment, they are preferably installed at the end to simplify the pipeline. Through the above automated cleaning process, this system achieves periodic automatic drainage of the feed pipeline, preventing long-term accumulation and mold growth of feed, reducing the labor intensity of manual cleaning, and simultaneously recovering residual feed, thus improving feed utilization.

[0068] Furthermore, in another embodiment, based on the described spatial distribution sensing and tilt adjustment, the vibration suspension mechanism is further expanded into multiple independently controlled units, each set in different sections of the feed pipeline. Each vibration suspension mechanism includes an independent vibrator to achieve segmented differentiated vibration control, solving the problem of localized feed differences caused by feed resistance, pipeline length, and uneven distribution of chickens in long-distance feed lines. To achieve this function, the feed pipeline needs to be divided into several sections along its length. A set of vibration suspension mechanisms is set at both ends or specific positions of each section. Its structure is the same as the suspension assembly described above, including a sleeve, elastic layer, vibration damping rod, vibration isolation joint, and independent vibrator. For example, for a 20-meter-long feed pipeline, it can be divided into 4 sections, each 5 meters long, with a set of vibration suspension mechanisms (i.e., independently lifting first and second lifting units) set at both ends of each section, for a total of 8 sets of mechanisms. The vibrator can be a low-power vibration motor, such as the MV series miniature vibration motor with a power of 0.05-0.1kW and adjustable excitation force, controlled by an independent frequency converter or driver. The controller needs to be expanded with a corresponding number of analog output modules or relay output modules to independently control the start / stop, frequency, and amplitude of the vibrator in each section. The controller internally presets several parameters related to zone control, including the boundary division of each section, the method for calculating local feeding activity, the baseline value of the group's average feeding activity, and the deviation threshold used to determine whether the feed supply is insufficient or sufficient.

[0069] In terms of mechanical assembly, the vibration suspension mechanism for each section is installed at both ends or designated locations (such as the middle) of that section. Due to the long pipeline and independent vibration of each section, vibration isolation needs to be considered at the section boundaries to avoid mutual interference between adjacent sections. Flexible vibration isolation joints can be installed inside the pipeline at the section boundaries, or rubber vibration damping pads can be installed outside the pipeline. However, considering the integrity of the feed pipeline and the continuity of feed delivery, this embodiment prefers a control strategy rather than mechanical isolation. That is, resonance and interference are avoided by controlling adjacent sections to not vibrate at high intensity simultaneously. All vibrator cables are centrally led out along the feed pipeline and connected to the electrical control cabinet via cable trays or cable ducts. Each vibrator corresponds to a set of control lines, which are connected to the corresponding frequency converter or driver. The frequency converter or driver is controlled by the analog or digital output module of the PLC to achieve independent adjustment. To simplify wiring, a fieldbus distributed I / O scheme can be adopted, with remote I / O slave stations set up near each section to collect signals from the sensors in that section and control the vibrators in that section, communicating with the main controller via the bus.

[0070] In terms of control logic, the controller first divides the feed pipeline into several vibration control sections based on the real-time calculated feeding activity index of each feed tray. The division of sections can be based on a fixed length (e.g., one section every 20 meters) or on the number of feed trays (e.g., one section every 10 feed trays). The division method can be set and stored in the controller during system installation and commissioning. For each vibration control section, the controller statistically analyzes the local feeding activity of all feed trays within that section, using the average or median as the representative value for that section. Simultaneously, the controller obtains the current global feeding status of the flock based on the calculated group feeding activity (the overall level of all feed trays). Then, the controller compares the local feeding activity of each section with the group average feeding activity and calculates the deviation value. When the local feeding activity of a section is lower than the group average feeding activity and exceeds a preset deviation threshold (e.g., -20%), it is determined that the feed supply in that section is insufficient, and the chickens in that section experience reduced feeding activity due to untimely feed supply. The controller accordingly increases the excitation frequency and / or amplitude of the vibrator in that section, for example, increasing the frequency from the base value of 30Hz to 45Hz, to enhance local vibration and promote the feeding speed of the feed tray in that section. When the local feeding activity in a certain section is higher than the average feeding activity of the group and exceeds a preset deviation threshold (e.g., +20%), it is determined that the feed supply in that section is sufficient, but there may be an oversupply or a risk of feed accumulation. The controller accordingly reduces the excitation frequency and / or amplitude of the vibrator in that section, for example, from 30Hz to 15Hz or stops vibration, to avoid overfeeding and waste. The current excitation parameters remain unchanged for other sections with deviations within the threshold range.

[0071] In a specific poultry farming example, a 20-meter-long chicken house has its feed pipeline divided into four vibration control sections: Section A (0-5 meters), Section B (5-10 meters), Section C (10-15 meters), and Section D (15-20 meters). The current average feeding activity of the flock is 60. The local feeding activity levels for each section are: Section A 75, Section B 65, Section C 55, and Section D 45. The preset deviation threshold is ±20% (i.e., ±12 relative to the flock average). Calculate the deviation for each section: Section A deviation +25% (75-60=15, 15 / 60=25%), exceeding the +20% threshold, is considered sufficient; Section B deviation +8.3%, within the threshold, is considered normal; Section C deviation -8.3%, within the threshold, is considered normal; Section D deviation -25% (45-60=-15, -15 / 60=-25%), exceeding the -20% threshold, is considered insufficient.

[0072] The controller then performs independent zone adjustments: for zone A (sufficient feed), the vibration frequency is reduced from the base 30Hz to 15Hz, and the vibration amplitude is reduced from 50% to 30% to slow down the feeding speed in this zone and prevent feed accumulation; for zone D (insufficient feed), the vibration frequency is increased from 30Hz to 45Hz, and the vibration amplitude is increased from 50% to 80% to strengthen the vibration in this zone and promote feed discharging from the far-end feed tray; zones B and C remain unchanged at 30Hz and 50%. After the adjustment begins, the strong vibration in zone D promotes the feeding speed of the far-end feed tray, allowing feed to enter the tray more quickly from the feed pipe, meeting the needs of the flock in this area; the weak vibration in zone A slows down the feeding speed and prevents feed from accumulating at the near end.

[0073] Meanwhile, the tilt adjustment may have slightly tilted the pipeline towards the distal end, for example, lowering the end of section D by 2 cm and raising the end of section A by 2 cm, creating a downward slope with a height difference of 4 cm from the proximal end to the distal end. The tilt adjustment guides the macroscopic flow of feed towards the distal end, while the zoned vibration enhances the microscopic feeding at the distal end. The two work together to optimize feed distribution within the 20-meter feed line range. After 2 hours of operation, the feeding activity of each section gradually tended to be balanced: the section A dropped to 68, and the section D rose to 55, with the difference significantly reduced. The controller then readjusted the vibration parameters based on the new activity distribution.

[0074] Through the aforementioned segmented independent vibration control, this system achieves precise adjustment of different sections of the 20-meter feed line, solving the problems of insufficient feed supply at the far end and excessive feed supply at the near end. Based on different chicken house lengths and flock distribution characteristics, the number of sections and deviation thresholds can be optimized on-site. For example, a 30-meter feed line can be divided into 5-6 sections (each 5-6 meters long); for flocks with significant differences in feeding behavior, the deviation threshold can be appropriately relaxed to 30%. The independent control of multiple vibrators, along with overall vibration and tilt adjustments, forms a multi-level collaborative control system, jointly achieving intelligent and precise management of feed distribution.

[0075] Furthermore, in another embodiment, based on the described adaptive automatic feeding system for chicken flock growth, a flexible connection module is introduced between the fixed feed hopper outlet and the feed inlet of the liftable feed pipeline. This module ensures continuous feed delivery and reliable sealing during the lifting and lowering of the feed pipeline. The flexible connection module mainly includes an outer flexible cylindrical bag, a lightweight spring frame inside the bag, and elastic clamps connecting the two ends of the bag to the feed hopper outlet and the feed pipeline inlet. The flexible cylindrical bag can be made of wear-resistant canvas or polymer composite materials, such as a composite canvas with polyester filament base fabric double-coated with polyurethane, which has excellent wear resistance, flexibility, and tensile strength, while also being able to withstand the high humidity, ammonia corrosion, and frequent washing in the chicken house. The lightweight spring frame can be made of multiple helical springs spaced apart along the axial direction of the bag, or an integral helical spring can be used. The spring material is made of stainless steel wire (such as 304 or 316) to ensure corrosion resistance. The spring diameter is selected according to the bag opening diameter. For example, for a DN75 bag, the spring wire diameter can be 1.5-2.0mm, and the outer diameter of the spring is slightly smaller than the inner diameter of the bag.

[0076] In terms of mechanical assembly, the lightweight spring skeleton is first installed inside the flexible cylindrical cloth bag. Both ends of the spring skeleton are connected to the elastic clamps at both ends of the cloth bag, which can be fixed by welding, hooking, or riveting to ensure the spring remains axially positioned within the bag and does not shift. For multiple spaced helical springs, a positioning ring needs to be installed axially at regular intervals (e.g., 100-150mm) inside the bag to fix the spring and prevent displacement. Then, both ends of the bag are respectively fitted onto the hopper outlet and the feed line inlet, with sufficient overlap (e.g., 50-80mm) allowed at the bag ends. Pressure is gradually applied through the locking mechanism of the elastic clamps, pressing the bag ends tightly against the outlet and inlet surfaces to form a reliable seal and fixation. During installation, ensure the bag's length exceeds the maximum straight-line distance between the fixed hopper outlet and the liftable feed line inlet. This means that when the feed line is at its lowest position, the bag remains naturally relaxed, without tensile stress. When the feed line is at its highest position, the bag is appropriately stretched but not exceeding its elastic limit, with the spring frame providing rebound force to help the bag regain its shape. The ball joint's spherical structure allows the bag to swing freely within a ±20° range, adapting to changes in the relative angle between the two ends during feed line lifting. Simultaneously, the radial locking force of the elastic retaining ring automatically compensates for minor changes in bag thickness, ensuring consistent sealing performance over long-term use.

[0077] In terms of working principle, when the feed line is raised or lowered, the flexible connection module achieves dynamic adaptation in the following ways: As the feed line descends to a lower position, the filter bag gradually relaxes and may bend. At this time, the internal lightweight spring skeleton provides support, preventing the filter bag from collapsing and blocking the flow channel due to negative pressure or its own weight. The spring skeleton can be locally compressed or stretched as the filter bag bends, always maintaining the basic unobstructed flow channel. When the feed line rises to a higher position, the filter bag is stretched and elongated. The spring skeleton extends accordingly and provides a certain rebound force to help the filter bag return to its original shape, avoiding permanent deformation due to excessive stretching. During the raising and lowering process, the relative angle between the feed line inlet and the hopper outlet changes, and the filter bag remains in a naturally curved state, avoiding dead bends or excessive bending at the interface.

[0078] In a specific installation example, the center elevation of the fixed hopper outlet is 1.5 meters, and the elevation of the feed line inlet is 1.0 meter at its highest position and 0.2 meters at its lowest position. Calculations show that the maximum straight-line distance is 1.3 meters (1.5-0.2), and the minimum straight-line distance is 0.5 meters (1.5-1.0). The flexible connection module's filter bag length is designed to be 1.6 meters, greater than the maximum straight-line distance of 1.3 meters. This ensures that when the feed line is at its lowest position (0.2 meters), the filter bag still has a 0.3-meter slack, preventing tensile stress. When the feed line is at its highest position (1.0 meter), the filter bag is compressed to 0.5 meters (actually folded, but the spring frame supports and maintains the flow channel). The spring frame is compressed to approximately 1 / 3 of its original length, still within its elastic working range, providing sufficient elastic reserve.

[0079] The bag is made of double-layer polyurethane coated canvas, 2mm thick, with abrasion resistance exceeding 100,000 cycles, capable of withstanding long-term erosion and frequent bending of feed pellets. Internally, it uses an integral helical spring with a spring wire diameter of 2mm, an outer diameter of 70mm (bag inner diameter 75mm), and a free length of 1.6 meters, matching the bag length. Both ends are welded and secured to elastic clamps.

[0080] As the feed line descends from its highest position of 1.0 meter to its lowest position of 0.2 meters, the filter bag gradually relaxes from a slightly stretched state, and the spring frame compresses accordingly. However, due to the support of the springs, the flow channel of the filter bag remains unobstructed, with the cross-section not less than 80% of the original cross-section, preventing collapse due to negative pressure or its own weight. As the feed line rises from its lowest position to its highest position, the spring frame gradually extends, assisting the filter bag in restoring its shape and preventing dead folds caused by excessive folding. During the lifting and lowering process, it automatically adapts to the angle changes between the feed line inlet and the hopper outlet. The maximum swing angle has been measured to be ±25°, meeting the angle change requirements within a 1.3-meter lifting stroke.

[0081] Through the above structural design, this flexible connection module effectively solves the dynamic connection problem between the fixed hopper and the lifting feed line. Compared with traditional canvas sleeves, the internal spring skeleton prevents collapse and blockage, ensuring smooth feed flow; the redundant length design avoids tearing of the filter bag due to excessive stretching; filter bag replacement is convenient and quick, and maintenance is easy. The overall structure is simple and cost-controllable, providing a reliable hardware foundation for the entire adaptive feeding system. Depending on different pipe diameters and lifting strokes, the filter bag length and spring specifications can be adjusted accordingly. For example, for DN90 pipelines, when the lifting stroke is larger, the filter bag length can be increased to 1.8 meters, and the spring wire diameter can be thickened to 2.5mm.

[0082] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. An automatic feeding system that adapts to the growth of a chicken flock, characterized in that, include: A liftable feeding pipeline with multiple material trays on it; A lifting mechanism that drives the feed pipeline to rise and fall; Multiple ultrasonic sensors are fixedly installed at intervals above the feed pipeline along its length. Each ultrasonic sensor is used to continuously emit ultrasonic waves at a sampling frequency of not less than 20Hz and receive echoes, and output a time-series distance signal. A controller is connected to both the ultrasonic sensor and the lifting mechanism. The controller has a preset ground reference distance value and is configured to perform the following steps: For each ultrasonic sensor, M consecutive distance measurements are acquired within a sliding time window, where M ≥ 20. The variance of the distance measurements within the sliding time window is calculated. When the variance is less than a preset stability threshold, it is determined that there is a feeding target below the sensor within the current time window, and the mean of the distance measurements within the window is used as a candidate height value. The candidate height value is compared with the ground reference distance value, and candidate values ​​with a difference less than a preset ground threshold are eliminated to obtain the effective chicken back height value. All effective chicken back height values ​​acquired by all ultrasonic sensors within a set time period are statistically analyzed, and the median or mode is taken as the average back height of the current flock. The target feed line height is calculated based on the average back height, and the lifting mechanism is controlled to adjust the feed pipeline to the target feed line height.

2. The automatic feeding system for adaptive chicken flock growth according to claim 1, characterized in that: The system also includes a passive infrared sensor installed corresponding to each ultrasonic sensor, wherein the detection area of ​​the passive infrared sensor and the beam area of ​​the ultrasonic sensor overlap in space. The controller is also configured to: synchronously acquire the distance measurement sequence of the ultrasonic sensor and the heat source signal sequence of the passive infrared sensor within the sliding time window; calculate the variance of the distance measurement sequence and calculate the fluctuation characteristic value of the heat source signal sequence; When the variance is less than a preset stability threshold and the fluctuation characteristic value of the heat source signal sequence exceeds a preset live animal threshold, it is determined that there are live chickens in the current window in a feeding state, and the average distance measurement value of the window is used as the candidate chicken back height value. When the variance is less than the preset stability threshold, but the fluctuation characteristic value of the heat source signal sequence does not exceed the preset live threshold, it is further determined whether the mean of the heat source signal sequence exceeds the ambient temperature threshold: if yes, it is determined to be a static non-chicken heat source interference, and the data of this window is discarded; if no, it is determined to be a period without chickens feeding, the data of this window is discarded and the sleep mode is triggered, and sampling is paused until the heat source signal is restored. When the variance is greater than or equal to a preset stability threshold, but the fluctuation characteristic value of the heat source signal sequence exceeds a preset live animal threshold, it is determined to be a live chicken in motion. The current distance value is not accepted, but the activity intensity of chickens in the area is recorded. The controller calculates the average back height based on all valid candidate chicken back height values ​​and dynamically adjusts the width of the sliding time window according to the activity intensity of the chickens.

3. The adaptive automatic feeding system for chicken flock growth according to claim 2, characterized in that, The controller is also configured to: Obtain the current actual height of the feeding pipeline, and calculate the theoretical distance value of the material tray below the corresponding position of each ultrasonic sensor in real time based on the fixed installation height of each ultrasonic sensor. When the variance within the sliding time window is less than a preset stability threshold and the fluctuation characteristic value of the heat source signal sequence exceeds a preset living threshold, the mean of the distance measurement value within the window is further compared with the theoretical distance value of the material tray at the corresponding position. If the difference between the mean value and the theoretical distance value of the tray is greater than or equal to the preset occlusion threshold, then the mean value is used as the candidate chicken back height value. If the difference between the mean value and the theoretical distance value of the material tray is less than the preset occlusion threshold, it is determined that the current sensor is occluded by the material tray, the data in that window is discarded, and the event of the sensor being occluded is recorded.

4. The automatic feeding system for adaptive chicken flock growth according to claim 3, characterized in that, It also includes pressure sensors installed in each tray for detecting the load weight on the tray; the controller is further configured to: Obtain the load weight signal within the time window corresponding to the candidate chicken back height value, and calculate its average value; Based on the relationship between the standard weight and back height of chickens at the current age, determine the theoretical back height range corresponding to the average load weight. If the candidate chicken back height value exceeds the upper limit of the theoretical back height range, and the excess is greater than the preset stacking threshold, then the candidate chicken back height value is determined to originate from chicken stacking and is discarded. For the discarded measurement points, the historical height value effectively measured by the sensor within adjacent time windows, or the weighted average of the effective height values ​​of sensors at adjacent locations within the same time window, is used as the representative value of the chicken back height for that area.

5. The adaptive automatic feeding system for chicken flock growth according to claim 4, characterized in that, The controller is also configured to: Within the sliding time window, the fluctuation variance of the load weight signal detected by the pressure sensor is calculated synchronously. When the variance of the ultrasonic distance measurement value sequence is greater than or equal to a preset stability threshold, but the fluctuation variance of the load weight signal is less than a preset body stability threshold, and the fluctuation characteristic value of the heat source signal sequence exceeds a preset live threshold, it is determined that the chickens in the current window are in a feeding state, and the median of the ultrasonic distance measurement value sequence is used as the candidate chicken back height value. When the variance of the ultrasonic distance measurement value sequence is greater than or equal to a preset stability threshold, and the fluctuation variance of the load weight signal is also greater than or equal to a preset body stability threshold, it is determined that the chickens in the current window are in a moving state, and the data of that window is discarded.

6. The adaptive automatic feeding system for chicken flock growth according to claim 5, characterized in that, The controller is also configured to: Record the candidate chicken back height values ​​and their corresponding timestamps generated by multiple ultrasonic sensors sequentially distributed along the length of the feed pipeline within multiple consecutive sliding time windows; Within a preset short time interval Δt, it is determined whether at least three adjacent ultrasonic sensors sequentially generate candidate chicken back height values, and the maximum difference between these candidate values ​​is less than a preset height consistency threshold. If the above situation is detected, it is determined to be an event of chickens running by quickly, and all candidate chicken back height values ​​generated within the corresponding time window are marked as pseudo data and discarded. If the above situation is not detected, the candidate chicken back height value determined according to claim 5 shall be retained for subsequent statistical average back height.

7. The automatic feeding system for adaptive flock growth according to any one of claims 1 to 6, characterized in that, It also includes suspension assemblies installed at both ends of the supply pipeline for support and vibration; the suspension assembly includes a sleeve, an elastic layer, a vibration damping rod, a vibration isolation joint, and a vibrator; the sleeve is a split annular structure that hugs the end of the supply pipeline, and the elastic layer is provided between its inner wall and the outer wall of the supply pipeline. The elastic layer is a ring-shaped rubber or polyurethane bushing used to transfer the static load of the supply pipeline to the sleeve and isolate the vibration generated by the vibrator from being transmitted upward; the vibration damping rod is a rigid rod, the upper end of which is connected to the suspension point of the lifting mechanism, and the lower end is connected to the sleeve through the vibration isolation joint; the vibration isolation joint is an elastic element used to further attenuate the vibration transmitted upward; the vibrator is fixedly installed on the supply pipeline to generate periodic excitation force, which drives the supply pipeline to vibrate through the sleeve and the elastic layer; The controller is also configured to control the start / stop, vibration frequency, and vibration amplitude of the vibrator based on the average back height of the flock, feeding activity, or a preset cleaning mode.

8. The automatic feeding system for adaptive flock growth according to claim 7, characterized in that, The controller is also configured to: calculate the feeding activity index of each feed tray in real time based on the load weight signal detected by the pressure sensor, wherein the feeding activity index is the high-frequency fluctuation energy value of the load weight signal within the sliding time window; The feeding activity index of all feed trays within the same time period is counted to obtain the current group feeding activity of the flock. The group's feeding activity level is compared with a preset activity benchmark value to obtain the activity deviation; The operating parameters of the exciter are dynamically adjusted based on the activity deviation: When the activity level deviation is positive and exceeds the first threshold, the flock is judged to be actively feeding. The vibration frequency and / or vibration amplitude are increased to promote feed flow and feeding. When the activity deviation is negative and exceeds the second threshold, the flock is judged to be feeding passively. The vibration frequency and / or vibration amplitude should be reduced or the vibrator should be stopped to avoid vibration disturbing the flock. When the activity deviation is between the first threshold and the second threshold, the current excitation parameters are maintained.

9. The automatic feeding system for adaptive flock growth according to claim 8, characterized in that, The lifting mechanism includes a first lifting unit and a second lifting unit that independently drive the two ends of the feeding pipeline to lift and lower. The controller is further configured to: Based on the feeding activity index of each feed tray, the differences in the spatial distribution of feeding activity are identified, and the spatial distribution characteristics of activity are obtained. Based on the spatial distribution characteristics of the activity level, the height difference between the two ends of the feed pipeline is dynamically adjusted to make the feed pipeline tilt. When it is identified that the feeding activity index of the feed tray in the first area is significantly higher than that in the second area, the feed pipeline end corresponding to the first area is lowered and the feed pipeline end corresponding to the second area is raised, forming a downward slope from the second area to the first area. The adjustment range of the height difference between the two ends is limited to a preset tilt angle threshold to ensure that the overall height of the feed pipeline is still within the optimal height range determined based on the average back height of the flock.

10. The automatic feeding system for adaptive flock growth according to claim 9, characterized in that, It also includes a vent valve located at the end of the feed line; the controller is further configured to execute a cleaning mode upon receiving a cleaning command. First, based on the load weight signal detected by the pressure sensor and the heat source signal of the passive infrared sensor, it is determined whether all feed trays are in a chicken-free state. Once it is confirmed that all feed trays are free of chickens, the lifting mechanism is controlled to raise the feed pipeline to a preset cleaning height, which is higher than the optimal height range determined based on the average back height of the flock. After the feed pipeline is raised to the cleaning height, the drain valve is opened and the vibrator is simultaneously controlled to operate in a preset cleaning mode, which includes high-frequency vibration and / or variable frequency vibration, for a first preset time, so that the residual feed in the feed pipeline is discharged from the valve under the combined action of gravity and vibration. After the cleaning mode ends, the drain valve is closed and the lifting mechanism is used to restore the feed line to the optimal height determined based on the average back height of the current flock.

Images