A cage poultry house odor anti-disturbance detection and tracing method and system

Abstract

This invention discloses a method and system for disturbance-resistant detection and source tracing of odors in caged poultry houses. The method includes using a proximity sniffer module to physically isolate the cage, acquiring targeted odor signals from the cage's microenvironment through negative pressure sampling, and collecting background odor signals, robot speed, and probe height. The targeted odor signal and the background odor signal are then subjected to dual-channel adaptive differential processing to eliminate background interference. Based on frequency domain inverse filtering, the response hysteresis of the gas sensor in the gas sensor chamber 2 is reconstructed to obtain a timestamp. Kinematic integral inversion is performed based on the actual peak timestamp and the robot speed to calculate the odor source coordinates and achieve precise source tracing of individual cages. This invention achieves high-fidelity acquisition of pathological odor signals from caged poultry houses and centimeter-level precise source tracing of individual cages through a hardware and software co-design of physical disturbance-resistant sampling, dynamic differential denoising, frequency domain hysteresis correction, and kinematic integral source tracing.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring technology for caged poultry houses, and in particular to a method and system for detecting and tracing the source of odors in caged poultry houses without disturbance. Background Technology

[0002] In modern intensive and large-scale poultry farming, tiered cage rearing has become the mainstream method due to its advantages of high space utilization, high stocking density, and high standardization. During the incubation period and early stages of typical poultry diseases such as respiratory and digestive tract diseases, the body's respiratory metabolism, shed skin products, and trace excretions continuously release specific volatile organic compounds, forming pathological odors distinct from normal farming environments. Typically, abnormal mutations in these odor signals precede visible clinical symptoms such as abnormal body temperature, lethargy, and depression, making them a core indicator for early warning and intervention in poultry diseases. Therefore, precise odor detection and lesion tracing technologies for caged poultry houses have significant engineering application value and livestock disease prevention significance for achieving early detection and treatment of poultry diseases, cutting off disease transmission chains, reducing poultry mortality rates, and improving the intelligent management level of large-scale farming.

[0003] With the iterative upgrades of agricultural IoT and intelligent inspection technologies, mobile inspection robots equipped with gas sensing devices are gradually replacing traditional manual inspection methods. They are being used for environmental monitoring in long-distance, narrow-channel caged poultry houses, effectively compensating for the shortcomings of manual inspections, such as low efficiency, strong subjectivity, and inability to provide continuous monitoring. However, in the complex environment of multi-layered caged poultry farming, the existing conventional detection solution of "mobile inspection equipment + gas sensor" suffers from significant technical defects due to the coupling effects of the complex airflow field inside the house, the inherent response characteristics of the sensors, and the dynamic operating conditions of the equipment. This makes it difficult to achieve high-precision odor detection and lesion tracing at the individual cage level.

[0004] On the one hand, the concentration of pathological odors produced by poultry metabolism is extremely low, and they are only concentrated in local microenvironments around cages and feed troughs, resulting in weak signals and concentrated spatial distribution. To ensure ventilation and environmental balance within caged poultry houses, longitudinal ventilation equipment is typically installed, creating continuous lateral convection airflow. Simultaneously, mobile inspection equipment generates wake disturbances as it travels through aisles. Under the combined effect of these two airflows, the weak odor plumes within the cage microenvironment are easily diluted, dispersed, and mixed. This causes conventionally deployed gas sensors to only collect the mixed background gas, with the effective characteristic odor signal being masked by background noise. This easily leads to missed characteristic signals and insufficient detection accuracy, failing to accurately reflect the actual odor state of individual cages.

[0005] On the other hand, the physicochemical processes of gas sensor chamber filling and discharging, pipeline gas transmission, and adsorption / desorption of the sensing medium inherently exhibit first-order inertial hysteresis, resulting in a fixed response delay in the sensor's output. When the inspection equipment is in dynamic operation, by the time the sensor detects a peak odor concentration signal, the equipment has already deviated from the actual location of the odor release cage. This inherent hysteresis not only causes tailing and waveform distortion of the odor pulse signal in the time domain but also translates into spatial positioning offset as the equipment moves. This leads to a mismatch between the timestamp of the detected signal and the actual spatial location, making it difficult to determine the true source cage location of the odor signal, ultimately resulting in poor source tracing accuracy and positioning failure.

[0006] In summary, existing poultry house odor inspection technologies lack targeted micro-environment sampling and airflow isolation methods, cannot avoid disturbances caused by convective airflow and equipment airflow within the house, and do not correct for dynamic detection errors caused by the inherent hysteresis characteristics of sensors at the algorithm level. As a result, it is difficult to achieve disturbance-resistant and accurate detection of odor signals in caged poultry houses and targeted tracing of individual cages, and cannot meet the practical application needs of early and accurate early warning of diseases in large-scale poultry farming.

[0007] The disclosure of the above background technical content is only for the purpose of assisting in understanding the concept and technical solution of this application, and does not necessarily provide technical instruction. Summary of the Invention

[0008] The purpose of this invention is to provide a method and system for detecting and tracing the source of odors in caged poultry houses without disturbance.

[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for detecting and tracing the source of odor in caged poultry houses with resistance to disturbance, comprising the following steps: S1. Directional extraction of air from the micro-environment of the cage, simultaneously collecting target odor concentration, ambient background odor concentration, instantaneous travel speed of the collection device, and cumulative displacement data and gimbal height data output by the digital odometer; S2. Configure dynamic compensation weighting coefficients based on the operating conditions of the poultry house longitudinal ventilation fan, perform differential calculations using the odor concentration data to obtain the net concentration signal and remove environmental noise. S3. Model the hysteresis characteristics of the sensor chamber and pipeline as a first-order inertial system with an inherent time constant; perform frequency domain conversion on the net concentration signal, solve the real signal spectrum through frequency domain inverse filtering, and then restore it to the time domain transient odor pulse through inverse transformation, and extract the real peak timestamp corresponding to the pulse; S4. Combining the real peak timestamp, the driving speed, the gimbal height data, and the digital odometer data, perform continuous time-space integral inversion to eliminate calibration errors caused by vehicle speed fluctuations and calculate the absolute lateral coordinates of the odor release point; fuse the absolute lateral coordinates with the height coordinates to output a single-unit cage-position targeted tracing command.

[0010] Furthermore, following any one or a combination of the aforementioned technical solutions, the directional extraction of microenvironmental air in step S1 is performed by using a proximity sampling method close to the cage, which isolates the odor dilution interference caused by the lateral airflow in the poultry house and the wake of the collecting device, ensuring that the collected microenvironmental odor signal is authentic and effective.

[0011] Furthermore, following any one or a combination of the aforementioned technical solutions, the dynamic compensation weighting coefficient in step S2 is adaptively adjusted in real time based on the real-time operating speed of the poultry house longitudinal ventilation fan, the current ventilation speed, and the air convection intensity inside the house. When the fan speed increases, the wind speed increases, and the air convection inside the building is enhanced, the dynamic compensation weighting coefficient is increased.

[0012] Furthermore, following any one or a combination of the aforementioned technical solutions, the net concentration signal in step S2 is calculated using the following formula:

[0013] in, To dynamically compensate for the weighting coefficients, the control unit reads the current fan speed of the chicken house environmental control system and dynamically looks up the table to assign weights. Different weights; When the fans are fully running and the air convection inside the building is intense, the dynamic compensation weighting coefficient is increased to enhance background suppression, thereby offsetting the high-concentration common-mode environmental noise and extracting the net concentration signal sequence representing the metabolites of sudden pathological events.

[0014] Furthermore, based on any one or a combination of the aforementioned technical solutions, the frequency domain transfer function of the first-order inertial system in step S3 is determined according to the following formula:

[0015] in, j The imaginary unit, f For signal frequency, τ This is the inherent physical time constant of the sensor chamber and pipeline; The frequency domain transfer function is used to characterize the amplitude attenuation and phase hysteresis characteristics of the sensor system for odor signals.

[0016] Furthermore, based on any one or a combination of the aforementioned technical solutions, the true odor concentration spectrum in S3 is calculated using the following formula:

[0017] Where H*(f) is the complex conjugate of the frequency domain transfer function. This is the regularization noise suppression constant.

[0018] Furthermore, following any one or a combination of the aforementioned technical solutions, the continuous time-space integral inversion described in S4 calculates the absolute lateral coordinates of the odor release point according to the following formula:

[0019] in, The initial lateral coordinates are obtained by integrating the instantaneous travel speed of the data acquisition device over time, using the actual peak timestamp as a reference. This refers to the fixed installation offset of the sampling probe relative to the reference origin of the acquisition device; The fixed installation offset is superimposed on the integral value to correct the positioning deviation caused by the misalignment between the probe and the center of the acquisition device, and finally the accurate absolute lateral coordinates of the odor release point are obtained.

[0020] According to another aspect of the present invention, the present invention provides an anti-disturbance detection and tracing system for odors in caged poultry houses, a proximity-type anti-disturbance snoring sampling module, which is equipped with a flexible windproof isolation structure and a negative pressure gas collection structure, for forming a locally enclosed sampling space close to the cage, isolating external airflow disturbances and collecting targeted odor signals of the cage microenvironment; The environmental baseline sensing unit is deployed in an unobstructed location on the leeward side of the acquisition equipment to collect background odor signals of the entire poultry house environment in real time, providing baseline data for differential noise reduction processing. The motion and position sensing unit is used to collect the instantaneous speed of the device and the real-time lifting height of the pan-tilt unit in real time, providing motion parameters and three-dimensional position parameters for kinematic integration tracing of odor sources; The edge computing processing unit is communicatively connected to the proximity-type anti-disturbance sniffing and sampling module, the environmental reference sensing unit, and the motion and position sensing unit, respectively. It has built-in signal denoising, sensor hysteresis compensation, waveform reconstruction, and kinematic source tracing algorithms to execute the anti-disturbance detection and source tracing method for odors in caged poultry houses as described above, and calculates accurate source tracing results for individual cage odors.

[0021] Furthermore, based on any or a combination of the aforementioned technical solutions, the proximity-type anti-disturbance sniffing sampling module includes a horn-shaped flexible windproof shroud, a coaxial dual-ring sampling probe, and a negative pressure suction pump. The horn-shaped flexible windproof hood is used to form a partially enclosed active airflow barrier between the sampling probe and the cage grid, effectively isolating the transverse convective wind in the poultry house and the wake of the sampling equipment, preventing the dilution of characteristic odors in the cage microenvironment, and providing a stable environment with low disturbance for the sampling process. A coaxial dual-ring sampling probe is used to be nested inside the flow guide hood. It achieves directional collection and transmission of target odor through a dual-path structure, while maintaining the stability of the sampling airflow, ensuring that the microenvironment gas in the cage can enter the sensor detection cavity in a low-delay and high-fidelity manner. The negative pressure suction pump is used to provide stable negative pressure power for the coaxial dual-ring sampling probe, continuously extracting characteristic odor gases in the cage microenvironment, ensuring a constant sampling gas flow rate, and avoiding gas transmission lag from affecting the real-time performance and accuracy of the detection.

[0022] Furthermore, based on any or a combination of the aforementioned technical solutions, the motion and position sensing unit includes a chassis speed encoder and a gimbal lifting absolute encoder. The chassis speed encoder of the acquisition device is used to collect the instantaneous driving speed and cumulative displacement data of the acquisition device in real time; The absolute encoder for pan-tilt-zoom (PTZ) lifting is used to provide real-time feedback on the current absolute height of the acquisition PTZ and the vertical position information corresponding to different stacked cage positions, providing a height dimension positioning basis for three-dimensional coordinate fusion and cage position matching.

[0023] The beneficial effects of the technical solution provided by this invention are as follows: a. This solution adopts a coaxial positive and negative pressure coupled air curtain structure, which improves the traditional passive gas collection sampling method into a microenvironment sampling mode with active barrier function. By using physical structure to isolate the disturbance caused by the cross airflow in the poultry house and the wake of the robot, the integrity and fidelity of the pathological characteristic odor signal collection are guaranteed from the hardware level. b. This solution applies frequency domain signal processing algorithms to gas detection scenarios in aquaculture environments. Addressing the inherent response hysteresis caused by the sensor's gas chamber and pipeline, it reconstructs the waveform of the distorted signal through mathematical model inversion, restoring the true transient odor pulse signal and thoroughly improving problems such as missed detections and odor source location errors caused by signal hysteresis. c. This solution does not require the deployment of a large-scale sensor array. Relying on the robot's kinematic integral mapping mechanism, it can improve the odor source positioning accuracy to the centimeter level during continuous equipment inspection, accurately matching it to a single cage position. The overall hardware cost of the solution is low, the operation reliability is high, and it has good prospects for industrial application. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A schematic diagram of an anti-disturbance detection and source tracing method provided as an exemplary embodiment of the present invention; Figure 2 A schematic diagram of the algorithm flow provided for an exemplary embodiment of the present invention; Figure 3 A schematic diagram of the overall hardware structure provided for an exemplary embodiment of the present invention; Figure 4 A schematic diagram of an anti-disturbance detection and tracing system provided for an exemplary embodiment of the present invention.

[0026] Among them, 1-inspection robot, 2-gas sensor chamber, 3-environmental reference sensor, 4-air pump, 5-lifting platform, 6-sniffing module, 7-windproof cover, 8-speed encoder. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0029] In one embodiment of the present invention, see Appendix Figure 1As shown, a disturbance-resistant detection and source tracing method for odor in caged poultry houses is provided, including the following steps: S1, proximity-based physical isolation and synchronous acquisition based on flexible mechanical shielding: During the operation of the inspection robot, a proximity-based sniffing module is used to directionally extract the micro-environmental air of the cage position under the physical barrier protection of the flexible windproof guide hood, and simultaneously collects the target odor concentration, ambient background odor concentration, instantaneous driving speed, and gimbal height information; S2, dual-channel spatial difference and environmental baseline elimination: a dynamic compensation weight coefficient is set according to the operating conditions of the poultry house longitudinal ventilation fan, and the net concentration signal is calculated through a differential algorithm to remove the background noise of the poultry house environment; S3, based on fast Fourier transform and frequency domain inverse filtering Hysteresis Waveform Reconstruction: The physical hysteresis characteristics of the sensor chamber and pipeline are modeled as a first-order inertial system with an inherent time constant. After converting the net concentration signal to the frequency domain, a frequency domain inverse filtering equation with a noise suppression constant is introduced to back-calculate the true signal spectrum. The signal is then restored to a transient odor pulse in the time domain through inverse transformation, and the true peak timestamp of the pulse is extracted. S4, Nonlinear Kinematic Back-calculation and Three-Dimensional Targeting Tracing: Based on the extracted true peak timestamp and digital odometer data, the calibration error caused by vehicle speed fluctuations is eliminated through continuous time-space integration inversion. The absolute lateral coordinates of the odor release point are calculated and fused with the height coordinates to output a single-cell cage-position targeting tracing command.

[0030] This invention addresses the pain points of mobile gas sampling in caged poultry houses, such as airflow dilution, background noise interference, sensor signal hysteresis, and insufficient traceability accuracy, by developing a complete detection and traceability solution. At the hardware level, a proximity-type anti-disturbance sampling module creates a locally isolated space, significantly improving sampling reliability under complex wind conditions. At the algorithm level, adaptive differential computation, frequency domain waveform reconstruction, and kinematic integral calculation with offset compensation are used to progressively optimize signal quality and positioning results. All algorithm parameters and calculation formulas are clearly defined, ensuring strong technical reproducibility. Furthermore, the system architecture is simple, requiring only an inspection robot to complete the entire process, resulting in minimal hardware investment and convenient deployment. Ultimately, it can stably achieve high-fidelity odor collection, noise suppression, signal correction, and accurate traceability of individual cages, with overall performance significantly superior to existing technologies, meeting the practical needs of early disease warning in large-scale poultry farming.

[0031] In one embodiment of the present invention, see Appendix Figure 1-4 As shown, a disturbance-resistant detection and source tracing method for odors in caged poultry houses is provided. This method is applied to stacked caged poultry houses and relies on autonomous inspection robots for continuous inspection operations. It comprehensively processes multiple interference factors such as ventilation airflow, robot wake, high-concentration environmental noise, sensor inherent hysteresis, and inspection displacement error to achieve high-fidelity acquisition of pathological odor signals, noise filtering, signal correction, and accurate source tracing of individual cages. The specific execution steps include the following: S1. A proximity-based physical isolation and synchronous data collection inspection robot, based on flexible mechanical shielding, moves along the poultry house aisle according to a preset travel path and speed. The robot's multi-degree-of-freedom lifting platform adjusts its position according to the cage height. For example, the lifting platform extends the proximity-based sniffing module to a position 2-5cm above the grid above the cage feed trough, aligning the sniffing module at the end of the platform with the area to be tested. For example, a horn-shaped flexible windproof guide is provided on the outside of the proximity-based sniffing module. The guide is made of flexible material and fits tightly against the cage grid surface during operation, creating a partially semi-enclosed space around the sampling area, forming an active physical barrier. This can forcibly block the lateral airflow of the chicken house fan with a flow rate of up to 2-3m / s and the wake generated by the robot's own movement, preventing the dilution of weak pathological odors in the cage microenvironment by the airflow.

[0032] The built-in negative pressure suction pump of the sniffing module is activated to directionally extract local micro-environment air in the cage; multiple types of data are collected synchronously at a fixed sampling frequency: the targeted odor concentration of the cage micro-environment is collected through the targeted gas sensor, the background odor concentration of the entire poultry house is collected through the independently deployed environmental reference sensor, the real-time instantaneous driving speed of the robot is collected through the chassis speed measuring device, and the current absolute height of the gimbal is collected through the gimbal encoder.

[0033] All collected data is transmitted synchronously to the edge computing unit in real time to complete the unified caching and preprocessing of multi-source sensing data, providing raw data support for subsequent signal processing.

[0034] S2. The dual-channel spatial difference and environmental baseline elimination edge calculation unit establishes real-time data interaction with the poultry house environmental control system, reading the operating speed of the longitudinal ventilation fan, real-time wind speed, and overall air convection intensity inside the house. Based on these operating parameters, it dynamically generates real-time changing compensation weight coefficients. This step uses a dual-channel spatial difference algorithm to remove environmental baseline noise, according to the formula... The net concentration signal was calculated; where To target odor concentration, Background odor concentration, This is the dynamic compensation weighting coefficient.

[0035] The weighting coefficient is adaptively adjusted by using a lookup table: when the fan is at a low speed, the air velocity inside the building is low, and the convection intensity is weak, the basic weighting value is used for compensation; when the fan speed is increased, the air velocity increases, and the air convection inside the building is intense, the system automatically increases the dynamic compensation weighting coefficient to enhance the suppression effect on background gases such as ammonia and hydrogen sulfide that are common to the entire area.

[0036] For example, dynamic compensation weight coefficient The value range is set to 0.2~0.9. The upper limit is used when the ventilator is running at full load, and the lower limit is used when it is running at low load. The high concentration of environmental noise in the poultry house is canceled by differential operation, and the signal sequence containing only the local features of the cage is separated from the mixed signal, so as to accurately extract the effective odor signal representing the metabolites of sudden pathological diseases in poultry.

[0037] S3. Hysteresis Waveform Reconstruction Based on Fast Fourier Transform and Inverse Frequency Domain Filtering: Due to the gas transmission delay in the gas sensor's detection chamber and connecting pipeline, the sensor as a whole exhibits significant response hysteresis characteristics. In this embodiment, the physical hysteresis characteristics of the sensor chamber and pipeline are uniformly modeled as a first-order inertial system. The frequency domain transfer function of this system is: In the formula The imaginary unit, Where τ is the signal frequency, and τ is the inherent physical time constant corresponding to the sensor chamber and the pipeline. For example, the inherent physical time constant τ of the sensor ranges from 0.8 to 2.5 s. This transfer function is used to quantitatively characterize the amplitude attenuation and phase lag of the odor signal caused by the sensor system.

[0038] The net concentration time-domain signal output in step S2 is converted into a frequency-domain signal using a Fast Fourier Transform (FFT). The conversion formula is as follows: .

[0039] To eliminate signal tailing and peak hysteresis caused by sensor hysteresis, a frequency domain inverse filtering algorithm with a noise suppression constant is introduced, according to the formula... Calculate the spectrum of true odor concentration; where, Let be the complex conjugate of the frequency domain transfer function. This is a regularization noise suppression constant, for example, a regularization noise suppression constant. Value set to This can effectively suppress the problem of high-frequency noise amplification during the inverse operation and ensure the stability of the algorithm.

[0040] For the compensated spectrum Performing an inverse fast Fourier transform (IFFT) restores the frequency domain signal to the time domain signal; the corresponding formula is: After inversion, the originally gentle wave signal with a tail lasting several seconds was forcibly "compressed" into an extremely steep transient pulse by a mathematical algorithm; the system extracted the moment of the maximum value of this reconstructed pulse and accurately pinpointed the real timestamp of the odor molecule leaving the sick chicken. This timestamp serves as the core time reference for subsequent tracing operations.

[0041] S4. Nonlinear Kinematics Backpropagation and 3D Targeted Source Tracing: Using the actual peak timestamp obtained in step S3 as the starting reference, and combining the driving data output by the digital odometry, continuous time-space integral inversion is performed on the robot's instantaneous driving speed to preliminarily calculate the initial lateral coordinates corresponding to the odor source. Considering that the proximity sniffer module is eccentrically mounted on the robot, and the probe position has a fixed deviation from the robot's reference origin, the fixed installation offset of the sniffer module relative to the robot's reference origin is included in the calculation. For example, fixed installation offset The value range is 5~15cm according to the formula. The absolute lateral coordinates of the odor release point are obtained by solving the problem, and the calibration error caused by hardware installation offset and vehicle speed fluctuation is corrected.

[0042] The calculated absolute lateral coordinates are fused with the PTZ height coordinates Z collected synchronously in step S1 to form the three-dimensional coordinates (X, Z) of the odor source. Then, address matching is performed in the system's preset "layered poultry house digital twin matrix" to match the three-dimensional coordinates with the poultry house cage topology layout one by one, and a precise alarm command is directly output. For example, "abnormal respiratory odor detected in cage at the second row, third layer, 15.6 meters deep on the left." The final source tracing and positioning error can be controlled within ±3cm, achieving the standard for individual cage identification, locking the individual cage corresponding to the odor release, and completing the entire process of detection, correction and source tracing.

[0043] This method relies on the continuous inspection mode of the robot without stopping throughout the process. It solves industry pain points such as airflow dilution, background interference, sensor hysteresis, and positioning deviation from multiple levels, including physical structure, signal processing, algorithm correction, and coordinate calculation. Ultimately, it achieves centimeter-level positioning accuracy and is fully adapted to the routine monitoring needs of large-scale caged poultry farming.

[0044] In one embodiment of the present invention, see Appendix Figure 3-4 As shown, an anti-disturbance detection and tracing system for odors in caged poultry houses is provided. The system is mounted on a mobile inspection robot 1 that can autonomously move along the aisles of the poultry house. The hardware and software work together to fully execute the anti-disturbance detection and tracing method for odors in caged poultry houses described in the above embodiments. The system consists of four main parts: a proximity-type anti-disturbance sniffing sampling module 6, an environmental reference sensing unit (exemplarily an environmental reference sensor 3), a motion and position sensing unit, and an edge computing processing unit. The units communicate bidirectionally via industrial communication lines for data transmission and command interaction. The overall structure is compact, adaptable to the confined working space of caged poultry houses, eliminates the need for a large-scale distributed sensor network, and offers low hardware cost and convenient deployment.

[0045] The proximity-type anti-disturbance sniffing sampling module is fixedly installed at the end of the liftable gimbal / lifting platform 5 of robot 1. It is the core hardware for achieving micro-environment isolation sampling, specifically including a horn-shaped flexible windproof shield (wind shield 7), a coaxial dual-ring sampling probe, and a negative pressure suction pump 4. The flexible windproof shield is made of elastic and wear-resistant flexible material, which has the ability to deform and fit. During robot inspection, the lifting gimbal delivers the sampling probe to a position 2-5cm away from the cage grid. The shield can fit tightly against the cage grid, forming a local semi-enclosed space around the sampling area, constructing an active airflow barrier, effectively resisting the impact of 2-3m / s lateral airflow and robot wake, solving the problem of fluid airflow diluting the odor signal during mobile gas measurement from the hardware level. The coaxial dual-ring sampling probe is integrated inside the shield, and performs the functions of gas conduction and signal acquisition. The negative pressure suction pump is the active gas sampling power source, continuously extracting air from the micro-environment of the cage, ensuring stable sampling airflow, and completing the targeted odor signal acquisition.

[0046] The environmental reference sensing unit (also known as the environmental reference sensor) 3 is independently installed in an unobstructed area on the leeward side of the robot, physically isolated from the proximity-type anti-disturbance sniffing sampling module to avoid mutual interference of sampling airflows. This unit integrates a high-sensitivity gas detection sensor specifically designed to collect environmental background odor signals across the entire poultry house in real time. This provides reference data for subsequent dual-channel differential calculations, accurately reflecting the basic gas concentration level of the entire poultry house and ensuring effective background noise suppression.

[0047] The motion and position sensing unit is integrated into the robot chassis and the lifting gimbal, and includes a robot chassis speed encoder and a gimbal lifting absolute encoder. The chassis speed encoder collects motion data in real time, such as the robot's instantaneous speed and cumulative mileage; the gimbal lifting absolute encoder collects the current absolute height of the gimbal in real time, corresponding to the height information of different stacked cage positions. Both types of sensing devices continuously output data, providing complete motion and position parameters for the edge computing unit to perform kinematic integration, 3D coordinate fusion, and cage position matching.

[0048] The edge computing processing unit is the core of the system's computation, control, and scheduling. It is electrically and communicatively connected to the proximity-based anti-disturbance sniffing and sampling module, the environmental reference sensing unit, and the motion and position sensing unit, and can also establish data interface with the existing environmental control system of the poultry house. The edge computing processing unit internally contains a complete algorithm program, including a dual-channel adaptive differential operation module, a first-order inertial system modeling module, a fast Fourier transform and frequency domain inverse filtering module, a waveform reconstruction and timestamp extraction module, a kinematic integral inversion and coordinate correction module, and a cage topology matching module.

[0049] During operation, the edge computing processing unit first receives the raw data uploaded by each sensing unit and reads operating parameters such as the fan speed and fan position in the poultry house. Then, it calls the differential operation module to remove environmental noise and dynamically adjusts the compensation weight coefficient α(t) according to the range of 0.2 to 0.9, adjusting the compensation weight coefficient based on the ventilation conditions. Finally, it uses the frequency domain processing module to model the sensor hysteresis, employing an inherent time constant τ of 0.8 to 2.5 seconds (e.g., ...). The regularization constant of ) Inverse filtering and signal waveform reconstruction are performed to extract the true peak timestamp. Finally, relying on the kinematic integral algorithm and combined with the probe installation offset ΔL of 5-15cm to correct the coordinates, the height information is fused to match the poultry house cage topology matrix, ultimately controlling the traceability error within ±3cm and outputting traceability results and early warning instructions accurate to the individual cage position. This system, based on an integrated hardware and software design, combines an active barrier sampling structure, adaptive differential algorithm, frequency domain signal correction technology, and kinematic tracing mechanism to fully adapt to the agronomic conditions and inspection modes of caged poultry houses. Under high-speed, non-stop robotic inspection conditions, it can stably achieve high-fidelity odor signal acquisition, noise filtering, hysteresis correction, and centimeter-level individual cage location tracing. It boasts high operational reliability and simple maintenance, offering significant cost and deployment advantages compared to traditional dense sensor networking solutions. It can be widely applied to early disease warning and odor tracing scenarios in various tiered caged poultry farms.

[0050] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0051] The above description is only a specific embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for detecting and tracing the source of odor in caged poultry houses with resistance to disturbance, characterized in that, Includes the following steps: S1. Directional extraction of air from the micro-environment of the cage, simultaneously collecting target odor concentration, ambient background odor concentration, instantaneous travel speed of the collection device, and cumulative displacement data and gimbal height data output by the digital odometer; S2. Configure dynamic compensation weighting coefficients based on the operating conditions of the poultry house longitudinal ventilation fan, perform differential calculations using the odor concentration data to obtain the net concentration signal and remove environmental noise. S3. Model the hysteresis characteristics of the sensor chamber and pipeline as a first-order inertial system with an inherent time constant; perform frequency domain conversion on the net concentration signal, solve the real signal spectrum through frequency domain inverse filtering, and then restore it to the time domain transient odor pulse through inverse transformation, and extract the real peak timestamp corresponding to the pulse; S4. Combining the real peak timestamp, the driving speed, the gimbal height data, and the digital odometer data, perform continuous time-space integral inversion to eliminate calibration errors caused by vehicle speed fluctuations and calculate the absolute lateral coordinates of the odor release point; fuse the absolute lateral coordinates with the height coordinates to output a single-unit cage-position targeted tracing command.

2. The method for detecting and tracing the source of odor in caged poultry houses with resistance to disturbance according to claim 1, characterized in that, The directional extraction of micro-environmental air in step S1 is carried out by using a proximity sampling method close to the cage, which isolates the odor dilution interference caused by the lateral airflow in the poultry house and the wake of the collection device, ensuring that the collected micro-environmental odor signal is true and effective.

3. The method for detecting and tracing the source of odor in caged poultry houses with resistance to disturbance according to claim 1, characterized in that, The dynamic compensation weighting coefficient mentioned in step S2 is adjusted in real time based on the real-time operating speed of the longitudinal ventilation fan in the poultry house, the current ventilation speed, and the air convection intensity inside the house. When the fan speed increases, the wind speed increases, and the air convection inside the building is enhanced, the dynamic compensation weighting coefficient is increased.

4. The method for detecting and tracing the source of odor in caged poultry houses according to claim 1 or 3, characterized in that, The net concentration signal mentioned in step S2 is calculated using the following formula: in, To dynamically compensate for the weighting coefficients, the control unit reads the current fan speed of the chicken coop's environmental control system and dynamically assigns weights based on a lookup table. Different weights; When the fans are fully running and the air convection inside the building is intense, the dynamic compensation weighting coefficient is increased to enhance background suppression, thereby offsetting the high-concentration common-mode environmental noise and extracting the net concentration signal sequence representing the metabolites of sudden pathological events.

5. The method for detecting and tracing the source of odor in caged poultry houses with resistance to disturbance according to claim 1, characterized in that, The frequency domain transfer function of the first-order inertial system described in step S3 is determined according to the following formula: in, j The imaginary unit, f For signal frequency, τ This is the inherent physical time constant of the sensor chamber and pipeline; The frequency domain transfer function is used to characterize the amplitude attenuation and phase hysteresis characteristics of the sensor system for odor signals.

6. The method for detecting and tracing the source of odor in caged poultry houses with resistance to disturbance according to claim 1, characterized in that, The true odor concentration spectrum described in S3 is calculated using the following formula: Where H*(f) is the complex conjugate of the frequency domain transfer function. This is the regularization noise suppression constant.

7. The method for detecting and tracing the source of odor in caged poultry houses with resistance to disturbance according to claim 1, characterized in that, The continuous-time-space integral inversion described in S4 calculates the absolute lateral coordinates of the odor release point according to the following formula: in, The initial lateral coordinates are obtained by integrating the instantaneous travel speed of the data acquisition device over time, using the actual peak timestamp as a reference. This refers to the fixed installation offset of the sampling probe relative to the reference origin of the acquisition device; The fixed installation offset is superimposed on the integral value to correct the positioning deviation caused by the misalignment between the probe and the center of the acquisition device, and finally the accurate absolute lateral coordinates of the odor release point are obtained.

8. A disturbance-resistant detection and traceability system for odors in caged poultry houses, mounted on a collection device traveling along the aisle of the poultry house, characterized in that, include: The proximity-type anti-disturbance sniffing and sampling module is equipped with a flexible windproof isolation structure and a negative pressure gas sampling structure, which is used to form a local closed sampling space close to the cage, isolate external airflow disturbances and collect targeted odor signals from the cage's microenvironment; The environmental baseline sensing unit is deployed in an unobstructed location on the leeward side of the acquisition equipment to collect background odor signals of the entire poultry house environment in real time, providing baseline data for differential noise reduction processing. The motion and position sensing unit is used to collect the instantaneous speed of the device and the real-time lifting height of the pan-tilt unit in real time, providing motion parameters and three-dimensional position parameters for kinematic integration tracing of odor sources; The edge computing processing unit is communicatively connected to the proximity-type anti-disturbance sniffing sampling module, the environmental reference sensing unit, and the motion and position sensing unit, respectively. It has built-in signal denoising, sensor hysteresis compensation, waveform reconstruction, and kinematic source tracing algorithms to execute the anti-disturbance detection and source tracing method for odor in caged poultry houses as described in any one of claims 1-7, and calculates accurate source tracing results for individual cage odor sources.

9. The anti-disturbance detection and traceability system for odor in caged poultry houses according to claim 8, characterized in that, The proximity-type anti-disturbance sniffing sampling module includes a horn-shaped flexible windproof guide, a coaxial dual-ring sampling probe, and a negative pressure suction pump. The horn-shaped flexible windproof hood is used to form a partially enclosed active airflow barrier between the sampling probe and the cage grid, effectively isolating the transverse convective wind in the poultry house and the wake of the sampling equipment, preventing the dilution of characteristic odors in the cage microenvironment, and providing a stable environment with low disturbance for the sampling process. A coaxial dual-ring sampling probe is used to be nested inside the flow guide hood. It achieves directional collection and transmission of target odor through a dual-path structure, while maintaining the stability of the sampling airflow, ensuring that the microenvironment gas in the cage can enter the sensor detection cavity in a low-delay and high-fidelity manner. The negative pressure suction pump is used to provide stable negative pressure power for the coaxial dual-ring sampling probe, continuously extracting characteristic odor gases in the cage microenvironment, ensuring a constant sampling gas flow rate, and avoiding gas transmission lag from affecting the real-time performance and accuracy of the detection.

10. The anti-disturbance detection and traceability system for odor in caged poultry houses according to claim 8, characterized in that, The motion and position sensing unit includes a chassis speed encoder and a gimbal lifting absolute encoder. The chassis speed encoder of the acquisition device is used to collect the instantaneous driving speed and cumulative displacement data of the acquisition device in real time; The absolute encoder for pan-tilt-zoom (PTZ) lifting is used to provide real-time feedback on the current absolute height of the acquisition PTZ and the vertical position information corresponding to different stacked cage positions, providing a height dimension positioning basis for three-dimensional coordinate fusion and cage position matching.

Images