Lifting type multi-parameter environment sensing module, communication device and intelligent pig farm track inspection robot and inspection method

Abstract

The application discloses a lifting type multi-parameter environment sensing module and communication device, and a smart pig farm track inspection robot and an inspection method, and belongs to the technical field of smart livestock breeding equipment and livestock house environment intelligent regulation and control, and aims to solve the problems of blind area and data distortion in traditional fixed sensor monitoring. The application adopts a lifting type multi-parameter sensing module matched with a track inspection robot, realizes accurate partition positioning by relying on an RFID landmark, and carries a dual-channel redundant communication device to guarantee stable data transmission. The device completes temperature drift compensation, abnormal value elimination, image definition detection and thermal imaging data compression through a built-in edge preprocessing algorithm, realizes accurate and lightweight collection and uploading of environment data. Based on a partition differentiated PID control logic, the environmental control host is combined with a pig growth stage matching threshold, and is matched with a multiple anti-interference constraint mechanism and a timing feedback parameter self-correction strategy, so that a control system of global perception, intelligent decision, accurate execution and closed-loop optimization is constructed.

Description

Technical Field

[0001] This invention belongs to the field of intelligent livestock breeding equipment and intelligent control technology of livestock and poultry house environment, specifically involving a lifting multi-parameter environmental sensing module and communication device, as well as an intelligent pig farm track inspection robot and inspection method. Background Technology

[0002] As the pig farming industry continues to develop towards large-scale, intensive, and standardized operations, enclosed, high-density pig houses have become the mainstream farming model. Environmental parameters within these pig houses, such as temperature, ammonia, and carbon dioxide levels, directly determine the growth rate, health status, and disease incidence of pigs. Refined environmental management is crucial for improving pig farm efficiency and reducing losses. Accurate, comprehensive, and real-time environmental sensing and differentiated ventilation control are fundamental prerequisites for achieving intelligent management in smart pig farms. Currently, most large-scale pig farms' environmental monitoring systems generally adopt a fixed sensor network monitoring scheme. This involves deploying temperature, humidity, and gas concentration sensors at fixed points on the pig house walls and pillars, and uploading the collected data to the environmental control host via wired or wireless transmission. Global unified ventilation control is then based on this fixed-point data. However, this traditional technical solution still has the following problems in practical applications, making it unsuitable for the environmental management needs of high-density, intensive farming.

[0003] Firstly, fixed sensors suffer from monitoring blind spots and data lag: The environment inside pigsties is significantly affected by factors such as pigs congregating and lying down, localized accumulation of manure, uneven airflow distribution, and equipment obstruction, resulting in significant spatial variations in environmental parameters. This easily creates localized ventilation dead zones with high temperatures, high humidity, and high concentrations of harmful gases. Fixed sensors can only monitor single-point environmental data and cannot cover the entire pigsty space, making it difficult to detect localized environmental anomalies. This leads to a large discrepancy between the environmental data acquired by the environmental control host and the actual overall environmental state, severely limiting control decisions.

[0004] Secondly, traditional equipment lacks adaptability and stability: existing inspection sensor equipment is mostly fixed installation structure, which cannot flexibly adjust the monitoring height according to the growth height of pigs and the needs of airflow monitoring points, and the monitoring points are limited; the communication method is limited and lacks a redundant transmission mechanism, resulting in poor communication stability in complex pig farm environments; the opening and closing, protection, and walking limit structures of the equipment are imperfect, making them susceptible to corrosion and damage from dust and moisture inside the pigsty, resulting in short equipment lifespan and high operation and maintenance costs.

[0005] Third, traditional control models suffer from a serious "one-size-fits-all" drawback: existing environmental control systems mostly rely on a unified global threshold for overall ventilation control, failing to address the differentiated needs of different areas within the pigsty and different growth stages of pigs for zoned management. If the ventilation equipment for the entire pigsty is activated when the environmental conditions in a localized area exceed the standard, it can easily lead to over-ventilation in normal areas, causing problems such as pigs getting chilled, stress, and diarrhea, while also significantly increasing equipment energy consumption. On the other hand, if the ventilation equipment is only activated slightly, it cannot completely solve the problem of localized environmental conditions exceeding the standard, resulting in extremely poor accuracy and adaptability of environmental control.

[0006] Fourth, traditional monitoring equipment has insufficient data transmission and preprocessing capabilities: the raw data collected by fixed sensors has not undergone noise reduction, compensation, and compression processing, and is easily affected by factors such as ambient temperature fluctuations, electromagnetic pulse interference, and lens dirt and obstruction, resulting in frequent data distortion and noise interference problems; at the same time, the data transmission volume of high-definition data such as thermal imaging is large, and the bandwidth of traditional transmission links is limited, resulting in data transmission delays, packet loss, and high transmission costs, which cannot meet the needs of real-time closed-loop control. Summary of the Invention

[0007] The purpose of this invention is to provide a lifting multi-parameter environmental sensing module and communication device, as well as a smart pig farm track inspection robot and inspection method, to achieve differentiated, intelligent, and automated management and control of the pig house environment, improve the comfort of the pig farming environment, reduce farming risks and equipment energy consumption, and solve the problems of spatial blind spots, low data accuracy, poor transmission stability, one-size-fits-all ventilation control, and lack of closed-loop self-optimization capability in existing pig house environmental monitoring.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A liftable multi-parameter environmental sensing module mainly consists of a module housing, a gas acquisition unit, and vision and thermal imaging units. The gas acquisition unit, vision unit, and thermal imaging unit are all integrated inside the module housing, used to collect gas concentration parameters such as ammonia and carbon dioxide, environmental image data, and thermal imaging temperature data in the pigsty, respectively, achieving simultaneous acquisition of multi-dimensional environmental parameters. At least two sets of lower tripods are fixedly mounted at the bottom of the module housing, which are then fixedly mounted on a lifting mechanism. Driven by the lifting mechanism, the module housing can be vertically adjusted, allowing for flexible adjustment of the sensing module's monitoring height according to the pigsty's floor height, pig growth height, and airflow monitoring point requirements, adapting to the monitoring needs of different farming scenarios. The liftable structural design achieves full vertical space coverage monitoring, effectively eliminating blind spots in vertical environmental monitoring of the pigsty. The integrated gas, vision, and thermal imaging multi-dimensional acquisition units can simultaneously acquire environmental physicochemical parameters and image temperature characteristics, providing comprehensive data support for subsequent precise environmental analysis and ventilation control.

[0009] To achieve stable lifting, the lifting mechanism adopts a dual-column lifting adapter structure, including two sets of vertically fixed base columns and two sets of corresponding lifting columns arranged in front of the base columns. The module housing is fixedly installed on the front side of the lifting columns via a lower tripod, and completes the vertical lifting action synchronously with the lifting columns. Each set of base columns and the corresponding lifting column are equipped with a guide rail and slide block mating structure. The guide rail is vertically fixed to the side of the base column, and the slide block is fixed to the inside of the lifting column. The guide rail and slide block slide to fit together, guiding and limiting the up and down movement of the lifting column to ensure smooth and non-deviation-free lifting. A servo motor is fixedly installed on the rear side of the base column. The output shaft of the servo motor is coaxially connected to two sets of wire rope winding and unwinding winches, which can realize the winding and unwinding actions of the two sets of wire ropes simultaneously. Two sets of fixed pulleys are rotatably installed at the top of the base column, and two sets of movable pulleys are rotatably installed at the bottom of the lifting column. The steel wire rope wound around the winch passes through the fixed pulleys in sequence and is fixedly connected to the movable pulleys at the end. The winch is driven to rotate by a servo motor, which drives the steel wire rope to wind up and down, thereby pulling the movable pulleys and the entire lifting column to slide up and down, realizing precise lifting and lowering adjustment of the sensing module. The dual-rail guide and pulley group steel wire rope transmission structure ensures smooth lifting and high positioning accuracy, effectively avoiding shaking and deviation during the lifting of the sensing module and ensuring the stability of the collected data. The dual-set synchronous lifting structure is adapted to the overall weight of the sensing module, with strong load-bearing capacity, low operating noise, and fast response speed. The servo motor is precisely controllable and can realize arbitrary height positioning adjustment, adapting to the monitoring needs of multiple scenarios.

[0010] To achieve position adjustment, multiple sets of height adjustment holes are linearly arrayed along the side of the base column. The mounting shaft of the fixed pulley can be detachably installed inside any set of height adjustment holes. The installation height of the fixed pulley can be flexibly adjusted according to the actual lifting stroke requirements, thus adapting to different lifting adjustment ranges. By changing the mounting position of the fixed pulley, it can match the operational needs of different pig house heights and different monitoring heights, achieving adaptive adjustment of the lifting stroke. This achieves modular adjustable lifting stroke, adapting to monitoring operations in pig houses of different sizes without replacing the entire mechanism, greatly improving the equipment's versatility and adaptability. The linear array hole positioning is precise, the adjustment operation is simple and convenient, and the cost of subsequent debugging, modification, and upgrades is low.

[0011] A lifting communication device consists of two parts: a mobile communication box and a fixed communication base station. The mobile communication box integrates a LoRa RF module, a leaky cable coupler, and a protocol conversion gateway. The fixed communication base station integrates a leaky cable terminal, an industrial Ethernet interface, and a leaky cable, constructing a redundant wireless + wired transmission link. The mobile communication box is fixedly mounted on the front of the lifting column using at least two sets of upper tripods, positioned above the sensor module housing. It can rise and fall synchronously with the lifting mechanism, adaptively adjusting its communication height according to the monitoring point. The fixed communication base station is fixedly installed at the pigsty environmental control host, achieving stable connection for data reception, parsing, and command issuance. The protocol conversion gateway can perform protocol standardization conversion of the device's collected data, adapting to the environmental control host's communication protocol. The LoRa RF module and leaky cable work together to achieve long-distance, interference-resistant data transmission. The communication device rises and falls synchronously with the sensing module, always maintaining the optimal communication height to avoid obstruction and interference, and improve communication stability. It adopts a dual-channel redundant transmission structure of LoRa wireless and leaky wired transmission, which can effectively resist dust, water vapor and electromagnetic interference in pig farms, and solve the problems of packet loss, delay and disconnection in traditional single-channel transmission. The protocol conversion gateway realizes standardized data encapsulation, has strong adaptability and can quickly connect to various environmental control host systems.

[0012] A smart pig farm track-based inspection robot features a chassis as its core support structure. A lifting communication device and a lifting multi-parameter environmental sensing module are fixedly mounted on top of the chassis, integrating sensing, communication, and transmission functions. A hinged upper cover is mounted at the front of the chassis, and a hinged lower cover is mounted at the rear, facilitating the inspection, maintenance, and cleaning of internal components. A large-capacity battery is integrated within the chassis, providing continuous and stable power to all electrical components, enabling unmanned autonomous inspection operations. Symmetrically mounted track drive mechanisms on the left and right sides of the chassis enable smooth movement and inspection along the tracks, suitable for installation on pig farm rooftops. The machine features high integration and a compact layout, enabling integrated operation of sensing, communication, and mobile inspection. The bidirectional openable flip-top structure significantly reduces the difficulty of equipment maintenance, facilitating daily dust removal, repair, and parts replacement. Powered by a built-in independent battery, it eliminates the need for external cables, freeing inspection operations from wiring constraints and achieving a high degree of automation. The dual-track drive structure provides strong walking stability, making it suitable for the complex track operation environment of pig farms.

[0013] To achieve propulsion, the robot's drive mechanism includes support rollers, chain rollers, and guide rollers symmetrically mounted on the left and right outer sides of the chassis. These three types of wheels are located in the same vertical plane, and a transmission track is fitted around their outer sides, forming a closed-loop drive structure. Each link of the track has a limiting groove at its bottom that matches the track structure. During movement, the limiting groove engages with the track sidewall, achieving lateral limitation and preventing deviation. A lower servo motor is fixedly installed on the inner side of the chassis. The output shaft of the lower servo motor is fixedly connected to the chain roller, driving the chain roller to rotate, which in turn drives the entire track, enabling the robot to move forward, backward, and stop at designated points along the track. Utilizing a multi-wheel cooperative transmission and track limiting structure, the robot's movement is free from deviation and derailment risks, ensuring high positioning accuracy and guaranteeing the accuracy of data from point-to-point inspections. The servo motor drive offers fast response and precise start-stop, enabling uniform speed inspection and point-to-point monitoring throughout the track. The track-type transmission structure provides excellent shock absorption, effectively reducing the impact of walking vibrations on the accuracy of the sensor module's data acquisition.

[0014] For precise positioning, an RFID reader is fixedly installed on the bottom of the chassis. Multiple RFID landmark tags are deployed at equal intervals or as needed along the pigsty inspection track. Each RFID landmark tag corresponds to a unique location code, accurately identifying different zones within the pigsty. During the robot's inspection along the track, the RFID reader reads the location information of the landmark tags in real time, achieving real-time and accurate positioning of the entire machine. This provides precise location coordinates for subsequent environmental data binding and zoned control. RFID landmark positioning achieves centimeter-level precise location identification, solving the problems of vague positioning and inaccurate correspondence to regional environmental parameters in traditional inspection equipment. It provides core location support for zoned environmental data collection, zoned deviation calculation, and precise zoned ventilation control, ensuring the accuracy of zoned management.

[0015] An inspection method for a smart pig farm track-based inspection robot includes four core steps, forming a complete closed-loop management process: S1. Integrated perception and edge preprocessing: During the robot's inspection along the track, the integrated multi-parameter sensing module simultaneously completes environmental parameter acquisition and RFID location positioning. At the same time, it relies on the built-in MCU to complete edge preprocessing of the raw data, eliminating interference, compensating for errors, and compressing data volume. S2. Location data binding and encapsulation: The main control unit of the equipment will encapsulate the pre-processed gas, temperature and image feature data, combined with real-time timestamps and RFID location codes, in a unified manner according to a custom environmental data frame format to achieve a three-in-one binding of "data-location-time". S3, Dual-channel redundant transmission and interaction: The encapsulated data frame is uploaded to the environmental control host through the dual-channel redundant link of the lift-type communication device, and at the same time, it receives downlink commands such as transmission control, breakpoint resume transmission, and equipment control issued by the host in real time, realizing two-way real-time interaction; S4. Host Analysis and Closed-Loop Control: The environmental control host receives and analyzes data frames, restores accurate environmental data of each zone, constructs a two-dimensional / three-dimensional heat map of the pig house environment, and completes the control of ventilation equipment based on the zone-differentiated PID precise ventilation logic. At the same time, the control results are fed back to the robot. Through continuous inspection and monitoring, the control effect is verified and iteratively optimized, and a complete closed-loop management and control system is constructed. The above steps achieve fully automated operation from data collection, preprocessing, transmission, analysis and decision-making, equipment execution to feedback optimization, without the need for manual intervention; data and location are precisely bound, solving the drawback of traditional full-domain data lacking spatial attributes, and providing core support for precise regional control; two-way interaction and breakpoint resume transmission ensure the stability of system operation, and the closed-loop control mode continuously optimizes the accuracy of environmental management.

[0016] During the inspection and perception phase of S1, the equipment MCU has a built-in edge-side data preprocessing algorithm to optimize the raw sensor data in real time. It includes four core processing modules to solve the data distortion problems caused by sensor temperature drift, pulse interference, and lens dirt from the source, and significantly improve the accuracy of environmental data acquisition. Through feature extraction and compression technology, it greatly reduces the data transmission bandwidth usage and avoids data transmission delay, congestion, and packet loss. Edge local preprocessing reduces the computing pressure on the environmental control host and improves the overall system response speed.

[0017] I. Adaptive Temperature Drift Compensation for Gas Concentration: Addressing the temperature drift issue in electrochemical gas sensors caused by large diurnal temperature variations in pigsties, a built-in compensation model is implemented. Perform data correction. The original concentration reading at time t. For real-time ambient temperature, Temperature coefficient of ammonia sensor for calibrating reference temperature Temperature coefficient of carbon dioxide sensor Temperature compensation eliminates temperature drift errors, outputting accurate and effective gas concentration data.

[0018] II. Thermal Imaging Feature Vector Extraction and Processing: To reduce transmission bandwidth pressure, the transmission method of a complete 256×192 thermal imaging matrix (49KB / frame) was abandoned. Instead, five-dimensional feature vectors were extracted locally at the edge. These correspond to the highest temperature, lowest temperature, average temperature, temperature standard deviation, and number of hotspot pixels exceeding 39℃ in the field of view, respectively. The feature vector is only 10 bytes, with a compression rate of 99.98%, which greatly reduces transmission overhead.

[0019] III. Optical Window Contamination Assessment and Processing: An image sharpness evaluation model was constructed using the Tenengrad gradient method. ,in For pixel grayscale values, The response value of the Sobel operator is set to a threshold of 30% of the sharpness of a normal clear image. When S is less than the threshold, the window is considered dirty, triggering the self-cleaning process to ensure the accuracy of image acquisition.

[0020] IV. Outlier Removal via Sliding Window: A sampling sliding window of length 5 is set to acquire 5 consecutive sets of gas concentration sampling data, which are then processed using the Grubbs test. Identify outliers in impulse interference, among which The sample mean. The standard deviation of the sample is used. Under the conditions of significance level α=0.05 and n=5, outlier samples that exceed the critical value are removed, and the mean of the remaining samples is taken as the valid data for uploading, which effectively suppresses impulse interference.

[0021] In S4, the environmental control host executes zone-differentiated PID closed-loop ventilation control based on precise environmental data with location tags uploaded by the robot. Combined with multiple constraints and self-correction mechanisms, it achieves refined ventilation management, overturning the traditional one-size-fits-all global ventilation model and enabling differentiated and precise ventilation based on region and livestock species. Multiple constraint mechanisms effectively mitigate environmental risks and reduce equipment energy consumption. The parameter self-correction closed-loop logic continuously improves control accuracy, adapting to complex and ever-changing pigsty environments. The intelligence and adaptability of environmental management are significantly enhanced. The specific process is as follows: I. Pigsty Space Zoning: Based on RFID track tags, the pigsty is divided into N independent control areas along the track. Each zone corresponds to an independent side wall fan, ventilation window, and wet curtain actuator, enabling independent and controllable hardware zones; II. Phased Threshold Configuration: Differentiated environmental thresholds are set for different pig breeds. For fattening pigs, the target temperature is 22±2℃, the maximum NH3 concentration is 10ppm, and the maximum CO2 concentration is 1500ppm; for pregnant sows, the target temperature is 20±2℃, the maximum NH3 concentration is 8ppm, and the maximum CO2 concentration is 1300ppm, to meet the breeding needs of different growth stages. III. Calculation of Zoning Deviation: Using the formula Calculate the temperature deviation between zones; use the formula Calculate the overall gas deviation; IV. Partitioned PID Calculation Output: Output via partitioned independent PID formula Calculate the 0-100% opening output of the ventilation equipment in each zone; V. Anti-one-size-fits-all constraint mechanism: Set up triple control constraints, only activate dedicated equipment for areas exceeding the standard, disable the whole-area ventilation equipment when a single area seriously exceeds the standard, and ensure that the difference in ventilation opening between adjacent areas does not exceed 30%, so as to prevent excessive ventilation, cross ventilation, and regional interference. VI. Closed-loop feedback self-correction: Five minutes after the control command is issued, the partition environmental data is re-collected and verified. If the target is not met, the proportional coefficient Kp,i is automatically increased to iteratively optimize the control parameters and achieve adaptive closed-loop optimization.

[0022] The beneficial effects of this invention are: 1. Comprehensive monitoring with no blind spots, significantly improving data acquisition accuracy: Adopting a track-mounted mobile inspection and a multi-dimensional sensing structure with adjustable vertical lifting, it breaks through the limitations of traditional fixed sensors in terms of location and blind spots, realizing environmental monitoring of the entire horizontal and vertical space of the pig house; At the same time, through edge-side temperature drift compensation, outlier removal, and turbidity self-detection algorithms, it eliminates data errors caused by temperature drift, pulse interference, and lens obstruction from the source, significantly improving the accuracy, authenticity, and reliability of environmental data acquisition, providing a solid data foundation for precise control; 2. Lightweight transmission and redundant communication for higher system stability: Through thermal imaging feature extraction and compression technology, the size of a single frame of thermal imaging data is compressed from 49KB to 10 bytes, with a compression rate of 99.98%, greatly reducing transmission bandwidth pressure and eliminating data congestion and latency issues; adopting a LoRa+leaky cable dual-channel redundant transmission architecture, it effectively resists dust, moisture, and electromagnetic interference from pig farms, solving the defects of traditional single-channel transmission that are prone to packet loss and disconnection, and ensuring stable and reliable data transmission throughout the entire process; 3. Differentiated zone control to solve the problem of one-size-fits-all approaches in the industry: This solution is based on RFID location zoning technology to achieve modular zoned control of pig house space. It configures exclusive environmental thresholds for different growth stages of fattening pigs and pregnant sows. Each zone has independent PID parameter tuning and independent equipment control, which can accurately solve the problems of local high temperature and high concentration of harmful gases. It avoids the problems of excessive ventilation in normal areas, pig stress and energy waste caused by traditional whole-area ventilation. The adaptability and accuracy of environmental control are greatly improved. 4. Multiple constraints and closed-loop self-optimization ensure high safety and intelligence in regulation: Through a triple constraint mechanism of zone-specific regulation, global device disabling, and adjacent opening balance, problems such as cross ventilation, regional interference, and ineffective ventilation are effectively avoided, ensuring a comfortable and safe pig farming environment; Relying on the closed-loop logic of "perception-execution-feedback-correction", the system can adaptively optimize PID parameters according to the actual regulation effect without manual debugging, and has the ability to learn and optimize itself, adapting to the complex and ever-changing dynamic environment of the pig house; 5. High equipment versatility and low operation and maintenance costs: The lifting stroke of the sensor module is adjustable and the modular integrated design of the equipment can be adapted to the monitoring and control needs of different pig houses and different breeding stages; the two-way openable flip-top structure of the robot facilitates daily operation and cleaning, the tracked limit walking structure is stable in operation and has a low failure rate, and the whole machine has a long service life, which greatly reduces the operation and maintenance costs of intelligent management and control of pig farms. 6. Full-process automated management and control, significantly reducing labor costs: It realizes unmanned autonomous operation of the entire process, including environmental perception, data processing, transmission and interaction, intelligent decision-making, equipment execution, and feedback optimization. It eliminates the need for manual fixed-point monitoring and manual control, greatly saving the cost of manual inspection and environmental management, and adapting to the production and operation needs of large-scale, intensive smart pig farms. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A three-dimensional structural diagram of a liftable multi-parameter environmental sensing module and communication device; Figure 2 This is a three-dimensional structural diagram of the lifting mechanism at the first angle; Figure 3 This is a three-dimensional structural diagram of the lifting mechanism at the second angle; Figure 4 This is a diagram of the internal structure of the lifting mechanism; Figure 5 A three-dimensional structural diagram of the smart pig farm track inspection robot from the first angle; Figure 6 A three-dimensional structural diagram of the smart pig farm track inspection robot from a second angle; Figure 7 This is a three-dimensional structural diagram of the main chassis section; Figure 8 This is a diagram showing the internal structure of the main chassis section; Figure 9 for Figure 7 A magnified view of a section at point A in the middle; Figure 10 This is a front view of the main body of the chassis; Figure 11 This is a flowchart illustrating the process of a smart pig farm track inspection robot performing inspections within the pig farm.

[0025] Reference numerals: 1. Module housing; 2. Gas collection unit; 3. Vision and thermal imaging unit; 4. Lower tripod; 5. Lifting mechanism; 6. Base column; 7. Lifting column; 8. Guide rail; 9. Slide; 10. Upper servo motor; 11. Winch; 12. Fixed pulley; 13. Moving pulley; 14. Steel wire rope; 15. Height adjustment hole; 16. Mobile terminal communication box; 17. Upper tripod; 18. Track; 19. Chassis body; 20. Upper chassis cover; 21. Lower chassis cover; 22. Battery; 23. Track drive mechanism; 24. Track roller; 25. Carrier roller; 26. Guide wheel; 27. Track; 28. Limiting groove; 29. ​​Lower servo motor; 30. RFID reader. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0027] Example 1, as Figure 1 As shown, a lifting multi-parameter environmental sensing module is provided, mainly composed of a module housing 1, a gas acquisition unit 2, and a vision and thermal imaging unit 3. Its working principle can be found in existing literature. The gas acquisition unit 2 and the vision and thermal imaging unit 3 are all integrated and installed inside the module housing 1, respectively used to collect gas concentration parameters such as ammonia and carbon dioxide in the pigsty, environmental image data, and thermal imaging temperature data, achieving simultaneous acquisition of multi-dimensional environmental parameters. At least two sets of lower tripods 4 are fixedly mounted at the bottom of the module housing 1, and the entire module is fixedly installed on a lifting mechanism 5 via the lower tripods 4. Driven by the lifting mechanism 5, the vertical lifting and lowering of the entire module housing 1 can be achieved. This allows for flexible adjustment of the sensing module's monitoring height according to the pigsty height, pig growth height, and airflow monitoring point requirements, adapting to the monitoring needs of different farming scenarios. Through its liftable structural design, it achieves full-area vertical space monitoring coverage, effectively eliminating blind spots in vertical environmental monitoring of pigsties; it integrates multi-dimensional acquisition units for gas, vision, and thermal imaging, which can simultaneously acquire environmental physicochemical parameters and image temperature characteristics, providing comprehensive data acquisition dimensions and complete data support for subsequent precise environmental analysis and ventilation control.

[0028] like Figures 2 to 4As shown, to achieve stable lifting, the lifting mechanism 5 adopts a double-column lifting adapter structure, including two sets of vertically fixed base columns 6 and two sets of corresponding lifting columns 7 arranged in front of the base columns 6. The module housing 1 is fixedly installed on the front side of the lifting columns 7 via a lower tripod 4, and completes the vertical lifting action synchronously with the lifting columns 7. Each set of base columns 6 and corresponding lifting columns 7 are equipped with a guide rail 8 and a slide block 9. The guide rail 8 is vertically fixed to the side of the base column 6, and the slide block 9 is fixed to the inside of the lifting column 7. The guide rail 8 and the slide block 9 slide and adapt to each other, which guides and limits the up and down movement of the lifting column 7, ensuring that the lifting process is smooth and without deviation. A servo motor 10 is fixedly installed on the rear side of the base column 6. The output shaft of the servo motor 10 is coaxially connected to two sets of wire rope 14 winding and unwinding winches 11, which can realize the winding and unwinding actions of the two sets of wire rope 14 simultaneously. Two sets of fixed pulleys 12 are rotatably installed at the top of the base column 6, and two sets of movable pulleys 13 are rotatably installed at the bottom of the lifting column 7. The steel wire rope 14 wound around the winch 11 passes around the fixed pulleys 12 in sequence, and its end is fixedly connected to the movable pulley 13. The winch 11 is driven to rotate by a servo motor, which drives the steel wire rope 14 to wind up and down, thereby pulling the movable pulley 13 and the lifting column 7 to slide up and down as a whole, realizing the precise lifting and lowering adjustment of the sensing module. The double-rail guide and pulley group steel wire rope 14 transmission structure ensures smooth lifting and high positioning accuracy, effectively avoiding shaking and deviation during the lifting and lowering of the sensing module, and ensuring the stability of the collected data. The dual-set synchronous lifting structure is adapted to the overall weight of the sensing module, with strong load-bearing capacity, low operating noise, and fast response speed. The servo motor is precisely controllable and can realize arbitrary height positioning adjustment, adapting to the monitoring needs of multiple scenarios.

[0029] like Figure 4 As shown, to achieve position adjustment, multiple sets of height adjustment holes 15 are linearly arrayed along the height direction on the side of the base column 6. The mounting shaft of the fixed pulley 12 can be detachably installed inside any set of height adjustment holes 15. The installation height of the fixed pulley 12 can be flexibly adjusted according to the actual lifting stroke requirements, thereby adapting to different lifting adjustment ranges. By changing the mounting position of the fixed pulley 12, it can match the operational requirements of different pig house heights and different monitoring heights, and complete the adaptive adjustment of the lifting stroke. It realizes modular adjustable lifting stroke, which can adapt to the monitoring operations of different pig houses without replacing the entire mechanism, greatly improving the equipment's versatility and adaptability; the linear array hole positioning is accurate, the adjustment operation is simple and convenient, and the cost of later debugging, modification, and upgrading of the equipment is low.

[0030] Example 2 is based on Example 1, with some modifications, such as... Figure 1 As shown, a liftable communication device is provided, comprising a mobile communication box 16 (such as...). Figure 1The system consists of two parts: a mobile communication box 16 and a fixed communication base station (not shown in the diagram). The mobile communication box 16 integrates a LoRa RF module, a leaky cable coupler, and a protocol conversion gateway; its working principle can be found in existing literature. The fixed communication base station integrates a leaky cable terminal, an industrial Ethernet interface, and a leaky cable; its working principle can also be found in existing literature. This constructs a redundant wireless + wired transmission link. The mobile communication box 16 is fixedly installed on the front of the lifting column 7 via at least two sets of upper tripods 17, positioned above the sensor module housing 1. It can rise and fall synchronously with the lifting mechanism, adaptively adjusting the communication height according to the monitoring point. The fixed communication base station is fixedly installed on the pigsty environmental control host, achieving stable connection for data reception, parsing, and command issuance. The protocol conversion gateway can perform protocol standardization conversion of the data collected by the equipment, adapting to the environmental control host's communication protocol. The LoRa RF module and the leaky cable work together to achieve long-distance, interference-resistant data transmission. The communication device rises and falls synchronously with the sensing module, always maintaining the optimal communication height to avoid obstruction and interference, and improve communication stability. It adopts a dual-channel redundant transmission structure of LoRa wireless and leaky wired transmission, which can effectively resist dust, water vapor and electromagnetic interference in pig farms, and solve the problems of packet loss, delay and disconnection in traditional single-channel transmission. The protocol conversion gateway realizes standardized data encapsulation, has strong adaptability and can quickly connect to various environmental control host systems.

[0031] Example 3 is based on Example 2, with some modifications, such as... Figures 5 to 9 As shown, a smart pig farm track-based inspection robot (18) is provided. The core supporting structure is a chassis body 19. The aforementioned lifting communication device and lifting multi-parameter environmental sensing module are fixedly installed on top of the chassis body 19, integrating sensing, communication, and transmission functions. A hinged upper chassis cover 20 is hinged to the front of the chassis body 19, allowing it to be opened upwards, and a hinged lower chassis cover 21 is hinged to the rear, allowing it to be opened downwards. This double-cover opening structure facilitates the inspection, maintenance, and dust removal of the internal components. A large-capacity battery 22 is integrated inside the chassis body 19, providing continuous and stable power to all electrical components, enabling unmanned autonomous inspection operations. Tracks 27 and drive mechanisms 23 are symmetrically mounted on the left and right sides of the chassis body 19, enabling smooth movement and inspection of the entire machine via the track 18, adapting to the installation scenario of the track 18 on the top of the pigsty. The machine features high integration and a compact layout, enabling integrated operation of sensing, communication, and mobile inspection. The bidirectional openable flip-top structure significantly reduces the difficulty of equipment maintenance, facilitating daily dust removal, repair, and parts replacement. Powered by a built-in independent battery, it eliminates the need for external cables, freeing inspection operations from wiring constraints and achieving a high degree of automation. The dual-sided track drive structure provides strong walking stability and is suitable for complex track environments in pig farms.

[0032] like Figure 10As shown, to achieve propulsion, the robot's drive mechanism includes support rollers 24, chain rollers 25, and guide rollers 26 symmetrically mounted on the left and right outer sides of the chassis body 19. These three types of wheels are in the same vertical plane, and a transmission track 27 is fitted around their outer sides, forming a closed-loop drive structure. Each link of the track 27 has a limiting groove 28 at its bottom that matches the structure of the track 18. During movement, the limiting groove 28 engages with the sidewall of the track 18, achieving lateral limiting and preventing deviation. A lower servo motor 29 is fixedly installed on the inner side of the chassis body 19. The output shaft of the lower servo motor 29 is fixedly connected to the chain roller 25. The servo motor drives the chain roller 25 to rotate, which in turn drives the entire track 27, thereby enabling the robot to move forward, backward, and stop at designated points along the track 18. Employing a multi-wheeled collaborative transmission and a track 27 limiting structure, its movement process is free from deviation and derailment risks, with high positioning accuracy, ensuring the accuracy of data for fixed-point inspection; the servo motor drive has a fast response speed and precise start and stop, enabling uniform speed inspection and fixed-point monitoring along the entire track 18; the track 27 transmission structure has good shock absorption effect, effectively reducing the impact of walking vibration on the data acquisition accuracy of the sensor module.

[0033] like Figure 8 As shown, for precise positioning, an RFID reader 30 is fixedly installed on the bottom surface of the chassis body 19. Multiple RFID landmark tags are deployed at equal intervals or as needed along the pigsty inspection track 18. Each RFID landmark tag corresponds to a unique location code, accurately corresponding to different zones within the pigsty. During the robot's inspection along the track 18, the RFID reader 30 reads the location information of the landmark tags in real time, achieving real-time precise positioning of the entire machine. This provides accurate location coordinates for subsequent binding of environmental data and location information, and for zoned control. RFID landmark positioning achieves centimeter-level precise location identification, solving the problems of vague positioning and inaccurate correspondence to regional environmental parameters in traditional inspection equipment. It provides core location support for zoned environmental data collection, zoned deviation calculation, and precise zoned ventilation control, ensuring the accuracy of zoned management.

[0034] Example 4, as Figure 11 As shown, an inspection method for a smart pig farm track inspection robot is provided. This embodiment adopts the third embodiment and includes four core steps to form a complete closed-loop control process.

[0035] S1. Intensive perception and edge preprocessing: During the robot's inspection along the track, the integrated multi-parameter sensing module simultaneously completes environmental parameter acquisition and RFID location positioning. At the same time, it relies on the built-in MCU to complete edge preprocessing of the raw data, eliminating interference, compensating for errors, and compressing data volume.

[0036] S2. Location data binding and encapsulation: The main control unit of the device will encapsulate the pre-processed gas, temperature and image feature data, combined with real-time timestamps and RFID location codes, in a unified manner according to a custom environmental data frame format to achieve a three-in-one binding of "data-location-time".

[0037] S3, Dual-channel redundant transmission and interaction: The encapsulated data frame is uploaded to the environmental control host through the dual-channel redundant link of the lift-type communication device. At the same time, it receives downlink commands from the host in real time, such as transmission control, breakpoint resume transmission, and equipment control, to achieve two-way real-time interaction.

[0038] S4. Host Analysis and Closed-Loop Control: The environmental control host receives and analyzes data frames, restores accurate environmental data for each zone, constructs a two-dimensional / three-dimensional heat map of the pig house environment, and completes the control of ventilation equipment based on the zone-differentiated PID precise ventilation logic. At the same time, the control results are fed back to the robot. Through continuous inspection and monitoring, the control effect is verified and iteratively optimized, and a complete closed-loop management and control system is constructed.

[0039] The above steps achieve fully automated operation from data collection, preprocessing, transmission, analysis and decision-making, equipment execution to feedback optimization, without the need for manual intervention; data and location are precisely bound, solving the drawback of traditional full-domain data lacking spatial attributes, and providing core support for precise regional control; two-way interaction and breakpoint resume transmission ensure the stability of system operation, and the closed-loop control mode continuously optimizes the accuracy of environmental management.

[0040] Example 5 differs from Example 4 in that, during the inspection and sensing phase of S1, the device MCU incorporates an edge-side data preprocessing algorithm to optimize the raw sensor data in real time. This algorithm includes four core processing modules, addressing data distortion caused by sensor temperature drift, pulse interference, and lens contamination at the source, thereby significantly improving the accuracy of environmental data acquisition. Feature extraction and compression technology greatly reduces data transmission bandwidth usage, avoiding data transmission delays, congestion, and packet loss. Edge-based local preprocessing reduces the computational burden on the environmental control host, improving the overall system response speed.

[0041] I. Adaptive Temperature Drift Compensation for Gas Concentration: Addressing the temperature drift issue in electrochemical gas sensors caused by large diurnal temperature variations in pigsties, a built-in compensation model is implemented. Perform data correction. The original concentration reading at time t. For real-time ambient temperature, Temperature coefficient of ammonia sensor for calibrating reference temperature Temperature coefficient of carbon dioxide sensor Temperature compensation eliminates temperature drift errors, outputting accurate and effective gas concentration data.

[0042] II. Thermal Imaging Feature Vector Extraction and Processing: To reduce transmission bandwidth pressure, the transmission method of a complete 256×192 thermal imaging matrix (49KB / frame) was abandoned. Instead, five-dimensional feature vectors were extracted locally at the edge. These correspond to the highest temperature, lowest temperature, average temperature, temperature standard deviation, and number of hotspot pixels exceeding 39℃ in the field of view, respectively. The feature vector is only 10 bytes, with a compression rate of 99.98%, which greatly reduces transmission overhead.

[0043] III. Optical Window Contamination Assessment and Processing: An image sharpness evaluation model was constructed using the Tenengrad gradient method. ,in For pixel grayscale values, The response value of the Sobel operator is set to a threshold of 30% of the sharpness of a normal clear image. When S is less than the threshold, the window is considered dirty, triggering the self-cleaning process to ensure the accuracy of image acquisition.

[0044] IV. Outlier Removal via Sliding Window: A sampling sliding window of length 5 is set to acquire 5 consecutive sets of gas concentration sampling data, which are then processed using the Grubbs test. Identify outliers in impulse interference, among which The sample mean. The standard deviation of the sample is used. Under the conditions of significance level α=0.05 and n=5, outlier samples that exceed the critical value are removed, and the mean of the remaining samples is taken as the valid data for uploading, which effectively suppresses impulse interference.

[0045] Example 6 differs from Example 4 in that, in S4, the environmental control host executes zone-differentiated PID closed-loop ventilation control based on precise environmental data with location tags uploaded by the robot. Combined with multiple constraints and a self-correction mechanism, it achieves refined ventilation management. This overturns the traditional one-size-fits-all global ventilation mode, realizing differentiated and precise ventilation by region and by livestock species. The multiple constraint mechanism effectively avoids environmental risks and reduces equipment energy consumption. The parameter self-correction closed-loop logic continuously improves control accuracy, adapting to complex and ever-changing pigsty environments. The intelligence and adaptability of environmental management are significantly enhanced. The specific process is as follows: I. Pigsty Space Zoning: Based on RFID track tags, the pigsty is divided into N independent control areas along the track. Each zone corresponds to an independent side wall fan, ventilation window, and wet curtain actuator, enabling independent and controllable hardware zones; II. Phased Threshold Configuration: Differentiated environmental thresholds are set for different pig breeds. For fattening pigs, the target temperature is 22±2℃, the maximum NH3 concentration is 10ppm, and the maximum CO2 concentration is 1500ppm; for pregnant sows, the target temperature is 20±2℃, the maximum NH3 concentration is 8ppm, and the maximum CO2 concentration is 1300ppm, to meet the breeding needs of different growth stages. III. Calculation of Zoning Deviation: Using the formula Calculate the temperature deviation between zones; use the formula Calculate the overall gas deviation; IV. Partitioned PID Calculation Output: Output via partitioned independent PID formula Calculate the 0-100% opening output of the ventilation equipment in each zone; V. Anti-one-size-fits-all constraint mechanism: Set up triple control constraints, only activate dedicated equipment for areas exceeding the standard, disable the whole-area ventilation equipment when a single area seriously exceeds the standard, and ensure that the difference in ventilation opening between adjacent areas does not exceed 30%, so as to prevent excessive ventilation, cross ventilation, and regional interference. VI. Closed-loop feedback self-correction: Five minutes after the control command is issued, the partition environmental data is re-collected and verified. If the target is not met, the proportional coefficient Kp,i is automatically increased to iteratively optimize the control parameters and achieve adaptive closed-loop optimization.

[0046] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A liftable multi-parameter environmental sensing module, comprising a module housing and a gas sampling unit and a visual and thermal imaging unit located inside the module housing; characterized in that: The module housing is mounted on the lifting mechanism via at least two sets of lower tripods, and the entire module housing can be raised and lowered vertically with the lifting mechanism.

2. The lifting multi-parameter environmental sensing module according to claim 1, characterized in that: The lifting mechanism includes two sets of vertically arranged base columns and two sets of lifting columns located in front of the base columns. The module housing is mounted on the front side of the lifting columns via a lower tripod. A guide rail and slide block for assisting the lifting columns to slide up and down are installed between each set of base columns and lifting columns. An upper servo motor is provided behind the base columns. Two sets of winches for winding and releasing steel wire ropes are coaxially connected to the output shaft of the upper servo motor. Two sets of fixed pulleys and two sets of movable pulleys are rotatably installed at the top of the base columns and the bottom of the lifting columns. The steel wire rope located in the winches passes around the fixed pulleys and is fixedly connected to the movable pulleys.

3. The lifting multi-parameter environmental sensing module according to claim 2, characterized in that: Height adjustment holes are provided through the side of the base column, arranged in a linear array along its height direction, with the pulley shaft containing the fixed pulley located inside the height adjustment holes.

4. A liftable communication device, comprising a mobile communication box and a fixed communication base station, wherein the mobile communication box includes a LoRa radio frequency module, a leaky cable coupler, and a protocol conversion gateway; and the fixed communication base station includes a leaky cable terminal, an industrial Ethernet interface, and a leaky cable; characterized in that: The mobile communication box is mounted on the front side of the lifting column of the lifting multi-parameter environmental sensing module according to claim 2 via at least two sets of upper tripods, wherein the mobile terminal is located above the module housing; the fixed communication base station is mounted on the environmental control host.

5. A smart pig farm track inspection robot, comprising a chassis body, characterized in that: A lifting communication device according to claim 4 is installed above the chassis body. A chassis upper cover that can be opened by flipping up is hinged to the front of the chassis body. A chassis lower cover that can be opened by flipping down is hinged to the rear of the chassis body. A storage battery that supplies power to all electrical components is installed inside the chassis body. Track drive mechanisms that move along the track are installed on the left and right sides of the chassis body.

6. The intelligent pig farm track inspection robot according to claim 5, characterized in that: The drive mechanism includes a support roller, a chain roller, and a guide roller rotatably mounted on the outer side of the chassis body and located in the same vertical plane. A track is connected to the support roller, chain roller, and guide roller for transmission. A limiting groove matching the track is opened at the bottom of the track link. A lower servo motor for driving the chain roller to rotate is installed on the inner side of the chassis body.

7. The intelligent pig farm track inspection robot according to claim 6, characterized in that: An RFID reader is also installed on the bottom of the chassis body, and multiple RFID landmark tags are laid on the track for the RFID reader to read location information.

8. An inspection method for a smart pig farm track-based inspection robot, characterized in that: This method uses the smart pig farm track inspection robot according to claim 7 to inspect the pig farm. The working steps are as follows: S1. Intensive perception and edge preprocessing: During the inspection of the track, the track inspection robot synchronously collects the gas, temperature and thermal imaging image data of the pig house environment through the multi-parameter environmental sensing module. At the same time, it obtains the location information in real time through the RFID card reader and performs edge preprocessing on the raw sensing data based on the built-in MCU. S2. Location and Data Binding Protocol Encapsulation: The main control unit encapsulates the pre-processed valid environmental data and image feature data, combined with the real-time timestamp and RFID location code, according to the custom environmental data frame format, to achieve a three-in-one binding of data, location, and time. S3, Dual-channel redundant transmission and two-way interaction: Through the LoRa wireless and leaky wired dual-channel links of the lift-type communication device, the encapsulated data frame is uploaded to the environmental control host, and at the same time, the device control, breakpoint resume, and parameter correction downlink commands issued by the environmental control host are received in real time. S4. Host Analysis and Closed-Loop Precise Control: The environmental control host receives and analyzes data frames, restores the precise environmental parameters of each zone, and constructs a heat map of the pig house environment. Based on the zoned environmental data with location tags, the system executes zoned differentiated and precise ventilation control logic, and verifies the control effect and self-corrects parameters based on subsequent inspection feedback data, forming a closed-loop intelligent control process of "perception-preprocessing-transmission-decision-execution-feedback correction".

9. The inspection method of the smart pig farm track inspection robot according to claim 8, characterized in that: The edge preprocessing in step S1 includes gas concentration adaptive temperature drift compensation processing, thermal imaging feature vector extraction processing, optical window turbidity assessment processing, and sliding window outlier removal processing deployed inside the MCU. I. Adaptive Temperature Drift Compensation for Gas Concentration: This is achieved through a temperature drift compensation model. Temperature compensation is performed on the raw sensor readings, where The value represents the initial gas concentration reading at time t. Real-time ambient temperature, reference temperature ammonia sensor temperature coefficient Temperature coefficient of carbon dioxide sensor The compensated concentration serves as effective decision-making data; II. Thermal Imaging Feature Vector Extraction and Processing: Discarding the transmission of the complete thermal imaging matrix, five-dimensional feature vectors are extracted from the edge side: Only feature vector data is uploaded, achieving a data compression rate of 99.98%; III. Optical Window Smudge Assessment and Processing: A sharpness evaluation model is constructed using the Tenengrad gradient method. Based on the comparison between the image sharpness index S and the preset threshold, the smudge status of the window is determined and the self-cleaning process is triggered. IV. Outlier Removal via Sliding Window: A sampling sequence with a window length of 5 is used. Impulse interference outliers are identified and removed using the Grubbs test. The mean of the remaining samples is used as the final valid uploaded data.

10. The inspection method of the smart pig farm track inspection robot according to claim 8, characterized in that: In step S4, the zone-specific precise ventilation control logic is deployed within the environmental control unit, specifically including: I. Pigsty Space Zoning: Based on RFID landmark tags, the pigsty is divided into N independent ventilation control zones along the track. Each zone has its own independent side wall fan, ventilation window, and wet curtain actuator; II. Threshold configuration for different growth stages: For fattening pigs, the target temperature is set at 22±2℃, the maximum NH3 concentration is 10ppm, and the maximum CO2 concentration is 1500ppm; for pregnant sows, the target temperature is set at 20±2℃, the maximum NH3 concentration is 8ppm, and the maximum CO2 concentration is 1300ppm. III. Calculation of Zoning Deviation: Using the formula Calculate the temperature deviation between zones; use the formula Calculate the overall gas deviation; IV. Zone-Independent PID Control: This is achieved through the zone-independent PID formula. Calculate the 0-100% opening output of the ventilation equipment in each zone; V. Zonal Control Constraints: If a single zone malfunctions, only the equipment in that zone will be activated; if the deviation of a single zone is much greater than that of other zones, the ventilation equipment for the entire area will be disabled; the difference in the opening degree of ventilation windows between adjacent zones shall not exceed 30%. VI. Closed-loop feedback self-correction: After the control command is issued, the environmental data of the partition is re-collected 300 seconds later. If the deviation is not reduced to less than 50% of the original deviation, the PID proportional coefficient of the corresponding partition is automatically increased to iteratively optimize the control accuracy.

Images