Intelligent control method and system of livestock transport vehicle-mounted disinfection and killing system

Abstract

The application discloses an intelligent control method and system of a livestock transport vehicle-mounted disinfection and sterilization system, and belongs to the technical field of vehicle hygiene control. The method comprises the following steps: constructing a dynamic disturbance observation matrix reflecting the overall pollution degree of a carriage according to real-time environmental parameters and livestock distribution density data in the carriage; performing time domain decomposition on the dynamic disturbance observation matrix to extract a low-frequency principal component signal representing the attachment trend of pollutants and a high-frequency residual signal representing instantaneous disturbance; calculating an initial disinfection and sterilization dose adjustment rate according to the two signals through an integral sliding mode surface function; combining the initial disinfection and sterilization dose adjustment rate with the actual output control quantity of the last control period to generate a smooth control instruction; and converting the smooth control instruction into drive signals of the flow valve opening degree and the atomizer frequency to adjust the spraying device. Through dynamic decoupling and smooth control of the pollution state, precise adjustment of the disinfection and sterilization dose is realized.

Description

Technical Field

[0001] This invention relates to the field of vehicle hygiene control technology, specifically to an intelligent control method and system for an onboard disinfection system for livestock transport vehicles. Background Technology

[0002] During long-distance transport of livestock, the interior of the vehicle compartment is affected by livestock excrement, body temperature, and respiration, creating a complex dynamic pollution field with varying environmental parameters such as ammonia concentration, temperature, and humidity. Existing vehicle-mounted disinfection systems mostly employ fixed-cycle or single-threshold triggering spraying strategies. These methods fail to detect spatial distribution differences and temporal accumulation trends of pollutants. Different areas within the vehicle compartment have varying livestock densities, resulting in significant differences in localized pollution levels. Uniform spraying dosages can easily lead to insufficient disinfection in high-density areas and excessive pesticide residues in low-density areas, affecting disease control effectiveness, potentially causing stress to livestock, and increasing operating costs. Existing technologies include a scheme that adjusts the disinfection dosage based on the average feedback from a single ammonia sensor. However, this approach treats the entire vehicle compartment as a homogeneous space, ignoring the non-uniform distribution of pollutants and the structural differences between trend components and random disturbance components in dynamic changes. When the vehicle travels on uneven surfaces, instantaneous ammonia pulses caused by livestock activity directly couple into the sensor signal, resulting in severe and high-frequency dose fluctuations in the control system. Frequent mechanical adjustments not only shorten the lifespan of flow valves and atomizers in the spray system, but also cause excessive spraying of pesticides in localized areas after disturbance peaks. Furthermore, some dosage calculation methods based on simple error feedback perform poorly in suppressing the inherent complex disturbances in vehicle-mounted scenarios, failing to simultaneously address both the need to compensate for slow adhesion processes and the sensitive suppression of instantaneous disturbances. Ultimately, this results in disinfection actions lagging behind the actual evolution of the pollution situation.

[0003] The problem this application aims to solve is how to effectively separate the low-frequency characteristics reflecting the long-term adhesion trend of pollutants and the instantaneous characteristics reflecting short-term random disturbances from the spatially non-uniform and transiently polluted carriage environment data, and to form a disinfection dosage control mechanism based on this to both smooth and regulate fluctuations and accurately track the pollution situation. Summary of the Invention

[0004] To address the shortcomings of existing vehicle-mounted disinfection systems, which cannot distinguish between pollution trends and instantaneous disturbances, resulting in frequent fluctuations in dosage adjustment and inaccurate disinfection, the present invention aims to provide an intelligent control method that can decouple the main components of pollutant adhesion and random disturbance components from multi-dimensional environmental data of the vehicle compartment, and generate smooth and accurate disinfection control commands based on the decoupling results, so as to achieve coordinated adaptive adjustment of the spray device flow rate and atomization degree.

[0005] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides an intelligent control method for a livestock transport vehicle-mounted disinfection system, comprising: constructing a dynamic disturbance observation matrix reflecting the overall pollution level of the current transport vehicle based on real-time environmental parameters inside the vehicle and livestock distribution density data; performing time-domain decomposition on the dynamic disturbance observation matrix to extract low-frequency principal component signals characterizing pollutant adhesion trends and high-frequency residual signals characterizing instantaneous disturbances; calculating the initial disinfection dose adjustment rate of the current control cycle using an integral sliding mode surface function based on the low-frequency principal component signals and the high-frequency residual signals; generating a smooth control command for the current control cycle based on the initial disinfection dose adjustment rate of the current control cycle and the actual output control quantity of the previous control cycle; and converting the smooth control command into a drive signal for an actuator to adjust the flow valve opening and atomizer frequency of the spray device.

[0006] This method constructs a dynamic disturbance observation matrix by unifying multi-source environmental information and livestock spatial distribution within the vehicle compartment. This matrix comprehensively depicts the diffusion and accumulation of pollutants, avoiding the incomplete reflection of overall vehicle pollution levels by single sensors or single-point detection. The dynamic disturbance observation matrix is ​​decomposed in the time domain, separating the pollution evolution process into low-frequency principal component signals and high-frequency residual signals. The low-frequency principal component signals correspond to the adhesion and accumulation trends of pollutants on the interior walls and floor of the vehicle compartment, while the high-frequency residual signals reflect instantaneous pollution disturbances caused by livestock activity, ventilation, and other factors. This allows the disinfection control to both track long-term pollution trends and promptly detect sudden pollution events. The disinfection dosage adjustment rate is calculated by fusing low-frequency change rates and high-frequency cumulative disturbances using an integral sliding mode surface function. This ensures that the control quantity responds quickly to changes in pollution trends and instantaneous disturbances. Furthermore, the inherent robustness of sliding mode control effectively suppresses parameter uncertainties and external interference, guaranteeing accurate application of the disinfection dosage. When generating control commands, the actual output control quantity of the previous control cycle is used to smooth the current calculated value. When the change in control quantity exceeds a preset gradual change threshold, it is reduced to avoid drastic jumps in the flow valve and atomizer, thereby reducing mechanical shock and wear on the actuators, improving system reliability and service life, and ensuring a smooth transition in the disinfection process, reducing waste and excessive local spraying. The smoothing control command is divided into a flow valve opening adjustment component and an atomizer frequency adjustment component according to the allocation coefficient. These components adjust the flow valve opening and atomizer vibration frequency of the spray device respectively, achieving a coordinated match between droplet size and spray volume. This allows the atomized disinfectant to adaptively diffuse according to the distribution of pollution and airflow in the vehicle compartment, enhancing the penetration of the disinfectant in complex vehicle spaces and improving the disinfection coverage of blind spots and vehicle walls.

[0007] As a technical solution of the present invention, the step of constructing a dynamic disturbance observation matrix reflecting the overall pollution level of the carriage based on real-time environmental parameters and livestock distribution density data inside the carriage includes: collecting environmental parameters and livestock distribution density data at multiple sampling points inside the carriage, wherein the environmental parameters include temperature, humidity, and ammonia concentration; normalizing the environmental parameters at each sampling point to obtain a normalized value for each sampling point; multiplying the livestock distribution density data at each sampling point by the corresponding normalized value for the environmental parameters to obtain the local pollution contribution of each sampling point; and arranging the local pollution contribution values ​​of all sampling points according to spatial coordinates to generate the dynamic disturbance observation matrix. Preferably, the sampling points are evenly distributed according to a spatial grid inside the carriage, and the distance between adjacent sampling points does not exceed a preset sampling distance threshold. Therefore, a high spatial resolution pollution situation description can be obtained with low computational overhead, allowing the matrix elements to directly reflect the differences in pollution intensity caused by the coupling of livestock metabolism and environmental factors at each location, providing a high-quality data foundation for subsequent spatiotemporal feature separation.

[0008] As a further technical solution of the present invention, the dynamic disturbance observation matrix is ​​decomposed in the time domain to extract low-frequency principal component signals characterizing the pollutant adhesion trend and high-frequency residual signals characterizing instantaneous disturbances. This includes: performing singular value decomposition on the dynamic disturbance observation matrix to obtain a left singular vector matrix, a singular value diagonal matrix, and a right singular vector matrix; extracting multiple principal singular values ​​greater than a preset singular value threshold from the singular value diagonal matrix, and truncating column vectors corresponding to the multiple principal singular values ​​in the left and right singular vector matrices to construct a reduced-dimensional principal component space; projecting the dynamic disturbance observation matrix onto the reduced-dimensional principal component space to obtain the low-frequency principal component signals; and performing matrix subtraction between the dynamic disturbance observation matrix and the low-frequency principal component signals to obtain the high-frequency residual signals. Preferably, the preset singular value threshold is dynamically determined based on the weighted sum of the mean and standard deviation of all singular values ​​in the singular value diagonal matrix. By adaptively extracting the energy-dominant principal components from pollution data through singular value decomposition to construct low-frequency signals, the dominant modes of global pollution propagation and attachment behavior can be preserved, while the residual signals retain local high-frequency disturbances caused by short-term livestock movement, door opening and closing impacts, etc., thus realizing multi-scale analysis of pollution dynamics.

[0009] As another technical solution of the present invention, the initial disinfection dose adjustment rate of the current control cycle is calculated by means of an integral sliding mode surface function based on the low-frequency principal component signal and the high-frequency residual signal, including: performing differentiation on the low-frequency principal component signal to obtain a low-frequency rate of change vector; performing absolute value integration on the high-frequency residual signal to obtain a high-frequency cumulative disturbance scalar; calculating the norm value of the low-frequency rate of change vector, and performing a weighted summation of the norm value and the high-frequency cumulative disturbance scalar to obtain the integral sliding mode surface value of the current control cycle; substituting the integral sliding mode surface value into the exponential reaching law function to calculate the initial disinfection dose adjustment rate of the current control cycle. Preferably, the norm value is the Euclidean norm of the low-frequency rate of change vector, and the weighting coefficients in the weighted summation are pre-calibrated according to the volume of the carriage and the amount of livestock loaded. The control rate calculation based on the integral sliding surface incorporates the evolution trend of low-frequency principal components and the cumulative effect of high-frequency disturbances into the unified sliding surface design, so that the state of the disinfection system converges to the neighborhood of the sliding surface within a finite time. When the degree of pollution increases, the disinfection intensity is rapidly increased, and when the pollution subsides, the dosage is steadily reduced to avoid overdose.

[0010] As a preferred embodiment of the present invention, generating a smooth control command for the current control cycle includes: acquiring the actual output control quantity of the previous control cycle and calculating the difference between the initial disinfection dose adjustment rate and the actual output control quantity of the previous control cycle; determining the control quantity change amplitude for the current control cycle based on the difference; when the control quantity change amplitude is greater than a preset gradual change threshold, reducing the initial disinfection dose adjustment rate according to a preset proportional coefficient to obtain the adjusted control quantity for the current control cycle; when the control quantity change amplitude is less than or equal to the preset gradual change threshold, determining the initial disinfection dose adjustment rate as the adjusted control quantity for the current control cycle; and limiting the amplitude of the adjusted control quantity to generate a smooth control command for the current control cycle. This smoothing mechanism, while ensuring responsiveness to pollution dynamics, filters out high-frequency fluctuations in the control quantity, making the spraying device's operation continuous and gradual, effectively reducing pipeline pressure fluctuations and valve impacts, and extending the maintenance-free cycle of the vehicle-mounted spraying equipment.

[0011] As a more preferred embodiment of the present invention, converting the smoothing control command into a drive signal for the actuator to adjust the flow valve opening and atomizer frequency of the spray device includes: splitting the smoothing control command according to a preset first allocation coefficient to obtain a flow valve opening adjustment component and an atomizer frequency adjustment component; generating a flow valve opening adjustment signal based on the flow valve opening adjustment component, the flow valve opening adjustment signal being used to control the opening angle of the flow valve; generating an atomizer frequency adjustment signal based on the atomizer frequency adjustment component, the atomizer frequency adjustment signal being used to control the vibration frequency of the atomizer; and superimposing the flow valve opening adjustment signal and the atomizer frequency adjustment signal and outputting them to the actuator. The coordinated adjustment of flow rate and atomization frequency can automatically adjust the droplet size and spray width according to real-time operating conditions. When the pollution is severe and densely distributed, the flow rate is increased and matched with a higher atomization frequency to obtain fine, high-speed droplets that can penetrate the livestock's hair and cover the gaps in the carriage structure. When the pollution is light or concentrated in a localized area, the flow rate is reduced and matched with an appropriate frequency to form larger droplets that can settle regionally onto the required surface, balancing the effectiveness of disinfection and the economy of the disinfectant.

[0012] This invention also provides an intelligent control system for a livestock transport vehicle-mounted disinfection system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the intelligent control method for the livestock transport vehicle-mounted disinfection system. This system integrates the aforementioned intelligent control method into the vehicle controller, continuously collecting information on the vehicle's environment and livestock distribution during transport, generating real-time disinfection dosage commands, and driving the flow valve and atomizer to coordinate their actions. This ensures that the disinfection operation requires no manual intervention throughout, and the control output dynamically matches the actual pollution level in the vehicle, significantly reducing water and pesticide consumption while achieving efficient epidemic prevention and disinfection.

[0013] The technical effects and advantages provided by the present invention in the above technical solution are as follows: A scheme is adopted to fuse environmental parameters and livestock distribution density data from multiple sampling points inside the carriage to generate the local pollution contribution of each spatial location and arrange them into a dynamic perturbation observation matrix. This matrix structure simultaneously includes coupled information of temperature, humidity, ammonia concentration, and livestock density at a single time cross-section, transforming the originally isolated sensor readings into a spatially correlated expression of pollution status. When the livestock density in a certain area of ​​the carriage increases and ventilation is poor, the corresponding element value in the matrix will be amplified synchronously. The control system thus obtains a direct basis reflecting the spatial heterogeneity of pollution, avoiding local omissions and dose mismatches caused by taking the average value of the entire carriage. A scheme is adopted to construct a dimensionality-reduced principal component space by performing singular value decomposition on the dynamic perturbation observation matrix and extracting multiple principal singular values ​​exceeding the dynamic threshold from the singular value diagonal matrix. The low-frequency principal component signal obtained by projecting the original matrix onto this space retains the slow changing trend of pollution adhesion and accumulation on the carriage surface. Its information is concentrated on a few dominant modes, stripping away the scattered data fluctuations caused by instantaneous livestock activity or vehicle bumps. The high-frequency residual signal obtained through matrix subtraction carries the amplitude information of these short-term random disturbances. This decomposition allows subsequent dose calculations to apply differentiated dynamic processing to signal components with different characteristics. Differentiating the low-frequency principal component yields the low-frequency rate of change vector, and integrating the high-frequency residual by absolute value yields the high-frequency cumulative disturbance scalar. The weighted sum of the Euclidean norm of the low-frequency rate of change vector and the high-frequency cumulative disturbance scalar constitutes the integral sliding mode surface value. The norm value dominates the basic dose output required to eliminate the contaminant adhesion trend, while the cumulative disturbance scalar provides the damping compensation against random shocks. Through the structured synthesis of the sliding mode surface and the mapping of the exponential reaching law, the initial disinfection dose adjustment rate for the current control cycle is calculated. This adjustment rate can synchronously track the evolution gradient of the low-frequency adhesion trend in terms of response speed, and will not overshoot isolated high-frequency disturbance peaks in terms of steady-state quality, achieving precise component-level countermeasures against the spatiotemporal dynamics of contaminants.

[0014] A scheme is adopted to determine whether the change in control quantity exceeds a preset gradual change threshold based on the difference between the initial disinfection dosage adjustment rate and the actual output control quantity of the previous control cycle. When the change exceeds the limit, the adjustment rate is reduced proportionally, so that the jump amplitude of valve position and frequency is constrained within a smooth range, eliminating water hammer in the spray pipeline and atomizer resonance instability caused by sudden changes in control quantity. The smooth control command generated after amplitude limiting is split into flow valve opening component and atomizer frequency component according to the allocation coefficient, which drive the two actuators to work together, so that the disinfection operation can maintain the stability of spray particle size distribution and coverage range while meeting the dosage requirements, avoiding the waste of agent dripping caused by high flow rate and low atomization and the drift loss caused by low flow rate and excessive atomization. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a flowchart of the intelligent control method for the livestock transport vehicle-mounted disinfection system; Figure 2 This is a flowchart of the dynamic disturbance observation matrix generation process; Figure 3 This is a flowchart of singular value decomposition and principal component extraction of the dynamic perturbation observation matrix; Figure 4 This is a flowchart of the integral sliding mode surface value calculation and the generation of the initial disinfection dose adjustment rate; Figure 5 This is a flowchart of the method for smoothing the adjustment rate of disinfection dosage; Figure 6 It is a flowchart of smooth control instruction splitting and actuator drive signal generation. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0018] See Figure 1 This invention provides an intelligent control method for a livestock transport vehicle-mounted disinfection system, comprising: constructing a dynamic disturbance observation matrix reflecting the overall pollution level of the current transport vehicle based on real-time environmental parameters inside the vehicle and livestock distribution density data; performing time-domain decomposition on the dynamic disturbance observation matrix to extract low-frequency principal component signals characterizing pollutant adhesion trends and high-frequency residual signals characterizing instantaneous disturbances; calculating the initial disinfection dose adjustment rate for the current control cycle using an integral sliding mode surface function based on the low-frequency principal component signals and the high-frequency residual signals; generating a smooth control command for the current control cycle based on the initial disinfection dose adjustment rate for the current control cycle and the actual output control quantity of the previous control cycle; and converting the smooth control command into a drive signal for the actuator to adjust the flow valve opening and atomizer frequency of the spray device.

[0019] Example 1: In specific implementation, please refer to Figure 2The interior of the vehicle is equipped with multiple temperature sensors, humidity sensors, ammonia concentration sensors, and image acquisition devices or infrared thermal imaging devices for acquiring livestock distribution density data. The interior space is pre-divided into a uniform spatial grid. The spatial distance between adjacent grid nodes is the spacing between adjacent sampling points, and this spacing does not exceed a preset sampling spacing threshold. The preset sampling spacing threshold is determined based on the actual length, width, and height dimensions of the vehicle and the spray disinfection coverage radius. In this embodiment, the preset sampling spacing threshold is set to 0.5 meters to achieve a balance between computational resource consumption and the accuracy of pollution distribution characterization. A sampling point is set at each grid node, and all sampling points are uniformly numbered according to the row and column order of the spatial grid, forming a sampling point array. During each control cycle, the processor reads the temperature, humidity, ammonia concentration, and livestock distribution density data at each sampling point location.

[0020] For each sampling point, the collected temperature, humidity, and ammonia concentration values ​​need to be normalized to eliminate the influence of dimensional differences in different environmental parameters on subsequent calculations. Linear normalization is used, and the specific calculation formula is as follows: in, The index is the sequence number of the sampling point. Take from 1 to the total number of sampling points. ; For environment parameter type indexing, The value can be 1, 2, or 3. Corresponding temperature value, Corresponding humidity value, Corresponding ammonia concentration value; Indicates the first The first sampling point collected the first The original values ​​of class environment parameters; This indicates the number of samples collected from all sampling points within the current control cycle. The minimum value among the class environment parameter values; This indicates the number of samples collected from all sampling points within the current control cycle. The maximum value among the class environment parameter values; Indicates the first The sampling point of the first sampling point The normalized values ​​of the environmental parameters, after the above linear normalization process, are... The range of values ​​is mapped to Interval.

[0021] After obtaining the number Normalized value of temperature corresponding to each sampling point Humidity value normalized value Normalized value of ammonia concentration After that, the first The livestock distribution density data at each sampling point are denoted as follows: Livestock distribution density data Indicates at the sampling point The number of livestock individuals detected within the corresponding unit spatial grid, or the percentage of thermal radiation-weighted area. This represents the livestock distribution density data. Multiplying each of these values ​​by the normalized values ​​of the three environmental parameters yields three pollution contribution components: Contribution of temperature pollution Humidity pollution contribution Contribution of ammonia pollution .

[0022] Next, the three pollution contribution components mentioned above are added together to obtain the first... Local pollution contribution of each sampling point ,Right now Contribution of local pollution This comprehensively reflects the situation at the sampling points. The relative degree of pollution caused by the combined effects of livestock activities and environmental conditions.

[0023] After obtaining the local pollution contribution of all sampling points, the row direction of the internal spatial grid of the carriage is used as the first dimension of the matrix, and the column direction is used as the second dimension. The local pollution contribution of each sampling point is then filled into the matrix according to its corresponding spatial coordinates to generate a dynamic disturbance observation matrix. Each element of the dynamic disturbance observation matrix corresponds to the pollution contribution at a specific spatial location within the carriage at the current moment. The spatial distribution of this matrix reflects the overall spatial distribution characteristics of pollution in the carriage in real time.

[0024] Example 2: In specific implementation, please refer to Figure 3 After the processor obtains the dynamic disturbance observation matrix generated in the previous processing stage, it performs matrix singular value decomposition on the dynamic disturbance observation matrix. Let the dynamic disturbance observation matrix be... Its dimensions are ,in, This indicates the number of sampling points in the row direction of the grid inside the carriage. Indicates the number of sampling points in the column direction. For After performing singular value decomposition, three decomposition results are obtained: the left singular vector matrix. Singular value diagonal matrix and right singular vector matrix ,satisfy Left singular vector matrix It is An orthogonal matrix whose each column is a left singular vector; a singular value diagonal matrix. It is A diagonal matrix whose diagonal elements are arranged in descending order. Called the dynamic disturbance observation matrix singular values, for rank; right singular vector matrix It is An orthogonal matrix, in which each column is a right singular vector.

[0025] In obtaining the singular value diagonal matrix After identifying all singular values, a preset singular value threshold needs to be determined. This is to enable singular value truncation. A preset singular value threshold is set. It is not a fixed value, but rather based on the diagonal matrix of singular values ​​in the current control cycle. The statistical characteristics of all singular values ​​are dynamically calculated. A preset singular value threshold is calculated. The formula is: in, Represents a singular valued diagonal matrix The arithmetic mean of all non-zero singular values ​​in the set. ; Represents a singular valued diagonal matrix The standard deviation of all non-zero singular values ​​in the , ; and These are weighting coefficients, all of which are positive constants. The range of values ​​is to Weighting coefficient The range of values ​​is to In this embodiment, after analyzing and calibrating the measured environmental parameter data of the carriage under different seasons and loading densities, the... Set as ,Will Set as This makes the preset singularity threshold It can effectively filter out minute disturbance components caused by sensor noise or instantaneous local activities of livestock while retaining the main components that reflect the overall distribution trend of pollution.

[0026] Determine the preset singularity threshold Then, from the singular value diagonal matrix Extract all that meet the conditions The singular values, assuming there are a total of The singular values ​​satisfy the condition, this The singular values ​​are labeled as major singular values, and are denoted in order as follows: Corresponding to this One principal singular value, from the left singular vector matrix Before the middle cut The column vectors form a dimensionless left singular vector matrix. , The dimension is From the right singular vector matrix Before the middle cut The column vectors form a dimensionless right singular vector matrix. , The dimension is .Will The principal singular values ​​are arranged in their original order to form a principal singular value diagonal matrix. , The dimension is From the reduced-dimensional left singular vector matrix Principal singular value diagonal matrix And the dimension-reduced right singular vector matrix The linear space spanned by both is the reduced principal component space.

[0027] After constructing the dimensionality-reduced principal component space, the dynamically perturbed observation matrix is... Projecting onto the dimension-reduced principal component space, the projection operation is implemented through matrix multiplication to obtain the low-frequency principal component signal. ,Right now Low-frequency principal component signal It is a low-rank approximation of the original dynamic disturbance observation matrix, which preserves the energy-dominated spatiotemporal distribution pattern in the dynamic disturbance observation matrix, corresponding to the trend distribution characteristics of pollutants in the carriage due to long-term accumulation and adhesion.

[0028] After obtaining the low-frequency principal component signal Then, the dynamic disturbance observation matrix will be... With low-frequency principal component signals Subtracting corresponding elements yields the high-frequency residual signal. ,Right now High-frequency residual signal It represents the rapidly changing components of pollution distribution caused by short-term disturbances such as instantaneous excretion by livestock, vigorous activity, or local temperature drift of sensors. The absolute values ​​of most elements in its matrix are close to zero, and non-zero values ​​are only observed at the spatial locations where sudden disturbances occur.

[0029] Example 3: In specific implementation, please refer to Figure 4After acquiring the low-frequency principal component signal and the high-frequency residual signal, the processor performs different operations on the two types of signals respectively. The low-frequency principal component signal is a matrix with the same dimensions as the dynamic perturbation observation matrix, containing... lines and Each element in the matrix represents the value of the trend component of pollutants at the corresponding spatial grid location. Differentiation is performed on the low-frequency principal component signal along the time dimension. Specifically, the low-frequency principal component signal matrix acquired in the current control cycle is subtracted one by one from the corresponding spatial location elements in the low-frequency principal component signal matrix acquired in the previous control cycle. The result is then divided by the time interval between two adjacent control cycles to obtain the low-frequency rate of change vector. The low-frequency rate of change vector is a vector of length... The column vectors, each component of the low-frequency rate of change vector corresponds to the rate of change of the pollutant trend component over time at a spatial grid location.

[0030] After obtaining the low-frequency rate of change vector, its norm is calculated. The Euclidean norm is used, calculated by squaring each component of the low-frequency rate of change vector, summing the squares of all components, and then taking the square root of the sum. The result is the Euclidean norm of the low-frequency rate of change vector. The Euclidean norm of the low-frequency rate of change vector comprehensively reflects the overall drastic change of the trend components of pollutants in the entire cabin space during adjacent control cycles.

[0031] While processing the low-frequency principal component signal, the processor performs absolute value integration on the high-frequency residual signal. The high-frequency residual signal is also a... The matrix contains elements with both positive and negative values. The steps for absolute value integration are as follows: take the absolute value of each element in the high-frequency residual signal matrix, sum all the elements after taking their absolute values, and then multiply the sum by the spatial volume representative value of the sampling grid cell to obtain the high-frequency cumulative disturbance scalar. The high-frequency cumulative disturbance scalar is a non-negative scalar that represents the cumulative sum of non-trend disturbance components caused by sudden animal behavior or instantaneous sensor fluctuations in the entire carriage space within the current control cycle.

[0032] After calculating the Euclidean norm of the low-frequency rate of change vector and the high-frequency cumulative disturbance scalar, the processor performs a weighted summation to obtain the integral sliding mode surface value for the current control cycle. The weighting coefficients used in the weighted summation include a first weighting coefficient and a second weighting coefficient. The first weighting coefficient is multiplied by the Euclidean norm of the low-frequency rate of change vector, and the second weighting coefficient is multiplied by the high-frequency cumulative disturbance scalar; the sum of these two products is the integral sliding mode surface value. The values ​​of the first and second weighting coefficients are pre-calibrated based on the vehicle compartment volume and livestock loading capacity. The calibration method involves recording the control response curve when the disinfection effect meets the standard using real-vehicle disinfection experiment data under different combinations of vehicle compartment volume and livestock loading capacity. The values ​​of the first and second weighting coefficients are iteratively adjusted to match the sensitivity of the integral sliding mode surface value change with the response speed and overshoot suppression requirements of the control system. The calibration results are stored in the on-board controller memory in the form of a two-dimensional lookup table. During actual control cycle operation, the processor retrieves the corresponding first and second weighting coefficient values ​​from a two-dimensional lookup table based on the currently detected cargo volume parameters and the livestock loading volume reported by the loading sensor. In one implementation scenario, when the cargo volume is 25 cubic meters and the livestock loading volume is 50 head, the first weighting coefficient is calibrated to 0.7, the second weighting coefficient is calibrated to 0.3, and the sum of the two weighting coefficients is 1.0. In another implementation scenario, when the cargo volume increases to 40 cubic meters and the livestock loading volume decreases to 20 head, the dominant role of the pollutant trend components in the overall disinfection requirements becomes more prominent, the first weighting coefficient is adjusted upward to 0.85, and the second weighting coefficient is correspondingly calibrated to 0.15.

[0033] After calculating the integral sliding mode surface value, the processor substitutes it into the exponential reaching law function to calculate the initial disinfection dose adjustment rate for the current control cycle. The exponential reaching law function contains two adjustable parameters: the reaching law gain coefficient and the reaching law exponential coefficient. The reaching law gain coefficient determines the approach speed of the control quantity when the sliding mode surface value is large. The value of the reaching law gain coefficient is determined by analyzing the dynamic response time constant of the disinfection spray actuator and the power limitation conditions of the vehicle power supply. In this embodiment, the reaching law gain coefficient is set to 0.5. The reaching law exponential coefficient is used to smooth the change process of the control quantity when the sliding mode surface value is close to zero, avoiding actuator chatter caused by abrupt changes in the sign function. In this embodiment, the reaching law exponential coefficient is set to 1.2. The processor calculates the product of the integral sliding mode surface value and the reaching law exponential coefficient, then calculates the product of the negative of the sign function value and the reaching law gain coefficient. The two products are added together, and the resulting value is the initial disinfection dose adjustment rate for the current control cycle. The initial disinfection dose adjustment rate is a real value that can be positive or negative. A positive value indicates that the disinfection dose output needs to be increased, while a negative value indicates that the disinfection dose output needs to be decreased.

[0034] Example 4: In specific implementation, please refer to Figure 5 The processor obtains the actual output control quantity applied to the actuator in the previous control cycle, and simultaneously obtains the initial disinfection dose adjustment rate calculated for the current control cycle. The actual output control quantity of the previous control cycle is stored in the non-volatile memory of the vehicle controller. At the end of each control cycle, the processor writes the actual output control quantity value of the current cycle into the memory, and reads it as the actual output control quantity of the previous control cycle when the next control cycle starts.

[0035] The processor calculates the difference between the initial disinfection dose adjustment rate and the actual output control quantity of the previous control cycle. The difference is calculated by subtracting the actual output control quantity of the previous control cycle from the initial disinfection dose adjustment rate of the current control cycle. The resulting difference has a positive or negative sign: a positive sign indicates that the currently calculated disinfection demand tends to increase the dose output, and a negative sign indicates that the currently calculated disinfection demand tends to decrease the dose output.

[0036] The magnitude of the control quantity change in the current control cycle is determined based on the above difference. The magnitude of the control quantity change is the absolute value of the difference, representing the magnitude of the intended change in the disinfection dosage adjustment between two adjacent control cycles, without including information on the direction of change.

[0037] The processor compares the magnitude of the control quantity change with a preset gradual change threshold. The preset gradual change threshold is a constant pre-set based on the physical response characteristics of the actuator and the overall stability requirements of the disinfection system. There is a mechanical response time for the flow valve in the actuator to adjust from its current opening to the target opening, and a motor speed adjustment delay for the atomizer to switch from its current vibration frequency to the target vibration frequency. The preset gradual change threshold is set as the maximum control quantity change value that the actuator can respond smoothly to within one control cycle. Exceeding this maximum control quantity change value will cause flow valve opening overshoot, atomizer frequency oscillation, or pipeline pressure surge. In this embodiment, the preset gradual change threshold is set to 15% of the maximum disinfection dose output value.

[0038] When the change in control quantity exceeds a preset gradual change threshold, it is determined that the jump in control quantity between adjacent cycles exceeds the smooth response capability of the actuator. At this point, the processor reduces the initial disinfection dose adjustment rate according to a preset proportional coefficient to obtain the adjusted control quantity for the current control cycle. The preset proportional coefficient is a positive real number between 0 and 1. The preset proportional coefficient is set based on the following: during system debugging, the actual output following curves of the actuator under different control quantity step amplitudes are collected through step response testing. The reduction ratio corresponding to the largest step amplitude that causes output overshoot less than 5% without continuous oscillation is fixed as the preset proportional coefficient. In this embodiment, the preset proportional coefficient is 0.6. The adjusted control quantity is calculated by multiplying the difference between the initial disinfection dose adjustment rate and the actual output control quantity of the previous control cycle by the preset proportional coefficient, and then adding the result to the actual output control quantity of the previous control cycle. This method preserves the directional information of the control quantity change, compressing the change amplitude only to a range that the actuator can smoothly withstand.

[0039] When the change in the control quantity is less than or equal to the preset gradual change threshold, it is determined that the change in the control quantity in the adjacent cycle is within the smooth response capability of the actuator. The processor directly determines the initial disinfection dose adjustment rate as the adjusted control quantity for the current control cycle, that is, it does not reduce the initial disinfection dose adjustment rate.

[0040] After obtaining the adjusted control quantity, the processor performs a limiting process on it. The specific operation of the limiting process is as follows: the adjusted control quantity is compared with the pre-set upper and lower limits of the control quantity for the disinfection system. The upper limit corresponds to the disinfection dose control quantity required when the spray device's flow valve is fully open and the atomizer operates at its maximum rated frequency. The lower limit corresponds to the minimum disinfection dose control quantity required for the disinfection system to maintain its minimum standby state; this minimum disinfection dose control quantity is set to zero in this embodiment. When the adjusted control quantity is greater than the upper limit, the upper limit is used as the smoothing control command for the current control cycle; when the adjusted control quantity is less than the lower limit, the lower limit is used as the smoothing control command for the current control cycle; when the adjusted control quantity is between the upper and lower limits, it is directly used as the smoothing control command for the current control cycle. The smooth control commands generated by the limiting process can ensure that the control signals output to the actuator are always within the hardware safety limits, preventing mechanical damage or control instability of the flow valve or atomizer due to receiving commands that exceed its operating range.

[0041] Example 5: In specific implementation, please refer to Figure 6After generating the smooth control command for the current control cycle, the processor reads the preset first allocation coefficient from the parameter storage area of ​​the vehicle controller. The preset first allocation coefficient is a real number between 0 and 1, represented as... , The preset first allocation coefficient The value is pre-calibrated based on the contribution weight of the flow valve and atomizer in the unit disinfection dose output of the spraying device. The calibration method is as follows: under the condition that the carriage is fully loaded and the environmental parameters are stable, the total disinfection dose adjustment value is fixed as the calibration benchmark value, the flow valve opening and the atomizer frequency are adjusted respectively, the effective coverage area of ​​the disinfectant and the sedimentation velocity of the disinfectant particles at each sampling point in the carriage are measured, the sensitivity coefficients of the flow valve adjustment and the atomizer frequency adjustment on the disinfection effect are calculated, and the sensitivity coefficients of the flow valve sensitivity coefficient and the atomizer sensitivity coefficient are divided by the sum of the flow valve sensitivity coefficient and the atomizer sensitivity coefficient. The resulting ratio is the preset first allocation coefficient. In this embodiment, the flow valve sensitivity coefficient is calibrated to 0.65, and the atomizer sensitivity coefficient is calibrated to 0.35. Therefore, the preset first distribution coefficient... The value is 0.65.

[0042] The processor will smoothly assign control instructions according to a preset first allocation coefficient. The process involves splitting the data using multiplication. The value of the smooth control command is denoted as... The value of the smooth control command will be adjusted. With the preset first allocation coefficient Multiply them, and the product is the flow valve opening adjustment component, denoted as . ,Right now The value of the smooth control command will be adjusted. minus the preset first allocation coefficient from the value 1 Multiply the difference by the second distribution coefficient. Multiplying them together, the product is the atomizer frequency adjustment component, denoted as... ,Right now In this embodiment, if the smoothing control command value... If it is 100, then the flow valve opening adjustment component The atomizer frequency adjustment component is 65. It is 35.

[0043] After obtaining the flow valve opening adjustment component The processor then generates a flow valve opening adjustment signal based on the flow valve opening adjustment component. The numerical range of the flow valve opening adjustment component is mapped to the physical opening range of the flow valve in a linear mapping manner. In one embodiment, the physical opening range of the flow valve is 0 degrees to 90 degrees, where 0 degrees corresponds to a fully closed state and 90 degrees corresponds to a fully open state. The upper limit value of the flow valve opening adjustment component... The lower limit value corresponds to the fully open state of the flow valve. Corresponding to the fully closed state of the flow valve, the processor adjusts the flow valve opening by a component. Divide by the upper limit of the flow valve opening adjustment component The resulting ratio is then multiplied by the maximum physical opening of the flow valve (90 degrees) to obtain the target opening angle value. The processor converts the target opening angle value into a flow valve opening adjustment signal. This signal is a pulse width modulation (PWM) signal, and its duty cycle is linearly proportional to the target opening angle value. The flow valve opening adjustment signal is transmitted to the flow valve's drive motor through the analog output port of the vehicle controller, controlling the rotation angle of the valve core and thus achieving precise adjustment of the disinfectant flow rate.

[0044] After obtaining the atomizer frequency adjustment component The processor then generates an atomizer frequency adjustment signal based on the atomizer frequency adjustment component. The numerical range of the atomizer frequency adjustment component is mapped to the physical operating range of the atomizer's vibration frequency in a linear mapping manner. In one embodiment, the atomizer's vibration frequency operating range is 20 kHz to 50 kHz, with 20 kHz corresponding to the lowest atomization frequency and 50 kHz corresponding to the highest atomization frequency. The upper limit of the atomizer frequency adjustment component... The lower limit value corresponds to the highest vibration frequency of the atomizer. Corresponding to the atomizer's lowest vibration frequency. The processor adjusts the atomizer frequency component. Divide by the upper limit of the atomizer frequency adjustment component. The resulting ratio is then multiplied by the atomizer's operating frequency range of 30 kHz. The product is then added to the atomizer's minimum vibration frequency of 20 kHz to obtain the target vibration frequency. The processor converts this target vibration frequency into an atomizer frequency adjustment signal. This signal is a frequency-adjustable square wave signal, with its fundamental frequency equal to the target vibration frequency. The atomizer frequency adjustment signal is transmitted to the atomizer's piezoelectric drive circuit via the vehicle controller's digital output port, controlling the oscillation frequency of the atomizer's ceramic oscillator, thereby adjusting the particle size and spray rate of the disinfectant solution.

[0045] After generating the flow valve opening adjustment signal and the atomizer frequency adjustment signal, the processor superimposes these signals and outputs them to the actuator. This superimposed output means that the processor sends the flow valve opening adjustment signal and the atomizer frequency adjustment signal in parallel within the same control cycle, acting simultaneously on the flow valve drive motor and the atomizer piezoelectric drive circuit through different hardware output channels, rather than simply superimposing the two signals at the electrical level. The flow valve drive motor in the actuator receives the flow valve opening adjustment signal and performs angle adjustment, while the atomizer piezoelectric drive circuit in the actuator synchronously receives the atomizer frequency adjustment signal and performs frequency adjustment. These two mechanisms work together to achieve joint control of the disinfection dosage and atomization effect.

[0046] In one embodiment, the aforementioned intelligent control method is executed by an intelligent control system of a livestock transport vehicle-mounted disinfection system. The control system includes a memory, a processor, and a computer program stored in the memory and running on the processor. The memory is an on-board industrial-grade flash memory, and the processor is an on-board embedded microcontroller. When executing the computer program, the processor sequentially implements the following steps: constructing a dynamic disturbance observation matrix based on real-time environmental parameters inside the vehicle and livestock distribution density data; performing time-domain decomposition on the dynamic disturbance observation matrix to extract low-frequency principal component signals and high-frequency residual signals; calculating the initial disinfection dose adjustment rate using an integral sliding mode surface function; generating a smooth control command based on the initial disinfection dose adjustment rate and the actual output control quantity of the previous control cycle; and converting the smooth control command into actuator drive signals to adjust the flow valve opening and atomizer frequency.

[0047] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. An intelligent control method for a livestock transport vehicle-mounted disinfection system, characterized in that, include: Based on real-time environmental parameters inside the carriage and livestock distribution density data, a dynamic disturbance observation matrix reflecting the current overall pollution level of the carriage is constructed. The dynamic disturbance observation matrix is ​​decomposed in the time domain to extract the low-frequency principal component signal characterizing the pollutant adhesion trend and the high-frequency residual signal characterizing the instantaneous disturbance. Based on the low-frequency principal component signal and the high-frequency residual signal, the initial disinfection dose adjustment rate for the current control cycle is calculated using an integral sliding mode surface function; Based on the initial disinfection dose adjustment rate of the current control cycle and the actual output control quantity of the previous control cycle, a smooth control command for the current control cycle is generated. The smooth control command is converted into a drive signal for the actuator to adjust the opening of the flow valve and the frequency of the atomizer in the spray device.

2. The intelligent control method for a livestock transport vehicle-mounted disinfection system according to claim 1, characterized in that, The process involves constructing a dynamic disturbance observation matrix reflecting the overall pollution level of the carriage based on real-time environmental parameters and livestock distribution density data inside the carriage, including: Environmental parameters and livestock distribution density data were collected from multiple sampling points inside the carriage. The environmental parameters included temperature, humidity, and ammonia concentration. The environmental parameters at each sampling point are normalized to obtain the normalized values ​​of the environmental parameters at each sampling point. Multiply the livestock distribution density data of each sampling point by the corresponding normalized value of the environmental parameter to obtain the local pollution contribution of each sampling point; The local pollution contribution of all sampling points is arranged according to spatial coordinates to generate the dynamic disturbance observation matrix.

3. The intelligent control method for a livestock transport vehicle-mounted disinfection system according to claim 2, characterized in that, The dynamic disturbance observation matrix is ​​decomposed in the time domain to extract low-frequency principal component signals characterizing the pollutant adhesion trend and high-frequency residual signals characterizing instantaneous disturbances, including: Singular value decomposition is performed on the dynamic disturbance observation matrix to obtain a left singular vector matrix, a singular value diagonal matrix, and a right singular vector matrix; Extract multiple principal singular values ​​that are greater than a preset singular value threshold from the singular value diagonal matrix, and extract the column vectors corresponding to the multiple principal singular values ​​from the left singular vector matrix and the right singular vector matrix to construct a dimension-reduced principal component space; The dynamic perturbation observation matrix is ​​projected onto the reduced-dimensional principal component space to obtain the low-frequency principal component signal; The high-frequency residual signal is obtained by performing matrix subtraction between the dynamic disturbance observation matrix and the low-frequency principal component signal.

4. The intelligent control method for a livestock transport vehicle-mounted disinfection system according to claim 3, characterized in that, Based on the low-frequency principal component signal and the high-frequency residual signal, the initial disinfection dose adjustment rate for the current control cycle is calculated using an integral sliding mode surface function, including: The low-frequency principal component signal is differentiated to obtain the low-frequency rate of change vector; The high-frequency residual signal is integralized by absolute value to obtain the high-frequency cumulative disturbance scalar. Calculate the norm value of the low-frequency rate of change vector, and then perform a weighted summation of the norm value and the high-frequency cumulative disturbance scalar to obtain the integral sliding mode surface value of the current control cycle; Substituting the integral sliding mode surface value into the exponential reaching law function, the initial disinfection dose adjustment rate of the current control cycle is calculated.

5. The intelligent control method for a livestock transport vehicle-mounted disinfection system according to claim 4, characterized in that, The generation of smooth control instructions for the current control cycle includes: Obtain the actual output control quantity of the previous control cycle, and calculate the difference between the initial disinfection dose adjustment rate and the actual output control quantity of the previous control cycle; The range of change in the control quantity for the current control cycle is determined based on the difference. When the change in the control quantity is greater than the preset gradual change threshold, the initial disinfection dose adjustment rate is reduced according to the preset proportional coefficient to obtain the adjusted control quantity for the current control cycle. When the change in the control quantity is less than or equal to the preset slow change threshold, the initial disinfection dose adjustment rate is determined as the adjusted control quantity for the current control cycle. The adjusted control quantity is subjected to amplitude limiting processing to generate a smooth control command for the current control cycle.

6. The intelligent control method for a livestock transport vehicle-mounted disinfection system according to claim 5, characterized in that, Converting the smooth control command into a drive signal for the actuator to adjust the flow valve opening and atomizer frequency of the spray device includes: The smooth control command is split according to a preset first allocation coefficient to obtain the flow valve opening adjustment component and the atomizer frequency adjustment component; A flow valve opening adjustment signal is generated based on the flow valve opening adjustment component, and the flow valve opening adjustment signal is used to control the opening angle of the flow valve. Atomizer frequency adjustment signal is generated based on the atomizer frequency adjustment component, and the atomizer frequency adjustment signal is used to control the vibration frequency of the atomizer; The flow valve opening adjustment signal and the atomizer frequency adjustment signal are superimposed and output to the actuator.

7. The intelligent control method for a livestock transport vehicle-mounted disinfection system according to claim 2, characterized in that, The sampling points are evenly distributed according to the spatial grid inside the carriage, and the distance between adjacent sampling points does not exceed the preset sampling distance threshold.

8. The intelligent control method for a livestock transport vehicle-mounted disinfection system according to claim 3, characterized in that, The preset singular value threshold is dynamically determined based on the weighted sum of the mean and standard deviation of all singular values ​​in the singular value diagonal matrix.

9. The intelligent control method for a livestock transport vehicle-mounted disinfection system according to claim 4, characterized in that, The norm value is the Euclidean norm of the low-frequency rate of change vector, and the weighting coefficients in the weighted summation are pre-calibrated based on the volume of the carriage and the amount of livestock loaded.

10. An intelligent control system for a livestock transport vehicle-mounted disinfection system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent control method for the livestock transport vehicle disinfection system according to any one of claims 1 to 9.

Images