Research apparatus and usage method for simulating the emergence and dispersal of natural enemies of pine scale insects using natural airflow

Abstract

This invention discloses a device and method for studying the emergence and dispersal of natural enemies of the pine scale insect, simulating natural airflow, belonging to the field of insect behavior and biological control technology. The device includes an integrated flight channel, an array fan module, a multi-point wind speed sensor array, an emergence release module, an observation and recording module, and a control and calculation module. The integrated flight channel simulates natural airflow; the array fan module generates and regulates the average wind speed, pulsation, and spatial distribution within the integrated flight channel; the multi-point wind speed sensor array collects real-time wind speed data from multiple points within the integrated flight channel; and the observation and recording module collects the flight trajectories of emerging individuals. This device reproduces the multi-scale time-varying characteristics of field wind fields, significantly enhancing the extrapolation ability and ecological explanatory power of experimental results for field scenarios. It facilitates the precise quantification of the takeoff rate, takeoff delay, dispersal nucleus, and source-finding behavior of natural enemies of the pine scale insect, expanding the applicability of the experimental results.

Description

Technical Field

[0001] This invention relates to a device and method for studying the emergence and dispersal of natural enemies of the pine bark scale, simulating natural airflow, and belongs to the field of insect behavior and biological control technology. Background Technology

[0002] Studies on the emergence-flight-dissemination behavior of forest pests such as the pine scale and their natural enemies are of great significance for formulating biological control strategies, predicting the spread of natural enemies, and evaluating the effectiveness of releases in the forest. Currently, researchers often conduct experiments on insect flight and source-finding behavior in the laboratory using small wind tunnels or flight channels with fixed wind speeds; at the same time, they obtain ecological parameters through field observations or release-recapture of insects in the wild.

[0003] The aforementioned experimental methods can reveal basic behavioral characteristics of insects to some extent, but they have significant limitations when extrapolating the results to complex field environments. The main reason is that existing small wind tunnels and flight paths typically operate with constant wind speeds or simple periodic wind fields, failing to reproduce the multi-scale time-varying characteristics of field wind fields, including low-frequency wind direction oscillations, mid-to-high frequency pulsations, and the energy distribution of turbulence spectra. Therefore, behavioral data such as insect takeoff rate, takeoff delay, and source-seeking paths obtained using these devices often cannot reliably reflect the true responses of insects under field conditions, resulting in poor experimental extrapolation and limited applicability of the results. Based on this, this invention proposes a device and method for studying the emergence and diffusion of natural enemies of the pine scale insect, simulating natural airflow. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a device and method for studying the emergence and diffusion of natural enemies of the pine bark scale, simulating natural airflow.

[0005] This invention is achieved through the following technical solution: A device for studying the emergence and dispersal of pine scale predators by simulating natural airflow includes an integrated flight channel, an array fan module, a multi-point wind speed sensor array, an emergence release module, an observation and recording module, and a control and computing module. The integrated flight channel is used to simulate natural airflow. The array fan module, located on the integrated flight channel, is used to generate and regulate the average wind speed, pulsation, and spatial distribution within the integrated flight channel. The multi-point wind speed sensor array, located on the integrated flight channel, is used to collect multi-point wind speed data within the integrated flight channel in real time. The emergence release module, located within the integrated flight channel, is used to place the pupae of pine scale predators to emerge and to mark and release the emerged individuals formed after the pupae emerge. The observation and recording module, located on the integrated flight channel, is used to collect the flight trajectories of the emerged individuals. The control and computing module is electrically connected to the array fan module, the multi-point wind speed sensor array, the emergence release module, and the observation and recording module.

[0006] The integrated flight channel includes a flight section, a slow-flow section, and a controllable main fan. One end of the slow-flow section is connected to one end of the flight section and is in communication with it. An adjustable grid is provided at the end of the slow-flow section near the flight section. Both the end of the flight section away from the slow-flow section and the end of the slow-flow section away from the flight section are provided with honeycomb end plates. The controllable main fan is located at the end of the slow-flow section away from the flight section and is electrically connected to the control computing module. The flight section is equipped with multiple tail sampling units at equal intervals near the slow-flow section, and the tail sampling units are sticky insect cards or traps.

[0007] The array fan module is located on the side wall of the slow flow section and includes several fans that can be independently controlled by PWM and at least one variable speed main fan. Both the fans and the variable speed main fan are electrically connected to the control computing module. The multi-point wind speed sensor array is located on the top and side walls of the slow-flow section and is arranged along the axial and lateral directions of the integrated flight channel. The feathering release module is located on the bottom plate at one end of the flight section away from the slow flow section. It includes a cuboid box, a feathering detection unit, and a marking unit. An automatic door is provided on the side of the cuboid box near the slow flow section. The feathering detection unit and the marking unit are both located on the top of the cuboid box. The automatic door, the feathering detection unit, and the marking unit are all electrically connected to the control calculation module. The observation and recording module includes multiple high-speed cameras and multiple supplementary lights. The high-speed cameras are mounted on the top of the flight segment via brackets, and the lenses of the high-speed cameras extend into the flight segment. The multiple supplementary lights are all located on the top of the inner side of the flight segment. The high-speed cameras and supplementary lights are all electrically connected to the control and computing module.

[0008] The control calculation module includes: The array-field identification module applies a known excitation sequence to the array fan module and records the wind speed response at multiple points within the integrated flight channel. A linear approximation model is then identified using either the least squares method or recursive least squares method. , In the formula, express Wind speed vector at the measuring point at time [time]. Let be the self-state transfer matrix, representing the linear influence of the wind speed at the previous measurement point on the wind speed at the next measurement point. express Wind speed vector at the measuring point at time [time]. The array-field influence matrix represents the linear effect of the input of each fan on the wind speed at each measuring point. express The PWM input vector of the fan at any given time. for Time-of-flight noise term; The closed-loop control module is used to control the array fan module in real time based on the target wind spectrum and the array-field mapping identified by the array-field identification module, so that the wind speed time sequence measured in the integrated flight channel approximates the target wind spectrum. The data storage module is used for data storage.

[0009] The closed-loop control module includes: The state estimation module, which is a Kalman filter, is used to estimate the wind speed field state within the integrated flight channel based on data collected by a multi-point wind speed sensor array. The tracking controller module is for either a model predictive controller or a multi-channel adaptive PID controller. The phase wave generation unit sends a space-time sinusoidal traveling wave drive signal to the array fan module to generate controllable low-frequency wind direction oscillation and local vortices. The phase wave it generates is given by the formula... Specify; In the formula, Indicates the first A fan in time Control input; Indicates time The baseline bias value is used to maintain the overall average wind speed within the integrated flight path; It represents the time-varying amplitude envelope, used to control the intensity of the space-time sinusoidal traveling wave, so as to adjust the generation of different levels of pulsation or oscillation at different stages; The space wave number is used to control the wavelength λ=2π / k of the space-time sinusoidal traveling wave on the array, which determines the vortex width or swing width at the spatial scale. For the first The spatial coordinates of each fan in the array are used to calculate the phase difference; The time angular frequency corresponds to the time period T=2π / ω of the space-time sinusoidal traveling wave, and is used to control the low-frequency oscillation period. The initial phase is used to align or offset the starting position of the space-time sinusoidal traveling wave.

[0010] The target wind spectrum is a measured wind speed time series, or a target wind speed time series generated using a two-dimensional Ornstein–Uhlenbeck process or spectral synthesis method, wherein the Ornstein–Uhlenbeck process is given by the following stochastic differential equation: , In the formula, The vector represents the local wind speed over continuous time, where x is the mainstream component and y is the lateral component. As the reference velocity vector, ; Θ is the recovery matrix or decay matrix, used to control the autocorrelation time scale of the Ornstein–Uhlenbeck process; Σ is the noise intensity matrix, used to control the pulsation amplitude and the covariance of each component; This represents two-dimensional Brownian motion, driven by white noise. Then, the stochastic differential equation is discretized to obtain the sampling time step. The following reference sequence .

[0011] It also includes a volatile organic compound (VOC) emission monitoring module, which is used to simulate the odor concentration field. This module is located at the top inside the flight section and near the slow-flow section. It includes a VOC atomizer, a VOC sensor array, and an odor field numerical solution unit electrically connected to the control calculation module. The odor field numerical solution unit is based on the convection-diffusion equation. Numerical discretization is performed on the odor concentration field driven by a controlled wind speed field, and the measured values ​​from the VOC sensor array are incorporated into the joint cost function of the closed-loop control module for synchronous control and tracking of wind speed and odor plume morphology; the convection-diffusion equations in... Represents the odor concentration field. Indicates time, This is a localized wind speed field used to drive odor convection. Represents the Hamiltonian operator. Let be the diffusion coefficient, representing the diffusion rate of the odor concentration field. Represents the Laplace operator. The source term represents the release rate of volatiles.

[0012] A method for using a device simulating the emergence and dispersal of natural enemies of the pine scale includes the following steps: S1. Obtain or set the target wind spectrum: Use the measured wind speed time series, or the target wind speed time series generated by the two-dimensional Ornstein–Uhlenbeck process or spectrum synthesis method as the target wind spectrum; S2, Array-Field Identification: The array-field identification module applies a known excitation sequence to the array fan module and records the wind speed response at multiple points in the integrated flight channel. The least squares method or recursive least squares method is used to identify the linear approximate model. S3. Calibration and Initialization: Calibrate the sensors in the multi-point wind speed sensor array and the high-speed camera in the observation and recording module, synchronize the clock, and initialize the parameters of the Kalman filter and tracking controller module. S4. Emergence and Release Control: Place the pupa of the pine bark scale predator to be emerged into a rectangular box. When the emergence detection unit detects the emergence event, the marking unit marks the emerged individual, and then the automatic door opens to allow the emerged individual to enter the integrated flight channel. S5. Closed-loop operation: In the online loop, wind speed field data is collected in real time through a multi-point wind speed sensor array, odor concentration field data is collected in real time through a VOC sensor array, the wind speed field state of the integrated flight channel is estimated through the closed-loop control module, and the control input is calculated and sent to the array fan module based on the linear approximation model and the target wind spectrum. At the same time, the phase wave drives the array fan module and controls the volatile release monitoring module to regulate the odor concentration field. S6. Extract the flight trajectory of the molted individuals using a high-speed camera, and collect and recapture the molted individuals in segments within the flight segment using a tail sampling unit. Then, save all control sequences and sensor data for offline analysis.

[0013] The closed-loop operation in step S5 adopts chance-constrained model predictive control, which ensures that the control input and output constraints meet the probability constraints with a predetermined confidence level, taking into account process noise and turbulence uncertainty.

[0014] It also includes online estimation of turbulence intensity, adaptive adjustment of PID gain or MPC process noise covariance and weights based on the estimated turbulence intensity, and periodic updating of the array-field influence matrix using recursive least squares or online learning methods.

[0015] The beneficial effects of this invention are as follows: 1. It reproduces the multi-scale time-varying characteristics of field wind fields, including low-frequency wind direction oscillation, mid-to-high frequency pulsations, and energy distribution of turbulence spectrum. It significantly enhances the extrapolation ability and ecological explanatory power of experimental results for field scenarios, facilitates the accurate quantification of the take-off rate, take-off delay, diffusion nuclei, and source-finding behavior of natural enemies of pine scale, and provides reliable experimental basis for the design of biological control strategies and field release experiments, thus expanding the applicability of experimental results.

[0016] 2. By combining segmented recapture of molted individuals by multiple tail sampling units and the flight trajectories of molted individuals captured by multiple high-speed cameras, the diffusion distribution of molted individuals can be quantified into spatial grid data. This facilitates the construction of diffusion kernels, calculation of arrival rate and recapture rate, and thus combines visual trajectory data with physical recapture data, improving experimental statistical efficiency and supporting rigorous mark-release-recapture analysis.

[0017] 3. By establishing a linear approximation model, the complex wind field response can be simplified into a mathematical mapping that can be used for real-time control. This enables the model-based tracking controller module to calculate and issue fan control quantities in real time, thereby significantly improving the accuracy of target wind spectrum reproduction, reducing the number of manual parameter adjustments, and providing an estimable parameter basis for subsequent robust or adaptive control.

[0018] 4. Introducing phase wave drive can generate controllable moving vortices and low-frequency oscillations in space, which can better simulate the spatial structure of natural wind than the traditional method of uniformly adding or subtracting all fans; at the same time, by combining the state estimation of Kalman filter and the closed-loop strategy of MPC / PID, the wind speed field can form a realistic physical field shape, while ensuring numerical tracking accuracy and system stability.

[0019] 5. The Ornstein–Uhlenbeck process or spectral synthesis method can approximate the low-frequency oscillations and mid-to-high-frequency pulsations of natural wind on both time scale and spectrum. This makes the generated target wind spectrum physically interpretable and easy for the tracking controller module to track directly, thereby improving the reproducibility of the experimental scene to the characteristics of the field wind field and the extrapolation credibility of biological experiments.

[0020] 6. By incorporating odor transport into numerical modeling and closed-loop decision-making, the device can not only reproduce wind spectrum morphology, but also the spatial structure and time-varying nature of odor plumes, meeting the high simulation requirements for insect source-finding behavior and significantly enhancing the ecological relevance and explanatory power of experimental results. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the structure of the present invention; Figure 2 This is a partial cross-sectional view of the present invention after the control calculation module has been removed; Figure 3 This is a flowchart of the method of the present invention.

[0022] In the diagram: 1-Integrated flight channel, 11-Flight section, 12-Slow flow section, 13-Controllable main fan, 14-Adjustable grid, 15-Honeycomb end plate, 2-Array fan module, 3-Multi-point wind speed sensor array, 4-Volatile release monitoring module, 5-Pluming release module, 51-Rectangular box, 52-Automatic door, 6-Observation and recording module, 61-High-speed camera, 62-Support, 7-Control and calculation module, 8-Tail sampling unit. Detailed Implementation

[0023] The technical solution of the present invention is further described below, but the scope of protection is not limited to what is described.

[0024] like Figures 1 to 3As shown, the present invention discloses a research device for simulating the emergence and dispersal of pine scale predators using natural airflow, comprising an integrated flight channel 1, an array fan module 2, a multi-point wind speed sensor array 3, an emergence release module 5, an observation and recording module 6, and a control and calculation module 7. The integrated flight channel 1 is used to simulate natural airflow. The array fan module 2 is located on the integrated flight channel 1 and is used to generate and regulate the average wind speed, pulsation, and spatial distribution of the wind within the integrated flight channel 1. The multi-point wind speed sensor array 3 is located on the integrated flight channel 1 and is used to collect multi-point wind speed data within the integrated flight channel 1 in real time. The emergence release module 5 is located within the integrated flight channel 1 and is used to place the pupae of pine scale predators to emerge, and to mark and release the emerged individuals formed after the pupae emerge. The observation and recording module 6 is located on the integrated flight channel 1 and is used to collect the flight trajectory of the emerged individuals. The control and calculation module 7 is electrically connected to the array fan module 2, the multi-point wind speed sensor array 3, the emergence release module 5, and the observation and recording module 6.

[0025] Specifically, this invention simulates natural airflow through an integrated flight channel 1. An array fan module 2 generates and regulates the average wind speed, pulsation, and spatial distribution within the integrated flight channel 1. Simultaneously, a multi-point wind speed sensor array 3 collects real-time wind speed data from multiple points within the integrated flight channel 1. A known excitation sequence is applied to the array fan module 2 via a control calculation module 7, and the multi-point wind speed response within the integrated flight channel 1 is recorded. A linear approximation model is identified using the least squares method or recursive least squares method. Then, based on the target wind spectrum and the linear approximation model, the array fan module 2 is controlled in real-time, ensuring that the wind speed time series measured within the integrated flight channel 1 approximates the target wind spectrum. This reproduces the multi-scale time-varying characteristics of the field wind field, including low-frequency wind direction oscillations, mid-to-high-frequency pulsations, and the energy distribution of the turbulence spectrum. This significantly enhances the extrapolation capability and ecological explanatory power of the experimental results for field scenarios, facilitating the precise quantification of the takeoff rate, takeoff delay, diffusion nuclei, and source-finding behavior of the pine scale natural enemies. It also provides reliable experimental evidence for the design of biological control strategies and field release experiments, expanding the applicability of the experimental results.

[0026] The molting and release module 5 is used to place the pupae or nymphs of the natural enemies of the pine scale insects that are about to molt.

[0027] The integrated flight channel 1 includes a flight section 11, a slow-flow section 12, and a controllable main fan 13. One end of the slow-flow section 12 is connected to one end of the flight section 11 and communicates with it. An adjustable grid 14 is provided at the end of the slow-flow section 12 near the flight section 11. A honeycomb end plate 15 is provided at the end of the flight section 11 away from the slow-flow section 12 and at the end of the slow-flow section 12 away from the flight section 11. The controllable main fan 13 is located at the end of the slow-flow section 12 away from the flight section 11 and is electrically connected to the control calculation module 7. The flight section 11 is provided with multiple tail sampling units 8 at equal intervals near the slow flow section 12, and the tail sampling units 8 are sticky insect cards or traps.

[0028] Specifically, the tail sampling unit 8 is used to record the number of feathered individuals arriving or remaining at different distances or angles. The integrated flight channel 1 has a length of 2.0~6.0 m, a cross-sectional dimension of 0.4~0.8 m × 0.4~0.8 m, and an average wind speed within the channel ranging from 0 to 2.0 m / s. The sampling period is... The MPC prediction step size is 0.01~0.1s. The preferred number of steps is 10–40.

[0029] The array fan module 2 is located on the side wall of the slow flow section 12, and includes several fans that can be independently controlled by PWM and at least one variable speed main fan. Both the fans and the variable speed main fan are electrically connected to the control calculation module 7. The multi-point wind speed sensor array 3 is located on the top and side walls of the slow-flow section 12 and is arranged along the axial and lateral directions of the integrated flight channel 1. The feathering release module 5 is located on the bottom plate at one end of the flight section 11 away from the slow flow section 12. It includes a cuboid box 51, a feathering detection unit, and a marking unit. An automatic door 52 is provided on the side of the cuboid box 51 near the slow flow section 12. The feathering detection unit and the marking unit are both located on the top of the cuboid box 51. The automatic door 52, the feathering detection unit, and the marking unit are all electrically connected to the control calculation module 7. The observation and recording module 6 includes multiple high-speed cameras 61 and multiple supplementary lights. The multiple high-speed cameras 61 are respectively mounted on the top of the flight segment 11 via brackets 62, and the lenses of the high-speed cameras 61 extend into the flight segment 11. The multiple supplementary lights are all located on the top of the inner side of the flight segment 11. The high-speed cameras 61 and the supplementary lights are all electrically connected to the control and calculation module 7.

[0030] Specifically, the emergence detection unit is a machine vision system used to monitor whether the pupae of the pine scale insect's natural enemy have emerged. The marking unit is used to attach micro-tags to the emerged individuals; the micro-tags are barcodes or fluorescent dot arrays. The emergence detection unit, marking unit, and automatic door 52 work together to achieve a closed-loop linkage of emergence detection → marking → release.

[0031] By combining segmented recapture of 8 pairs of molted individuals from multiple tail sampling units and flight trajectories of molted individuals captured by multiple high-speed cameras 61, the diffusion distribution of molted individuals can be quantified into spatial grid data. This facilitates the construction of diffusion kernels and the calculation of arrival and recapture rates. By combining visual trajectory data with physical recapture data, the experimental statistical efficiency can be improved and rigorous analysis of mark-release-recapture can be supported, such as estimating ecological indicators like migration rate, survival rate, and arrival probability.

[0032] The control calculation module 7 includes: The array-field identification module is used to apply a known excitation sequence to the array fan module 2 and record the wind speed response at multiple points within the integrated flight channel 1. A linear approximate model is then identified using the least squares method or recursive least squares method. , In the formula, express Wind speed vector at the measuring point at time [time]. Let be the self-state transfer matrix, representing the linear influence of the wind speed at the previous measurement point on the wind speed at the next measurement point. express Wind speed vector at the measuring point at time [time]. The array-field influence matrix represents the linear effect of the input of each fan on the wind speed at each measuring point. express The PWM input vector of the fan at any given time. for Time-of-flight noise term; The closed-loop control module is used to control the array fan module 2 in real time based on the target wind spectrum and the array-field mapping identified by the array-field identification module, so that the wind speed time sequence measured in the integrated flight channel 1 approximates the target wind spectrum. The data storage module is used for data storage.

[0033] Specifically, the array-field identification module applies a known excitation sequence to the array fan module 2, which can be a pseudo-random binary sequence such as PRBS; the identification result of the array-field identification module is used as the internal model of the closed-loop control module for MPC or other model-based control.

[0034] , which is the wind speed observation vector composed of m measuring points (e.g., hot-wire wind speed probes arranged along the axial and transverse directions) within the integrated flight channel 1 at the k-th sampling time. The components usually represent the instantaneous or local average wind speed at the measuring point, with the unit being m / s. , for the kth sampling time, is the control input vector of the n independently controllable fan units in the array, such as PWM duty cycle or equivalent voltage, and the component value range is subject to hardware limiting constraints; , is the self-state transfer matrix, representing the linear influence of the wind speed at the previous measurement point on the wind speed at the next measurement point, including sensor dynamics and field inertia; , is the array-field influence matrix, which represents the linear influence coefficient of the input of each fan on the wind speed at each measuring point, that is, the "fan→field" coupling matrix; , which is the process noise term, representing random effects such as unmodeled dynamics, turbulent random disturbances, and measurement errors. It can be regarded as a zero-mean random term and is used to estimate noise statistics during identification. The discrete time k corresponds to the sampling time t = kΔt, where Δt is the system sampling period.

[0035] By establishing the above linear approximation model, the complex wind field response can be simplified into a mathematical mapping that can be used for real-time control. This enables model-based tracking controller modules (such as MPC) to calculate and issue fan control quantities in real time, thereby significantly improving the accuracy of target wind spectrum reproduction, reducing the number of manual parameter adjustments, and providing an estimable parameter basis for subsequent robust or adaptive control.

[0036] The closed-loop control module includes: The state estimation module, which is a Kalman filter, is used to estimate the wind speed field state within the integrated flight channel 1 based on the data collected by the multi-point wind speed sensor array 3. The tracking controller module is for either a model predictive controller or a multi-channel adaptive PID controller. The phase wave generation unit sends a space-time sinusoidal traveling wave drive signal to the array fan module 2 to generate controllable low-frequency wind direction oscillation and local vortices. The phase wave it generates is given by the formula... Specify; In the formula, Indicates the first A fan in time Control input; Indicates time The reference bias value is used to maintain the overall average wind speed (corresponding to the reference field U0) within the integrated flight channel 1. It represents the time-varying amplitude envelope, used to control the intensity of the space-time sinusoidal traveling wave, so as to adjust the generation of different levels of pulsation or oscillation at different stages; The space wave number is used to control the wavelength λ=2π / k of the space-time sinusoidal traveling wave on the array, which determines the vortex width or swing width at the spatial scale. For the first The spatial coordinates of each fan in the array are used to calculate the phase difference; The time angular frequency corresponds to the time period T=2π / ω of the space-time sinusoidal traveling wave, and is used to control the low-frequency oscillation period. The initial phase is used to align or offset the starting position of the space-time sinusoidal traveling wave.

[0037] Specifically, It can be a PWM duty cycle or a voltage command.

[0038] For a Kalman filter, if X k This represents the system state to be estimated (optionally, the wind speed vector at the measurement point can be directly taken). Kalman filters fuse the model and measurements through a prediction-update step, providing noise-suppressed state estimates for use by MPC or PID. When the system exhibits nonlinearity (e.g., nonlinear fan dynamics or nonlinear sensors), extended Kalman filtering (EKF) or unscented Kalman filtering (UKF) can be used. Introducing phase wave drive can generate controllable moving vortices and low-frequency oscillations in space, which can better simulate the spatial structure of natural wind than the traditional method of uniformly adding or subtracting all fans. At the same time, by combining the state estimation of Kalman filter and the closed-loop strategy of MPC / PID, the wind speed field can form a realistic physical field shape, while ensuring numerical tracking accuracy and system stability.

[0039] The target wind spectrum is a measured wind speed time series, or a target wind speed time series generated using a two-dimensional Ornstein–Uhlenbeck process or spectral synthesis method, wherein the Ornstein–Uhlenbeck process is given by the following stochastic differential equation: , In the formula, The vector represents the local wind speed over continuous time, where x is the mainstream component and y is the lateral component. The reference velocity vector (long-term average or design average wind speed). ; Θ is the recovery matrix or decay matrix, used to control the autocorrelation time scale of the Ornstein–Uhlenbeck process; Σ is the noise intensity matrix, used to control the pulsation amplitude and the covariance of each component; This represents two-dimensional Brownian motion, driven by white noise. Then, the stochastic differential equation is discretized to obtain the sampling time step. The following reference sequence .

[0040] Specifically, Θ∈R 2×2 , which is the recovery matrix or decay matrix, used to control the autocorrelation time scale of the Ornstein–Uhlenbeck process; the diagonal elements θx and θy determine the memory time of the x and y components, respectively, and the off-diagonal terms can represent the main lateral coupling (generating wind oscillation coupling). Σ∈R 2×2 , is the noise intensity matrix, used to control the pulsation amplitude and the covariance of each component; This represents two-dimensional Brownian motion (standard Wiener process), indicating white noise driving; Discretization: Using the sampling period Δt, an exact discretization solution or an approximate discrete update formula can be obtained (used for numerical generation and MPC reference sequences).

[0041] Implementation notes: If there is a field-measured wind speed time series, it can be played back directly; if there is no measured data, the Ornstein–Uhlenbeck process can generate a synthetic sequence with statistical properties by estimating Θ and Σ (e.g., from field data or by setting TI); Discretization often employs analytical discrete solutions or approximate AR(1) models: for example ,in ; The Ornstein–Uhlenbeck process or spectral synthesis method can approximate the low-frequency oscillations and mid-to-high-frequency pulsations of natural wind on both time scale and spectrum. This makes the generated target wind spectrum both physically interpretable and easy for tracking controller modules (such as MPC) to track directly, thereby improving the reproducibility of experimental scenarios for field wind field characteristics and the extrapolation credibility of biological experiments.

[0042] It also includes a volatile matter release monitoring module 4, which is used to simulate the odor concentration field. This module is located at the top inside the flight section 11 and near the slow-flow section 12. It includes a volatile matter atomizer, a VOC sensor array, and an odor field numerical solution unit electrically connected to the control calculation module 7. The odor field numerical solution unit is based on the convection-diffusion equation. Numerical discretization is performed on the odor concentration field driven by a controlled wind speed field, and the measured values ​​from the VOC sensor array are incorporated into the joint cost function of the closed-loop control module for synchronous control and tracking of wind speed and odor plume morphology; the convection-diffusion equations in... Represents the odor concentration field. Indicates time, This is a localized wind speed field used to drive odor convection. Represents the Hamiltonian operator. Let be the diffusion coefficient, representing the diffusion rate of the odor concentration field. Represents the Laplace operator. The source term represents the release rate of volatiles.

[0043] Specifically, This indicates an odor concentration field, such as VOC concentration, and the unit can be ppm or μg / m³. 3 ; Specifically, the local wind speed field generated after control is provided or interpolated by the array control and identification model and is used to drive odor convection. Specifically, it is the effective diffusion coefficient of molecular or turbulent mixing, which depends on the scale and turbulence intensity, and is used to control the diffusion rate of the concentration field; The source term represents the release rate of volatiles, which can be pulsed or continuous release, and includes location and time dependence.

[0044] For the convection-diffusion equation The first term on the left is the time derivative, the second term is the convection term, the first term on the right is the diffusion term, and the second term is the source term, which are used to fully describe the transport and diffusion process of odor in the velocity field.

[0045] Key points for numerical implementation: Discretize the convection-diffusion equations on a two-dimensional or three-dimensional grid using the finite difference or finite volume method; use upwind finite difference or higher-order stream schemes for the convection term to ensure numerical stability; use central difference for the diffusion term; By using the sensors in the VOC sensor array as sparse observation points, and integrating the numerical model with the measurements using filters (Kalman or particle filters), a more complete concentration field estimate can be obtained. While controlling the target, wind speed tracking error and VOC field error are considered (weighted synthesis) to achieve joint reproduction of "wind + smell"; By incorporating odor transport into numerical modeling and closed-loop decision-making, the device can not only reproduce wind spectrum morphology, but also the spatial structure and temporal variability of odor plumes, meeting the high simulation requirements for insect source-finding behaviors (such as host or pheromone) and significantly enhancing the ecological relevance and explanatory power of experimental results.

[0046] A method for using a device simulating the emergence and dispersal of natural enemies of the pine scale includes the following steps: S1. Obtain or set the target wind spectrum: Use the measured wind speed time series, or the target wind speed time series generated by the two-dimensional Ornstein–Uhlenbeck process or spectrum synthesis method as the target wind spectrum.

[0047] S2, Array-Field Identification: The array-field identification module applies a known excitation sequence to the array fan module 2 and records the wind speed response at multiple points within the integrated flight channel 1. The linear approximate model is then identified using the least squares method or the recursive least squares method.

[0048] The self-state transfer matrix A and array-field influence matrix B identified in step S2 will be used for control and prediction in step S5.

[0049] S3. Calibration and initialization: Calibrate the sensors in the multi-point wind speed sensor array 3 and the high-speed camera 61 in the observation and recording module 6, synchronize the clock, and initialize the parameters of the Kalman filter and tracking controller module. S4. Emergence and Release Control: Place the pupa of the pine bark scale predator to be emerged into the rectangular box 51. When the emergence detection unit detects the emergence event, the marking unit marks the emerged individual, and then the automatic door 52 opens to allow the emerged individual to enter the integrated flight channel 1.

[0050] S5. Closed-loop operation: In the online loop, wind speed field data is collected in real time through the multi-point wind speed sensor array 3, and odor concentration field data is collected in real time through the VOC sensor array. The wind speed field state of the integrated flight channel 1 is estimated through the closed-loop control module, and the control input is calculated and sent to the array fan module 2 based on the linear approximation model and the target wind spectrum. At the same time, the phase wave drives the array fan module 2 and controls the volatile release monitoring module 4 to regulate the odor concentration field.

[0051] The closed-loop operation in step S5, combined with the decision execution of the Kalman filter and tracking controller module in step S3, ensures that the wind speed field and odor field approximate the target time sequence.

[0052] S6. Extract the flight trajectory of the molted individuals using a high-speed camera 61, and collect and recapture the molted individuals in segments using a tail sampling unit 8 within the flight segment 11. Then save all control sequences and sensor data for offline analysis.

[0053] Specifically, systematizing and modularizing the apparatus and methods can form a reproducible experimental standard operating procedure (SOP): from target spectrum setting, identification, and calibration to eclosion release and closed-loop operation, all are integrated, greatly improving experimental efficiency and repeatability, and reducing human intervention on insect behavior, making the results more scientifically comparable and statistically robust.

[0054] The closed-loop operation in step S5 adopts a chance-constrained model predictive control, which ensures that the control input and output constraints meet the probability constraints with a predetermined confidence level, taking into account process noise and turbulence uncertainties.

[0055] Specifically, using a chance-constrained model for predictive control can improve robust tracking performance under conditions of high turbulence intensity.

[0056] The basic idea of ​​the opportunity constraint tracking controller module is that in optimization problems, constraints are not required to be strictly satisfied on every sample path, but rather the constraints are guaranteed to be satisfied in a probabilistic form. Under the Gaussian noise assumption, random constraints can be transformed into deterministic inequalities using the mean and variance; In practice, it can be achieved through one of the following methods: analytically converting the linear Gaussian model into convex constraints, approximating based on sampling (scenario method), or using a conservative weighted variance boundary to reduce computational complexity. The chance-constrained tracking controller module allows for probabilistic tolerance of process noise while ensuring high probability constraint satisfaction. This enables it to maintain good tracking performance and equipment safety (e.g., avoid excessively frequent large control actions) even under conditions of strong turbulence and significant random disturbances, thereby improving system robustness and reducing vulnerability to controller parameters.

[0057] It also includes online estimation of turbulence intensity, adaptive adjustment of PID gain or MPC process noise covariance and weights based on the estimated turbulence intensity, and periodic updating of the array-field influence matrix using recursive least squares or online learning methods.

[0058] Specifically, the adaptive adjustment of turbulence intensity based on real-time estimation and online updating of the array-field influence matrix B enables the tracking controller module to maintain performance under environmental changes and equipment characteristic drift, reduces the frequency of offline re-identification, and improves long-term operational stability and consistency of experimental results. Online learning can also gradually improve the array-field mapping accuracy, thereby improving the repeatability and comparability of subsequent experiments.

[0059] This invention can reproduce the multi-scale time-varying characteristics of natural airflow in the field with high fidelity under laboratory conditions. It couples and reproduces the wind speed field with odor plumes, achieving closed-loop tracking. A target wind spectrum is generated using a two-dimensional Ornstein–Uhlenbeck model or spectral synthesis, and the array-field mapping matrix is ​​obtained through PRBS identification. Kalman filtering estimation and model-based control are employed, enabling the array fan to approximate the predetermined wind spectrum in real-time in a controllable, replayable, or randomized manner. This significantly improves the matching degree between the wind speed time series and the spectrum, as well as the experimental repeatability. Furthermore, phase wave... Array-driven and chance-constrained / stochastic MPC can maintain robust tracking in the presence of turbulence and uncertainty. Furthermore, online updates based on online estimated turbulence intensity TI and recursive least squares can adaptively compensate for environmental changes and equipment drift, ensuring long-term stable operation. Moreover, by incorporating odor convection-diffusion numerical solutions and VOC sensor observations into a joint control loop, and integrating feathering-marking-release and tail-end segmented recapture sampling, the entire process from feathering triggering to diffusion observation, trajectory recording, and mark recapture is automated and data-driven, reducing behavioral biases caused by human intervention and improving statistical efficiency.

Claims

1. A device for studying the emergence and diffusion of natural enemies of the pine scale insect using simulated natural airflow, characterized in that: The system includes an integrated flight channel (1), an array fan module (2), a multi-point wind speed sensor array (3), a feathering release module (5), an observation and recording module (6), and a control and calculation module (7). The integrated flight channel (1) is used to simulate natural airflow. The array fan module (2) is located on the integrated flight channel (1) and is used to generate and regulate the average wind speed, pulsation, and spatial distribution of the wind within the integrated flight channel (1). The multi-point wind speed sensor array (3) is located on the integrated flight channel (1) and is used to collect data in real time. The multi-point wind speed data in the flight channel (1) is collected. The feather release module (5) is located in the integrated flight channel (1) and is used to place the pupa of the pine bark beetle to be feathered, and to mark and release the feathered individuals formed after the pupa of the pine bark beetle to be feathered. The observation and recording module (6) is located on the integrated flight channel (1) and is used to collect the flight trajectory of the feathered individuals. The control calculation module (7) is electrically connected to the array fan module (2), the multi-point wind speed sensor array (3), the feather release module (5), and the observation and recording module (6).

2. The apparatus for studying the emergence and diffusion of natural enemies of the pine bark scale, as described in claim 1, is characterized in that: The integrated flight channel (1) includes a flight section (11), a slow flow section (12), and a controllable main fan (13). One end of the slow flow section (12) is connected to one end of the flight section (11), and an adjustable grid (14) is provided at the end of the slow flow section (12) near the flight section (11). A honeycomb end plate (15) is provided at the end of the flight section (11) away from the slow flow section (12) and at the end of the slow flow section (12) away from the flight section (11). The controllable main fan (13) is located at the end of the slow flow section (12) away from the flight section (11) and is electrically connected to the control calculation module (7). The flight section (11) is provided with multiple tail sampling units (8) at equal intervals near the slow flow section (12), and the tail sampling units (8) are sticky insect cards or traps.

3. The apparatus for studying the emergence and diffusion of natural enemies of the pine bark scale, as described in claim 2, is characterized in that: The array fan module (2) is located on the side wall of the slow flow section (12), including several fans that can be independently controlled by PWM and at least one variable speed main fan, and the fans and the variable speed main fan are electrically connected to the control calculation module (7). The multi-point wind speed sensor array (3) is located on the top and side walls of the slow-flow section (12) and is arranged along the axial and transverse directions of the integrated flight channel (1). The feathering release module (5) is located on the bottom plate at one end of the flight section (11) away from the slow flow section (12). It includes a cuboid box (51), a feathering detection unit and a marking unit. An automatic door (52) is provided on the side of the cuboid box (51) near the slow flow section (12). The feathering detection unit and the marking unit are both located on the top of the cuboid box (51). The automatic door (52), the feathering detection unit and the marking unit are all electrically connected to the control calculation module (7). The observation and recording module (6) includes multiple high-speed cameras (61) and multiple supplementary lights. The multiple high-speed cameras (61) are respectively mounted on the top of the flight segment (11) via brackets (62), and the lenses of the high-speed cameras (61) extend into the flight segment (11). The multiple supplementary lights are all located on the top of the inner side of the flight segment (11). The high-speed cameras (61) and supplementary lights are all electrically connected to the control and calculation module (7).

4. The apparatus for studying the emergence and diffusion of natural enemies of the pine bark scale, simulating natural airflow as described in claim 3, is characterized in that: The control calculation module (7) includes: The array-field identification module is used to apply a known excitation sequence to the array fan module (2) and record the wind speed response at multiple points within the integrated flight channel (1). A linear approximate model is obtained by using the least squares method or the recursive least squares method: ， In the formula, express Wind speed vector at the measuring point at time [time]. Let be the self-state transfer matrix, representing the linear influence of the wind speed at the previous measurement point on the wind speed at the next measurement point. express Wind speed vector at the measuring point at time [time]. The array-field influence matrix represents the linear effect of the input of each fan on the wind speed at each measuring point. express The PWM input vector of the fan at any given time. for Time-of-flight noise term; The closed-loop control module is used to control the array fan module (2) in real time based on the target wind spectrum and the array-field mapping identified by the array-field identification module, so that the wind speed time sequence measured in the integrated flight channel (1) approaches the target wind spectrum. The data storage module is used for data storage.

5. The apparatus for studying the emergence and diffusion of natural enemies of the pine bark scale, as described in claim 4, is characterized in that: The closed-loop control module includes: The state estimation module is a Kalman filter used to estimate the wind speed field state in the integrated flight channel (1) based on the data collected by the multi-point wind speed sensor array (3); The tracking controller module is for either a model predictive controller or a multi-channel adaptive PID controller. The phase wave generation unit is used to send a space-time sinusoidal traveling wave drive signal to the array fan module (2) to generate controllable low-frequency wind direction oscillation and local vortices. The phase wave generated by it is given by the formula. Specify; In the formula, Indicates the first A fan in time Control input; Indicates time The baseline bias value is used to maintain the overall average wind speed within the integrated flight path (1); It represents the time-varying amplitude envelope, used to control the intensity of the space-time sinusoidal traveling wave, so as to adjust the generation of different levels of pulsation or oscillation at different stages; The space wave number is used to control the wavelength λ=2π / k of the space-time sinusoidal traveling wave on the array, which determines the vortex width or swing width at the spatial scale. For the first The spatial coordinates of each fan in the array are used to calculate the phase difference; The time angular frequency corresponds to the time period T=2π / ω of the space-time sinusoidal traveling wave, and is used to control the low-frequency oscillation period. The initial phase is used to align or offset the starting position of the space-time sinusoidal traveling wave.

6. The apparatus for studying the emergence and diffusion of natural enemies of the pine bark scale, as described in claim 4, is characterized in that: The target wind spectrum is a measured wind speed time series, or a target wind speed time series generated using a two-dimensional Ornstein–Uhlenbeck process or spectral synthesis method, wherein the Ornstein–Uhlenbeck process is given by the following stochastic differential equation: ， In the formula, The vector represents the local wind speed over continuous time, where x is the mainstream component and y is the lateral component. As the reference velocity vector, ; Θ is the recovery matrix or decay matrix, used to control the autocorrelation time scale of the Ornstein–Uhlenbeck process; Σ is the noise intensity matrix, used to control the pulsation amplitude and the covariance of each component; This represents two-dimensional Brownian motion, driven by white noise. Then, the stochastic differential equation is discretized to obtain the sampling time step. The following reference sequence .

7. The apparatus for studying the emergence and diffusion of natural enemies of the pine bark scale, as described in claim 5, is characterized in that: It also includes a volatile release monitoring module (4), which is used to simulate the odor concentration field. It is located at the top inside the flight section (11) and close to the slow flow section (12). It includes a volatile atomizer, a VOC sensor array, and an odor field numerical solution unit that are electrically connected to the control calculation module (7). The odor field numerical solution unit is based on the convection-diffusion equation. Numerical discretization is performed on the odor concentration field driven by a controlled wind speed field, and the measured values ​​from the VOC sensor array are incorporated into the joint cost function of the closed-loop control module for synchronous control and tracking of wind speed and odor plume morphology; the convection-diffusion equations in... Represents the odor concentration field. Indicates time, This is a localized wind speed field used to drive odor convection. Represents the Hamiltonian operator. Let be the diffusion coefficient, representing the diffusion rate of the odor concentration field. Represents the Laplace operator. The source term represents the release rate of volatiles.

8. A method of using the apparatus for studying the emergence and diffusion of natural enemies of the pine bark scale as described in claim 7, characterized in that: Includes the following steps: S1. Obtain or set the target wind spectrum: Use the measured wind speed time series, or the target wind speed time series generated by the two-dimensional Ornstein–Uhlenbeck process or spectrum synthesis method as the target wind spectrum; S2, Array-Field Identification: Apply a known excitation sequence to the array fan module (2) through the array-field identification module, record the wind speed response at multiple points in the integrated flight channel (1), and use the least squares method or recursive least squares method to identify the linear approximate model; S3. Calibration and initialization: Calibrate the sensors in the multi-point wind speed sensor array (3) and the high-speed camera (61) in the observation and recording module (6), synchronize the clock, and initialize the parameters of the Kalman filter and tracking controller module; S4. Emergence and Release Control: Place the pupa of the pine bark cricket to be emerged into a rectangular box (51). When the emergence detection unit detects the emergence event, the marking unit marks the emerged individual, and then the automatic door (52) opens to allow the emerged individual to enter the integrated flight channel (1). S5. Closed-loop operation: In the online loop, wind speed field data is collected in real time through the multi-point wind speed sensor array (3), odor concentration field data is collected in real time through the VOC sensor array, the wind speed field state of the integrated flight channel (1) is estimated through the closed-loop control module, and the control input is calculated and sent to the array fan module (2) based on the linear approximation model and the target wind spectrum. At the same time, the phase wave drives the array fan module (2) and controls the volatile release monitoring module (4) to regulate the odor concentration field. S6. Extract the flight trajectory of the molted individuals using a high-speed camera (61), and collect and recapture the molted individuals in segments using a tail sampling unit (8) within the flight segment (11). Then save all control sequences and sensor data for offline analysis.

9. The method of using the simulated natural airflow research device for the emergence and diffusion of natural enemies of the pine bark scale as described in claim 8, characterized in that: The closed-loop operation in step S5 adopts chance-constrained model predictive control, which ensures that the control input and output constraints meet the probability constraints with a predetermined confidence level, taking into account process noise and turbulence uncertainty.

10. The method of using the simulated natural airflow research device for the emergence and diffusion of natural enemies of the pine bark scale as described in claim 8, characterized in that: It also includes online estimation of turbulence intensity, adaptive adjustment of PID gain or MPC process noise covariance and weights based on the estimated turbulence intensity, and periodic updating of the array-field influence matrix using recursive least squares or online learning methods.

Images