Method and system for estimating number of termite populations by utilizing diet special effects

By alternating the use of perturbation estimation and baseline calibration modes, and combining different bait designs in chambers B and A, adaptive decoupling and calibration of termite population size and feeding activity were achieved. This solved the problems of inaccurate estimation and model drift in existing technologies, and improved the long-term accuracy and adaptability of monitoring.

CN122046750AActive Publication Date: 2026-05-15DA BA WEI SHI (BEIJING) NETWORK TECH CO LTD +1
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202610510120.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-17
Publication Date
2026-05-15
Estimated Expiration
2046-04-17

AI Technical Summary

Technical Problem

Existing technologies struggle to effectively decouple termite population size from feeding activity, and the models are susceptible to changes in the environment and the physiological state of termites, leading to inaccurate estimation results and long-term monitoring failure.

Method used

By alternating between perturbation estimation mode and baseline calibration mode, and using a writable perturbation preference model and a read-only baseline activity model, combined with different feed designs in chambers B and A, data is collected and model parameters are updated online to achieve adaptive decoupling and calibration of population size and feeding activity.

Benefits of technology

It improves the long-term accuracy and adaptability of termite population estimation, overcomes model drift, and enables unattended and robust monitoring.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122046750A_ABST
    Figure CN122046750A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of pest monitoring, and discloses a termite population number estimation method and system using diet special effects, and the system comprises a writable disturbance preference model and a read-only reference vitality model; in the disturbance estimation mode, opening a B chamber channel, and calculating an estimated population value by using a disturbance preference model; switching to a reference calibration mode, closing a B chamber channel, and calculating a calibration anchor point by using a reference vitality model; updating the disturbance preference model on line according to the calibration anchor point; and switching back to the disturbance estimation mode by using the updated model. The system comprises a data acquisition terminal and a data processing system, and the data processing system is provided with a state machine module, a dual-model storage module, a data decoupling calculation module and a dynamic response baseline calibration module so as to realize mode switching, closed-loop calibration and population value calculation of the method. By means of dual-mode switching and closed-loop calibration, the problem of model drifting is solved, the population number and the ingestion activity are decoupled, and high-precision estimation of the termite population is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of pest monitoring technology, specifically to a method and system for estimating termite population size using dietary effects. Background Technology

[0002] In the field of termite control, accurately estimating the size of a target termite population is a crucial prerequisite for assessing the extent of damage and developing effective control strategies. Currently, a common method is to indirectly infer the population size by monitoring the rate at which termites consume bait.

[0003] However, this method has inherent uncertainties because the total food consumption is the result of the combined effect of two parameters: the termite population size and its overall feeding activity. Traditional techniques struggle to effectively separate these two highly coupled parameters, making it impossible to clearly distinguish between low-activity feeding by large-scale colonies and high-activity feeding by small-scale colonies, thus limiting the reliability of the estimation results.

[0004] Furthermore, termite feeding activity is not a constant static value; it is dynamically influenced by environmental factors (such as temperature and humidity) and the physiological state of the termite colony itself (such as seasonal rhythms and health status). This dynamic change causes pre-defined estimation models to gradually become ineffective over time, resulting in so-called model drift, which severely impacts the accuracy and effectiveness of long-term, continuous monitoring. Therefore, existing technologies require an estimation method that can automatically decouple population size from feeding activity and adaptively compensate for model drift to achieve robust assessment of termite population size. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method and system for estimating termite population size using dietary effects. This solves the problems of existing estimation methods, which are unable to effectively decouple population size from feeding activity and whose models are easily affected by changes in the environment and the physiological state of termites, resulting in low long-term monitoring accuracy.

[0006] The first aspect of this invention provides a method for estimating termite population size using dietary effects. This method first configures a writable perturbation preference model and a read-only baseline vitality model.

[0007] In the perturbation estimation mode, the entrance passage to chamber B is opened to allow termites to simultaneously access both chamber A, which contains the baseline bait, and chamber B, which contains the perturbation bait. At this time, an estimated population value is calculated using the writable perturbation preference model. This calculation specifically includes: Consumption data and activity density data of Room A and Room B are collected, and a preference selection index is calculated based on these data.

[0008] Collect environmental parameters.

[0009] Using the preference selection index and the environmental parameters as inputs to the perturbation preference model, a perturbation vitality is obtained by querying or calculating the perturbation preference model.

[0010] Collect the total feeding flux of chamber A and chamber B.

[0011] The estimated population value is calculated by dividing the total feeding flux, the perturbation activity, and the preset feeding constant by the product of the perturbation activity and the preset feeding constant.

[0012] In the baseline calibration mode, the entrance passage to chamber B is closed to prevent termites from accessing chamber B. At this time, a calibration anchor point is calculated using the read-only baseline viability model. This calculation specifically includes: After the entrance passage to Room B is closed, only the consumption data and environmental parameters of Room A are collected.

[0013] A baseline feeding flux is calculated based on the consumption data of chamber A.

[0014] Using the environmental parameters as input to the baseline vitality model, a baseline vitality is calculated.

[0015] The calibration anchor point is calculated using the baseline feeding flux and the baseline vitality.

[0016] To overcome model drift, the method further includes: updating the internal parameters of the writable perturbation preference model online based on the calculated calibration anchor point. This online update step specifically includes: Extract a set of the last valid historical data packets before the system switches to the baseline calibration mode. This data packet includes the last total food intake flux and the last preference selection index.

[0017] Using the calibration anchor point and the final total feed flux, a desired vitality is calculated.

[0018] A current vitality is calculated using the final preference selection index and the perturbation preference model before the update.

[0019] Calculate a vitality drift error between the desired vitality and the current vitality.

[0020] Based on the vitality drift error, the parameters of the perturbation preference model are corrected.

[0021] In one implementation, the perturbation preference model is represented as a multidimensional lookup table. In this case, the parameter correction step includes: using the last preference selection index and the last environmental parameter from the last valid historical data packet as indexes to locate a target storage cell in the multidimensional lookup table; and using the product of the vitality drift error and a preset learning rate to update the value of the target storage cell, thereby correcting the multidimensional lookup table.

[0022] After the internal parameters of the perturbation preference model are updated, the method switches back to the perturbation estimation mode, reopens the entrance channel of chamber B, and begins to calculate the estimated population value using the updated perturbation preference model, thereby forming a closed-loop calibration.

[0023] The method can trigger the benchmark calibration mode according to specific conditions, which may include: the expiration of a preset calibration period, or the short-term variance of the estimated population value continuously exceeding a preset threshold.

[0024] To ensure the accuracy of the data in the baseline calibration mode, after closing the entrance channel of chamber B and before calculating the baseline feed flux, the method may further include entering a preset stabilization waiting period, and performing the step of calculating the baseline feed flux only after the stabilization waiting period has ended.

[0025] The disturbing bait in chamber B may be a standard baseline bait with an added feeding preference modifier; the feeding preference modifier is a non-lethal, non-repellent chemical substance used to modify the palatability of the bait.

[0026] A second aspect of the present invention provides a termite population estimation system utilizing dietary effects, the system comprising a data acquisition terminal and a data processing system.

[0027] The data acquisition terminal includes: room A, room B, and a room B channel actuator configured to physically open or physically close the entrance channel of room B. The data acquisition terminal also includes: a consumption monitoring unit and an activity density monitoring unit respectively located in room A and room B, and an environmental parameter unit.

[0028] The data processing system includes a processor and a memory. When the instructions stored in the memory are executed by the processor, they achieve the following: A dual-model storage module is used to store a writable perturbation preference model and a read-only baseline vitality model; A data decoupling computation module is configured to calculate the estimated population value using the perturbation preference model in a perturbation estimation mode; A dynamic response baseline calibration module is configured to calculate calibration anchors using the baseline vitality model in baseline calibration mode, and is further configured to update the internal parameters of the perturbation preference model online based on the calibration anchors; A state machine module is configured as follows: Switching between the disturbance estimation mode and the reference calibration mode; In the disturbance estimation mode, the B-room channel actuator is controlled to open the entrance channel of the B-room; In the reference calibration mode, the B chamber channel actuator is controlled to close the entrance channel of the B chamber; After the dynamic response baseline calibration module updates the disturbance preference model, it switches back to the disturbance estimation mode.

[0029] This invention provides a method and system for estimating termite population size using dietary specialties. It has the following beneficial effects: 1. This invention achieves system self-calibration by configuring a perturbation estimation mode and a baseline calibration mode, and periodically switching between these two modes. Using the calibration anchor points obtained by closing chamber B in baseline calibration mode and using a read-only baseline viability model, the writable perturbation preference model used when chamber B is open in perturbation estimation mode is updated online. This closed-loop calibration mechanism can continuously correct model parameters, effectively overcoming model drift caused by environmental changes or changes in termite physiological state, and improving the accuracy and adaptability of estimated population values ​​in long-term monitoring. 2. This invention constructs a preference selection index by introducing disturbed bait in chamber B, and maps this index and environmental parameters to disturbed activity using a writable disturbed preference model. When calculating the estimated population value, the total feeding flux is divided by the product of the disturbed activity and a preset feeding constant. This method successfully decouples the highly coupled population size parameter from the complex feeding activity parameter, solving the problem of traditional methods' difficulty in distinguishing between colony size and feeding activity, thus making the estimation results more reliable. 3. This invention, through the establishment of a state machine module, automatically controls the actuator in the B-chamber channel to switch modes based on preset calibration cycles or triggering conditions such as the estimated short-term variance of the population value. Furthermore, after switching to the baseline calibration mode, the system includes a preset stabilization waiting period to ensure data stability. This automated process design, combined with hardware actuators and timing logic control, achieves unattended and robust monitoring of termite population numbers, ensuring the stability and effectiveness of the calibration process without human intervention. Attached Figure Description

[0030] Figure 1 This is a block diagram of the overall system architecture of the present invention; Figure 2This is a schematic diagram of the data acquisition terminal of the present invention; Figure 3 This is a flowchart of the system recovery and loop of the present invention.

[0031] Among them, 10. G06F data processing system; 101. State machine module; 102. Dual model storage module; 103. Data decoupling calculation module; 104. Dynamic response baseline calibration module; 20. Data acquisition terminal; 201. Room A; 202. Room B; 203. Consumption monitoring unit; 204. Activity density monitoring unit; 205. Environmental parameter unit; 206. Room B channel actuator. Detailed Implementation

[0032] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] Reference Figure 1 The present invention provides a termite population estimation system that utilizes dietary effects. The system includes a G06F data processing system 10 and a data acquisition terminal 20.

[0034] The G06F data processing system 10 and the data acquisition terminal 20 are connected via wired or wireless communication.

[0035] The data acquisition terminal 20 is responsible for collecting real-time data on termite feeding and activity, and sending the data to the G06F data processing system 10. The G06F data processing system 10 is responsible for receiving and processing the data, performing population estimation calculations, and sending control commands to the data acquisition terminal 20.

[0036] In one embodiment, the G06F data processing system 10 includes a processor, a memory, a communication interface, and an I / O interface. The memory stores computer program instructions, which, when executed by the processor, implement the functions of the following modules: A state machine module 101 is used to drive the G06F data processing system 10 to switch between disturbance estimation mode and benchmark calibration mode according to a preset time period or detected data anomalies.

[0037] A dual-model storage module 102 is used to store and manage two independent vitality calculation models, specifically including: a writable perturbation preference model for perturbation estimation patterns (…). ); and a read-only benchmark vitality model for benchmark calibration mode ( ).

[0038] A data decoupling calculation module 103 is used to calculate the preference selection index in real time under the disturbance estimation mode. ), and combined The model decouples and calculates the estimated population value ( ).

[0039] A dynamic response baseline calibration module 104 is used to combine, in reference calibration mode, Model calculation calibration anchor point ( ), and further calculate the expected vitality ( ) and current vitality ( The drift error between ) ), and based on Online updates The model's internal parameters.

[0040] In one embodiment, the data acquisition terminal 20 includes: chamber A 201 (reference bait chamber), chamber B 202 (disturbance bait chamber), consumption monitoring unit 203, activity density monitoring unit 204, environmental parameter unit 205, and chamber B channel actuator 206.

[0041] Chamber A 201 and Chamber B 202 are physically isolated from each other. Chamber A 201 is filled with standard reference bait. Chamber B 202 is filled with disturbing bait containing a non-lethal, non-repellent feeding preference modifier.

[0042] Consumption monitoring unit 203 is used to monitor the amount of feed consumed in chamber A 201 and chamber B 202 in real time, and to acquire consumption data. and .

[0043] Activity density monitoring unit 204 is used to monitor the termite activity density in chamber A 201 and chamber B 202 in real time, and to acquire activity density data. and .

[0044] Environmental parameter unit 205 is used to monitor the ambient temperature inside the terminal. and humidity .

[0045] The B-room passage actuator 206 is located at the entrance passage of B-room 202 and is connected to the I / O interface of the G06F data processing system 10. The B-room passage actuator 206 performs the operation of physically opening or closing the entrance passage of B-room 202 according to the control commands sent by the state machine module 101 of the G06F data processing system 10.

[0046] During system operation, the data acquisition terminal 20 transmits data through its communication interface. , , , , and Data is continuously sent to the G06F data processing system 10.

[0047] Reference Figure 2 The data acquisition terminal 20 is configured as a physical monitoring device, which provides real-time, multi-dimensional biological and environmental data input to the G06F data processing system 10 and serves as the execution end of the physical control commands of the G06F data processing system 10.

[0048] In one embodiment, the data acquisition terminal 20 includes chamber A 201 and chamber B 202. Chamber A 201 serves as a reference bait chamber, and chamber B 202 serves as a disturbance bait chamber. Chamber A 201 and chamber B 202 are structurally physically isolated from each other to ensure that the reference bait in chamber A 201 and the disturbance bait in chamber B 202 do not physically interfere with each other, but both have independent entrances for the termite colony to access.

[0049] Chamber A, 201, is provided with standard reference bait, which is a standard cellulose matrix edible by termites. Chamber B, 202, is provided with disturbing bait, which is the standard reference bait with the addition of a predetermined concentration of a feeding preference regulator. The feeding preference regulator is a non-lethal, non-repellent chemical substance whose function is only to alter the palatability of the bait (e.g., slightly reduce it) without causing termite death or avoidance behavior, thus ensuring that termites still make preferred choices between chambers A, 201 and B, 202.

[0050] The data acquisition terminal 20 further includes a consumption monitoring unit 203. In one embodiment, the consumption monitoring unit 203 includes high-precision weighing sensors respectively disposed below the bait-bearing structure of chamber A 201 and the bait-bearing structure of chamber B 202. The weighing sensors are used to continuously or periodically measure the real-time weight of the bait in chamber A 201 and chamber B 202. The G06F data processing system 10 receives the real-time weight data and obtains the cumulative consumption data by performing a differential operation with the initial weight or the weight at the previous moment. and .

[0051] The data acquisition terminal 20 further includes an activity density monitoring unit 204. In one embodiment, the activity density monitoring unit 204 includes passive infrared sensors respectively disposed inside chamber A 201 and chamber B 202 or at their entrance passages. The passive infrared sensors are used to detect changes in infrared thermal radiation caused by termites entering or moving. In another embodiment, the activity density monitoring unit 204 may also be a piezoelectric vibration sensor, used to detect weak vibrations generated when termites feed or move. The activity density monitoring unit 204 converts the detected analog signals into digitized activity levels or pulse counts as dimensionless activity density data. and The relative foraging intensity of termites in chambers A 201 and B 202 is used to characterize the relative foraging intensity of termites.

[0052] The data acquisition terminal 20 further includes an environmental parameter unit 205. The environmental parameter unit 205 includes at least one temperature sensor and at least one humidity sensor (e.g., an integrated digital temperature and humidity sensor module) disposed within the terminal. The environmental parameter unit 205 is responsible for acquiring the ambient temperature inside the monitoring station in real time. and relative humidity The G06F data processing system 10 uses this data as a vitality calculation model. and ) environment input parameters.

[0053] The data acquisition terminal 20 further includes a B-chamber channel actuator 206. The B-chamber channel actuator 206 is an electromechanical control component located at the entrance channel of chamber B 202, electrically connected to the I / O control interface of the G06F data processing system 10. In one embodiment, the B-chamber channel actuator 206 is a low-power servo motor that drives a physical baffle to open or close the entrance channel of chamber B 202. In another embodiment, the B-chamber channel actuator 206 may also be a solenoid valve. The B-chamber channel actuator 206 strictly executes actions according to the instructions of the state machine module 101 of the G06F data processing system 10; that is, it keeps the channel open in disturbance estimation mode and performs a closing action to physically isolate chamber B 202 in reference calibration mode.

[0054] The G06F data processing system 10 constitutes the estimation host of the present invention, which is configured as a central control unit for receiving data, performing calculations, and sending control commands. In one embodiment, the G06F data processing system 10 includes one or more processors, a memory, a communication interface connected to the data acquisition terminal 20, and an I / O interface connected to the B-room channel actuator 206.

[0055] The processor can be a microcontroller, an embedded system-on-a-chip (SoC), or a general-purpose central processing unit (CPU). The communication interface is used to enable bidirectional data exchange between the G06F data processing system 10 and the data acquisition terminal 20. In one embodiment, the communication interface is an RS-485 bus interface, a CAN bus interface, or a low-power wireless communication module such as LoRa or ZigBee. The I / O interface can be a general-purpose input / output port or a relay control port, used to output high / low level or switching control signals to the B-chamber channel actuator 206.

[0056] The memory is used to store computer program instructions that can be executed by the processor. When the instructions are executed by the processor, the G06F data processing system 10 implements the functions of the following logic modules: State machine module 101, configured as the top-level logic controller of G06F data processing system 10, manages the operating states of G06F data processing system 10. State machine module 101 internally maintains at least two states: disturbance estimation mode and reference calibration mode. State machine module 101 operates based on an internal periodic timer (e.g., upon reaching a preset calibration period). ) or based on data monitoring results (e.g., the output of the data decoupling calculation module 103). The short-term variance of the value continuously exceeds a preset threshold. The state machine module 101 sends a shutdown command to the B-chamber channel actuator 206 via the I / O interface when switching to the reference calibration mode; when switching back to the disturbance estimation mode, it sends an open command.

[0057] The dual-model storage module 102 is a specific data area partitioned in memory for persistently storing two independent vitality calculation models. Specifically, the dual-model storage module 102 includes: a perturbation preference model (…). This storage area is defined as writable and is used to store... The multidimensional lookup table data or its equivalent model parameters, The content can be accessed and updated by the dynamic response baseline calibration module 104; and a baseline vitality model ( This storage area is defined as read-only and is used to store... The environmental model parameters are determined to ensure their long-term stability as a calibration benchmark in the reference calibration mode.

[0058] The data decoupling calculation module 103 is a calculation program activated by the state machine module 101 in the disturbance estimation mode. The data decoupling calculation module 103 receives data reported in real time by the data acquisition terminal 20. , , , , , Data. The data decoupling computation module 103 accesses the dual-model storage module 102 to read data. storage area ( The current parameters of ) are used to execute. , , The calculations are then decoupled, and the estimated population value is output. .

[0059] The dynamic response baseline calibration module 104 is a calculation program activated by the state machine module 101 in the baseline calibration mode. The dynamic response baseline calibration module 104 receives... , , Data, and access the dual-model storage module 102 to read. storage area ( The read-only parameters of the dynamic response baseline calibration module 104 are executed. , The calculation is performed to obtain the calibration anchor point. Subsequently, the dynamic response baseline calibration module 104 retrieves historical data from the disturbance estimation mode and performs inverse calculations to derive the drift error. The core function of the dynamic response baseline calibration module 104 is to adjust the baseline calibration according to the actual situation. The calculated correction value is written to and updates the dual-model storage module 102. storage area ( ), to complete the Online calibration of the model.

[0060] This phase describes the initial configuration process of the G06F data processing system 10 for core computing resources and control parameters before it begins to execute the termite population estimation task.

[0061] The processor of the G06F data processing system 10 first starts the state machine module 101 and initializes the system state to the disturbance estimation mode. At the same time, the processor executes program instructions to initialize the dual model storage module 102 and load model parameters.

[0062] The dual-model storage module 102 needs to load and maintain two independent vitality calculation models: the perturbation preference model ( ) and benchmark vitality model ( ). The model is used to calculate baseline vitality. The model is configured as read-only in the system. The model is defined as an environment parameter function. The parameters are obtained by fitting the relationship between termite feeding activity and environmental temperature and humidity on a single standard reference bait (corresponding to bait 201 in chamber A) under standard laboratory conditions.

[0063] Model The parameters are pre-programmed into the read-only area of ​​the dual-model storage module 102 before system deployment to ensure its stability as a calibration benchmark. Benchmark vitality. The calculation formula is as follows: ; in: Real-time ambient temperature; Real-time ambient humidity; function It can be fitted using known biological models such as the Arrhenius equation or exponential regression.

[0064] The model is used to calculate perturbation vitality. The model is configured to be writable in the system. The model is defined as a multidimensional function. ,in The (preference selection index) is its core input parameter.

[0065] Model initial parameters It is based on the pre-calibration and storage of data from A / B double-bait selection experiments conducted on termites of known quantity and viability in a laboratory environment. This can be represented as a multidimensional lookup table (LUT), which uses... , , As an index, its corresponding value.

[0066] The calculation formula is as follows: ; in: The parameters are the targets for online correction by the dynamic response baseline calibration module 104.

[0067] In addition, the G06F data processing system 10 also needs to load the known feeding constant during initialization. . It is a physical constant related to the type of bait, used to determine the vitality value. and population size Transformed into observable total food flux .constant It is also stored in memory.

[0068] This phase describes the data processing flow of the G06F data processing system 10 performing continuous population estimation in its default operating state.

[0069] The state machine module 101 of the G06F data processing system 10 maintains the system in disturbance estimation mode. In this mode, the state machine module 101 sends and maintains an open command to the B-chamber channel actuator 206 of the data acquisition terminal 20 via the I / O interface, ensuring that the entrance channel of the B-chamber 202 is open, allowing the termite colony to freely access the A-chamber 201 and the B-chamber 202, thereby enabling feeding preference selection behavior to occur.

[0070] In this mode, the data decoupling calculation module 103 of the G06F data processing system 10 is activated and continuously receives real-time data streams reported by the data acquisition terminal 20, including: , , , , and .

[0071] The data decoupling calculation module 103 first executes the vitality decoupling algorithm. The first step of this algorithm is to calculate the preference selection index. Data decoupling calculation module 103 extracts... , , and Data were used to calculate the approximate average individual feeding efficiency (i.e., ...) for chambers A 201 and B 202, respectively. and Then calculate the ratio of the two to obtain... : ; In one embodiment, to avoid computational instability caused by a denominator that is zero or too small, the data decoupling calculation module 103 performs... During calculation, if detected or The value is zero or below a minimum activity threshold within a specific time window. Then the data decoupling calculation module 103 is paused. The update calculation is performed, and the previous valid one is used. Value, or use a preset default value (e.g.) ), until and All restored to above.

[0072] The second step of the vitality decoupling algorithm is to calculate the perturbation vitality. The data decoupling calculation module 103 uses the data obtained in the previous step. and the environmental parameter unit 205 obtained and As input, access the dual-model storage module 102, and from... Read the current perturbation preference model from the memory area (writable area). ( It is the initial model Or it has been updated by the dynamic response baseline calibration module 104. The data decoupling calculation module 103 performs queries or calculations. The decoupled perturbation vitality value is obtained. : ; In obtaining Then, the data decoupling calculation module 103 executes the population estimation algorithm. The first step of this algorithm is to calculate the total feeding flux. Data decoupling calculation module 103 extracts... and The data is summed to obtain the total consumption. Then to Perform time differentiation (e.g., through a preset time window) (Calculate the difference between the incremental consumption within the time limit) to obtain : ; The second step of the population estimation algorithm is to calculate the estimated population. The data decoupling calculation module 103 extracts the data obtained in the previous step. The vitality decoupling algorithm obtained and the feeding constant read from memory. The data decoupling calculation module 103 solves the feeding equation to... Divide by and The product of these factors is used to decouple and calculate the final estimated population value. : ; The G06F data processing system 10 will calculate the... The population size estimate at the current moment is output or stored, and the calculation of this stage (stage two) is continuously executed in a loop until it is triggered by state machine module 101 to switch to stage three.

[0073] This phase describes how the G06F data processing system 10 pauses the regular perturbation estimation mode and actively switches to a controlled, undisturbed state to obtain a reliable population baseline.

[0074] The state machine module 101 of the G06F data processing system 10 continuously monitors the mode triggering mechanism. In one embodiment, the triggering mechanism is periodic triggering, that is, when the timer inside the state machine module 101 reaches a preset calibration period. (For example, 168 hours). In another embodiment, the triggering mechanism may be an exception trigger, i.e., the state machine module 101 detects the output of the data decoupling calculation module 103. The short-term variance or drift rate of the value continuously exceeds a preset threshold. .

[0075] When any trigger condition is met, the state machine module 101 immediately switches the system state from the disturbance estimation mode to the reference calibration mode.

[0076] After entering the baseline calibration mode, the state machine module 101 first executes hardware control. The state machine module 101 sends a shutdown command to the B-chamber channel actuator 206 of the data acquisition terminal 20 via the I / O interface. Upon receiving the command, the B-chamber channel actuator 206 physically closes the entrance channel of B-chamber 202, physically isolating the termite colony and preventing it from accessing B-chamber 202 and its disturbing bait. The termite colony is forced to access only A-chamber 201 and its baseline bait.

[0077] After chamber B, 202, is closed, state machine module 101 enters a preset stable waiting period. During this period, the system does not perform calculations, but simply waits for the termites' foraging behavior to adapt to this environment without choice.

[0078] The duration is preset and stored in the memory of the G06F data processing system 10. Its value is determined experimentally based on the biological habits of the target termite population (e.g., the behavioral stabilization time after switching from a multi-choice environment to a single-choice environment) to ensure that the calibration calculation window is entered. Previously, the feeding flux in chamber A, room 201 was... It has reached a relatively stable state.

[0079] exist After completion, state machine module 101 enters the calibration calculation window. During this window period, the G06F data processing system 10 activates the dynamic response baseline calibration module 104 and collects only the consumption data of chamber A 201. (From consumption monitoring unit 203) and environmental data and (From environmental parameter unit 205).

[0080] The dynamic response baseline calibration module 104 executes the calibration anchor point calculation algorithm: First, the dynamic response baseline calibration module 104 calculates the baseline feeding flux. The dynamic response baseline calibration module 104 pairs in... Consumption data of Room A collected during the window period Perform time differentiation (e.g., difference calculation): ; Secondly, the dynamic response baseline calibration module 104 calculates the baseline vitality. The dynamic response baseline calibration module 104 accesses the dual-model storage module 102 from its read-only memory. Storage area read baseline vitality model and use Environmental data collected during the window period and Calculated as input: ; Finally, the dynamic response baseline calibration module 104 calculates the calibration anchor point. The dynamic response baseline calibration module 104 extracts the results from the previous two steps. and And read the feeding constant from memory. The dynamic response baseline calibration module 104 solves the feeding equation to obtain... : ; The value is temporarily stored by the dynamic response baseline calibration module 104, and this value is considered to be in the current environment. and Next, by The reliable population benchmark anchored to (read-only model). After this stage is completed, state machine module 101 prepares to enter stage four.

[0081] This phase describes the core calibration steps automatically executed by the G06F data processing system 10 after the baseline calibration mode ends and before returning to the disturbance estimation mode. This phase is performed by the dynamic response baseline calibration module 104, and its goal is to utilize the calibration anchor points obtained in phase three. To correct the perturbation preference model ( The parameter drift of ).

[0082] The dynamic response baseline calibration module 104 first retrieves a set of the last valid historical data packets in the perturbation estimation mode state before switching to the reference calibration mode from the memory (e.g., a circular buffer) of the G06F data processing system 10. . include: ; in: This represents the final total ingestion flux under the perturbation estimation model; The final preference selection index is chosen under the perturbation estimation model; and This refers to the corresponding environmental data.

[0083] Subsequently, the dynamic response baseline calibration module 104 performs drift error calculation. The first step of this calculation is to calculate the expected vitality. The dynamic response baseline calibration module 104 uses the results obtained in phase three. (Calibration anchor point) serves as the benchmark for the current population's true value, and is combined with... In and feeding constant By reverse calculation, it was found that... In state The correct vitality value that the model should have output: ; The second step in drift error calculation is to calculate the current vitality. The dynamic response baseline calibration module 104 accesses the dual-model storage module 102 and reads... Model parameters in the storage area before the update and use In As input, calculate The actual vitality value output by the model before the update: ; The third step in drift error calculation is to calculate the vitality drift error. Dynamic response baseline calibration module 104 calculates... and The difference between them: ; Quantified Model( )exist The degree and direction of drift at the corresponding data points.

[0084] In obtaining Then, the dynamic response baseline calibration module 104 performs... Model parameter update operation. The dynamic response baseline calibration module 104 accesses the dual model storage module 102 and updates the parameters. The storage area (writable area) performs a parameter write operation to generate a new calibrated model. .

[0085] In one embodiment, when When it is a multidimensional lookup table (LUT), the update operation is as follows: the dynamic response baseline calibration module 104 uses... In As an index, locate The corresponding storage cell is then used. Next, the dynamic response baseline calibration module 104 extracts a preset learning rate stored in memory. Then apply the following formula to update the value of the cell: ; Learning rate It is a preset scalar value stored in memory, with a value range greater than 0 and less than or equal to 1. In one embodiment, This is set to a fixed constant (e.g., 0.1) to ensure a consistent step size for each correction. In another embodiment, It can be a dynamically adjusted parameter, for example, Values ​​and drift error absolute value or variance This is correlated to use a larger correction step size when the system drifts significantly and a smaller correction step size when the system tends to stabilize.

[0086] In another embodiment, if It is a sparse lookup table, which the dynamic response baseline calibration module 104 can use when performing an update. The correction value is based on Coordinates are assigned to using a multidimensional interpolation algorithm (e.g., trilinear interpolation). The parameters are adjusted to be more smoothly adjusted by using the nearest neighboring storage cells at the coordinates.

[0087] After the parameter update is completed, the dynamic response baseline calibration module 104 will... cover As The new version of the model. After this stage is completed, state machine module 101 prepares to enter stage five.

[0088] Reference Figure 3 This phase describes the completion of phase four of the G06F data processing system 10. After the model is corrected online, it is restored to its default working state and a closed-loop operation is formed during the conversion process.

[0089] The state machine module 101 of the G06F data processing system 10 detects that the dynamic response baseline calibration module 104 has successfully executed... Model parameter update operation (i.e.) After the data has been written to the dual-model storage module 102, a state switch is immediately performed to switch the system state from the benchmark calibration mode back to the disturbance estimation mode.

[0090] Upon switching back to the perturbation estimation mode, the state machine module 101 sends an open command to the B-chamber channel actuator 206 of the data acquisition terminal 20 via its I / O interface. The B-chamber channel actuator 206 receives the command and executes a corresponding action (e.g., driving a servo motor to reset the baffle), physically opening the entrance channel of chamber B 202. This operation restores the termite colony's free access to chambers A 201 and B 202, thereby re-establishing the feeding preference selection perturbation conditions necessary for performing perturbation estimation mode estimation.

[0091] While the hardware is restored, the state machine module 101 reactivates the data decoupling calculation module 103 within the G06F data processing system 10, enabling it to resume executing the continuous population estimation task in the perturbation estimation mode.

[0092] When the data decoupling calculation module 103 recovers its computational disturbance activity When it accesses the dual-model storage module 102 Access the storage area (writable area) When configuring model parameters, the parameters read are those that have been corrected and written by the dynamic response baseline calibration module 104 in stage four. (Right now ).

[0093] The G06F data processing system 10 was subsequently used after calibration. Model parameters, continuously executing perturbation estimation mode Estimation continues until state machine module 101 detects the next reference calibration mode trigger condition (e.g., calibration cycle). Expiration or An anomaly has occurred.

[0094] By cyclically executing the disturbance estimation mode stage, the baseline calibration mode stage, the model parameter calibration stage, and the system recovery and loop stage, the present invention achieves a closed-loop, adaptive calibration process, ensuring... The parameters of the model can be based on The baseline anchor points provided by the (read-only model) are used for periodic or triggered online corrections.

Claims

1. A method for estimating termite population size using dietary specialties, characterized in that, include: Configure a writable perturbation preference model and a read-only baseline vitality model; In the disturbance estimation mode, the entrance passage of chamber B is opened, allowing termites to access chamber A, which is equipped with baseline bait, and chamber B, which is equipped with disturbance bait. The estimated population value is calculated using the writable perturbation preference model. Switch to baseline calibration mode and close the entrance passage to chamber B to prevent termites from accessing chamber B; In the reference calibration mode, the calibration anchor point is calculated using the read-only reference vitality model; The internal parameters of the perturbation preference model are updated online based on the calibration anchor points. Switch back to the perturbation estimation mode, reopen the entrance channel of Chamber B, and calculate the estimated population value using the updated perturbation preference model.

2. The method for estimating termite population size using dietary specialties according to claim 1, characterized in that, The calculation of the estimated population value under the perturbation estimation mode includes: Consumption data and activity density data for chamber A and chamber B are collected; Calculate the preference selection index based on the consumption data and the activity density data; Collect environmental parameters; Using the preference selection index and the environmental parameters as inputs to the disturbance preference model, the disturbance vitality is calculated by querying or calculating the disturbance preference model; Collect the total feeding flux of chamber A and chamber B; The estimated population value is calculated by dividing the total feeding flux, the perturbation activity, and the preset feeding constant by the product of the perturbation activity and the preset feeding constant.

3. The method for estimating termite population size using dietary specialties according to claim 1, characterized in that, The calculation of the calibration anchor point in the reference calibration mode includes: After the entrance passage to Room B is closed, only the consumption data and environmental parameters of Room A are collected; The baseline feed flux is calculated based on the consumption data of chamber A. Using the environmental parameters as input to the baseline vitality model, the baseline vitality is calculated; The calibration anchor point is calculated using the baseline feeding flux and the baseline vitality.

4. The method for estimating termite population size using dietary specialties according to claim 1, characterized in that, The online update of the internal parameters of the perturbation preference model includes: Extract the last set of valid historical data packets before switching to the baseline calibration mode, the last set of valid historical data packets including the last total food intake flux and the last preference selection index; Expected vitality is calculated using the calibration anchor point and the final total feed flux; The current vitality is calculated using the final preference selection index and the perturbation preference model before the update; Calculate the vitality drift error between the desired vitality and the current vitality; The parameters of the perturbation preference model are corrected based on the vitality drift error.

5. The method for estimating termite population size using dietary specialties according to claim 4, characterized in that, The parameters of the perturbation preference model corrected based on the vitality drift error include: The perturbation preference model is represented as a multidimensional lookup table; The target storage cell in the multidimensional lookup table is located using the last preference selection index and the last environmental parameter in the last valid historical data packet as an index. The product of the vitality drift error and the preset learning rate is used to update the value of the target storage cell, thereby correcting the multidimensional lookup table.

6. The method for estimating termite population size using dietary specialties according to claim 1, characterized in that, The triggering conditions for switching to the reference calibration mode include at least one of the following: The preset calibration period has expired; The short-term variance of the estimated population value continuously exceeds a preset threshold.

7. The method for estimating termite population size using dietary specialties according to claim 2, characterized in that: The consumption data is collected by consumption monitoring units respectively located in room A and room B; The activity density data is collected by activity density monitoring units respectively located in chamber A and chamber B; The environmental parameters are collected by the environmental parameter unit.

8. The method for estimating termite population size using dietary specialties according to claim 1, characterized in that, The disturbance bait in chamber B is a standard baseline bait with an added feeding preference modifier; the feeding preference modifier is a non-lethal, non-repellent chemical substance used to modify the palatability of the bait.

9. The method for estimating termite population size using dietary specialties according to claim 3, characterized in that, After closing the entrance passage to Chamber B and before calculating the baseline feed flux, the method further includes: Enter a preset stable waiting period; After the stabilization period ends, the step of calculating the baseline food flux is performed.

10. A termite population estimation system utilizing dietary special effects, characterized in that, include: A data acquisition terminal includes a room A, a room B, and a room B passage actuator, wherein the room B passage actuator is configured to physically open or physically close the entrance passage of the room B; A G06F data processing system includes a processor and a memory, wherein instructions stored in the memory, when executed by the processor, implement: A dual-model storage module is used to store a writable perturbation preference model and a read-only baseline vitality model; A data decoupling computation module is configured to calculate the estimated population value using the perturbation preference model in a perturbation estimation mode; A dynamic response baseline calibration module is configured to calculate calibration anchors using the baseline vitality model in baseline calibration mode, and is further configured to update the internal parameters of the perturbation preference model online based on the calibration anchors; A state machine module is configured as follows: Switching between the disturbance estimation mode and the reference calibration mode; In the disturbance estimation mode, the B-room channel actuator is controlled to open the entrance channel of the B-room; In the reference calibration mode, the B chamber channel actuator is controlled to close the entrance channel of the B chamber; After the dynamic response baseline calibration module updates the disturbance preference model, it switches back to the disturbance estimation mode.