Dynamic regulation and control method for working parameters of environment self-adaptive detector

Abstract

The invention provides a dynamic regulation and control method for working parameters of an environment self-adaptive detector, and relates to the technical field of operation parameter optimization, and the method comprises the steps: obtaining a non-intervention operation state based on the model identifier of the detector and multi-modal environment feature vector retrieval; based on the non-intervention operation state and the multi-mode environment feature vector, working condition adaptability analysis is carried out, and reference control parameters are output; and after the jump control parameters of the updated environment feature vectors are matched to gradually adjust the detector from the reference control parameters to the jump control parameters, working condition suitability analysis is carried out according to the real-time operation working conditions and the updated environment feature vectors, and closed-loop dynamic regulation and control of the working parameters of the detector are carried out. The technical problem that in the prior art, due to the fact that working parameter setting of a detector in a complex working scene cannot be updated in a self-adaptive mode along with the environment, running reliability and detection precision of the detector are insufficient is solved. The technical effects of autonomously optimizing the working parameters of the detector in a complex and changeable environment and remarkably improving the detection precision of the detector are achieved.

Description

Technical Field

[0001] This invention relates to the field of operating parameter optimization technology, and in particular to a method for dynamic control of operating parameters of an environment-adaptive detector. Background Technology

[0002] Environmentally adaptive detectors, such as group III nitride detectors and SiC ultraviolet detectors, have become key devices in high-end fields such as space environment monitoring and biochemical detection due to their wide bandgap characteristics and high sensitivity in the ultraviolet band.

[0003] However, the photoelectric conversion performance of this type of detector is extremely sensitive to fluctuations in environmental parameters (such as temperature, humidity, and light intensity). In complex and dynamic working scenarios (such as vehicle-mounted mobile platforms, extreme outdoor climates, or industrial environments with strong interference), sudden changes in environmental parameters (such as abrupt changes in temperature and humidity, or abrupt changes in light intensity) can lead to performance degradation phenomena such as signal-to-noise ratio attenuation and response time drift.

[0004] Traditional parameter optimization methods mainly rely on two types of techniques: one is rule-based control based on fixed thresholds, which matches static parameters through a preset environment-parameter mapping table, but it is difficult to adapt to continuous environmental changes; the other is manual parameter tuning based on experience, which requires operators to monitor the environment in real time and manually adjust parameters such as bias voltage and gain. The former can cause instantaneous performance crashes when the environment changes abruptly due to the lag in parameter updates, while the latter is constrained by subjective judgment errors and response delays, and cannot cope with high-frequency environmental changes.

[0005] In summary, existing environmental adaptive detectors often fail to adapt their operating parameters to the environment in complex working scenarios, resulting in insufficient reliability and detection accuracy. Summary of the Invention

[0006] To address the above-mentioned deficiencies or improvement needs of existing technologies, this invention provides a method for dynamically adjusting the operating parameters of an environment-adaptive detector. This method is used to solve the technical problem that the operating parameter settings of existing environment-adaptive detectors often cannot keep up with environmental updates in complex working scenarios, resulting in insufficient detector reliability and detection accuracy.

[0007] To achieve the above objectives, the present invention provides a method for dynamically adjusting the operating parameters of an environment-adaptive detector.

[0008] A first aspect of the present invention provides a method for dynamically adjusting the operating parameters of an environment-adaptive detector, the method comprising:

[0009] Based on the real-time position coordinates of the environmental adaptive detector, local environmental perception processing is performed in the workspace to obtain a multimodal environmental feature vector. The model identifier of the environmental adaptive detector and the multimodal environmental feature vector are used as unsupervised similarity matching retrieval conditions to search the local historical database and obtain the uninterrupted operation status. The uninterrupted operation status and the multimodal environmental feature vector are input into the equipment control analysis model for condition adaptability analysis, and benchmark control parameters are output. The environmental adaptive detector is started after being adjusted to the benchmark control parameters. Environmental fluctuation jumps are identified in the updated environmental feature vector of the local environmental perception to match the jump control parameters. After the environmental adaptive detector is gradually adjusted from the benchmark control parameters to the jump control parameters, condition adaptability analysis is performed based on the real-time operating conditions and the updated environmental feature vector, using the jump control parameters as a starting point, to obtain updated control parameters. Based on the environmental fluctuation jump identification results of the working environment, closed-loop dynamic adjustment of the operating parameters of the environmental adaptive detector is performed until the detection mission ends.

[0010] In one implementation, environmental fluctuation jumps are identified in the updated environmental feature vector of the local environment perception to match jump control parameters, and the following processing is also performed:

[0011] After performing short-term multimodal data backtracking on the working environment, dynamic fluctuation feature clustering analysis of environmental parameters is performed to obtain a multi-scale environmental fluctuation interval set; a jump backup parameter set is pre-set based on the multi-scale environmental fluctuation interval set; environmental parameter jump events are identified by comparing the updated environmental feature vector of the local environment perception with the multi-scale environmental fluctuation interval set in real time; and the jump control parameters are located and extracted from the jump backup parameter set according to the jump environmental fluctuation interval corresponding to the jump event.

[0012] In one implementation, based on the real-time position coordinates of the environment adaptive detector, local environment perception processing is performed in the workspace to obtain a multimodal environment feature vector, and the following processing is also performed:

[0013] Based on the model identifier, the device parameters are networked and called to obtain the effective detection radius; the effective detection boundary is defined based on the real-time location coordinates and the effective detection radius; based on the effective detection boundary, the multimodal sensor array is locally activated in the workspace to obtain a local environment sensing array, wherein each sensing node in the multimodal sensor array is composed of a temperature sensor, a humidity sensor, a light intensity sensor, and an ultraviolet sensor; the multimodal environmental feature vector is obtained by performing multimodal temporal feature fusion processing on the multimodal sensing data returned by the local environment sensing array.

[0014] In one implementation, after performing short-term multimodal data backtracking on the working environment, dynamic fluctuation feature clustering analysis of environmental parameters is performed to obtain a multi-scale environmental fluctuation interval set, and the following processing is also performed:

[0015] Starting from the acquisition timestamp of the multimodal environmental feature vector, short-term multimodal data backtracking is performed within a preset backtracking time window to obtain a multimodal historical sensing sequence. After aligning the multimodal historical sensing sequence in time, W multimodal sensing slices are extracted at equal intervals. Based on the W multimodal sensing slices, dynamic fluctuation feature clustering of environmental parameters is performed to extract a multimodal fluctuation interval scale sequence. The multimodal fluctuation interval scale sequence is combined and enumerated to output the multi-scale environmental fluctuation interval set.

[0016] In one implementation, the model identifier and multimodal environmental feature vector of the environment adaptive detector are used as unsupervised similarity matching retrieval conditions to search the local historical database, obtain the uninterrupted operation status, and then the following processing is performed:

[0017] Based on the model identifier, a partition index is performed on the local historical database to locate the target historical data sub-database; multiple cosine similarities are calculated between the multimodal environment feature vector and multiple historical environment feature vectors in the target historical data sub-database; using the multiple cosine similarities, N historical environment feature vectors that meet a preset similarity threshold are selected from the multiple historical environment feature vectors, and then N historical no-intervention operation records are extracted from the N historical environment feature vectors; using the N cosine similarities as weights, the N historical no-intervention operation records are weighted and fused to output the no-intervention operation status.

[0018] In one implementation, the non-interventional operation state and multimodal environment feature vectors are input into the device control analysis model for condition adaptability analysis, and baseline control parameters are output. Before this, the following processing is performed:

[0019] The system interactively obtains a sample non-interventional operation record set, a sample environmental feature record set, and a sample control output set; decomposes the sample control output set to obtain a K-dimensional control parameter subset; performs multivariate regression analysis on the sample non-interventional operation record set, the sample environmental feature record set, and the first-dimensional control parameter subset to output a first parameter tuning regression function; similarly, performs multivariate regression analysis on the sample non-interventional operation record set, the sample environmental feature record set, and the K-dimensional independent control parameter subset to output K parameter tuning regression functions; loads the K parameter tuning regression functions into K device control analysis channels of a parallel computing architecture, and then integrates the K device control analysis channels in parallel to obtain the device control analysis model.

[0020] In one implementation, the non-interventional operation state and multimodal environment feature vectors are input into the device control analysis model for condition adaptability analysis, and baseline control parameters are output. The following processing is also performed:

[0021] The non-interventional operation state and multimodal environment feature vectors are input in parallel into the K equipment control analysis channels in the equipment control analysis model. The K equipment control analysis channels independently perform operating condition adaptability analysis and output K-dimensional benchmark control components. Parameter tuning processing is performed on the K-dimensional benchmark control components, and the benchmark control parameters are output.

[0022] In one implementation, the following processing is also performed:

[0023] The offset distance between the updated position coordinates and the real-time position coordinates of the environment adaptive detector is dynamically monitored. If the offset distance meets the preset displacement threshold, the activation state of the multimodal sensing array is updated in the working space according to the updated position coordinates and the effective detection radius to obtain a local updated sensing array. The reference control parameters are dynamically adjusted and updated according to the feedback sensing data of the local updated sensing array.

[0024] The technical solution provided in this invention has at least the following technical effects or advantages:

[0025] The method provided in this invention performs local environmental perception processing in the workspace based on the real-time position coordinates of an environmental adaptive detector to obtain a multimodal environmental feature vector. The model identifier of the environmental adaptive detector and the multimodal environmental feature vector are used as unsupervised similarity matching retrieval conditions to search the local historical database and obtain the uninterrupted operating state. The uninterrupted operating state and the multimodal environmental feature vector are input into a device control analysis model for condition adaptability analysis, and benchmark control parameters are output. The environmental adaptive detector is started after being adjusted to the benchmark control parameters. Environmental fluctuation jumps are identified in the updated environmental feature vector of the local environmental perception to match the jump control parameters. After the environmental adaptive detector is gradually adjusted from the benchmark control parameters to the jump control parameters, condition adaptability analysis is performed based on the real-time operating conditions and the updated environmental feature vector, using the jump control parameters as a starting point, to obtain updated control parameters. Based on the environmental fluctuation jump identification results of the working environment, closed-loop dynamic adjustment of the operating parameters of the environmental adaptive detector is performed until the detection task ends. It achieves the technical effect of autonomously optimizing the operating parameters of the environment adaptive detector in complex and ever-changing environments, significantly improving the stability, reliability and detection accuracy of the environment adaptive detector. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This invention illustrates a flowchart of the method for dynamically adjusting the operating parameters of an environment-adaptive detector.

[0028] Figure 2 The diagram illustrates the process of obtaining the non-interventional operating state in the dynamic adjustment method of the operating parameters of the environmental adaptive detector provided by the present invention. Detailed Implementation

[0029] This invention provides a method for dynamically adjusting the operating parameters of an environment-adaptive detector. It addresses the technical problem in existing technologies where the operating parameters of environment-adaptive detectors often fail to adapt to environmental changes in complex operating scenarios, leading to insufficient reliability and detection accuracy. This method achieves the technical effect of autonomously optimizing the operating parameters of the environment-adaptive detector in complex and changing environments, significantly improving the detector's stability, reliability, and detection accuracy.

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0031] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0032] Unless otherwise expressly stated, throughout the specification and claims, the term "comprising" or its variations such as "including" or "comprises" shall be understood to include the stated elements or components without excluding other elements or other components.

[0033] An embodiment of the present invention provides a flowchart of a method for dynamically adjusting the operating parameters of an environment-adaptive detector. (See attached flowchart.) Figure 1 The method includes:

[0034] Step A100: Based on the real-time position coordinates of the environment adaptive detector, perform local environment perception processing in the workspace to obtain multimodal environment feature vectors.

[0035] In one implementation, based on the real-time position coordinates of the environment adaptive detector, local environment perception processing is performed in the workspace to obtain a multimodal environment feature vector. The method step A100 provided by this invention includes:

[0036] Step A110: Based on the model identifier, network call the device parameters to obtain the effective detection radius.

[0037] Step A120: Using the real-time location coordinates as the center, define the effective detection boundary based on the effective detection radius.

[0038] Step A130: Based on the effective detection boundary, perform local activation of the multimodal sensing array in the workspace to obtain a local environment sensing array, wherein the device configuration of each sensing node in the multimodal sensing array includes a temperature sensor, a humidity sensor, a light intensity sensor, and an ultraviolet sensor.

[0039] Step A140: The multimodal environmental feature vector is obtained by performing multimodal temporal feature fusion processing on the multimodal sensing data transmitted back by the local environment sensing array.

[0040] Specifically, this embodiment uses the model identifier (unique device ID) of the environmentally adaptive detector to remotely access the manufacturer's database or cloud parameter library to automatically obtain the effective detection radius value of that model under standard operating conditions. This radius value is determined by the detector's optical characteristics and signal processing capabilities; for example, the theoretical detection distance of a specific model in a 25°C environment is 10 meters. This network-based call ensures that the parameters strictly match the device's physical characteristics, avoiding errors from manual configuration.

[0041] Centered on the real-time location coordinates, an effective detection boundary (circular) is defined based on the effective detection radius. Within this defined circular boundary, a group of sensor nodes is deployed or activated. Each node integrates four sensors: temperature, humidity, light intensity, and ultraviolet radiation. For example, 10 sensing nodes are evenly distributed within the boundary, with the node spacing dynamically adjusted according to the radius. Upon activation, each node synchronously begins data acquisition, forming a dense sensing network centered on the detector, covering all dimensions of environmental parameters including heat, humidity, light, and ultraviolet radiation.

[0042] After receiving the multimodal sensing data (real-time data stream) returned by the local environment sensing array, timestamp alignment and missing value compensation are first performed; then key temporal features (such as temperature change rate within 1 minute and light fluctuation variance) are extracted; finally, the features of the four types of sensors are weighted and fused into a unified vector to obtain the multimodal environment feature vector. The multimodal environment feature vector condenses the dynamic characteristics of the local environment within the effective detection boundary, providing quantitative input for subsequent parameter optimization.

[0043] Step A200: Using the model identifier and multimodal environmental feature vector of the environmental adaptive detector as unsupervised similarity matching retrieval conditions, search the local historical database to obtain the uninterrupted operation status.

[0044] In one implementation, Figure 2 As shown, the method step A200 of the present invention includes using the model identifier of the environmental adaptive detector and the multimodal environmental feature vector as unsupervised similarity matching retrieval conditions to search the local historical database and obtain the uninterrupted operation status.

[0045] Step A210: Perform a partition index on the local historical database based on the model identifier to locate the target historical data sub-database.

[0046] Step A220: Calculate multiple cosine similarities between the multimodal environment feature vector and multiple historical environment feature vectors in the target historical data sub-database.

[0047] Step A230: After selecting N historical environment feature vectors that meet the preset similarity threshold from the multiple historical environment feature vectors using the multiple cosine similarity, N historical uninterrupted operation records of the N historical environment feature vectors are mapped and extracted.

[0048] Step A240: Using N cosine similarities as weights, the N historical uninterrupted operation records are weighted and fused together to output the uninterrupted operation state.

[0049] Specifically, in this embodiment, the local historical database is located in the local cloud. The database is divided into storage areas according to different models (e.g., independent partitions for models A / B / C), and historical operating data of different detectors are stored in partitions based on model identifiers. Based on this, a partition search is first performed in the local historical database using the detector's model identifier (e.g., a unique device ID), ensuring that the search scope is strictly limited to a subset of historical data matching the current detector's physical characteristics. For example, a detector with the model "UV-2030" only accesses its corresponding partition, excluding interference from data from other models and ensuring the accuracy of subsequent matching.

[0050] Calculate multiple cosine similarities between the multimodal environment feature vector and multiple historical environment feature vectors in the target historical data sub-database. The closer the value is to 1, the more similar the environmental states are.

[0051] Set a similarity threshold (e.g., ≥0.9) to filter the calculation results, selecting N historical environmental feature vectors that meet the similarity standard, where N is an unspecified positive integer. Each vector is associated with a historical uninterrupted operation record (e.g., parameters such as gain and bias). For example, select 5 historical records with a similarity >0.92 and extract their corresponding detector parameter configurations.

[0052] The mapping process extracts N historical uninterrupted operation records from the N historical environmental feature vectors. These records are performance data automatically recorded during continuous and stable operation in the corresponding historical environmental feature vector's working environment, under natural working conditions without human intervention (such as manual parameter adjustment, calibration, or debugging). This data objectively reflects the detector's working state without external interference, providing empirical basis for subsequent dynamic parameter adjustment. For example, when the detector operates stably at a gain of 25dB for 10 minutes in a high-temperature environment, this parameter combination and the corresponding detector performance are automatically recorded, forming a historical uninterrupted operation record.

[0053] Using N cosine similarities as weights, the N historical uninterrupted operation records are weighted and fused to output the uninterrupted operation state. The uninterrupted operation state provides a relatively objective prediction of the real-time performance indicators (such as signal-to-noise ratio and sensitivity) of the detector after it has been running stably for a specific period of time in the working environment of the multimodal environmental feature vector without human intervention (such as manual parameter adjustment, calibration or debugging).

[0054] This embodiment utilizes an unsupervised similarity matching mechanism to accurately reuse historical best operating experience, providing an objective and adaptive decision-making basis for optimizing detector parameters in the current environment, and laying a reliable foundation for subsequent dynamic parameter tuning.

[0055] Step A300: Input the non-interventional operation state and multimodal environment feature vectors into the device control analysis model to perform condition adaptability analysis and output the baseline control parameters.

[0056] In one implementation, the non-interventional operation state and multimodal environment feature vectors are input into the device control analysis model for condition adaptability analysis, and baseline control parameters are output. Prior to this, the method step A300 provided by the present invention includes:

[0057] Step A3001: Interact to obtain the sample non-intervention operation record set, the sample environment feature record set, and the sample control output set.

[0058] Step A3002: Decompose the sample control output set to obtain a K-dimensional control parameter subset.

[0059] Step A3003: Perform multiple regression analysis on the sample non-intervention operation record set, sample environmental feature record set, and first-dimensional control parameter subset, and output the first parameter tuning regression function.

[0060] Step A3004: Similarly, perform multiple regression analysis on the sample non-intervention operation record set, sample environmental feature record set, and K-dimensional independent control parameter subset, and output K parameter tuning regression functions.

[0061] Step A3005: After loading the K parameter-tuning regression functions into the K device control analysis channels of the parallel computing architecture, the device control analysis model is obtained by parallel integration of the K device control analysis channels.

[0062] In one implementation, the non-interventional operation state and multimodal environment feature vectors are input into the device control analysis model for condition adaptability analysis, and baseline control parameters are output. The method step A300 provided by this invention includes:

[0063] Step A310: Input the non-interventional operation state and multimodal environment feature vectors into the K equipment control analysis channels in the equipment control analysis model in parallel, and have the K equipment control analysis channels independently perform the working condition adaptability analysis and output K-dimensional benchmark control components.

[0064] Step A320: Perform parameter tuning processing on the K-dimensional reference control components and output the reference control parameters.

[0065] Specifically, in this embodiment, the reference control parameters are intervention control parameters, used to provide the detector with the optimal initial operating configuration for environmental adaptability. They are core control commands that connect historical experience with the real-time environment. By dynamically balancing the detector's performance indicators (signal-to-noise ratio, sensitivity, response time), they ensure that the detector starts up and operates with maximum efficiency during the environmental steady-state phase. The method for obtaining the reference control parameters is as follows:

[0066] First, a device control analysis model is constructed. The device control analysis model can perform working condition adaptability analysis based on the input non-interventional operating state and multimodal environmental feature vectors, and output the benchmark control parameters. Under the adjustment and intervention of the benchmark control parameters on the detector, the detector can ensure high-performance operation in the working environment.

[0067] Specifically, the system interactively obtains a sample uninterrupted operation record set, a sample environmental feature record set, and a sample control output set. The sample uninterrupted operation record set contains the automatically recorded operating parameters (such as gain and bias) and corresponding performance indicators (such as signal-to-noise ratio) of the same model detector under natural, uninterrupted conditions. The sample environmental feature record set synchronously collects corresponding environmental parameter feature vectors (such as fused data of temperature, humidity, and light intensity). The sample control output set covers the optimal set of operating parameters actually used by the detector in historical detection missions. These data originate from real operation logs, covering diverse environmental conditions (such as high temperature and strong light interference), providing comprehensive empirical evidence for model training.

[0068] Step A3002: Decompose the sample control output set to obtain a K-dimensional control parameter subset.

[0069] The sample control output set is split into K-dimensional subsets (independent subsets) of control parameters based on their physical characteristics. For example, the gain parameter is assigned to the dimensional 1 subset (historical gain value set), the bias voltage to the dimensional 2 subset (historical voltage value set), and the response time to the dimensional K subset (historical response time set).

[0070] The control results are split here to ensure that each adjustable parameter is modeled independently, avoiding the physical coupling effect between parameters from interfering with the analysis results. The split subsets each carry the optimization objective of a single parameter dimension, supporting the subsequent divide-and-conquer strategy.

[0071] Step A3003: Perform multiple regression analysis on the sample non-intervention operation record set, sample environmental feature record set, and first-dimensional control parameter subset, and output the first parameter tuning regression function.

[0072] Using a set of environmental feature records and a set of uninterrupted operation records as input, and a subset of first-dimensional parameters (such as gain parameters) as the output target, a regression model is constructed: This model integrates environmental features (such as temperature slope and light intensity) with the current uninterrupted gain value, and generates a prediction function for the first-dimensional parameters and a first-parameter-tuning regression function (such as gain = f(temperature, humidity, historical gain)) through multiple linear regression or support vector regression algorithms. This first-parameter-tuning regression function quantifies the mapping relationship between environmental variables and target parameters, laying the foundation for single-parameter optimization.

[0073] Similarly, multiple regression analysis is performed on the sample non-intervention operation record set, sample environmental feature record set, and K-dimensional independent control parameter subset to output K parameter tuning regression functions, fully covering the dynamic optimization logic of all adjustable parameters.

[0074] A parallel prediction framework is constructed by deploying K parameter-tuned regression functions to K independent computing channels on an FPGA or GPU. Each channel is dedicated to real-time computation of a single parameter dimension (e.g., channel 1 handles gain prediction), and serial latency is eliminated through hardware-level parallelism. The channel outputs integrate a unified interface to form a complete device control analysis model, achieving millisecond-level response (≤10ms) to meet the real-time decision-making requirements in dynamic environments.

[0075] The non-interventional operation state and multimodal environment feature vectors are synchronously distributed to the K equipment control analysis channels in the equipment control analysis model. The K equipment control analysis channels independently calculate based on their respective channel-specific regression functions, and finally summarize and output K-dimensional benchmark control components to characterize the collaborative optimization direction of multi-dimensional parameters.

[0076] The K-dimensional baseline control components (i.e., the multidimensional normalized adjustment tendency values ​​output by the model) are converted into physically executable parameters that can directly drive the environmentally adaptive detector through parameter tuning. First, the weight coefficients of each parameter are dynamically allocated according to the environmental sensitivity (e.g., the weight of shortening the response time is increased in high-temperature environments). Then, the normalized component values ​​are mapped to actual parameter values ​​in combination with the physical range of the device (e.g., the gain component 0.85 is mapped to 25dB). Finally, a set of baseline control parameters (e.g., gain, bias voltage, etc.) that can be directly configured for the detector is generated, ensuring that the detector starts up and runs with the optimal initial configuration under the current environmental conditions, and provides a reliable reference for dynamic parameter tuning in the event of subsequent environmental changes.

[0077] Step A400: After adjusting the parameters of the environmental adaptive detector to the reference control parameters, start the operation.

[0078] Step A500: Perform environmental fluctuation jump identification on the updated environmental feature vector of the local environment perception to match the jump control parameters.

[0079] In one implementation, environmental fluctuation jumps are identified in the updated environmental feature vector of the local environment perception to match jump control parameters. The method step A500 provided by this invention includes:

[0080] Step A510: After performing short-term multimodal data backtracking on the working environment, perform dynamic fluctuation feature clustering analysis of environmental parameters to obtain a multi-scale environmental fluctuation interval set.

[0081] Step A520: Preset the jump backup parameter set according to the multi-scale environmental fluctuation interval set.

[0082] Step A530: Identify environmental parameter jump events by comparing the updated environmental feature vectors of the local environment perception and the multi-scale environmental fluctuation interval set in real time.

[0083] Step A540: Based on the fluctuation range of the changing environment corresponding to the changing environmental parameter change event, locate and extract the changing control parameters from the set of changing backup parameters.

[0084] In one implementation, after performing short-term multimodal data backtracking on the working environment, dynamic fluctuation feature clustering analysis of environmental parameters is performed to obtain a multi-scale environmental fluctuation interval set. The method step A510 provided by this invention includes:

[0085] Step A511: Starting from the acquisition timestamp of the multimodal environment feature vector, perform short-term multimodal data backtracking within a preset backtracking time window to obtain the multimodal historical perception sequence.

[0086] Step A512: After temporally aligning the multimodal historical sensing sequence, extract W multimodal sensing slices at equal intervals.

[0087] Step A513: Based on the W multimodal sensing slices, perform dynamic fluctuation feature clustering of environmental parameters to extract the multimodal fluctuation interval scale sequence.

[0088] Step A514: Combine and enumerate the multimodal fluctuation interval scale sequence to output the multi-scale environmental fluctuation interval set.

[0089] It should be understood that if the local environment in which the detector is located changes significantly, it is necessary to intervene and adjust the detector's baseline control parameters according to the environmental changes, so as to ensure that the detector's performance is not affected even if the working environment changes.

[0090] Specifically, during the operation of the detector, based on the acquisition time of the current environmental feature vector, historical multimodal data (such as the temperature, humidity, light intensity, and ultraviolet intensity sequence of the past 3 hours) within a fixed time window are traced back.

[0091] By performing dynamic fluctuation feature clustering analysis on these time-series data, environmental fluctuation patterns at different time scales (such as second-level abrupt changes and minute-level gradual changes) are identified, ultimately generating a multi-scale environmental fluctuation interval set covering various fluctuation characteristics. This multi-scale environmental fluctuation interval set provides a quantitative classification standard for identifying environmental abrupt changes.

[0092] The specific calculation process for obtaining the multi-scale environmental fluctuation interval set is as follows:

[0093] Using the generation time of the real-time environmental feature vector as the time origin, historical multimodal data is traced back for a preset time period. The tracing range covers sensor data across all dimensions, including temperature, humidity, light intensity, and ultraviolet intensity, forming an original historical sensing sequence sorted by timestamp, i.e., the multimodal historical sensing sequence. The multimodal historical sensing sequence completely records the trajectory of environmental changes within the tracing window, providing a data foundation for fluctuation feature analysis.

[0094] The multimodal historical sensing sequences obtained through retrospection are timestamped to eliminate time discrepancies from different sensors. Subsequently, fixed time segments are extracted at equal intervals (e.g., one slice every 15 minutes) to generate W temporally continuous multimodal sensing slices. Each slice contains synchronized data from all sensors at the same time (e.g., temperature, humidity, illumination, and UV values), ensuring spatiotemporal consistency in the extraction of fluctuation features.

[0095] For each multimodal sensing slice, its dynamic fluctuation characteristics (such as temperature change rate, illumination fluctuation variance, and ultraviolet intensity range) are calculated. Based on these characteristics, an unsupervised clustering algorithm (such as DBSCAN) is used to classify the slice into different fluctuation mode categories (such as "sudden temperature rise type" and "illumination pulse type"). Each cluster represents a type of environmental fluctuation mode, and its physical feature boundaries (such as temperature change rate > 10% / s) are extracted as fluctuation interval scale sequences, forming a multimodal fluctuation interval scale sequence. Each modal fluctuation interval scale sequence represents the fluctuation amount of the corresponding modal environmental parameter at multiple scales from small to large.

[0096] By combining and enumerating the multimodal fluctuation interval scale sequence, the multi-scale environmental fluctuation interval set is output.

[0097] After extracting the median of the multi-scale environmental fluctuation interval set to obtain the multi-scale environmental fluctuation median set, the multi-scale environmental fluctuation median set is used to update the multi-modal environmental feature vector in step A300 to obtain the multi-modal environmental median feature vector set. The multi-modal environmental median feature vector set and the non-interventional operation state input device control analysis model are then used for condition adaptability analysis to output the jump backup parameter set.

[0098] The system receives the latest environmental feature vectors uploaded by the local sensing array in real time and matches them item by item with a set of multi-scale environmental fluctuation intervals. When a feature value breaks through the physical boundary of a certain multi-scale environmental fluctuation interval (such as the temperature change rate jumping from 5% / s to 20% / s), an environmental parameter jump event is triggered. For example, if the light intensity jumps by 50 Lux within 1 second, it matches the "light pulse type" interval and is determined to be a jump event.

[0099] Based on the fluctuation range of the changing environment corresponding to the changing environmental parameter change event, the changing control parameters are located and extracted from the set of changing backup parameters. The changing control parameters are directly used for dynamic parameter adjustment of the detector to achieve millisecond-level response during sudden environmental changes.

[0100] This embodiment solves the problem of response lag in traditional progressive parameter tuning under sudden environmental changes. Through a closed loop of "fluctuation mode pre-classification → parameter pre-loading → boundary trigger jump", it achieves the technical effect of ensuring that the detector maintains stable performance in sudden scenarios such as sudden changes in light and temperature, and avoids signal distortion or detector saturation.

[0101] Step A600: After gradually adjusting the environmental adaptive detector from the reference control parameter to the jump control parameter, take the jump control parameter as the starting point, perform a condition adaptability analysis based on the real-time operating conditions and the updated environmental feature vector to obtain the updated control parameter.

[0102] When environmental fluctuations trigger a transition event (such as a sudden 50% increase in illumination), the detector's operating parameters are gradually transitioned from baseline control parameters (such as a gain of 20dB and a voltage of 5V) to transition control parameters (such as a gain of 18dB and a voltage of 5.2V). This gradual adjustment is achieved through linear interpolation or exponential smoothing algorithms; for example, the gain is reduced from 20dB to 18dB within 0.5 seconds to avoid signal distortion caused by sudden parameter changes. After the transition is complete, using the transition control parameters as a new starting point, and combining real-time environmental feature vectors (such as updated temperature and illumination data) and detector operating conditions (such as the current signal-to-noise ratio), the device control analysis model (see step A300) dynamically analyzes and outputs updated control parameters (such as a gain of 17.8dB and a voltage of 5.3V) to ensure the detector continuously adapts to environmental changes.

[0103] Step A700: Based on the environmental fluctuation and change identification results of the working environment, perform closed-loop dynamic adjustment of the working parameters of the environmental adaptive detector until the detection mission ends.

[0104] Based on the environmental fluctuation and abrupt change identification results (such as the temperature step event identified in step A530), closed-loop dynamic control is continuously executed: after each environmental change triggers a parameter change, a new cycle of "real-time monitoring → operating condition analysis → parameter update" is immediately initiated. For example, if a fluctuation in illumination is detected again after the abrupt change, the environmental feature vector and the real-time operating performance input model of the detector at that time are re-extracted and updated to generate new control parameters and adjust the detector configuration. This closed-loop process responds to environmental changes in real time, dynamically balances detector performance indicators (signal-to-noise ratio, response time, etc.), until the detection mission is completed or the environment returns to steady state, ultimately ensuring that the detector outputs stable and reliable detection data in complex environments throughout the entire process.

[0105] This embodiment achieves the technical effect of autonomously optimizing detector operating parameters in complex and variable environments, significantly improving the stability, reliability, and detection accuracy of the environment-adaptive detector.

[0106] In one implementation, the method further includes:

[0107] Step A515: Dynamically monitor the offset distance between the updated position coordinates and the real-time position coordinates of the environment adaptive detector.

[0108] Step A516: If the offset distance meets the preset displacement threshold, then the activation state of the multimodal sensing array is updated in the working space according to the updated position coordinates and the effective detection radius to obtain the local update sensing array.

[0109] Step A517: Based on the feedback sensing data from the local update sensing array, dynamically adjust and update the reference control parameters.

[0110] It should be understood that the displacement of the detector in the working environment will also cause distortion of the environmental data transmitted back by the sensor array, which in turn will affect the setting of the detector's control parameters (reference control parameters) to counteract environmental influences.

[0111] Based on this, this embodiment continuously tracks the updated position coordinates of the detector and compares them with the previously recorded real-time position coordinates to calculate the spatial offset distance between them. When the calculated offset distance exceeds a preset threshold, a new circular monitoring area is delineated using the current updated position coordinates as the center and the effective detection radius corresponding to the detector model (e.g., 10 meters). Within this area, new sensor nodes (e.g., temperature and light sensors) are activated or deployed, redundant nodes outside the original boundary are deactivated, and a localized updated sensing array focused on the new position is generated. For example, after the detector moves 10 meters, nodes in the original array far from the new position are deactivated, and three new nodes are added around the new position, forming an updated dense sensing network.

[0112] After the local update sensing array is activated, it transmits environmental data from the new location in real time (e.g., a sudden temperature drop of 2°C, an increase in illumination of 100 Lux). Based on the transmitted data, it generates an updated multimodal environmental feature vector, inputs it into the device control analysis model (see A300), and re-executes the operational condition adaptability analysis, outputting baseline control parameters adapted to the new environment (e.g., gain reduced from 22dB to 20dB). This process ensures that the detector's operating parameters remain strictly matched to the local environment after relocation, achieving the technical effect of ensuring that the detector always acquires accurate local environmental characteristics during movement and maintaining strong consistency between parameter optimization and environmental changes.

[0113] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for dynamically adjusting the operating parameters of an environment-adaptive detector, characterized in that, include: Based on the real-time position coordinates of the environmental adaptive detector, local environmental perception processing is performed in the workspace to obtain multimodal environmental feature vectors. The model identifier and multimodal environmental feature vector of the environmental adaptive detector are used as unsupervised similarity matching retrieval conditions to search the local historical database and obtain the uninterrupted operation status. The non-interventional operation state and multimodal environment feature vectors are input into the device control analysis model to perform condition adaptability analysis and output baseline control parameters. After adjusting the parameters of the environmental adaptive detector to the reference control parameters, it is started to run. Environmental fluctuation jumps are identified in the updated environmental feature vectors of local environment perception in order to match jump control parameters; After the environmental adaptive detector is gradually adjusted from the reference control parameter to the jump control parameter, the working condition adaptability analysis is performed based on the real-time operating conditions and the updated environmental feature vector, starting from the jump control parameter, to obtain the updated control parameter. Based on the identification results of environmental fluctuations in the working environment, the working parameters of the environmental adaptive detector are dynamically adjusted in a closed loop until the detection mission ends.

2. The method for dynamically adjusting the operating parameters of an environment-adaptive detector as described in claim 1, characterized in that, Environmental fluctuation transitions are identified in the updated environmental feature vectors of the local environment perception system to match transition control parameters, including: After performing short-term multimodal data backtracking on the working environment, a cluster analysis of the dynamic fluctuation characteristics of environmental parameters is performed to obtain a multi-scale environmental fluctuation interval set. A set of backup parameters for jumps is pre-set based on the multi-scale environmental fluctuation range set; By comparing the updated environmental feature vectors of the local environment perception and the multi-scale environmental fluctuation interval set in real time, environmental parameter jump events can be identified. Based on the fluctuation range of the changing environment corresponding to the changing environmental parameter change event, the changing control parameters are located and extracted from the set of changing backup parameters.

3. The method for dynamically adjusting the operating parameters of an environment-adaptive detector as described in claim 1, characterized in that, Based on the real-time position coordinates of the environment adaptive detector, local environment perception processing is performed in the workspace to obtain a multimodal environment feature vector, including: The effective detection radius is obtained by connecting the device parameters to the network based on the model identifier. Using the real-time location coordinates as the center, the effective detection boundary is defined based on the effective detection radius; Based on the effective detection boundary, a multimodal sensor array is locally activated in the workspace to obtain a local environment sensing array. The device configuration of each sensing node in the multimodal sensor array includes a temperature sensor, a humidity sensor, a light intensity sensor, and an ultraviolet sensor. The multimodal environmental feature vector is obtained by performing multimodal temporal feature fusion processing on the multimodal sensing data transmitted back from the local environment sensing array.

4. The method for dynamically adjusting the operating parameters of an environment-adaptive detector as described in claim 2, characterized in that, After performing short-term multimodal data backtracking on the working environment, a cluster analysis of the dynamic fluctuation characteristics of environmental parameters was performed to obtain a set of multi-scale environmental fluctuation intervals, including: Starting from the acquisition timestamp of the multimodal environment feature vector, short-term multimodal data backtracking is performed within a preset backtracking time window to obtain a multimodal historical perception sequence; After aligning the multimodal historical sensing sequence in time, W multimodal sensing slices are extracted at equal intervals; Based on the W multimodal sensing slices, perform dynamic fluctuation feature clustering of environmental parameters to extract the multimodal fluctuation interval scale sequence; Combine and enumerate the multimodal fluctuation interval scale sequence to output the multi-scale environmental fluctuation interval set.

5. The method for dynamically adjusting the operating parameters of an environment-adaptive detector as described in claim 1, characterized in that, Using the model identifier and multimodal environmental feature vector of the environmental adaptive detector as unsupervised similarity matching retrieval conditions, the local historical database is searched to obtain the uninterrupted operation status, including: Based on the model identifier, perform a partition index on the local historical database to locate the target historical data sub-database; Calculate multiple cosine similarities between the multimodal environment feature vector and multiple historical environment feature vectors in the target historical data sub-database; After selecting N historical environment feature vectors that meet the preset similarity threshold from the multiple historical environment feature vectors using the multiple cosine similarity, N historical uninterrupted operation records of the N historical environment feature vectors are extracted by mapping. Using N cosine similarity scores as weights, the N historical uninterrupted operation records are weighted and fused to output the uninterrupted operation state.

6. The method for dynamically adjusting the operating parameters of an environment-adaptive detector as described in claim 1, characterized in that, The non-interventional operation state and multimodal environment feature vectors are input into the device control analysis model for condition adaptability analysis, and baseline control parameters are output. Prior to this, the following steps are included: Interactively obtain the sample non-intervention operation record set, the sample environmental feature record set, and the sample control output set; Decompose the sample control output set to obtain a K-dimensional subset of control parameters; A multivariate regression analysis was performed on the sample non-intervention operation record set, the sample environmental feature record set, and the first-dimensional control parameter subset to output the first parameter tuning regression function; Similarly, multiple regression analysis is performed on the sample non-intervention operation record set, sample environmental feature record set, and K-dimensional independent control parameter subset to output K parameter tuning regression functions; After loading the K parameter-tuned regression functions into the K device control analysis channels of the parallel computing architecture, the device control analysis model is obtained by integrating the K device control analysis channels in parallel.

7. The method for dynamically adjusting the operating parameters of an environment-adaptive detector as described in claim 6, characterized in that, The non-interventional operation state and multimodal environment feature vectors are input into the device control analysis model for condition adaptability analysis, and the baseline control parameters are output, including: The non-interventional operation state and multimodal environment feature vectors are input in parallel into the K equipment control analysis channels in the equipment control analysis model. The K equipment control analysis channels independently perform the working condition adaptability analysis and output K-dimensional benchmark control components. Perform parameter tuning processing on the K-dimensional reference control components and output the reference control parameters.

8. The method for dynamically adjusting the operating parameters of an environment-adaptive detector as described in claim 3, characterized in that, Also includes: The offset distance between the updated position coordinates and the real-time position coordinates of the environmental adaptive detector is dynamically monitored. If the offset distance meets the preset displacement threshold, the activation state of the multimodal sensing array is updated in the workspace according to the updated position coordinates and the effective detection radius to obtain a local update sensing array. Based on the feedback sensing data from the local update sensing array, the reference control parameters are dynamically adjusted and updated.

Images