6939results about How to "Ensure consistency" patented technology

The invention provides a digital power grid power cloud resource real-time hierarchical distribution method, which comprises the steps of dividing power cloud resources into a plurality of sub-resource pools according to a physical architecture and a topological structure of a digital power grid, setting a special resource pool for key equipment and sensitive loads of the power grid, and dynamically adjusting network topology and a routing strategy. A low-delay and high-reliability network transmission channel is provided through a priority queue and a bandwidth guarantee mechanism; for each sub-resource pool, extracting periodicity, tendency and randomness characteristics of a load by analyzing historical load data and a current system state of power grid equipment, and updating a load prediction result in real time; after incremental resources are migrated, a power grid control request is routed to a corresponding computing node by adopting a consistent Hash algorithm in a sub-resource pool, load balancing is realized by dynamically adjusting the number and distribution of virtual nodes of a Hash ring, and weights of the virtual nodes are dynamically adjusted according to load optimization indexes of the nodes so as to adapt to power grid load changes.

The invention belongs to the technical field of multi-agent collaboration, and provides an electric power cross-modal knowledge fusion multi-agent collaboration processing method and system, and the technical scheme is as follows: analyzing a task execution command instruction input by a user; key elements of the task are extracted, the task is decomposed into a plurality of sub-tasks according to the extracted key elements, and logic association and priorities among the sub-tasks are identified; according to the logical association and priority among the subtasks, allocating the subtasks to the corresponding agents; each agent executes the corresponding subtask according to a set execution path, the executed task result is shared to other agents to cooperatively complete the complex task, and finally a cooperative processing result is obtained through integration. Efficient execution of complex tasks such as document generation, review and management is realized, and smooth completion of the tasks is ensured.

The invention belongs to the technical field of artificial intelligence, and relates to a power document intelligent question and answer method and system based on a large language model. The method comprises the steps of obtaining document data according to a preset period; preprocessing the document data to obtain an adaptive document; constructing a vector retrieval library through an embedded model according to the adaptive document; obtaining a user query question, inputting the user query question into a vector retrieval library, performing retrieval based on a vector similarity retrieval method and a keyword retrieval method to obtain related text blocks, and screening out the most related text block from the related text blocks through a two-stage rearrangement technology; a user queries the question and the most relevant text block and inputs the question and the most relevant text block into the large language model to obtain preliminary intelligent answer information; and correcting the preliminary intelligent answer information based on a preset self-provincial inversion mechanism to obtain final intelligent answer information. Multiple rounds of questions and answers and context understanding are supported based on a correction mechanism, context consistency can be kept in continuous query of complex technical questions, and the accuracy of multiple rounds of questions and responses is ensured.

The invention provides a textile defect positioning method, which comprises the following steps: according to an extracted defect feature map, fusing feature maps of different scales, extracting multi-scale defect features, and comprehensively depicting space distribution and combination information of defects; aiming at the condition that multiple flaws exist at the same time, a multi-task learning model is constructed, flaw detection, classification and positioning tasks are carried out at the same time, and a back propagation algorithm is adopted to optimize the multi-task learning model; positioning the defect area by adopting an image segmentation algorithm, obtaining the contour of the defect area according to the positioning result of the defect area and by adopting a Canny edge detection algorithm, and calculating the area and position information of the defect; and based on an evaluation result, a dynamic threshold adjustment algorithm is adopted, a defect judgment threshold is adjusted according to a production batch and a product type, and a threshold parameter is dynamically updated by analyzing historical data and a current production condition.

The invention relates to a temperature control intelligent optimization method and system for adhesive tape production. The method comprises the following steps that S1, production data in the drying process are collected through a multi-mode sensor; s2, constructing a state prediction model based on GNN and LSTM, inputting the collected production data into the prediction model, and dynamically predicting the state of each temperature control area; s3, modeling each temperature control area into an intelligent agent based on the predicted state between the temperature control areas, and realizing dynamic distribution of heat resources of each heating interval by training a cooperative control strategy between the intelligent agents; s4, a fuzzy control optimization scheme incorporating environment compensation: for short-term changes of environment temperature and humidity, dynamically adjusting temperature zone gradient distribution between drying sections by using fuzzy control, and optimizing energy efficiency; and S5, the temperature control module adjusts the temperature and the wind speed of the temperature control area in real time according to the optimization schemes obtained in the steps S3 and S4. The intelligent level of the temperature control system in the adhesive tape production drying process is comprehensively improved.

The invention relates to the technical field of coordinate conversion, in particular to a CAD and BIM model coordinate automatic conversion method and system based on coordinate mapping, and the method comprises the following steps: extracting CAD building two-dimensional coordinates input by reference to a building drawing, constructing an initial two-dimensional triangular mesh, extracting key geometric feature points of a building, and combining the floor height parameter and corner information of the building to obtain a CAD and BIM model coordinate automatic conversion model; and mapping the two-dimensional grid to a three-dimensional coordinate space to generate initial building three-dimensional point cloud data. According to the method, a high-quality initial triangular mesh is constructed through a Delaunay triangulation algorithm, coordinate information of a sparse position is accurately captured, a Poisson surface reconstruction technology is combined, global curved surface fitting and smoothing processing are performed on point clouds of a sparse region, and density compensation and precision improvement of coordinate point distribution are realized. According to the optimization mode, the problem of model precision loss is effectively relieved in complex scenes such as building corners, special-shaped spaces and floor height changes.

The invention discloses a method and a system for realizing data synchronization between heterogeneous databases based on logs. The method comprises the following steps: acquiring connection information of a source database and a target database; querying a table structure of the source end database and generating an SQL statement to establish a data mapping rule; matching a corresponding rule model from a mapping rule model library; if the matching is not successful, the system abnormally quits and prompts the user to create a new mapping model and add the new mapping model to the mapping rule model base; converting a table structure of the source end database into a JSON character string based on a standard field type; an SQL statement required by the target database is output through the data mapping engine module; the message monitoring program reads the data in the temporary storage area and converts the data into data meeting the requirements of the target database; formatting the converted data into SQL statements supported by a target database; and sending the generated SQL statement to a target database to complete data migration. Accurate and controllable synchronization of the data and the model in a heterogeneous environment is realized.

The invention relates to the technical field of flexible screens, and discloses a flexible screen brightness control method and device based on region segmentation, and the method comprises the steps: carrying out the multi-view image collection of a plurality of bending regions of a flexible display screen, obtaining target image data, and calculating the deformation compensation partition data of the flexible display screen; gridding brightness sampling is carried out on the deformation compensation partition data to generate brightness uniformity characteristic data, LED driving parameter calculation is carried out based on the brightness uniformity characteristic data, and a driving compensation matrix is output; inputting the driving compensation matrix into a second-order state observer for real-time current estimation and temperature feedback, and generating a partition driving control signal; and performing multi-region collaborative optimization on the partition driving control signal to realize brightness uniformity control of the flexible display screen in different bending states. According to the method, the influence of temperature drift on the display uniformity is eliminated, the stability of the driving current is improved, and the display deviation caused by temperature change is remarkably reduced.

The invention provides an OracIeRAC-based CDC synchronization method and system, and relates to the technical field of OracleRAC, and the method comprises the steps: deploying an incremental log analyzer, a session state collector and a resource monitor, and respectively collecting atomic operation, concurrent transaction information and node performance indexes; and identifying data stream features by using an intelligent classification model, constructing a priority matrix, and performing fine-grained acquisition on high-priority data based on an adaptive sampling strategy. According to the method, multi-dimensional evaluation indexes are further utilized, a time sequence prediction model is constructed to predict node loads, data synchronization tasks are distributed to low-load nodes based on a load balancing matrix, and a two-stage submission protocol is adopted to ensure data consistency. By monitoring the synchronization state in real time and dynamically adjusting the allocation strategy, the CDC synchronization efficiency and stability in the Oracle RAC environment are effectively improved, and the synchronization delay is reduced.

The invention discloses a terrain-level real-scene three-dimensional entity and element dynamic updating method, which comprises the following steps: S1, data acquisition and preprocessing: acquiring data of a target area by an acquisition mechanism based on entity and element type priorities, and preprocessing the acquired data to obtain model data of the target area; s2, change detection: comparing newly collected model data with historical model data, and identifying a change region and type; s3, entities and elements are extracted and updated, wherein the entities and the elements of the change area are processed to obtain updated models; s4, model fusion and optimization: fusing the updated entity and element model with the original model, and realizing efficient acquisition and preprocessing of target area data by using the multi-source surveying and mapping equipment and an acquisition mechanism based on entity and element type priorities; by dividing the small regions according to the region characteristics and correcting the acquisition frequency, the resource allocation is optimized, and the pertinence and efficiency of data acquisition are improved.

The invention relates to a power station screen cabinet equipment temperature early warning method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring a dual-light image of power station screen cabinet equipment to be analyzed according to acquisition parameters corresponding to current acquisition; the dual-light image comprises an infrared thermal imaging image and a visible light image; performing preprocessing, feature extraction processing, spatial registration processing and information fusion processing on the dual-light image to generate a fused image; performing temperature field reconstruction processing and hot spot detection processing on the fused image to obtain temperature field distribution information and hot spot area information; and according to the temperature field distribution information and the hot spot area information, performing temperature analysis on the power station screen cabinet equipment to obtain temperature early warning information of the power station screen cabinet equipment. By adopting the method, the temperature early warning accuracy of the power station screen cabinet equipment can be improved.

The invention provides a time-space dynamic graph construction method and system for an engineering construction project, and relates to the field of railway tunnel engineering management.The method comprises the steps that construction data entities are extracted, and an entity set is generated; taking entities in the entity set as nodes to be added into the graph, and configuring attributes for each node; adding a relation edge for each node according to a logic structure between the construction data; adding a time label and a space label to each node so as to form a data graph; associating each entity in the data atlas with a time chain and a space chain to form a data atlas attached to the time-space chain; combining a node merging strategy and a graph neural network to optimize a graph structure; and dynamically modeling dimensional changes of entities and relationships in time by means of a time sequence diagram neural network. According to the method, the dynamic change graph of the relationship between the construction data entities is constructed, so that the spatial-temporal information and the entity relationship of each process are accurately expressed and efficiently associated, and the retrieval efficiency and the management level are favorably improved.

The invention relates to the technical field of process data optimization, and discloses a traditional Chinese medicine extraction process optimization method and system based on dynamic countercurrent extraction and a medium. The method comprises the following steps: carrying out online monitoring data processing on a traditional Chinese medicinal material digital feature data set and a dynamic countercurrent extraction parameter initial data set to obtain a countercurrent extraction process real-time monitoring data stream and parameter associated data, and carrying out multi-dimensional parameter optimization processing; a dynamic countercurrent extraction process optimization parameter group is obtained and is analyzed and processed in real time, and an automatic parameter regulation and control instruction sequence and process abnormity early warning data are obtained; and performing quality characteristic data processing on the extract after execution of the automatic parameter regulation and control instruction sequence to obtain a process quality evaluation data set and a process tracing data set, and performing association analysis processing to obtain a dynamic countercurrent extraction process parameter optimization rule base. According to the invention, intelligent monitoring and automatic parameter optimization of the extraction process are realized, so that the accuracy and controllability of the traditional Chinese medicine extraction process are improved.

The invention provides a multi-modal sentiment analysis method based on unwrapping representation learning, which comprises the following steps: acquiring a multi-modal sentiment analysis data set, and extracting features of texts, audios and videos; constructing a representation learning network, extracting private features and shared features, and processing through adversarial learning optimization and time smoothness constraint; under the guidance of different language scales, the private features of the audio and the video are fused; further fusing text, audio and video features through cross-modal attention and a gating mechanism; and finally, inputting the fusion features into an emotion analysis module, and generating an emotion prediction result in combination with an emotion classification or regression task. According to the method, the consistency of the audio and video features in continuous time is considered, redundant information is effectively reduced, the time dimension consistency is ensured, deep fusion of the audio and video features is realized under text guidance, and the cross-modal fusion effect is remarkably improved.

The invention discloses a knowledge tracking cold start optimization method and system based on a large language model, and belongs to the field of knowledge tracking. The method comprises the following steps: firstly, acquiring student answering records of an online learning platform, constructing a prompt template to enable a large language model to output question solving steps, labeling knowledge concepts, and generating a binary pairing relationship between the question solving steps and the knowledge concepts; secondly, taking the problem and the knowledge concept as heterogeneous graph nodes, establishing edge connection and generating a semantic embedding vector; then, a heterogeneous graph encoder is used in combination with a graph attention network and a jump knowledge mechanism, and final node embedding of knowledge concepts and problems is output. And finally, predicting the answer of the student to the target question by comparing, learning and training the answer prediction model. According to the method, the performance and accuracy of the knowledge tracking cold start stage on the online learning platform are effectively improved.

The invention relates to the technical field of data processing, in particular to a real estate surveying and mapping data processing method and system based on intelligent data. The method comprises the following steps: acquiring multi-source real estate surveying and mapping data; performing space-time synchronization on the multi-source real estate surveying and mapping data to generate multi-source real estate surveying and mapping space alignment data; real estate space feature extraction is carried out on the multi-source real estate surveying and mapping space alignment data to obtain real estate space feature data; performing cross-domain feature fusion on the multi-source real estate surveying and mapping data based on the real estate spatial feature data to generate real estate surveying and mapping fusion spatial feature data; public surveying and mapping point confirmation is carried out on the multi-source real estate surveying and mapping data through the real estate surveying and mapping fusion spatial feature data, and real estate surveying and mapping public surveying and mapping point data are obtained. Through multi-source data space-time synchronization, feature fusion, distortion correction, high-dimensional space reconstruction, semantic tag association and chain storage, the efficiency and accuracy of real estate surveying and mapping data processing are improved.

The invention relates to the technical field of image classification, in particular to a fetal heart monitoring image classification method and system based on a neural network. The method comprises the following steps: acquiring a fetal heart monitoring image, generating a standard monitoring image through Gaussian noise elimination and spatial domain normalization, then performing multi-channel convolution on the standard image to extract image convolution features, obtaining standard fetal heart features through compression regular optimization, performing morphological reconstruction to generate a fetal heart dynamic model, and finally obtaining the fetal heart dynamic model. Baseline features are identified based on historical data, then periodic variation quantification and safe floating range derivation are performed on fusion features, and finally a multi-scale discrimination framework is constructed to realize intelligent classification of fetal heart monitoring images. The fetal heart monitoring image classification method is more stable and more accurate.

The invention provides a distributed service intelligent master selection method and system based on double-engine cooperation, and relates to the technical field of intelligent master selection, and the method comprises the steps: collecting the processing capability index, the network throughput and the resource utilization rate of a service node to construct a performance data matrix, and generating an intelligent identification code based on a node scoring function and a physical clock value; nacos and Redis signals are monitored through a double-layer event monitor, the optimal incremental step length is determined according to the score sequence distribution trend, and a target score is generated to construct a dynamic priority sequence; and periodically selecting a main node and a preheating standby node, and when the main node fails, starting increment state synchronization based on Redis affairs after voting arbitration confirmation, so as to realize main and standby non-inductive switching. According to the method, the election accuracy of the main node and the switching reliability are improved, and the system interruption risk is reduced.

The invention relates to the technical field of tea leaf detection, and discloses a tea leaf quality grade detection method and system.The method comprises the steps that a tea leaf sample is collected and pretreated; extracting chemical component indexes and physical properties of the pretreated tea leaves, and performing fusion and analysis; and establishing a tea quality grade classification model, inputting characteristics of a tea sample to be detected, outputting a detection result, and generating a tea quality grade report. Through combination of the support vector machine, the gradient boosting decision tree and the improved logistic regression algorithm, the precision of the tea quality classification model is optimized. The dynamic threshold function and the segmented excitation function are introduced to further solve the boundary fuzzy problem of the high-grade tea leaves and the low-grade tea leaves, and accurate judgment of tea leaf grade classification is achieved. And data restoration is carried out through weighted compensation correction and a recheck process, so that the output tea grade report is ensured to have high accuracy and reliability.

The invention discloses a self-adaptive data caching and synchronizing method, system and device for an edge gateway of the Internet of Things and a storage medium, and the method comprises the steps: obtaining network quality parameters, such as link bandwidth, time delay and packet loss rate, between nodes, detecting and solving instruction conflicts based on a vector clock mechanism, generating an instruction execution sequence, and carrying out the self-adaptive data caching and synchronizing of the edge gateway of the Internet of Things; local caching and incremental synchronization strategies are adopted for hotspot data, a data access mode is dynamically predicted in combination with a machine learning algorithm, and a cache updating period is optimized. The future data is predicted according to the data characteristics of the Internet of Things application, and the data cache updating period of each node is dynamically adjusted, so that each node can keep a relatively consistent updating rhythm when caching a hotspot data copy; therefore, the hot spot data copies of different nodes can be ensured to have similar accuracy and timeliness at the same time point, so that the data consistency is maintained.

The invention relates to the technical field of data synchronization, and discloses a distributed account book remote data synchronization method and system, and the method comprises the steps: carrying out the Hash mapping calculation of all nodes of a distributed account book system, and obtaining a distributed storage architecture; performing hierarchical network construction on the nodes, and setting an inter-region remote synchronization link and an intra-region local synchronization link to obtain a double-layer communication architecture; performing parallel compression coding, digital signature generation, transmission scheduling arrangement and consistency verification processing on the data blocks to obtain a node synchronous data set; calculating a correlation coefficient of node processing delay and network transmission delay to obtain a synchronous delay analysis model; based on a data block corresponding to the synchronous delay analysis model, adding a time sequence identifier and calculating an integrity check code, and generating a synchronous reliability guarantee mechanism; according to the method, the consistency and the traceability of the distributed account book data are ensured.

The invention discloses a method and device for automatically converting a cross-format document into a Markdown document and a form enhancement method and system. The Markdown document conversion method comprises the following steps: firstly, obtaining a document image; mining hidden structure information from the document image, wherein the hidden structure information comprises target elements and hierarchical relationships among the target elements; performing semantic content analysis on the document image based on the structure information to obtain a semantic content analysis result; extracting key information in the document image according to an analysis result; the key information is converted into a Markdown format, preliminary formatting is carried out in a form or list mode, and a preliminary Markdown form is obtained; and finally, optimizing the preliminary Markdown form to form a target Markdown form, so that the cross-format document is automatically converted into the Markdown. According to the method, a complex nested structure and a form with irregular layout in the document can be effectively identified and converted, and the integrity and readability of form information when the form is converted into the Markdown format are ensured.

The invention provides an Internet of Things gateway control method fusing local communication and AI computing power, and relates to the technical field of data processing, the method comprises the following steps: receiving an original data packet sent by at least one terminal device; protocol analysis, noise filtering and format conversion are carried out on the received data, an AI analysis model corresponding to the equipment type is loaded, standardized data are input into the AI analysis model, an analysis result is sent to target terminal equipment, and the analysis result and a communication result are stored in a gateway cache module; according to a preset time interval, the analysis result is transmitted to a remote cloud server, and synchronization state information and synchronization error log information are generated; according to the synchronization error log information, synchronization error analysis data is generated, network synchronization optimization and data anomaly positioning are executed, and a synchronization optimization result is generated; according to the data type and the error type, abnormal data are processed, and synchronous updating is executed; according to the invention, the data processing efficiency and accuracy are improved.

The invention relates to the technical field of production process control, in particular to a PVC developing functional granule production control method and system, and the method comprises the steps: calling a spectrometer to capture spectral data emitted from PVC granules in real time based on the starting of a production line, analyzing the intensity and wavelength distribution in the spectral data in real time, and generating spectral data characteristic values; and comparing and analyzing the deviation of historical data and real-time data through the spectral data characteristic value, and generating a deviation dynamic prediction analysis result. According to the invention, by monitoring and analyzing the spectral data of the PVC granules in real time, the response speed and precision of a production line are improved, and the spectral data are compared in real time, so that production deviations such as light source intensity and reaction temperature can be quickly found and adjusted, material waste is reduced, the output quality is improved, and continuous data analysis helps to adjust chemical agent feeding and heating power; product consistency is guaranteed, automatic fault diagnosis is achieved, production parameters are adjusted in time, the service life of equipment is effectively prolonged, shutdown is reduced, and cost benefits are increased.

The invention provides a multi-core heterogeneous system-oriented startup deployment and cooperative processing method, which comprises the following steps of: when the system runs, obtaining startup codes after a master core and a slave core are electrified, respectively obtaining mirror image files of Linux and RTOS (Real Time Operating System) on the basis of synchronous running and storage partition, and running the mirror image files on the master core and the slave core after moving the mirror image files; and after the heterogeneous operating system is loaded, generating an interrupt signal based on inter-core interrupt and a corresponding register through a CLINT module in the main core, and realizing data communication between the main core and each slave core based on the interrupt signal and the shared memory. By means of the method, after the system is powered on, the multiple cores can synchronously run the starting codes, and time delay and uncertainty possibly caused by one-by-one starting of the single cores are avoided; communication and data transmission among multiple cores are realized through interruption and memory sharing, the synchronism and consistency of data are ensured, and misjudgment and decision errors caused by data delay and inconsistency are avoided, so that the real-time performance of the system is remarkably improved.

The invention discloses a system and a method for security enhancement of generated content information of a large model, and the method comprises the steps: firstly, carrying out the bidirectional LoRA fine tuning technology and knowledge distillation technology on the large model, and enabling the large model to adapt to a factual output task from the illusion dimension and the factual dimension; performing extraction of intermediate layer representation in a fact space on the fine-tuned large model, and obtaining different mappings from fine-grained illusion types to a fact direction through a comparative learning method; and finally, in the reasoning generation process of the large model, extracting the representation of the middle layer of the large model to judge the potential illusion type, then applying the learned mapping corresponding to different illusions to the middle representation of the reasoning process, correcting factual representation, reducing the illusion output probability, and in the pre-training stage and the reasoning generation stage, obtaining the illusion type of the large model. Parameters and generated contents of the large model are controlled, and the factuality and reliability of the generated contents of the large model are greatly improved while the generation capability of the large model is guaranteed.

The invention discloses an injection molding process for a plastic part of an air purifier, and relates to the technical field of injection molding processes. Multi-area dynamic mold temperature adjustment is performed, a plurality of temperature control areas are divided according to the thin wall and curved surface structure characteristics of the plastic part, the temperature is adjusted in real time, stress concentration and deformation are remarkably reduced, and the service life of the plastic part is prolonged. The problem of shrinkage stress of a thin-wall area is solved in a targeted mode, cooling of a thick area is accelerated, meanwhile, real-time temperature control and deformation compensation are controlled in a linkage mode, the temperature and the cooling rate are dynamically adjusted according to deformation data monitored in real time, and the deformation trend is predicted, so that pre-deformation compensation is provided, warping of the thin-wall area and the curved surface area is controlled, and deformation of the thin-wall area and the curved surface area is controlled. The stability of the plastic part in the forming process is kept, and the product precision is improved; in addition, simulation prediction and mold parameter adjustment are carried out, cooling requirements of different areas are adjusted, the temperature difference between the thin wall and the curved surface area is reduced, and the whole cooling process is more efficient.

The invention relates to the technical field of artificial intelligence and the field of financial science and technology, and discloses a text-to-video method based on deep semantic analysis, which comprises the following steps of: extracting keywords and named entities from a to-be-converted text, analyzing a core theme and a text structure of the text, and performing semantic analysis to generate a semantic analysis result; retrieving matched images, video clips, sound effects and background music according to a semantic analysis result, and performing spatial layout to generate a preliminary visual layout framework; and constructing a video timeline, integrating the layout scheme, the background music and the sound effect to generate a complete video content, and finally outputting the video content as a video file. Keywords, named entities, core topics and text structures in texts are accurately extracted through semantic analysis, images, video clips, sound effects and background music conforming to semantics are intelligently matched, and the problem that a large amount of manual intervention is needed is avoided. An adaptive animation effect is generated through a deep learning model, and dynamic presentation of video content and continuity of a visual effect are ensured.

The invention relates to the technical field of fault detection for a wireless communication network, and discloses a fault detection method for the wireless communication network. By collecting data of base station signal strength, network delay, data packet loss rate and user equipment signal quality, the problem of insufficient fault detection caused by a single index is solved. Through multi-dimensional data acquisition, the system can comprehensively evaluate network performance, the fault detection accuracy is improved, and different fault types can be effectively identified especially in a complex fault mode. By storing the data according to the time sequence, the problem of time sequence chaos is avoided, and accurate detection of time-varying faults is ensured. The comprehensive data vector generated by each time of sampling comprises a plurality of network performance indexes, so that the fault detection and positioning capability is improved. Abnormal values are eliminated through data preprocessing, the data quality is ensured, an accurate basis is provided for subsequent analysis, and the reliability of a detection result is ensured.

The invention provides a micro robot motion control method and system, and the method comprises the steps: carrying out the synchronization and calibration processing of environment state information, and generating a comprehensive perception image; analyzing the comprehensive perception image to obtain a classification result, and predicting the environment state change influencing the motion track of the micro robot in a period of time in the future based on the classification result; generating a candidate path by using a fast random tree, and optimizing the candidate path in combination with an AI search algorithm to obtain a target path; constructing a three-dimensional dynamic probability graph, continuously updating the three-dimensional dynamic probability graph by using a reinforcement learning algorithm, and generating an instant decision map; and estimating the position of the micro-robot by using a Kalman filter to obtain the position state of the micro-robot, adjusting the speed and steering angle of a driving motor of the micro-robot, and generating trajectory tracking of the micro-robot. According to the technical scheme provided by the invention, the flexibility and intelligence of the micro robot are improved.