42 results about "Spatiotemporal database" patented technology

The invention discloses a main beam formwork erection elevation adjustment method based on machine learning, and belongs to the technical field of elevation adjustment. The method comprises the steps that a space-time database is constructed by collecting data of the construction period of a cable-free section; processing the time sequence data by adopting an LSTM-Transform hybrid model, extracting long-period characteristics, and outputting an initial prediction value of the elevation; analyzing nonlinear variables such as a cable force change rate, a sunlight gradient and a material age by using an XG-Boost algorithm, and outputting a compensation factor; fusing and generating a joint prediction value; constructing a reinforcement learning agent, and outputting an adjustment instruction by taking a construction stage as a state space, taking an elevation adjustment amount as an action space, minimizing deviation between a predicted value and a measured value and taking construction stability as a reward function; the formwork erecting elevation is adjusted according to the driving hydraulic system, and the database is updated in real time. According to the invention, through multi-dimensional data fusion and intelligent optimization, the elevation adjustment precision and stability are improved.

The embodiment of the invention provides a rail transit 3D data defect detection method and system based on spatial indexing, and the method comprises the steps: collecting rail 3D original data, carrying out the preprocessing, and unifying the data into a unified rail coordinate system; dividing the preprocessed track 3D data into a plurality of regular grid units as space units according to spatial positions, extracting multi-dimensional features of data in each space unit, establishing a composite index for each space unit, and storing the composite index in a space-time database; after the acquired 3D data of the current track are subjected to identical processing, spatial position areas corresponding to the 3D data of the current track and historical data are quickly matched through a composite index, and a difference value of multi-dimensional features is calculated; dynamically adjusting the defect judgment threshold of each dimension according to the statistical distribution of the difference values, judging the defect type and severity of the region meeting the threshold, and generating a detection report. The problems that traditional 3D data processing is low in efficiency and long in comparison time consumption are solved.

The invention provides a real estate data visualization display method and system based on a GIS technology, and the method comprises the steps: carrying out the collection, cleaning, standardized mapping and entity alignment of multi-source real estate data, such as registration, transaction, planning approval, judicial, tax, sensing terminals, remote sensing images, and the like, building a space-time joint index with a space unit ID and a time slice ID as main keys, extracting a feature set and a target attribute set for modeling; constructing a structured causal model in the space-time database, and establishing a causal query index through a structure learning algorithm and parameter training; receiving the scene parameter set, analyzing the scene parameter set into an intervention instruction, executing a do-intervention operation on a causal model, collecting user interaction behaviors and external observation data, triggering online calibration of the model and a mapping rule when an error exceeds a threshold value, updating a causal query index and a mapping threshold value, and supporting version rollback; and a data, model, display and feedback closed-loop link is formed.

The invention discloses a park security multi-target path planning system and method, and belongs to the technical field of security, and the park security multi-target path planning system comprises a data layer, an algorithm layer and an application layer. The data layer constructs a park security space-time database and realizes multi-dimensional data fusion, and provides data support for security scheduling and risk assessment. The dynamic action range model of the algorithm layer can accurately determine the action range of the patrol route of the security and protection personnel and identify key points, warning areas and blind areas; a multi-target path optimization engine integrates the coverage degree, the time efficiency and the personnel fatigue degree to construct a three-dimensional dynamic optimization model, and patrol path dynamic adjustment is achieved by means of an improved NSGA-III algorithm. The application layer provides a six-dimensional analysis grid layer, and assists a command center and security patrol personnel in making decisions.

This invention discloses a deep learning method and system for generating trajectory data that supports heterogeneous data fusion, belonging to the fields of smart cities and smart transportation. The invention first acquires multi-source heterogeneous data such as GPS and traffic monitoring video, and performs a preprocessing process. Then, using a divide-and-conquer strategy, an attention-based RNN model predicts pedestrian trajectories, and a fusion model combining macroscopic traffic flow and GNN predicts vehicle trajectories, introducing multimodal features to generate a high-precision set of predicted trajectories. A spatiotemporal database is constructed to create a composite index for each trajectory. Finally, a digital twin system integrating these prediction results is built to achieve dynamic rendering and evolution. This invention effectively solves the problem of multi-source heterogeneous data fusion and significantly improves trajectory prediction accuracy.

The invention provides a data intelligent analysis method and system applied to carbon emission monitoring, and the method comprises the following steps: executing an NMF load decoupling algorithm to obtain a dynamic non-negative vector which comprises a resident coefficient and an industrial coefficient; according to the non-negative row vector and the space-time database, respectively calculating resident and industrial carbon emission intensities corresponding to the current time period, and updating the carbon emission intensities to the space-time database; calculating and generating future carbon emission predicted values corresponding to different future time periods according to the space-time database; and performing anomaly detection calculation according to the carbon emission intensity and a future carbon emission predicted value corresponding to the current time period. By adopting the method, the carbon emission intensity condition of the whole area can be intuitively and indirectly reflected in real time without tampering.

The application discloses a ground bearing capacity intelligent identification method and system based on a sensing network, relates to the technical field of geotechnical engineering detection, and comprises the following steps: a mixed perception network is constructed, and multi-source heterogeneous monitoring data of a foundation is acquired; a unified space-time datum of the multi-source heterogeneous monitoring data is established, and a space-time database is constructed; data in the space-time database is subjected to homogeneous fusion and heterogeneous fusion, comprehensive judgment of a ground bearing capacity state is carried out based on a fusion result, and a bearing capacity distribution field of the ground bearing capacity is generated. The application solves the technical problems in traditional ground bearing capacity identification, such as difficulty in giving consideration to coverage range and precision during data acquisition, lack of effective integration of multi-source data, and lagged response to abnormal conditions, achieves accurate identification of the ground bearing capacity, global coverage and long-term reliable optimization, and meets the technical effect of engineering safety monitoring and control of the foundation.

The application discloses a kind of based on machine learning's main girder erects elevation adjustment method, belong to elevation adjustment technical field, this method constructs space-time database by gathering the data of no cable section construction period;Adopt the mixed model of LSTM-Transformer to process time series data, extract long-period characteristic, output initial predicted value of elevation;Nonlinear variable such as the change rate of cable force, sunshine gradient and material age is analyzed using XG-Boost algorithm, and compensation factor is output;Fusion and generate joint predicted value;Construct reinforcement learning intelligent agent, with construction stage as state space, elevation adjustment quantity as action space, the deviation of predicted value and measured value and construction stability are minimized as reward function, and output adjustment instruction;According to driving hydraulic system, adjust erecting elevation, and real-time update database.The application improves the precision and stability of elevation adjustment by multi-dimensional data fusion and intelligent optimization.

The invention relates to the technical field of database construction, and particularly discloses a database construction management method for multi-source data. The method comprises the following steps of S1, acquiring multi-source data and performing preprocessing; S2, establishing an agricultural rural space-time database; S3, establishing a first construction management analysis model; S4, establishing a second construction management analysis model; according to the method, the multi-source data is acquired and preprocessed, the space-time database is constructed, and the construction management quality is evaluated by establishing the first construction management analysis model and the second construction management analysis model; according to the invention, integration and unified management of agricultural rural multi-source data are realized, data quality and application value are significantly improved, comprehensive and accurate data support is provided for agricultural analysis, and improvement of agricultural production efficiency, reduction of production risk and promotion of agricultural sustainable development are facilitated.

The invention discloses an airport aircraft taxiing efficiency redisk analysis method and system. The method comprises the following steps: performing space-time alignment on flight plan, aircraft position distribution, field monitoring radar and airport notification data to generate a fused space-time database; constructing a weighted directed graph model which takes the airport physical position as a node and the slidable path as an edge as an airport operation map; identifying flights with abnormal taxiing time, and mapping the actual track into an actual taxiing path sequence on the map through a hidden Markov map matching algorithm; calculating a theoretical optimal sliding path sequence by taking starting and ending nodes as constraints; the two paths are compared to position deviation nodes, and delay attribution is carried out in combination with node and edge attributes; and finally, carrying out visual display and parameterized simulation deduction. According to the method and the system provided by the invention, through a data fusion and map modeling technology, automatic and precise root cause tracing and strategy evaluation for the problem of low sliding efficiency are realized, and the intelligent level of airport ground operation management is improved.

This invention belongs to the field of stellar positioning and navigation technology, and particularly relates to an auxiliary altimetry method for stellar positioning. The method includes constructing a three-dimensional spatiotemporal coordinate system, collecting raw altimetry data, constructing a three-dimensional spatiotemporal database, analyzing the spatial and temporal correlations of measurement points, calculating the spatial and temporal correlations between different measurement point areas and surrounding measurement point areas in the latitudinal and meridional directions, identifying correlated data based on the correlation of the altimetry data, eliminating other invalid samples without correlation, and performing spatiotemporal weighted fusion based on the spatial correlation of adjacent measurement points and the temporal correlation at different times to obtain fused altimetry data. This application integrates the spatial correlation of measurement points during the altimetry data acquisition process, as well as the temporal correlation attributes of the same measurement point at different times, into the verification statistics of the altimetry data results, providing a data foundation for subsequent output verification of altimetry data and correction of erroneous data.

This invention discloses a method for extracting irrigation prescription maps based on BeiDou information of irrigation nodes. The method includes: acquiring a known large-scale water demand map and water shortage index map of the irrigation area under study; acquiring the BeiDou positioning information of the irrigation terminal nodes in the irrigation area under study and establishing a spatiotemporal database of all irrigation terminal nodes in the irrigation area under study; dividing the irrigation area under study into several partitions using fuzzy C-means clustering based on the spatiotemporal database, and using nodes with a membership degree greater than or equal to a set threshold in each partition as the initial boundary of the partition; smoothing the initial boundary of each partition to obtain the final partition boundary of each partition; determining the water demand relationship of the partitions based on the water demand map of the irrigation area under study; and calculating the gross irrigation quota of the partitions as the irrigation prescription map of the partitions based on the water demand map and water shortage index map of the irrigation area under study and the water demand relationship of the partitions.

The invention provides a natural resource element space-time process analysis method based on land type pattern spots. According to the method, a geographic vector pattern spot is used as a minimum space unit, spatial and temporal change types (newly increased, changed and killed) of land cover plaques are defined, and a corresponding algorithm is constructed. Firstly, multi-period basic geographic information data are collected and preprocessed, pattern spots are extracted and endowed with unique codes, and a topological structure and standardized codes are checked. And then realizing a spatio-temporal variation algorithm, judging the pattern spot variation type and calculating the area variation. And storing the data into a space-time database adopting an Oracle + ArcSDE engine, and carrying out space-time process analysis and visual expression, including a plurality of modes such as time axis animation and the like. The method is high in precision, avoids precision loss, is high in space-time complexity expression capability, is good in visualization effect, can provide scientific basis and decision support for the fields of natural resource management and the like, and is wide in application prospect.

This invention discloses a method and system for designing a spatiotemporal database of land value, applied in the fields of spatial geographic information data processing and image correction technology. The method involves acquiring spectral remote sensing image data; performing geographic coordinate matching on the spectral remote sensing image data to generate coordinate spectral remote sensing image data; digitally rasterizing the spectral remote sensing image data using a digital elevation model to generate point cloud data; performing 3D reconstruction based on the point cloud data to obtain a 3D terrain model; performing boundary region reconstruction processing on the 3D terrain model to obtain a reconstructed model; extracting ground features from the reconstructed model to obtain ground feature data; and constructing a spatiotemporal database by spatiotemporally indexing the ground feature data. This method can correct the brightness of edge defects of interval-type and linear ground features based on predicted brightness changes, significantly reducing grayscale distortion in 3D images and eliminating blurring caused by interference.

This invention relates to the field of database construction technology, specifically disclosing a database construction and management method for multi-source data. The method includes the following steps: S1: acquiring and preprocessing multi-source data; S2: establishing an agricultural and rural spatiotemporal database; S3: establishing a first construction management analysis model; S4: establishing a second construction management analysis model; S5: evaluating the construction management quality; and S6: establishing a construction management in-depth analysis model. This invention acquires and preprocesses multi-source data to construct a spatiotemporal database, and evaluates the construction management quality by establishing a first and a second construction management analysis model. This invention achieves the integration and unified management of multi-source agricultural and rural data, significantly improving data quality and application value, providing comprehensive and accurate data support for agricultural analysis, and helping to improve agricultural production efficiency, reduce production risks, and promote sustainable agricultural development.

The invention discloses a dynamic environmental effect grading method for a coastal concrete structure. The method comprises the following steps: firstly, constructing a coupling space-time database of the surface chloride ion concentration, the chloride ion diffusion coefficient and the steel bar corrosion rate; then, an improved space-time Kriging interpolation model (MSTK) is adopted to reconstruct space chloride ion distribution, and the weight coefficient of the space-time distance is optimized through a particle swarm; and meanwhile, a Hilbert dynamic probability interpolation model (HDPI) is combined with Bayesian optimization LSTM to realize time dimension continuous prediction, and the chloride ion concentration is decomposed into a deterministic trend, low-frequency environment fluctuation and high-frequency material fluctuation. Based on the interpolation result and the actually measured corrosion rate, high, medium and low corrosion risk grades are automatically divided through hierarchical clustering; and finally, quantizing the contribution degree of the detection points based on the diagonal elements of the prediction matrix, and optimizing the layout of the detection points. The method is high in precision (predicting R2gt; and the cost is low (detection points are reduced by 37.5%), and differentiated protection design of coastal electric power facilities is supported.

The application discloses a natural language spatiotemporal retrieval method and system based on a large model, relates to the technical field of big data resource services, and comprises the following steps: fine-tuning a preset large language model according to a spatiotemporal retrieval problem and a preset prompt word engineering template; dividing a target management area into geographical space grids, each of which has a unique grid code; constructing an entity knowledge base classified according to different data categories based on a business spatiotemporal database and dynamic semantic labels; performing natural language intention analysis on user input based on the fine-tuned large language model, identifying spatial constraints, time constraints and business attribute constraints; converting the spatial constraints into filtering conditions; querying the data categories of corresponding key entities in the entity knowledge base; and outputting spatiotemporal retrieval results based on the fine-tuned large language model according to the business attribute constraints, the data categories, the spatial constraints and the filtering conditions. The application is free from professional dependence, improves the interactive experience and is highly flexible.

The application discloses a kind of discrete manufacturing industry data representation methods based on deep reinforcement learning, including collecting discrete manufacturing industry data, construct space-time database;Discrete manufacturing industry data is divided into discrete features and continuous features, construct data coupling coding network, and the encoding vector in coding network is converted into representation vector, and construct data representation model;The distinguishing degree of the class of representation data is quantified by clustering evaluation index;Different dimensions of clustering evaluation index are weighted as dynamic reward, construct deep reinforcement learning model, update the neural network parameters of deep reinforcement learning through the interaction between representation model and discrete manufacturing decision analysis system.The application can represent dynamically changing mixed data, based on deep reinforcement learning, the discrete manufacturing industry data is represented and enhanced, interacts with discrete industrial system through clustering evaluation index, and constantly feedback dynamic reward information, to obtain the optimal data representation form.

The invention relates to the technical field of numerical simulation, in particular to an ultrasonic flowmeter measurement error and reason analysis method, system and device and a medium. According to an actual flowmeter structure, an ultrasonic flowmeter digital simulation model is constructed, and an error injection module is introduced into the ultrasonic flowmeter digital simulation model; based on the constructed ultrasonic flowmeter digital simulation model, actual flowmeter data are collected in real time, the actual flowmeter data and virtual simulation data are adopted for collaborative sampling and correction, parameters of the ultrasonic flowmeter digital simulation model are dynamically adjusted through an optimization algorithm, and the ultrasonic flowmeter digital simulation model with the dynamically adjusted parameters is used for simulating the ultrasonic flowmeter. The error injection module is used for carrying out a disturbance experiment on a key error source, generating a space-time database, establishing a machine learning model, inputting the generated space-time database and cooperative correction data into the machine learning model, analyzing and outputting a root cause of an error, and improving the accuracy of error identification and the root cause traceability.

The invention discloses an irrigation prescription map partition extraction method based on irrigation node Beidou information. The method comprises the following steps: obtaining a known large-scale irrigation district water demand diagram to be researched and a known large-scale irrigation district water shortage index diagram to be researched in an irrigation district to be researched; beidou positioning information of end domain irrigation terminal nodes of an irrigation area to be researched is acquired, and a time-space database of all the end domain irrigation terminal nodes of the irrigation area to be researched is established; according to the space-time database, the to-be-researched irrigation area is divided into a plurality of partitions by adopting fuzzy C-means clustering, and nodes with the membership degree larger than or equal to a set threshold value in each partition serve as partition initial boundaries; smoothing the initial boundary of each partition to obtain the final partition boundary of each partition; determining the water demand relationship of the subareas based on the water demand diagram of the irrigation area to be researched; according to the water demand diagram and the water shortage index diagram of the irrigation area to be researched and the water demand relation of the subareas, the gross irrigation quota of the subareas is calculated to serve as the irrigation prescription diagram of the subareas.

The invention discloses an intelligent foundation bearing capacity identification method and system based on a sensing network, and relates to the technical field of geotechnical engineering detection.The method comprises the steps that a mixed sensing network is constructed, and multi-source heterogeneous monitoring data of a foundation are obtained; establishing a unified space-time reference of the multi-source heterogeneous monitoring data, and constructing a space-time database; and carrying out homogeneous fusion and heterogeneous fusion on the data in the time-space database, carrying out comprehensive judgment on a foundation bearing capacity state based on a fusion result, and generating a bearing capacity distribution field of the foundation bearing capacity. The technical problems that in traditional foundation bearing capacity recognition, coverage range and precision are difficult to consider in data collection, effective integration of multi-source data is lacked, and response to abnormal conditions is delayed are solved, and the technical effects of accurate recognition, global coverage and long-term reliable optimization of the foundation bearing capacity and meeting of engineering for foundation safety monitoring and management and control are achieved.

The invention discloses a city update three-dimensional model lightweight and semantic segmentation fusion method, and relates to the technical field of three-dimensional geographic information processing, and the method comprises the steps: constructing a dual-branch multi-task deep learning network comprising a shared feature extraction layer, a lightweight branch and a semantic segmentation branch through the standardization preprocessing of multi-source data; feature extraction is enhanced by adopting dynamic graph convolution and an attention mechanism, and joint training is realized through a composite dynamic loss function; semantic tags are optimized through post-processing of conditional random fields, incremental updating is achieved based on a gridded space-time database, and multi-granularity intelligent query and cross-platform docking are supported. According to the method, collaborative optimization of light weight and semantic segmentation is achieved, the local updating efficiency is improved by 10 times, the key element labeling accuracy rate is larger than or equal to 95%, the complex query response time is smaller than or equal to 2 seconds, and efficient and accurate technical support is provided for city updating full-life-cycle management.

The present application relates to the field of traffic infrastructure digital twinning, and aims to provide a large traffic infrastructure full life cycle immersive traceability system and interaction method based on spatial imaging, so as to realize the fusion management, efficient traceability and collaborative decision of the full life cycle data of large traffic infrastructure. The technical scheme of the present application comprises: a multi-source data access and fusion engine for accessing and processing multi-source heterogeneous data from different sources; a three-dimensional space-time database for storing and associating multi-source heterogeneous data to form space-time objects; a progressive immersive interaction and rendering engine for reading data of the three-dimensional space-time database and rendering a three-dimensional scene in a spatial imaging environment and a collaborative interaction terminal; a spatial imaging environment and a collaborative interaction terminal for presenting the three-dimensional scene rendering through projection mode to realize the visual display of the infrastructure; and a server for running the multi-source data access and fusion engine, the three-dimensional space-time database and the progressive immersive interaction and rendering engine.

The invention provides a local updating system and method for a high-precision map of a strip mine, and the system comprises six modules: a multi-source data collection module, an intelligent change detection module, an incremental three-dimensional reconstruction module, a space-time database management module, a visual verification module and a mining and stripping amount measurement module. PointNet + + and U-Net networks are fused to realize change region extraction with the precision less than or equal to 1m, only local reconstruction is carried out on the change region, centimeter-level seamless fusion is realized by adopting graph optimization and Laplacian grid deformation, and updating is completed in a short time. The model supports version second-level rollback, WebGL30FPS rendering and TIN method automatic amount calculation, the error is less than 1.2%, the bandwidth is saved by 90%, the GPU occupation is reduced by 75%, the requirement of an unmanned mine card for the current trend less than or equal to 30 minutes under the pace of daily propulsion of a strip mine is met, mining while updating is realized, and a real-time data base is provided for intelligent scheduling and digital twin mines.

The application provides a data intelligent analysis method and system applied to carbon emission monitoring, and comprises the following method: a dynamic non-negative row vector is obtained by executing an NMF load decoupling algorithm, the dynamic non-negative row vector comprises a resident coefficient and an industrial coefficient, carbon emission intensities of residents and industries corresponding to a current period are calculated according to the non-negative row vector and a space-time database, and are updated to the space-time database; future carbon emission prediction values corresponding to future different periods are calculated and generated according to the space-time database; and abnormality detection calculation is performed according to the carbon emission intensity and the future carbon emission prediction values corresponding to the current period. By using the above method, the carbon emission intensity situation of the whole region can be indirectly reflected intuitively, in real time and in a non-falsifiable manner.

The invention relates to the technical field of data compilation, in particular to a list data compilation method for supporting carbon footprint accounting of various energy products, which comprises the following steps: establishing a space-time decoupling feature model, and generating a dynamic feature tensor set; constructing a credible data acquisition chain, injecting the dynamic feature tensor set into edge acquisition equipment, generating an encrypted data unit with a physical feature signature through a hardware fingerprint binding module, and writing the encrypted data unit into a space-time database; executing cross-domain association verification, performing time axis alignment on the multi-time-zone data by using a distributed consistency algorithm, and detecting a data tampering risk through a trusted anchor verification module; and dynamically generating an optimization strategy, adjusting weight distribution of the feature tensor according to a verification result, and outputting a carbon footprint accounting list with a space-time correction factor. The generated carbon footprint list has higher timeliness and accuracy, can provide a carbon emission accounting basis with high credibility and dynamic response for energy enterprises and supervision organizations, and has good popularization and application prospects.

This invention discloses a method and system for dynamic obstacle avoidance path planning for robots. The method includes: collecting data such as images and UWB tag coordinates from multiple mobile robots via an edge computing terminal, and using this data to construct a dynamic environment spatiotemporal database; preprocessing the collected images; identifying and classifying static and dynamic obstacles; outputting their feature information and motion vectors; obtaining the reliable weights of each robot; predicting the future position and trajectory of dynamic obstacles; forming predicted risk areas; using static landmarks to correct positioning errors; dynamically adjusting obstacle avoidance strategies and path planning; generating collaborative scheduling instructions to guide local obstacle avoidance or global task reallocation; and recording detailed execution logs. By implementing the method of this invention, robot swarms can achieve high-precision, low-latency, continuous, and safe autonomous obstacle avoidance and path planning in complex dynamic environments.

The invention discloses an electric power meteorology and extreme weather toughness early warning and enhancement system, and the system comprises a data collection layer which is used for collecting multi-source heterogeneous data; the data fusion and preprocessing layer is used for carrying out data cleaning, quality control and standardization processing on the multi-source heterogeneous data, carrying out multi-modal data fusion on the standardized multi-source heterogeneous data, and constructing a unified space-time database by utilizing a multi-modal data fusion result; the physical information fusion AI prediction layer is used for constructing a multi-modal physical information fusion AI prediction model combining a power system physical equation and a deep learning model based on the space-time database; and the multi-stage intelligent early warning and decision support layer is used for predicting the multi-modal physical information based on the multi-modal physical information fusion AI prediction model, performing graded early warning according to a prediction result, and generating an optimal power grid active control and toughness enhancement strategy, and the system can enhance the toughness and safety of the power system in extreme weather.

The application provides a method and system for inverting degree-day factors based on snow water equivalent difference, and relates to the technical field of hydrological remote sensing. The method of the application firstly acquires and unifies multi-source snow water equivalent data, air temperature, radiation, wind speed, relative humidity and other meteorological elements to form a standardized input data set, adopts data assimilation fusion to generate assimilated snow water equivalent and calculate uncertainty; combines meteorological conditions and uncertainty to determine the snow melting process, calculates the pixel-scale degree-day factor when the snow melting is determined, and constructs a time-space database; uses convolutional neural network and long short-term memory network to extract spatial and temporal characteristics, obtains historical sample degree-day factor distribution based on similarity matching for prediction, calculates the snow melting amount and estimates the uncertainty, so as to improve the snow melting simulation and prediction accuracy.

The application discloses a kind of trajectory data deep learning generation method and system supporting multimodal data integration, belong to intelligent traffic management field.The application is first obtained and preprocessed multimodal data, washes invalid point and interpolates missing value;Subsequently, the multi-source data is fused into standardized trajectory set according to unified JSON Schema;Attention-RNN, STGCN and ConvLSTM deep learning models are selected for different types of entities such as pedestrians and vehicles, and auxiliary features such as map events and weather are integrated for training and prediction;Finally, the predicted trajectory is efficiently stored, indexed and safely managed by the spatiotemporal database.The application can effectively integrate heterogeneous traffic data, improve the accuracy and robustness of trajectory prediction.