38 results about "Virtual structure" patented technology

The invention relates to the technical field of unmanned aerial vehicle cluster control, in particular to a multi-mode sensing fusion unmanned aerial vehicle cluster control method, device and equipment based on deep learning and a storage medium. According to the invention, gesture, voice, vision and touch data are collected in parallel through a multi-mode perception access layer and are preprocessed and synchronized; utilizing a special deep network to extract high-level features rich in cluster semantics; through cross-modal semantic alignment, graph neural network association learning and a multi-dimensional conflict detection and intelligent resolution mechanism, deep fusion and accurate intention analysis of four modals in a cluster control semantic space are realized; high-level instructions are efficiently converted into single-machine executable tasks and coordination strategies by adopting hierarchical task decomposition, virtual structure formation control and a multi-objective optimization algorithm; state feedback and adaptive parameter optimization are carried out in real time, and the accuracy, reliability, environmental adaptability and man-machine interaction efficiency of large-scale unmanned aerial vehicle cluster control are improved.

The invention relates to the technical field of building structure health monitoring, and discloses a BIM-based steel structure deformation monitoring processing method and system. According to the method, a BIM design model and a physical monitoring space are mapped through a digital twin interface, and a virtual structure mirror image is generated. And injecting real-time monitoring data into the mirror image, and driving the mirror image to evolve to generate the dynamic twin. And by carrying out reverse backtracking and feature deconstruction on the historical evolution trajectory of the twinborn body, an independent deformation component caused by an environment variable and a load variable is stripped out. And performing nonlinear time sequence extension deduction on each component, and constructing a future potential development path of each component. And carrying out spatial topological superposition and interference analysis on the paths of all the components, and synthesizing a prediction deformation field. The mapping rule is dynamically adjusted by comparing the deviation of the prediction field and the real-time twinborn body, and a closed-loop process is formed. According to the method, accurate separation and attribution of steel structure deformation causes are realized, and the spatial precision and reliability of deformation development prediction are improved.

The invention provides a multi-variety order dynamic scheduling management method and system for a liquid crystal display module production line, and relates to the technical field of industrial automation, and the method comprises the steps: collecting the size, resolution, interface type, batch and delivery time data of all orders, obtaining the process equipment, jig configuration, material inventory state and material supply plan data of the production line at the same time, obtaining an order data set and a resource data set; based on the order data set and the resource data set, an order and resource associated data point set is constructed, a three-dimensional virtual voxel space is defined in the virtual structure, and the order and resource associated data point set is mapped to corresponding voxels one by one; and recording the number of data points of each voxel falling into the three-dimensional virtual voxel space, the corresponding order type, the resource number and the material demand matching degree information to obtain a voxel sub-region with a data attribute mark. According to the invention, dynamic optimization of the scheduling scheme is realized.

The present application relates to the technical field of unmanned aerial vehicle cluster control, in particular to a multi-modal perception fusion unmanned aerial vehicle cluster control method and device based on deep learning, equipment and storage medium. The present application collects gesture, voice, visual and touch data in parallel through a multi-modal perception access layer, and pre-processes and synchronizes the data. A special deep network is used to extract high-level features rich in cluster semantics. Through cross-modal semantic alignment, graph neural network correlation learning, and multi-dimensional conflict detection and intelligent resolution mechanism, the present application realizes the deep fusion of the four modes in the cluster control semantic space and the accurate analysis of the intention. The present application uses hierarchical task decomposition, virtual structure formation control and multi-objective optimization algorithm to efficiently convert high-level instructions into single-machine executable tasks and coordination strategies. The present application also performs real-time state feedback and adaptive parameter optimization, improving the accuracy, reliability, environmental adaptability and human-computer interaction efficiency of large-scale unmanned aerial vehicle cluster control.

A design support device S011: uses an installable area calculation unit P01 to calculate, on the basis of spatial data D01, installation rule data R01, and virtual structure design data B01, which are stored in an external storage, areas that satisfy installation rules in the installation rule data R01 and design rules in the virtual structure design data B01, and thereby acquire installable area data T01; and uses a virtual structure installation unit M01 to install, on the basis of the installable area data T01 and the virtual structure design data B01, a virtual structure within an installable area based on the installable area data T01, and acquire and output design support data E01 to the outside.

A method for detecting and incorporating three-dimensional objects into a video stream reads an input video data stream is described. A method may read an input video and may accept hotspots defining at least one real-world object of interest shown within the video. A method may track movement of any hotspots generating a trajectory of any objects of interest in two-dimensional space of the video and may obtain a three-dimensional topology defining a three-dimensional volume of interest in a three-dimensional space. The method may translate the hotspots to the three-dimensional volume and may calculate motion of the hotspots in the three-dimensional space. A method may build virtual structures to relate any hotspots to the three-dimensional topology to create a three-dimensional geometric shape and finally project the output, if needed.

Embodiments of the present application provide a protein stability prediction method and device, computer equipment and a storage medium. The method comprises: obtaining a first to-be-tested sequence corresponding to an original protein and a second to-be-tested sequence corresponding to a mutant protein; determining a plurality of first target residue features from the first to-be-tested sequence according to the attention mechanism of the protein language model and extracting the global features thereof to obtain first virtual structure microenvironment features; similarly, obtaining second virtual structure microenvironment features corresponding to a plurality of second target residue features; determining a first difference based on the first virtual structure microenvironment features and the second virtual structure microenvironment features, and determining a second difference based on a first global sequence feature corresponding to the first to-be-tested sequence and a second global sequence feature corresponding to the second to-be-tested sequence; and predicting based on the first difference and the second difference to obtain a protein stability prediction result. In this way, the accuracy of predicting the stability of the protein can be improved.

The application provides a liquid crystal display module production line multi-variety order dynamic scheduling management method and system, relates to the technical field of industrial automation, and the method comprises the following steps: collecting the size, resolution, interface type, batch, and delivery date of all orders, simultaneously acquiring production line process equipment, jig configuration, material inventory state, and material supply plan data, obtaining an order data set and a resource data set; based on the order data set and the resource data set, an order and resource associated data point set is constructed, a three-dimensional virtual voxel space is defined in a virtual structure, the order and resource associated data point set is mapped into corresponding voxels one by one, and the number of data points, corresponding order type, resource number, and material demand matching degree information of each voxel falling into the three-dimensional virtual voxel space are recorded, so that a voxel sub-region with data attribute labels is obtained. The application realizes dynamic optimization of a scheduling scheme.

This invention discloses a drone swarm formation obstacle avoidance control method based on an improved artificial potential field method, comprising the following steps: Step 1, performing formation control of the drone swarm based on a virtual structure method; Step 2, for the i-th drone, during its movement towards the target point, determining whether the distance between the i-th drone and the obstacle is greater than the safe distance using the artificial potential field method. If the distance between the i-th drone and the obstacle is greater than the safe distance, it continues to move towards the target point; if the distance between the i-th drone and the obstacle is less than the safe distance, an improved artificial potential field method is introduced for obstacle avoidance until the target point is reached; the improved artificial potential field method for obstacle avoidance specifically involves: making F i ob With F i att Maintain orthogonality, when F i att With F i ob When the angle between them is 0, only F is retained. i ob China and F i att Orthogonal components F i obrep Remove F i att The parallel components make the attraction force dominate the overall motion trend, improving the efficiency of the drone in reaching the target point.

This invention discloses a method for testing the corrosion crack propagation path of floating structures based on digital twins, relating to the field of path testing. The method includes the following steps: acquiring basic information about the floating structure, setting up a three-dimensional virtual structure model, and marking key nodes; setting up an intelligent sensor network to acquire real-time monitoring data at each key node; performing labeling and analysis on the historical monitoring data at each node within the three-dimensional virtual structure model to obtain a longitudinal and transverse label dataset; performing longitudinal and transverse analysis on the longitudinal and transverse label dataset to obtain the interaction relationship between seawater corrosion and crack propagation at each key node in three dimensions: time, space, and degree of influence; constructing a monitoring verification model and a crack propagation model based on the analysis results; verifying and analyzing the real-time monitoring data using the monitoring verification model; and performing path analysis on the verification analysis results based on the crack propagation model. This invention improves the accuracy of the crack propagation path testing process.

The aim is to provide design support for the new construction or relocation locations of infrastructure facilities using virtual space data. [Solution] In a design support device S011, an installable area calculation unit P01 calculates an area that satisfies the installation rules in the installation rule data R01 and the design rules of the virtual structure design data B01 based on spatial data D01, installation rule data R01, and virtual structure design data B01 stored in external storage, to obtain installable area data T01, and a virtual structure installation unit M01 installs a virtual structure within the installable area based on the installable area data T01 based on the installable area data T01 and the virtual structure design data B01, to obtain design support data E01 and output it to the outside.

The invention provides an unmanned aerial vehicle formation keeping, transformation, obstacle avoidance and collision avoidance system based on an artificial potential field method. The system comprises a formation structure module, a parameter measurement and calculation module, an artificial potential field module and an instruction integration module. The formation modeling module is used for performing virtual modeling on the formation through a virtual structure method, giving a target position of the current unmanned aerial vehicle, calculating a relative position and a relative speed with the target position according to the target position, calculating each obstacle according to the obstacle information, setting an emergency degree for the obstacles, and determining the formation according to the emergency degree. And calculating the gravitation and the repulsive force of the target position and the obstacle, finally synthesizing an initial resultant force, limiting the resultant force, and finally outputting a control instruction which can be actually executed by the optimized unmanned aerial vehicle. According to the method, the overall stability of the formation is ensured, flexible and efficient formation change can be realized, and real-time and reliable obstacle avoidance and collision avoidance capabilities are realized, so that the autonomous collaborative operation level of a multi-unmanned aerial vehicle system in a complex environment is improved.

The present invention particularly relates to a safety element comprising a grid consisting of a set of adjacent or adjacent raised lines (2) oriented in the directions of their respective axes X and Y, while the raised lines are oriented in the direction of axis Z. At least one of the lines (2) is defined by two consecutive valleys or two consecutive peaks, thereby forming two opposing sides. At least one opposing side has an average slope in the Y-Z plane, which varies continuously between two extreme values. The set of slopes of the grid is such that at any given point, the slope P(x, y) is expressed as, in the equation: k is a constant between 0.2 and 0.8; and P1(x, y) and P2(x, y) are the corresponding local slopes of the sides of the first virtual structure (S1) and the second virtual structure at point (x, y).

The present invention relates to a method for maintaining the formation of a USV formation based on passive positioning under communication-restricted conditions, and belongs to the field of unmanned surface vessel formations. The method comprises: establishing a circular USV formation positioning model under passive conditions based on a three-point triangulation method, adopting an improved least squares filtering algorithm, and introducing a distributed state estimation mechanism to enable the USV to achieve self-positioning; combining a leader-follower model with a virtual structure method, establishing a virtual leader-follower circular formation model under passive conditions, setting an overall formation virtual pilot structure when the USV formation moves, and guiding the formation movement; integrating the distributed state estimation mechanism into a distributed model predictive control strategy, so that each USV follows the trajectory of the USV corresponding to itself in the virtual pilot structure, and performing real-time rolling optimization control on its own trajectory at each time step. The present invention can enhance the robustness of formation control under communication-restricted scenarios.

The invention provides a disassembly and assembly simulation method and a disassembly and assembly simulation system, and relates to the technical field of simulation. The method comprises the steps that after an activation instruction of a target detachable object is received, a virtual model of the object is presented on a terminal display interface; when a dismounting instruction of a target component is received, searching a dependent object having a dependency relationship with the target component from the virtual model, and marking the dependent object in the model; after the disassembly and assembly instruction of the dependent object is activated, the system demonstrates the disassembly and assembly process of the dependent object, and updates and displays the disassembled and assembled virtual model after demonstration is completed; and under the condition that all the dependent objects are judged to be disassembled and assembled, the system executes disassembly and assembly demonstration of the target part, and displays the virtual structure after the disassembly and assembly of the target part after the demonstration is completed. According to the invention, the dismounting sequence can be automatically determined according to the dependency relationship, the dismounting logic is truly presented, and the accuracy and operability of virtual dismounting teaching and simulation training are improved.

The invention provides a multi-agent formation method, system and device based on a virtual structure and an improved potential function, and the method comprises the steps: building a nonlinear multi-agent dynamic system model with N same agents, and making an obstacle avoidance strategy; designing a controller based on the nonlinear multi-agent dynamic system model; adding the controller into the nonlinear multi-agent dynamic system model to obtain a second-order multi-agent system; and according to a second-order multi-agent system stability method, controller parameters and system parameters of the multi-agent system meet discrimination conditions, and system errors of the multi-agent system are gradually converged to zero by using an obstacle avoidance strategy, so that the multi-agent formation is ensured to reach a stable expected trajectory. According to the invention, the multi-agent formation can avoid the obstacle and then reconstruct the formation, and the situation that the unmanned vehicle falls into a local extreme value and cannot continue to move is effectively avoided.

This invention belongs to the field of unmanned surface navigation technology, and particularly relates to a method, system, and medium for pre-defined time-based cooperative formation control of multiple unmanned surface vessels (USVs) with error constraints. The method includes: first, establishing kinematic and dynamic models of the USVs; then, combining virtual structure and leader-follower methods to establish a virtual leader-follower formation control framework, and designing a pre-defined time-based disturbance observer to observe disturbances; finally, designing a pre-defined time function and a pre-defined performance inequality to design a pre-defined time-based formation control protocol. The controller designed in this invention enables USVs to perform formation tasks within a preset time. Furthermore, the pre-defined time-based disturbance observer observes and compensates for external disturbances, exhibiting strong robustness and anti-interference capabilities. Compared to existing finite-time formation controllers, the controller of this invention does not exhibit chattering and its convergence performance is independent of the initial state; compared to fixed-time formation controllers, the system convergence time can be manually set, improving formation control efficiency.

The invention belongs to the technical field of computer vision and robots, and relates to a multi-modal semantic feature point ordering VSLAM method and system, and the method comprises the following steps: carrying out the analysis based on the image content of a video frame, and generating a semantic partition image containing an object type and a contour boundary; generating a mixed feature point cloud containing virtual structure anchor points and native texture features by inferring a geometric intersection relationship among different semantic regions; wherein each feature point is allocated with a structured semantic identity code containing physical affiliation information of the feature point; constructing a semantic physical constraint graph representing a scene physical rule; generating a weighted credibility parameter of the camera pose by evaluating the reliability of positioning contributions of different semantic entities; and triggering propagation calibration from the high-credibility pose to the low-credibility area, and generating a camera track and a map which are subjected to structured correction. The problem that physical consistency and accuracy of a global map are still difficult to guarantee in a large-scale scene is solved.

The invention discloses an unmanned aerial vehicle cluster virtual structure control method based on distributed consistency estimation. The method comprises the following steps: S1, setting a virtual navigator state and an expected relative position vector; s2, uniformly estimating the state of the virtual navigator based on the state information exchanged by the neighbor unmanned aerial vehicles through distributed state estimation; and S3, generating and executing a local control instruction of the unmanned aerial vehicle according to the estimated state of the virtual navigator and the expected relative position vector. According to the invention, the unmanned aerial vehicle group can keep a high-precision formation, and the method is high in expandability, supports large-scale deployment, and is higher in robustness.

This invention discloses a UAV formation cooperative obstacle avoidance method based on improved fuzzy decision-making, comprising: continuously acquiring the positions of UAVs and the virtual lead UAV using onboard sensors; obtaining position and attitude control commands for the UAV formation through a formation control method based on virtual structure and consistency theory; setting the distance obstacle avoidance range and rotation angle obstacle avoidance range of the UAVs according to their performance, and constructing a fuzzy decision-maker; determining the threat level of the UAVs based on the fuzzy decision-maker and obstacles, and outputting corresponding formation obstacle avoidance decisions; for UAVs that have broken out of formation, adjusting the obstacle avoidance control quantity of the UAVs that have broken out of formation based on the speed obstacle method and the expected obstacle avoidance speed. This invention ensures that the maneuverability of most members is not affected during inter-UAV collision avoidance, improves the stability and safety of the formation, and is more conducive to the execution of formation tasks.

The application discloses a database virtual physical structure optimization method, device and medium, and belongs to the technical field of databases. The method comprises the following steps: acquiring workload data of a database runtime and aggregating the workload data into sample sequences, performing normalization and structured analysis on SQL to generate a workload graph; inputting the workload graph into an artificial intelligence prediction model to generate a multi-dimensional prediction feature vector within a future time window; generating a candidate virtual physical structure set in the memory according to a prediction result, defining only metadata descriptors without actually creating physical entities; injecting the virtual structure into an optimizer candidate path through a database planner Hook to generate a hypothetical execution plan and a corresponding cost; calculating a multi-objective global score; and outputting an optimal physical structure adjustment scheme according to the global score and system constraints. The application realizes low-risk and forward-looking adaptive optimization of a database without creating real indexes and other physical structures, and in combination with AI load prediction and multi-objective evaluation.

The application discloses a kind of based on grouping multi-unmanned aerial vehicle formation control method, comprising the following steps: step 1: establishing the communication topology structure of unmanned aerial vehicle cluster;Step 2: according to the communication topology structure of unmanned aerial vehicle cluster, unmanned aerial vehicle is divided into three kinds of unmanned aerial vehicle, virtual leader, secondary leader, follower;Step 3: adjacent secondary leader is directly communicated, and the method of navigation and following is used to keep flight formation;Step 4: a secondary leader and several followers form a group, with secondary leader as reference point in group, using virtual structure method keeps flight formation;Step 5: according to the efficiency of unmanned aerial vehicle flight, the damage reconstruction strategy of formation is designed.The application can improve the formation accuracy of multi-unmanned aerial vehicle cluster formation, and enhance the robustness of cluster system.

This disclosure relates to a sound field simulation method, apparatus, device, medium, and product. The method includes: acquiring first structural data of a first virtual structure and array element excitation data of an ultrasonic transducer; performing forward simulation based on the first structural data and the array element excitation data using a first simulation mode to obtain first excitation data on the first virtual structure; using the first virtual structure as a first emission source, performing forward simulation based on the first structural data and the first excitation data using a second simulation mode to obtain sound field distribution data of an intracranial target point; wherein, the first virtual structure is located in the sound propagation domain between the ultrasonic transducer and the intracranial target point, and the simulation efficiency of the first simulation mode and the second simulation mode differs. This disclosure solves the problem of significant lag in sound field simulation and can adapt to medical application scenarios with high timeliness requirements.

The present invention relates to a computer implemented method for determining a gaze location of a human interacting with products in a store environment. The method comprising: obtaining, from a camera monitoring at least a portion of the store environment, one or more images of human interaction with products in the store environment, the camera having a known position and orientation within a coordinate system of the store environment; determining a position and orientation of a head of the human in the coordinate system of the store environment by analyzing the one or more images; and determining the gaze location of the human as an intersection between a raycast from a position and orientation of the head of the human in a virtual representation of the store environment corresponding to the position and orientation of the head of the human in the coordinate system of the store environment and a virtual structure in the virtual representation of the store environment.

This invention discloses an astronomical autonomous attitude measurement method based on virtual dual-field of view, belonging to the field of astronomical autonomous attitude measurement technology. The method includes: (1) acquiring historical time-series frames; (2) key frame screening; (3) star point coordinate optimization; (4) virtual structure parameter calculation; and (5) joint attitude measurement. The technical solution provided by this invention is based on multi-frame star images acquired by a monocular space camera in a specific time sequence. By screening key frames, optimizing star point coordinates, and calculating virtual structure parameters between the main field of view and the incremental field of view, a virtual dual-field of view observation model is constructed under monocular hardware constraints, thereby realizing joint attitude measurement. This can effectively improve the problem of unbalanced three-axis errors in traditional monocular attitude measurement results, and provide a new approach for astronomical autonomous attitude measurement that combines high-precision measurement requirements with lightweight hardware constraints for space missions.

The invention discloses an unmanned vehicle formation planning and obstacle avoidance method, and belongs to the field of engineering control, and the method comprises the steps: constructing a kinematics model and a dynamics model of an unmanned vehicle; based on the kinematics model and the dynamics model, a pilot following method and a virtual structure method are adopted to obtain a follower unmanned vehicle space attitude tracking error model; obtaining a multi-unmanned vehicle formation system model based on the follower unmanned vehicle attitude tracking error model; obtaining a target point distance factor and an adaptive adjustment factor, and establishing a potential field function model; based on the multi-unmanned vehicle formation system model and the potential field function model, obtaining a dynamic obstacle avoidance path planning model; and based on the dynamic obstacle avoidance path planning model, a formation tracking control model is obtained by adopting a backstepping sliding mode algorithm, and the formation tracking control model is used for controlling unmanned vehicle formation. The method solves the problems that an existing artificial potential field method is unreachable in target, prone to falling into local optimum, too large in gravitational force and the like.

A method for detecting and incorporating three-dimensional objects into a video stream reads an input video data stream is described. A method may read an input video and may accept hotspots defining at least one real-world object of interest shown within the video. A method may track movement of any hotspots generating a trajectory of any objects of interest in two-dimensional space of the video and may obtain a three-dimensional topology defining a three-dimensional volume of interest in a three-dimensional space. The method may translate the hotspots to the three-dimensional volume and may calculate motion of the hotspots in the three-dimensional space. A method may build virtual structures to relate any hotspots to the three-dimensional topology to create a three-dimensional geometric shape and finally project the output, if needed.

The invention relates to a shock absorber external characteristic online adjustment method based on structural domain recognition, and belongs to the technical field of automobile suspensions. Key geometric and mechanical characteristics of a valve system and a piston assembly are abstracted into virtual structure parameters, a reversible mapping model between a structural domain and an external characteristic curve is established, damping force-speed data measured in real time are used for reversely solving equivalent structure parameters, rapid updating from the structural domain to a curve domain is completed in a prediction model, and the equivalent structure parameters are obtained. And external characteristics can be adjusted on line by taking structural parameters as control variables. According to the method, real-time characterization of the internal state of the shock absorber is constructed through structural parameter recognition, external characteristic adjustment is converted into a curve shaping process based on a structural domain from traditional actuator input adjustment, the adjustment direction is determined through sensitivity analysis, the structural parameters are periodically updated under a closed-loop strategy, the external characteristics continuously approach a target curve, and the accuracy of adjustment is improved. Therefore, real-time and self-adaptive optimization of the external characteristics of the shock absorber can be realized without changing a hardware structure.

The present application relates to the technical field of power data communication transmission, in particular to a data communication method and system of a power parameter acquisition terminal, the method provided by the present application mainly includes constructing a virtual structure tree with a server terminal as a root node, an acquisition node as a child node and an acquisition terminal as a leaf node, establishing a child node evaluation model and marking a child node where a target operating parameter is located. When network congestion occurs, the target operating parameter with a higher evaluation index can be preferentially transmitted to the server terminal for evaluation, and if an abnormality occurs, emergency measures can be taken in advance without waiting for all data to be transmitted before evaluation. In addition, when a fault level is required for secondary evaluation and complete data of the child node is required again, the required child node can be quickly located through the virtual structure tree marked after the child node, so that the operating parameters of all leaf nodes of the child node can be quickly obtained.

A method of generating an audio signal for an extended reality (XR) scene. The XR scene comprises a first virtual space, a second virtual space, and a virtual structure. A virtual sound source is located in the first virtual space and the virtual structure at least partially divides the first and second virtual spaces. The method comprises obtaining a value of a total transmission coefficient (TT) for determining how much of virtual sound reaching the virtual structure from the first virtual space is transmitted to the second space through the virtual structure. The method further comprises, based on the value of the total transmission coefficient, deriving one or more of: a value of a directional transmission (DRT) coefficient for determining how much of the virtual sound reaching the virtual structure from the first virtual space is transmitted to the second space from a direction of the virtual sound source (i.e., for determining how much of virtual sound incident on the virtual structure passes through the virtual structure and radiates into the second space from a direction of the virtual sound source) and / or (ii) a value of a distributed transmission (DBT) coefficient for determining how much of the virtual sound reaching the virtual structure from the first virtual space is transmitted to the second space in a distributed way (i.e., for determining how much of virtual sound incident on the virtual structure passes through the virtual structure and radiates into the second space in a distributed way). The method further comprises generating the audio signal for the XR scene based on the DRT and / or DBT coefficients.