65 results about "Reachability" patented technology

The invention belongs to the crossing field of network space security, software engineering and artificial intelligence, and discloses a dynamic and static cooperative vulnerability full-link intelligent analysis and verification method based on a large language model and an MCP protocol, and the method comprises the steps: S1, task initialization and context construction; s2, target code deep analysis and static taint preanalysis are carried out; s3, carrying out Fuzzing Harness directional generation and iterative optimization on the basis of the stain paths; s4, performing dynamic and static collaborative sandbox execution and dynamic stain tracking; and S5, vulnerability accessibility judgment and report output of multi-source data fusion are carried out. According to the method, a dynamic and static combination mechanism is adopted, vulnerability analysis and verification can be automatically completed for software target codes, and accurate technical support is provided for vulnerability risk grading and repairing priority ranking.

This invention relates to a multi-constraint trajectory planning method for bouncing movement on the surface of complex small celestial bodies, belonging to the field of deep space exploration. The implementation method is as follows: Establishing a closed-form relationship between bouncing movement distance, velocity, takeoff angle, and energy consumption, and a two-layer expansion radius obstacle representation model; establishing escape velocity constraints, no-fall constraints, fly-through feasibility constraints, and convex obstacle fly-through constraints; using the two-layer expansion radius obstacle representation model and multiple constraints, establishing a fly-through feasibility determination mechanism for obstacles; using Poisson-disk sampling to establish the bouncing reachability domain and an analytical solution for the energy-optimal trajectory satisfying multiple constraints; establishing a reachability domain filtering method based on the bouncing reachability domain and the analytical solution for the energy-optimal trajectory; using energy consumption as the weight, employing the A* algorithm to search for a global energy-optimal bouncing path point sequence and the corresponding velocity and takeoff angle of each path point, generating a global energy-optimal trajectory, thus realizing multi-constraint trajectory planning for bouncing movement on the surface of complex small celestial bodies.

The application discloses a kind of based on multi-source data fusion's catering intelligent decision-making method and system, it belongs to site selection decision-making technical field, by using big language model to carry out knowledge extraction to multi-source data, and to this construction catering site selection knowledge graph, based on knowledge graph automatically calculates geographic position index, traffic accessibility index, consumption potential index and the objective weight of competitive environment index, simultaneously using big language model to the qualitative description of the region to be selected is retrieved to obtain membership vector, according to objective weight and membership vector calculation comprehensive suitability degree score, the recommended site selection region is obtained by ordering comprehensive suitability degree score, it quantifies index weight by catering site selection knowledge graph, avoids subjective dependence on expert experience, quantifies index using big language model, significantly improves the accuracy and efficiency of catering site selection, adapts to large-scale multi-source heterogeneous data scene, and in the face of dynamic factors can respond to update in time, convenient for application and promotion.

This invention relates to a method and system for demand regulation of transformer substations based on time-series learning and model prediction, comprising the following steps: S1, collecting historical electricity consumption data and predicting total peak electricity consumption events in future time periods; S2, decomposing the total peak electricity consumption events into multiple peak sub-events and determining the type and intensity of the peak sub-events; S3, establishing a peak event combination graph based on the multiple peak sub-events; S4, performing control resource reachability matching on the nodes and edges in the peak event combination graph based on the structural constraint verification; S5, writing the control resource reachability matching results as node attributes and edge constraints into the peak event combination graph to form a resource-enhanced peak event combination graph, and dynamically generating differentiated control strategy combinations based on the resource-enhanced peak event combination graph and distributing them to control resources.

This invention discloses a method for generating differentiated maintenance strategies for power equipment, belonging to the field of power system and automation technology. The method involves: constructing a two-layer graph model of the power grid based on power grid topology data and traffic network data; calculating key indicators and reachability costs for any power equipment node within the model to obtain a dynamically corrected criticality matrix; performing hierarchical response based on the dynamically corrected criticality matrix to obtain a candidate resource point scheduling set; performing path planning based on real-time traffic data and the candidate resource point scheduling set to obtain a target scheduling path set; obtaining material inventory data and path traffic data corresponding to any target scheduling path in the target scheduling path set; and performing weighted scoring on the material inventory data and path traffic data to obtain a maintenance resource allocation scheme that meets preset scoring conditions. This invention discloses a method for generating differentiated maintenance strategies for power equipment, enabling differentiated maintenance of power equipment in a power grid system.

A processor-implemented method for indoor navigation map generation with accessibility information is provided. The method includes collecting a plurality of colored point cloud maps by utilizing both mobile and stationary mapping systems. The map can be aligned with an outdoor map if the GNSS or the global coordinate is available. The method also includes performing semantic segmentation based on at least implicit representation and synthetic data- aided training network for generating a plurality of structural elements and a plurality of functional elements. The method also includes performing topology-aided structural modeling by generating a structural building information modeling (BIM) using topology information at a plurality of levels of topology and by automatically applying a plurality of levels of topological constraints associated with the structural elements. The method also includes performing parametric modeling by detecting a plurality of functional objects in the functional elements using one or more simulated datasets including simulations of lighting, and environmental conditions.

This invention discloses a method and apparatus for planning the end-effector path following configuration of a mobile robotic arm based on convex set graph optimization and penalty function constraints. The planning method includes the following steps: S1: After generating the reachability graph of the robotic arm, the reachability graph is projected onto the base plane of the robotic arm to generate the inverse reachability graph of the robotic arm; S2: Based on the inverse reachability graph and the end-effector target path of the robotic arm, the feasible configuration set of the mobile chassis corresponding to each end-effector target path point is solved; S3: Based on the feasible configuration set of the mobile chassis, the feasible configurations of the mobile chassis are initially planned to obtain the initial configuration sequence of the mobile chassis; S4: The initial configuration sequence of the mobile chassis is secondarily planned to obtain the optimal configuration sequence of the mobile chassis; S5: Based on the optimal configuration sequence of the mobile chassis, the end-effector target path, and the motion relationship between the mobile chassis and the end-effector of the robotic arm, the complete configuration sequence of the mobile robotic arm is calculated.

This invention relates to the field of urban sanitation and waste collection scheduling technology, and discloses a dual-mode intelligent scheduling and back-end management method based on an AI large-scale model. The method includes: acquiring candidate equipment status data and multi-source event information stream data; parsing event information and determining the regular or emergency scheduling mode; classifying candidate equipment by reachability and adjusting for urgency; using DBSCAN to perform dense overflow clustering of reachable equipment; dividing multi-vehicle collaborative load based on a weighted Voronoi diagram; calling the AI ​​large-scale model to generate a multi-vehicle collaborative operation sequence; and dynamically adjusting and updating scheduling parameters during execution via the back-end management module. This invention solves the problem that dual-mode scheduling logic cannot coordinately respond to event-driven dense overflow scenarios within the same work cycle.

The application discloses an end-to-end multi-target tracking method and system based on layered space-time memory and cross-modal priori. The application cooperatively deals with the problems of instantaneous instability and long-time appearance mutation in the tracking process by constructing a short-term memory module and a long-term memory module. The short-term memory adopts a ring buffer structure to cache recent trajectory representation, and provides continuous association clues through time proximity and geometric accessibility screening; the long-term memory adopts a novelty-driven strategy to maintain a sparse appearance prototype library, and provides diversified appearance anchor points for occlusion reappearance and re-identification. Further, the method integrates multi-modal positioning priori, encodes it into a feature sequence through space-time alignment and coordinate unification, and performs cross-modal attention fusion with the visual dominant trajectory query. The method unifies the short-term continuity, long-term recognition and absolute position constraint in an end-to-end framework, effectively improving the identity consistency and tracking continuity of target trajectories in complex scenes.

A system and method for predicting stats for different destinations is provided. The system and method can include receiving a dataset of stats in a spatial format and a request for stats in a geographical area. The system and method can include determining a grid of cells based on the geographical area, determining an origin point in each cell in the grid of cell, determining a graph based on the grid of cells, determining a proportion of the stat to assign to each node, edge or both in the graph in the geographical area based on an overlap between the dataset of stats and the graph, and determining a total for the stat for the geographical area.

This invention discloses a method and system for optimizing draft model inference for long-context speculative decoding, belonging to the field of electronic information technology. The method includes: dividing the historical context into multiple context units according to a predetermined strategy; evaluating the relevance of the context units to select the most relevant subset of contexts for the current generation step; adapting rotational position encoding to the tree-like draft tokens generated by the draft model, and simultaneously maintaining the position encoding state during sliding window pruning and rollback; jointly constraining attention reachability relationships, outputting results and scores, and generating a candidate token sequence; submitting the candidate token sequence to the target model for verification, and performing a consistent rollback of the generation state, context cache state, and position encoding state of the failed verification token sequences. This invention optimizes the inference process of the draft model by introducing a rollbackable sliding window mechanism and a relevance-based context retrieval mechanism.

The application relates to the technical field of intelligent agent cooperation and discloses a practical training behavior analysis method based on multi-intelligent agent cooperation, which comprises the following steps: first, performing time sequence alignment and aggregation on multi-source practical training data to generate a time sequence sequence containing scene states and operation instructions; using a behavior recognition intelligent agent to establish a scene evolution model, performing multi-step deduction on a virtual environment state based on the sequence; in view of a preset reference instruction set, evaluating the reachability of the instruction set relative to risk constraints and task targets in the deduced state, and calculating a decay index reflecting the degree of recoverable space contraction; then, using a target matching intelligent agent to calculate task completion degree which is deeply coupled with the decay index, and combining resource consumption data of the practical training process to generate a comprehensive score; finally, mapping the comprehensive score to a qualified or unqualified evaluation result based on an offline calibrated qualified benchmark value. The application can accurately identify high-risk operation trends and realize the standardization of practical training evaluation.

This invention discloses an RGB-D grasping candidate ranking calibration method for improving the reliability of a robot's single grasping operation. The method constructs a depth validity mask and depth boundary response from RGB-D input, and performs a reliability-weighted fusion of RGB features and depth features accordingly. After the grasping detection network outputs grasping quality, angle, size, and uncertainty results, the grasping score is calibrated using pixel-level uncertainty and candidate rectangle neighborhood uncertainty. During the training phase, a geometric inverse mapping consistency constraint and a ranking loss based on the maximum calibration score interval between positive and negative regions are constructed. During the inference phase, a candidate sequence is formed based on the calibration score, and a robot-executable grasping pose is generated through depth validity, gripper opening, and manipulator reachability checks.

The application relates to the technical field of electric digital data processing, and particularly discloses a product information risk assessment method and system based on a large model, which comprises the following steps: simultaneously scanning source code and products by a scanner to generate a source dependency list and a product component list, performing main name normalization and alignment through a large model, locking a warehouse and a version according to the source dependency list, analyzing original code by a parser to generate a cross-language path logic group, combining the product component list to correct a real package version, converging into a unified intermediate representation, performing reachability and joint analysis on the unified intermediate representation to obtain a risk candidate path set and complete path scoring, finally, according to a preset weight and a running period evidence, modifying the score, screening out a risk path, matching a disposal strategy, and generating executable strategy code by the large model for risk control of a gateway or a service side.

The application discloses a fruit autonomous picking method and system based on active perception and visual arm cooperative optimization, and relates to the technical field of robot control. The method comprises the following steps: acquiring a color image and a depth image of a working environment, and performing semantic segmentation on the color image by using a preset zero-sample learning model; performing semantic point cloud extraction on the depth image based on a mask corresponding to a target category, and constructing a fruit tree environment map containing spatial information and semantic information; performing candidate target screening and extraction according to the fruit tree environment map, and generating a candidate grasping pose according to the screened candidate target; performing kinematic reachability analysis on all candidate targets and candidate grasping poses, and calculating a joint score; determining a next grasping object according to the joint score; and picking the next grasping object according to a preset trajectory. The application deeply couples active visual perception and mechanical arm grasping action, and realizes efficient fruit autonomous picking of the robot.

The application discloses a desulfurization process optimal path optimization method based on deep learning, comprising the following steps: collecting and preprocessing working condition data, and constructing a multivariate working condition state vector; constructing a dynamic characteristic matrix, extracting trend fluctuation coupling characteristics and generating candidate actions; judging working condition accessibility based on a turn-back partition, and screening candidate control action sequences; constructing a process inertia perspective chain, layering according to response characteristics and establishing a node relationship network; constructing an improved StemGNN model, calculating space-time correlation and screening executable actions; generating candidate paths based on executable actions, and performing safety screening through uncertainty arbitration; and comprehensively optimizing candidate paths, and outputting optimal control action sequences and execution sequences. The application realizes intelligent generation and stable optimization of desulfurization process control paths under complex working conditions by constructing multivariate working condition characteristic analysis and control path optimization mechanisms.

This invention relates to a method for constructing a spatial graph network that realizes cross-layer coupling between the ecological layer and the public space layer, comprising: Step S1: obtaining the grid range covered by each ecological patch and public patch; Step S2: generating ecological nodes and public nodes, as well as the area and centroid position of all nodes; Step S3: using ecological nodes as nodes of the ecological layer graph network, generating inner edges of the ecological layer between ecological nodes whose centroid position distance is less than an ecological distance threshold, and generating the weights of each inner edge of the ecological layer graph network; Step S4: using public nodes as nodes of the public layer graph network, generating inner edges of the public layer between public nodes whose centroid position distance is less than a public distance threshold, and generating the weights of each inner edge of the public layer graph network; Step S5: constructing inter-layer edges between ecological nodes and public nodes whose centroid position distance is less than a public distance threshold, and generating the weights of the inter-layer edges based on the straight-line distance, traffic accessibility, ecological impact intensity, and human flow relationship between the corresponding ecological nodes and public nodes. Compared with existing technologies, this invention can deeply explore the mutual coupling relationship between the ecological layer and the public layer.

A compiler creates a dependency graph for functions in an input program. The dependency graph includes nodes corresponding to commands within the functions and edges corresponding to dependencies between commands. The compiler performs forward reachability analysis on the dependency graph to eliminate redundant dependencies. The compiler also adds a minimum set of back edges to the dependency graph to enforce circular resource dependencies in the input program. Then, based on the contents of the dependency graph, the compiler assigns synchronization primitives (such as semaphores) provided by the multiprocessor computing system to commands within the functions of the input program.

A qualitative and quantitative (Q2) method for verifying safety of movement of complex learning-enabled cyber-physical systems (Le-CPS) using a computer is provided. The method includes determining an initial set of states of the Le-CPS. The computer is used for determining a reachable set of the Le-CPS using a probstar reachability algorithm saved in memory of the computer. The computer includes a Q2 algorithm saved in the memory that is configured is used to simultaneously check (i) if the reachable set satisfies a predetermined safety constraint (qualitative verification) and (ii) determine a probability of satisfaction that the reachable set will satisfy the predetermined safety constraint (quantitative verification). Movement of the Le-CPS is planned based on the step of using the computer including the Q2 algorithm.

The application discloses a method for intelligent routing optimization based on machine learning in a software-defined network, and relates to the technical field of communication networks, comprising the following steps: obtaining link state encoding and alleyway topology data of a target object; performing topology skeleton mapping on the alleyway topology data to generate a twin topology base map; performing time series difference operation on the link state encoding to generate a sequence of encoding difference values, and performing abnormality measurement on the sequence of encoding difference values to generate a sequence of abnormality evidence. According to the method, communication abnormality is searched in the twin topology base map, a block probability map is generated by forward accumulation and backward smoothing, and is updated to a reachability update twin graph, so that noise misjudgment is inhibited, reachability is continuously evolved with abnormality, an action mask is generated by double-layer mask synthesis of the reachability update twin graph and the block probability map, and an actionable candidate solution is constructed, thereby, unreachable and high-risk actions are eliminated in advance, and invalid exploration is reduced.

This invention discloses a machine learning-based method for user operation and maintenance decision-making in an embedded tag e-commerce platform, comprising: collecting multi-source behavioral and operation and maintenance interaction data to construct a current session state set; constructing an intent impulse sequence and generating a tag reachability tensor, constructing and updating a tag debt state set; constructing a multi-layer hypergraph of tag position and action, and constructing a corresponding multi-scale input feature set; inputting the multi-scale features into an improved deep factorization machine model to generate a counter-evidence net effect value; selecting target operation and maintenance solutions based on the net effect value, and constructing a tag commitment execution chain across page positions; controlling the embedding position and timing of embedded tags and the execution of operation and maintenance actions, and updating the model and the tag commitment execution chain. This invention achieves accurate reach and continuous operation and maintenance decision-making for embedded tags in multi-page, multi-stage sessions by constructing a tag debt memory mechanism and an improved deep factorization machine dual-mirror recognition model.

The application provides a network fault repair method and system based on multi-protocol detection, and relates to the technical field of network communication, which comprises the following steps: collecting link layer discovery protocol, address resolution protocol, routing table and bidirectional forwarding detection session state information; constructing a directed graph model comprising a plurality of nodes and links; identifying a primary path of current traffic according to the bidirectional forwarding detection session state information, and determining a fault link segment according to the primary path; generating a dynamic state identifier set according to the address resolution protocol, the routing table and the fault link segment; and generating a candidate optimized path according to the directed graph model and the dynamic state identifier set. Three key states, i.e., forwarding reachability, policy blocking and link fault, are accurately identified. The dynamic state identifier set truly reflects the actual carrying capacity of a link to service traffic, and fundamentally avoids the problem of traffic black hole after switching caused by misjudgment of routing reachability, i.e., forwarding reachability.

The application relates to the technical field of military simulation data engineering, and discloses a multi-dimensional semantic knowledge graph framework construction method for military simulation, which comprises the following steps: analyzing an object model extension component of a high-level architecture to construct a simulation object model incremental state sequence, and mapping a multi-dimensional coordinate of data distribution management into a discretized grid data routing domain; analyzing a federal protocol session layer log to determine a member connection reliability index, and combining the grid data routing domain to calibrate a data reachability index of each grid unit; generating an extended semantic type identifier for a semantic assertion, constructing a semantic missing tendency evaluation model associated with the extended semantic type identifier and the data reachability index, and calculating a missing risk measurement value for an incremental semantic field; and generating a global consistency compensation factor based on the missing risk measurement value and fusing the semantic triple.

The application provides a reachability index updating method and system of a graph data structure satisfying time and step constraints, acquires a time two-part graph data, converts it into a time graph, and constructs an index based on a 2-hop label; a local incremental updating optimization strategy based on node degree sorting is used to update the index; and a parallel strategy is used to optimize the index updating. The local incremental updating strategy based on node degree sorting of the application combines a multi-level pruning mechanism, so that the updating process only performs a repair operation on the affected local area, avoiding repeated construction of a global index. Compared with a traditional full reconstruction method, the method can significantly reduce unnecessary propagation and checking operations, so that the index maintenance overhead is reduced by more than 50%. After identifying the independent affected area, a parallel strategy is used to synchronously update the local index, load decomposition of the updating task is realized, the overall updating throughput is greatly improved, and the real-time index maintenance requirement in a high dynamic rate scene can be met.

The present application relates to the technical field of data processing and spatial operations optimization, and particularly relates to a pension facility spatial layout optimization method based on multi-model and reachability analysis, which obtains the spatial distribution data of the elderly population by constructing a hexagonal grid, introduces a Gaussian distance decay function to improve the two-step moving search method, combines the multi-mode travel time cost and facility service capacity to measure the spatial reachability, identifies the supply-demand mismatch area based on the standard score model, constructs a hierarchical optimization strategy by fusing the P-barycenter location model and the reachability variance minimization model for the area, obtains the location scheme through the location optimization stage respectively, and obtains the configuration scheme by using the quadratic programming algorithm in the service capacity configuration stage, and the present application realizes the organic unification of the pension resource configuration efficiency and spatial fairness through the technical means of multi-model fusion and hierarchical progressive optimization.

This invention discloses a metadata-driven adaptive Text-to-SQL method, relating to the field of natural language processing technology. It involves: initially screening pattern items based on database metadata to obtain a preliminary pattern subset; then, obtaining a simplified pattern subset through disambiguation verification of candidate columns; constructing a weighted set of join constraints based on foreign key relation metadata; determining an anchor table set based on the simplified pattern subset; calculating the reachability strength of each table; selecting and forming a set of joinable closure tables and determining the structurally joinable retrieval domain; establishing an inverted index for columns in the retrieval domain; allocating recall quotas based on the relevance scores of each column and the reachability strength of its respective table; and retrieving and filtering to generate a predicate evidence set. Finally, organizing database prompts based on the simplified pattern subset, relevant join constraints, and the predicate evidence set, and inputting them into an SQL generation model to generate executable SQL. This method adaptively constrains and enhances the Text-to-SQL generation process using metadata.

The embodiment of the invention provides a holographic interaction intention prediction method based on multi-modal fusion and uncertainty perception, and belongs to the technical field of data processing, and the method specifically comprises the steps: collecting an environment, a user state and holographic data through a sensor, and generating an environment model and an occupied raster map; calculating the accessibility of the user and the interaction object, and generating a physical barrier mask matrix; mapping the data to a high-dimensional feature space, fusing features through an attention mechanism, and generating a dynamic fusion state vector; inputting the vector to a prediction model, and calculating an evidence vector and an uncertainty index of the intention; dynamically adjusting the Kalman filter to smooth the interaction state to obtain a final intention instruction; and executing different holographic interaction responses according to the instruction and the uncertainty. Through the scheme disclosed by the invention, physical environment constraints can be deeply fused, the uncertainty can be honest and quantitatively predicted, and the tracking stability can be adaptively adjusted according to the uncertainty.

The application discloses a ball and link system control method based on fuzzy sliding mode and periodic event triggering, first constructs a state space equation of the system and carries out regular transformation to obtain a standard system model; then, a fuzzy sliding mode switching function is established in combination with a fuzzy membership function, and a periodic event triggering mechanism is designed; on the basis, a periodic event triggering fuzzy sliding mode controller is constructed, an actual sliding mode band is constructed by analyzing a periodic sampling error, state trajectory is ensured to enter the band in a limited time and remain in the band, and the reachability of the sliding mode surface is ensured; then, a controller gain matrix is solved based on stability analysis of the sliding mode, is brought into the controller to complete closed-loop control on the ball and link system, and a system sampling period is established. The application uses the fuzzy membership function to design the sliding mode surface, has lower conservativeness, and uses the periodic sampling to design the event triggering mechanism, theoretically eliminates the Zeno phenomenon, and provides an efficient control scheme for a system with limited network and communication resources.

Embodiments of the present application provide an end-side model inference dynamic optimization method, device, equipment and medium. The method surrounds the end-side single-flow CPU inference path, establishes a performance, energy consumption and thermal state measurement system covering the prompt processing stage and the self-recurrence generation stage, and adopts a fixed word window unified index statistics and configuration decision. The offline measurement is performed on the CPU frequency, thread number and affinity configuration space, the feasible boundary and preferred priori are refined, and a discrete neighborhood search dynamic optimization method under the offline priori constraint is proposed. Thus, taking a unified comprehensive cost function as an optimization target, the online search is limited in the feasible neighborhood obtained through offline screening under the first word delay, thermal safety and frequency accessibility constraint, and combined with the minimum residence, anti-shake, limited memory and controlled jump-out mechanism, the low-overhead joint adjustment of the CPU frequency and thread number is realized, while the adaptability to state changes is improved.