40results about How to "Reduce overdependence" patented technology

The invention belongs to the technical field of artificial intelligence, discloses a human action understanding method based on a large language model and self-adaptive contrast decoding, and provides an FLARE framework, which enables actions and language semantics to be aligned by introducing a pre-trained frozen large language model LLM and a frozen action encoder, and utilizes a self-adaptive contrast decoding AdaCD method to understand actions and language semantics. According to the method, excessive dependence on language priori is reduced, so that the understanding ability of the model on actions is improved, the method can process various action understanding tasks, the generalization ability of the human action understanding method is improved, and meanwhile, the problems of task specificity and computing resource consumption in the prior art are reduced.

The invention discloses a valet parking method and system, and relates to the technical field of intelligent driving, and the valet parking method comprises the steps that a vehicle end responds to a connection request sent by a parking lot end, an encryption channel is established with the parking lot end, and a parking lot entrance positioning encryption data packet sent by the parking lot end is received through the encryption channel; when the parking lot entrance positioning encryption data packet is received, the initial pose of the vehicle is set according to the parking lot entrance positioning encryption data packet, and a receiving confirmation message is sent to the parking lot end; acquiring vehicle environment data, and constructing a real-time local map at the entrance of the parking lot based on the initial pose of the vehicle and the vehicle environment data; obtaining a valet parking route map, and matching the real-time local map with the valet parking route map to obtain a route matching result; and the vehicle is controlled to complete valet parking according to the route matching result, and the route matching success rate of the valet parking function can be improved with low cost.

The invention discloses a pipeline leakage point positioning method considering gas-liquid two-phase flow, and belongs to the technical field of pipeline leakage point positioning. In order to solve the problem that an existing pipeline system leakage detection method is single in monitoring form and inaccurate, the method comprises the steps that a horizontal pipeline gas-liquid two-phase stratified flow motion model is established; judging the relation between the gas phase and the liquid phase in the pipeline and the pressure of the gas phase and the liquid phase relative to the pipeline; expressing and calculating the leakage rate of the pipeline system; establishing a two-phase flow model for leakage point positioning; and finally positioning a leakage point. The method is suitable for leakage point positioning of a single-phase flow pipeline system and is also suitable for leakage point positioning of a gas-liquid two-phase flow pipeline system. Necessary data are provided for leakage point positioning through a model simulation method, excessive dependence on pipeline system data monitoring is reduced, the leakage point positioning efficiency is greatly improved, and the leakage point positioning cost is reduced.

The present invention discloses an enterprise operation supervision system for enterprise management, which specifically relates to the field of equipment supervision. It collects signal strength data of vehicle-mounted monitoring equipment, uses the signal average and standard deviation to dynamically evaluate the stability of the communication environment, and intelligently determines whether it is suitable for data transmission. In an unstable environment, ensure that data is securely stored locally; in a good environment, extract the data update change index and the fault impact trend index to calculate the pre-evaluation coefficient, and compare it with the early warning analysis threshold to generate a warning signal, so as to timely discover potential problems and improve equipment reliability. By calculating the distance from the maintenance point, the vehicle's motion status can be monitored, and the early warning strategy can be dynamically adjusted to improve the safety, efficiency and controllability of vehicle operations, and more intelligently supervise corporate property.

The invention discloses a cutting chain type transverse shaft type cutting mechanism suitable for a coal roadway and a half-coal-rock roadway and a heading machine. The cutting mechanism comprises a main machine frame, a motor, a cutting chain, a left cutting head and a right cutting head, the motor, the cutting chain, the left cutting head and the right cutting head are installed on the main machine frame, the left cutting head and the right cutting head are located on the two sides of the cutting chain respectively, and the motor drives the cutting chain to rotate and drives the left cutting head and the right cutting head to rotate; wherein a plurality of middle cutting teeth are distributed on the cutting chain; a plurality of left cutting teeth are distributed on the left cutting head; a plurality of right cutting teeth are distributed on the right cutting head; at least part of the middle cutting picks are matched with the left cutting picks to form a spiral line shape, and at least part of the middle cutting picks are matched with the right cutting picks to form a V shape. The problem that an existing cutting mechanism is low in cutting efficiency is solved.

The invention relates to the technical field of power grid operation, and provides a power grid operation safety distance accurate sensing device and method in a non-fixed scene, and the device comprises a hardware device and terminal software which is in communication connection with the hardware device. The hardware device is integrated and deployed on a movable support or a holder, and comprises a laser radar sensor which is used for rapidly scanning a power grid operation scene and obtaining three-dimensional point cloud data including three-dimensional coordinates, laser intensity and echo information; the high-definition visible light camera is approximately parallel to the optical axis of the laser radar sensor and is used for collecting real-time video images of the operation scene; an AI visual analysis module and a point cloud processing unit are arranged in the embedded processor; and the terminal software is deployed on the edge computing equipment or the background server and is used for receiving and processing the data from the hardware device. According to the invention, the safe distance can be accurately sensed.

The invention belongs to the technical field of wind power generation, and particularly relates to a wind turbine generator state evaluation method and a related device, and the method comprises the steps: obtaining an electromagnetic transient response signal of a to-be-evaluated wind turbine generator; performing empirical mode decomposition on the electromagnetic transient response signal of the wind turbine generator to be evaluated to obtain a plurality of image forming model components; performing feature extraction on each image forming model component, and performing normalization processing to obtain a feature vector; classifying the feature vectors by using a pre-constructed state classification model to obtain a state evaluation result of the to-be-evaluated wind turbine generator; wherein the pre-constructed state classification model is a pre-trained support vector machine; according to the method, the fault features in the electromagnetic transient response signals of the wind turbine generator can be effectively extracted, and the accuracy and robustness of state evaluation are improved; meanwhile, non-linear and non-stationary signals can be decomposed in a self-adaptive mode, the time domain features of the signals are reserved, and the information amount and accuracy of state evaluation are improved.

The invention discloses a distributed computing fault snapshot generation system and method and a storage medium, and relates to the technical field of distributed computing, and the distributed computing fault snapshot generation system comprises a master control node and computing nodes; the master control node is used for capturing a calculation task exception event and adding context information for the calculation task exception event; reporting the calculation task exception event added with the context information to the calculation node with the calculation task exception; the context information at least comprises abnormal information; the computing node is used for analyzing the type of abnormal information and executing a diagnosis strategy according to the type of the abnormal information when the computing task abnormal event is received; and collecting snapshot data, packaging the collected snapshot data, and uploading the packaged snapshot data to a distributed storage. The system can actively capture a complete state snapshot at the moment of a fault, and can effectively support deep root cause analysis. In addition, the system automatically executes a proper diagnosis strategy and persistently stores snapshot data, and has accurate triggering and intelligent decision-making capabilities.

The application provides a strain of Arthrobacter sp. ZJUHUYJH1 and application thereof in prevention and treatment of paratyphaceae root-knot nematodes, and belongs to the technical field of prevention and treatment of plant root-knot nematodes. The Arthrobacter sp. ZJUHUYJH1 provided by the application has a preservation number of CCTCC NO: M 2025922, is isolated from rice rhizosphere soil infected by paratyphaceae root-knot nematodes, has a good prevention and treatment effect on paratyphaceae root-knot nematodes, a rectification mortality of 91.4% on second instar larvae of root-knot nematodes, can reduce the number of root-knots and the ratio of the number of root-knots to dry weight of rice infected by paratyphaceae root-knot nematodes, and has potential and practical application value in the research of paratyphaceae root-knot nematodes.

The invention relates to the field of distilled liquor classification, and provides an intelligent distilled liquor classification method and device in order to improve the distilled liquor classification efficiency and reduce dependence on expert experience at the same time. According to the method, the color features and the compound features capable of being quantitatively expressed in the distilled liquor are spliced to serve as input of the intelligent distilled liquor classification model, classification is completed based on the intelligent distilled liquor classification model, the processing efficiency of the whole process is higher, excessive dependence on traditional artificial experience is reduced, and classification is more intelligent and standardized.

The invention discloses a pod propulsor structure optimization design method based on data driving. Comprising the steps of defining a design variable and an optimization target, constructing a parameterized geometric model, building a full-process automatic simulation analysis framework, generating a sample space through experimental design, obtaining a complete sample data set, constructing a high-precision agent model, carrying out algorithm optimization based on the agent model, verifying optimal design and calculating a data closed loop. And continuously updating the model. According to the method, the trained data driving agent model is utilized, in a huge design space, performance evaluation can be completed only in an extremely short time, the calculation period is greatly shortened, and a global optimal solution set can be rapidly obtained in combination with a global optimization algorithm; according to the method, various factors such as design requirements, test design, an agent model construction method and an optimization algorithm are comprehensively considered, the method can be suitable for structural design of other pod propellers with different production models or different design requirements, and effective guidance is provided for subsequent design of the pod propellers.

This invention provides an aortic pre-fenestration branch guidewire guidance system, comprising a central catheter, a capture ring, and a sliding cannula. The capture ring is connected to the outer wall of the central catheter. The sliding cannula is coaxially sleeved on the outside of the central catheter and can slide along the axial direction of the central catheter. When the sliding cannula moves proximally, the capture ring is released from the constraint of the sliding cannula and unfolds radially outward of the central catheter, thereby forming a capture opening that can accommodate the target guidewire. When the sliding cannula moves distally, the sliding cannula radially compresses the capture ring in its unfolded state, forcing the capture ring to undergo radial compression deformation. The radial compression deformation of the capture ring reduces the cross-sectional area of ​​the portion surrounding the target guidewire, thereby clamping the target guidewire and improving the success rate.

The invention discloses a cordyceps sinensis bionic culture method which comprises the following steps: determining a target area in which cordyceps sinensis is naturally distributed, arranging a signal tower A for each area, and collecting climate data of the target area; a plurality of independent compartments are arranged in the culture base, each compartment corresponds to one target area, a signal tower B is arranged, and the signal tower B receives data transmitted by the signal tower A; the inoculated hepialus armoricanus larvae are evenly distributed in all the compartments; monitoring the growth information of the hepialus armoricanus larvae in different time periods in each compartment in real time, and recording the growth information of the hepialus armoricanus larvae in different time periods; growth information in different time periods is analyzed, climate data are reconstructed and stored in a signal tower B, and the climate data in different compartments are controlled to be unified by using the signal tower B; according to the method, the natural environment can be accurately simulated, it is ensured that the growth of cordyceps sinensis is closer to the natural state, the ecological requirements of cordyceps sinensis hyphae and hepialus armoricanus larvae are met, and quality improvement and accumulation of active ingredients are facilitated.

The invention discloses a method for preparing cyclopentenone substances through furanol rearrangement-hydrogenolysis, and belongs to the field of fine organic chemicals. The process comprises the following steps: mixing furfuryl alcohol compounds (furfuryl alcohol, 5-methyl furfuryl alcohol and 2, 5-dimethylolfuran) with a water solvent according to different concentrations, adding a cobalt disulfide (CoS2) catalyst, and reacting in a hydrogen environment of 0.5-1.5 MPa at the reaction temperature of 160-200 DEG C; the 2-cyclopentenone / 3-methyl-2-cyclopentenone is obtained by adopting a rearrangement-hydrogenolysis method, the process is simple, the operation is convenient, the catalyst is cheap and easy to obtain and can be recycled for multiple times, and industrialization is easy. When the conversion rate of 2-cyclopentenone synthesized by furfuryl alcohol reaches up to 100%, the yield reaches up to 90%; when the conversion rate of the 3-methyl-2-cyclopentenone synthesized from the 5-methylfuranol and the 2, 5-dimethylolfuran is up to 100%, the yields are up to 80% and 78% respectively.

The invention belongs to the technical field of electrochemical in-situ characterization, and particularly relates to a packaging layer processing method for near-normal-pressure photoelectron spectroscopy analysis. The method comprises the following steps: constructing an initial packaging layer covering liquid electrolyte on the surface of a working electrode; integrating the samples on a sample carrying table and transferring the samples to an analysis equipment chamber; the thinning system of the equipment is used for carrying out in-situ thinning on a target area, the analysis beam of the equipment is synchronously or alternately used for monitoring the signal change of the thinned area in real time, and the thickness of the packaging layer is precisely regulated and controlled to the target nanoscale based on signal feedback. According to the method, through a strategy of first packaging and then in-situ optimization, the contradiction that effective sealing and high signal transmittance are difficult to consider in a traditional method is solved, high-quality in-situ and dynamic characterization of an electrode-electrolyte interface in a volatile liquid electrolyte environment is realized, and the accuracy and reliability of test data are remarkably improved.

The invention relates to a fine-adjustable side window mounting structure and method for a passenger train, and belongs to the manufacturing technology of rail transit vehicles. According to the technical scheme, the extending direction of an installation block long round hole in an installation block and the extending direction of a window frame long round hole in a window frame installation inclined face are in a crossed state, adjustment freedom degrees are provided in the vehicle width direction and the vehicle length direction, and fine adjustment in the vehicle width direction and the vehicle length direction is conducted respectively; adjusting freedom is provided by adjusting the matching position of the mounting block and the window frame mounting slope in the vertical direction of the vehicle, and the vehicle is adjusted in the vertical direction. The method has the beneficial effects that high precision and controllability are achieved, and the installation comprehensive error is controlled within the range of + / -4 mm through a unique three-way fine adjustment mechanism; 25-degree inclined plane matching is combined with tooth-shaped meshing, and the anti-micro-motion performance is excellent in a vibration environment. Adjustment is convenient and fast, and the adjustment direction is clear due to the design of the crossed oblong holes. Due to the design of the split type mounting blocks, stress concentration is avoided, the rigidity of the whole structure is good, and the service life is long.

The invention relates to the technical field of image segmentation, in particular to a tongue coating segmentation image and digestive disease classification method. According to the method, the position of a particle center point is determined by introducing multi-scale LoG filtering, and tongue particles and tongue coating particles are distinguished according to the law of spatial distribution and density change characteristics; according to the method, local adaptive threshold adjustment is carried out on other pixel points through the change of uneven particle density of different parts, more accurate tongue coating areas and tongue quality areas are obtained through division, and subsequent classification is carried out based on the tongue coating areas and the tongue quality areas. According to the method, the tongue coating area is divided more accurately through multi-scale LoG filtering and spatial statistical characteristics, a reliable data basis is provided for subsequent auxiliary classification of digestive diseases, and the classification accuracy is improved.

This invention belongs to the field of tunnel engineering, specifically relating to a data-mechanism-model-driven intelligent analysis method and system for rock mass strength parameters. It includes: generating a simulation dataset adapted to the current engineering rock mass strength and geological conditions through high-fidelity drilling numerical simulation, and training a multilayer perceptron as a physical information proxy model to achieve physical mapping of the rock drilling process. Secondly, the three-dimensional parameters are processed by three-dimensional spatial aggregation and segmentation. Combining graph neural network analysis and transformer model information extraction, a hybrid loss function for multi-dimensional data is defined. This hybrid loss function, together with the physical loss function, forms an optimized loss function with physical information and drilling data. An automatic optimization process for the loss function is then performed to achieve rock mass strength prediction. This invention can effectively analyze rock mass strength parameters.

This application discloses an in-vehicle operation guidance method and vehicle, applied in the field of smart cockpit technology. The method identifies at least two guidance channels with different sensory modalities as target channels from multiple guidance channels, and configures the visual guidance channel to present the guidance content visually in a position that does not overlap with the driver's current primary field of vision. This allows the user to receive visual guidance information without shifting their gaze, while guidance channels with other sensory modalities further assist in information transmission. As a result, the driver's gaze remains consistently focused on the primary field of vision ahead, reducing safety hazards caused by the driver's gaze leaving their primary field of vision. Furthermore, guidance information can be efficiently received through multimodal collaboration, improving the efficiency of guidance information transmission while ensuring driving safety.

The invention relates to the technical field of shield tunneling machine automatic control, in particular to an automatic control system and method for a shield tunneling machine muddy water conveying system, and the method comprises the following steps: when a shield tunneling machine starts tunneling or stops tunneling, executing a bypass mode, and transiting from the bypass mode to a tunneling mode or transiting from the bypass mode to a stop mode. According to the automatic control system and method, the tunneling mode, the stop mode and the bypass mode are set, the muddy water conveying system is automatically controlled to enter the bypass mode according to whether the shield tunneling machine conducts tunneling or not, then the muddy water conveying system enters the tunneling mode or the stop mode from the bypass mode, and through transition of the bypass mode, the muddy water conveying system is automatically controlled to enter the tunneling mode or the stop mode. And the stability of muddy water circulation and pressure adjustment can be ensured.

The invention provides an intelligent positioning and protocol recommendation method, device and system for magnetic resonance scanning and a medium, and the method comprises the steps: obtaining a positioning image of a patient, and recognizing the anatomical structure information in the positioning image through a pre-trained deep learning model; based on the identified anatomical structure information, according to a preset clinical rule base, generating at least one suggested scanning range frame on the positioning image; and displaying the generated scanning range frame in an interactive interface, and adjusting the scanning range frame by responding to a user operation instruction. According to the invention, the traditional manual operation is converted into a one-key confirmation mode or a fine adjustment mode, so that the scanning preparation time is greatly shortened, and the equipment turnover rate is improved. Based on an AI algorithm and a preset rule, man-made differences are eliminated, and consistency and normalization of scanning results in different scenes are guaranteed. Dependence on personal experience is reduced, new user training cost is reduced, low-level errors are avoided, and user experience is optimized through an intuitive and intelligent interaction interface.

The application relates to the technical field of drainage devices, and a multifunctional negative pressure drainage ball, which comprises a negative pressure ball body, liquid inlet and liquid outlet arranged at two ends of the negative pressure ball body, a liquid inlet pipeline and a pressure adjusting device, one end of the liquid inlet pipeline is connected with the liquid inlet through a three-way pipe, a pressure gauge is connected to the three-way pipe, the pressure adjusting device is connected to the liquid inlet pipeline, the pressure adjusting device comprises a scale cylinder and a piston rod, the piston rod is slidably connected with the scale cylinder, one end of the scale cylinder away from the piston rod is connected with the liquid inlet pipeline and a cleaning liquid container through a connecting pipe and a cleaning pipe, first and second valves are respectively arranged on the connecting pipe and the cleaning pipe, and third and fourth valves are respectively arranged at two ends of the connecting position of the liquid inlet pipeline and the connecting pipe. The application can effectively improve the stability and smoothness of the drainage operation.

The application discloses a gyrotron integrated test analysis method, and belongs to the field of high-power microwave device testing. The method designs a steady-state working condition debugging starting point according to a starting current curve; and combines experimental power measurement data to correct a simulated power characteristic surface. The method provided by the application combines theoretical simulation with experimental data, realizes rapid identification and optimization of gyrotron working characteristics, improves debugging efficiency and precision, and is suitable for engineering debugging and operation optimization of an electron cyclotron resonance heating system. The method has the advantages of process standardization, high safety and strong migration, and can improve debugging efficiency, enhance calibration precision and reliability.

The invention relates to the field of feed, in particular to complete grass pellet feed as well as a preparation method and application thereof. The complete grass pellet feed comprises a coarse feed and a concentrate supplement feed, and by improving the addition amount of the coarse feed and the length of the coarse feed, physical effective NDF is increased, rumination is promoted, and acidosis is reduced; the phenomenon of picky feeding of animals is effectively reduced, nutrition intake is more balanced and accurate, and growth and development are promoted; transportation and storage are more convenient, feeding is easy, and cost is reduced.

The utility model provides a novel funnel chest orthopedic system, include: orthopedic subassembly and be used for overturning orthopedic subassembly's overturning subassembly, orthopedic subassembly includes orthopedic main part and be used for fixing orthopedic main part's fixed component, orthopedic main part is arc, orthopedic main part includes first end and with first end opposite second end, first end's width is less than second end's width, fixed component includes first fixed part and second fixed part, first fixed part with second fixed part all is provided with recess structure, first end passes recess structure of first fixed part, and first fixed part with first end passes detachable mode connection, second end passes recess structure of second fixed part, and second fixed part with second end passes detachable mode connection, the utility model discloses novel funnel chest orthopedic system is more small, operation is more simple, and security is higher.

The application provides a tracking method for high-speed high-maneuvering target, through adaptive adjustment of model parameters in time sequence by using adaptive fence division, each motion model in the interactive model set can spontaneously adapt and better match the current target maneuvering mode, the application can effectively reduce excessive dependence of the prior information of model parameters in the model set of the traditional maneuvering target tracking algorithm, realize more accurate tracking of the maneuvering target; and the application faces highly complex target maneuvering mode, adopts the more robust dangerous sensitive filtering estimation method for model type and target motion measurement noise uncertainty, therefore can effectively reduce excessive dependence of the prior information of model type in the model set of the traditional maneuvering target tracking algorithm, and meet the tracking demand of high-speed high-maneuvering target.

The embodiment of the invention relates to the field of rehabilitation training, and discloses a rehabilitation training device based on multi-split muscle activation feedback and electronic equipment, and the device comprises a signal collection module which is arranged in a body surface projection area of multi-split muscle and is used for obtaining a multi-split muscle signal in a process of performing rehabilitation training for a target training action; and the processor is connected with the signal acquisition module and is used for calculating an activation level in real time according to the multi-split muscle signal and outputting a feedback signal for reflecting a relationship between the activation level and a target training interval. According to the rehabilitation training device based on the multifidus muscle activation feedback, the problems of accurately monitoring the multifidus muscle activation condition and providing instant feedback and action guidance in dynamic waist rehabilitation training are solved, so that the multifidus muscle activation condition can be accurately monitored in the dynamic waist rehabilitation training, and the rehabilitation training efficiency is improved. Immediate feedback and action guidance are provided, and real-time and understandable biological feedback is formed in daily rehabilitation training.

The invention is suitable for the field of hydraulic structure design of water conservancy and hydropower engineering, and provides a gravity dam economic section automatic design method and system based on an intelligent optimization algorithm, and the method comprises the steps: inputting design parameters, and initializing an algorithm population; performing stability calculation and constraint verification on each candidate section in the population; the candidate sections are evaluated through a preset fitness function, the function applies a great penalty value to the sections which do not meet the stability constraint, and the volumes of the sections which meet the constraint are directly used as fitness values; the population is iteratively evolved by using an intelligent optimization algorithm, and individuals with better fitness values are continuously screened out; and finally outputting an economic section which meets all safety constraints and has the minimum concrete volume. According to the method, a complex constrained engineering optimization problem is converted into an unconstrained optimization problem, automation and intellectualization of gravity dam section design are achieved, the design efficiency is greatly improved, and the engineering amount of a dam body can be remarkably reduced on the premise that safety is guaranteed.

The present application relates to the technical field of intelligent transportation and vehicle behavior prediction, in particular to a real-time trajectory prediction method for over-limit and over-load vehicles based on collaborative filtering, especially to a trajectory prediction method for over-limit and over-load vehicles, combining multi-dimensional behavior clustering and BiLSTM-Transformer hybrid model, introducing a DBSCAN clustering step based on multi-dimensional comprehensive similarity, which can automatically and accurately classify over-limit and over-load vehicles with similar behavior patterns into the same group. The most similar group is matched for the target vehicle for model training, effectively solving the problem of fuzzy model learning target caused by data heterogeneity. At the same time, the hybrid model architecture layer is good at capturing local time series dynamics and short-term driving rules, while the self-attention mechanism of the encoder can effectively model global long-distance dependence and complex trajectory patterns, and the two complement each other, thus achieving a leap in prediction accuracy as a whole. The algorithm can automatically identify noise points, further enhancing the robustness of the method to data noise.