20results about How to "Reduce high cost" patented technology

This invention provides a Chinese text-to-speech method based on an offline unit library constructed using a multimodal large model, relating to the field of text-to-speech technology. The method includes the following steps: receiving the target text to be synthesized; parsing the target text for scene tags, sentiment tags, and stress positions using a lightweight semantic parsing model; querying a pre-constructed offline unit library based on the parsing results, and retrieving corresponding candidate acoustic prosodic units for each Chinese character or word; constructing a state network with the candidate acoustic prosodic units as nodes; calculating the optimal path based on the inter-unit transition cost and semantic matching degree in the offline unit library using the Viterbi algorithm, and selecting the globally optimal acoustic prosodic unit sequence; and smoothly concatenating the candidate acoustic prosodic units in the optimal acoustic prosodic unit sequence to generate and output the target speech.

The embodiment of the present specification provides a kind of decision graph atlas graph data generation method and device.The method is executed by computer, comprising: reading first graph data from first storage area, wherein it includes the first node corresponding to the event attribute of first type business event.Several rules scripts are read from second storage area, rule script defines several decision rules, each decision rule includes identification field and expression field;Any variable included in the expression field of any decision rule is one of several event attributes or the identification value of the identification field of other decision rules.Through parsing rule script, determine second graph data, wherein it includes: parsing expression field obtains the dependent relationship between decision rule and event attribute and between each decision rule;Second graph data includes the second node corresponding to several rules, and the connection edge generated according to dependent relationship.Further, based on first graph data and second graph data, generate the graph data of decision graph atlas.

The utility model belongs to the technical field of ship protection equipment, and particularly relates to a corrosion-resistant rubber fender which comprises a mounting plate, and a mounting assembly is arranged on the periphery of the mounting plate. The installation assembly comprises installation grooves, the multiple installation grooves are evenly formed in the periphery of the installation plate, bolts are connected to the periphery of the installation plate in a threaded mode, the installation plate is sleeved with gaskets in an abutting mode, the bolts are sleeved with the gaskets, and the installation grooves are in a U shape. Due to the fact that the peripheries of the protective pad, the fender body and the bolts are all coated with the anti-corrosion coatings, corrosion of external corrosive media can be effectively resisted. And the corrosion-resistant buffer layer in the protection plate is made of a special corrosion-resistant material, so that the overall corrosion resistance of the fender is further enhanced, the service life of the rubber fender in a severe corrosion environment is greatly prolonged, the high cost caused by frequent replacement of the fender due to corrosion is reduced, and long-term safety protection of ship berthing is ensured.

The application discloses a kind of fine rod type valve opening recognition method, the method includes: making identification band, and it is pasted on multiple valve components, the structural dimension of each the valve component is measured;Valve number represented by identification band, size information, initial switch state of valve component represented by identification band of identification band, color feature of identification band;Collect the image I of target valve;Image processing is carried out to the image I of target valve and is recognized, and the image area where identification band is located is fitted, and the actual opening of target valve is calculated;The valve opening recognition method of the application can mark and data entry to the valve component to be identified in advance, compare by image acquisition processing of inspection robot and input data, to calculate the actual opening size of valve, reduce human intervention, improve calculation accuracy;Enhance the applicability of inspection robot in actual inspection process.

This invention relates to the field of vehicle OTA upgrade testing technology, and discloses a network simulation system and testing method for vehicle OTA upgrade testing. The system includes a cellular communication simulation network, a programmable attenuation and channel simulation device, and a user terminal. The cellular communication simulation network includes an access network, a bearer network, and a core network. The access network deploys a base station controller and multi-band base stations. The radio frequency signal output from the access network is connected to the test environment via the programmable attenuation and channel simulation device. The user terminal sends commands to the base station controller for base station activation and deactivation configuration, and base station transmit power adjustment. The user terminal also sends commands to the programmable attenuation and channel simulation device for channel simulation mode control and signal strength adjustment control. Finally, the user terminal sends commands to the bearer network device for network transmission link parameter configuration. This invention can simulate various typical network scenarios through parameter configuration, enabling non-road testing of OTA upgrades.

The application provides a large model multi-agent A2A virtual power grid dispatching and power selling system and method, which comprises a vertical field energy large model module, an MCP tool layer module, an A2A multi-agent negotiation layer module and an application service layer module; the vertical field energy large model module serves as a system core reasoning engine; the large model multi-agent A2A virtual power grid dispatching and power selling system and method; the large model multi-agent A2A virtual power grid dispatching and power selling system provided by the application improves system operation safety and reliability, constructs a continuous learning closed loop of execution, evaluation, optimization and updating, iteratively optimizes dispatching and power selling strategies through review and optimization of agents, strengthens system adaptive and autonomous evolution capabilities, guarantees long-term stable and efficient operation, and provides a new technical path for intelligent operation of a virtual power grid.

The invention discloses a multi-source heterogeneous data collecting and processing method and system and a storage medium. The method comprises the following steps that a configuration management layer loads and analyzes a configuration file; the data driving layer dynamically creates all enabled data source instances according to configuration; the data scheduling output layer starts an independent collection task for each instance, and periodically triggers collection of multi-source heterogeneous data; the data driving layer performs preliminary analysis on the collected original data; the data processing engine layer processes the data obtained through preliminary analysis according to a configured pipeline rule; and the data scheduling output layer reports the processed data to a target system. Unified access, configuration, acquisition and data processing of various heterogeneous data sources are realized, so that the system complexity is reduced, the development, operation and maintenance efficiency is improved, and the reliability and real-time performance of data acquisition are guaranteed.

This invention proposes a bidirectional synchronous high-speed imaging and diagnostic system for electric thruster plasma oscillation. It comprises a vacuum chamber, an electric thruster, a vacuum chamber observation window, electric thruster plasma, a reflector assembly, a side-viewing optical path, a high-speed camera, a data processing system, and a high-speed camera's imaging range and a front-viewing optical path. The electric thruster, vacuum chamber observation window, and reflector assembly are bolted together inside the vacuum chamber. The electric thruster plasma is generated by the electric thruster and adheres to its outer surface. The side-viewing and front-viewing optical paths are formed by the reflector assembly. The high-speed camera is placed outside the vacuum chamber, with its imaging range directly facing the vacuum chamber observation window to capture the side-viewing and front-viewing optical paths. The electric thruster and its plasma are the subjects of the imaging. Compared to traditional bidirectional synchronous high-speed imaging, this invention establishes a bidirectional optical path to simultaneously introduce images of the side and front of the electric thruster into the lens of a single high-speed camera, requiring only a single high-speed camera. This reduces the financial and time costs of experiments and improves research efficiency.

This invention discloses a modular VR / AR surgical teaching system and method based on a local large-scale model. The local large-scale model training module is deployed in a local computing environment, storing structured medical knowledge and dynamically generating teaching scripts and assessment question banks. A lightweight VR / AR hardware group collects multi-dimensional user operation data and provides multi-sensory feedback. A full-process training control module, with a built-in physical simulation engine and evaluation logic, constructs a virtual surgical scene, simulates tissue mechanical changes in real time, compares user operations with standard surgical procedure models and generates error correction guidance, and generates a quantitative ability assessment report based on the entire process data. A multi-scene data mapping interface adopts a modular architecture based on a standardized intermediate communication protocol, achieving low-coupling logical isolation and dynamic integration between functional modules through a unified data exchange format and abstract interface specifications. This application adopts an edge-side deployment strategy to protect patient privacy and ensure real-time force feedback and tissue deformation rendering in the VR / AR scene.

The application discloses an on-line monitoring device and method for dissolved gas in transformer oil based on infrared spectrum, which comprises a transformer oil gas separation part, an infrared spectrum gas detection part and a data processing part. The method comprises the following steps: firstly, carrying out oil gas separation; then, sending the obtained dissolved gas in oil into a gas chamber and scanning the gas chamber by using an infrared spectrometer to obtain a spectrum diagram; and finally, obtaining the content of various target gases in oil by using a gas quantitative analysis model. The application uses infrared spectrum method as the gas detection part, has the advantages of fast spectrum data acquisition, no consumption of gas samples, wide detection range, high stability, maintenance-free, long-term operation monitoring cost, etc. compared with the most widely used gas chromatography method for on-line monitoring of gases in transformer oil. The device adopts a structure based on single-chamber vacuum degassing, solves the problem of spectrum baseline offset distortion, and does not need to consume nitrogen as background, thereby improving the automation degree of the device.

The application relates to a satellite navigation narrow-band interference real-time suppression method based on power inversion. The method comprises the following steps: segmenting and windowing a continuous satellite navigation digital intermediate frequency signal, converting the signal to a frequency domain through FFT, realizing power inversion through unit circle projection to attenuate narrow-band interference, determining an effective frequency point index range according to a signal center frequency and a nominal bandwidth, zeroing out out-of-band spectral lines, recovering a time domain signal through IFFT, and finally recombining and outputting through overlapping post-processing. The method does not require prior information and threshold judgment throughout the process, can blindly and adaptively suppress interference, solves the problem of signal-to-noise ratio deterioration under a wide sampling bandwidth, has strong real-time performance, small signal distortion, and is suitable for multiple navigation systems.

The application discloses a label management method and device based on a configuration template driver. The method acquires a predefined configuration file to define field types, operators and SQL templates in a declarative manner; defines labels based on the configuration and user input, and stores and organizes logical relationships between fields in the labels in a first tree structure; combines business rules based on the defined labels, and stores and organizes set operation relationships between the labels in a second tree structure; and in response to a query request, dynamically generates and executes a structured query statement according to the configuration file and the two-level tree structure. The application can achieve the technical effects of greatly improving configuration efficiency, qualitatively changing support capability of logical complexity, sharply reducing system expansion and maintenance cost, and significantly enhancing overall stability and security.

This application discloses a cross-database dialect intelligent DDL conversion method based on G4 formal grammar comparison. This method uses ANTLR4 G4 formal grammar files as the semantic benchmark and establishes a precise grammatical mapping relationship between the source and target databases through four steps: grammar rule alignment, grammar element mapping, grammar difference marking, and conversion rule generation. Based on a unified AST model, the parsing and generation processes are decoupled. A cross-dialect standardized segment model is constructed through the parsing visitor, and then the target SQL is generated node by node by the translation visitor. It supports three configurable conversion strategies: strict, lenient, and warning, as well as real-time incremental DDL synchronization. This invention has advantages such as high conversion accuracy, comprehensive dialect coverage, low expansion cost, flexible and configurable strategies, and zero intrusion into business systems. It is adaptable to bidirectional DDL conversion of various mainstream databases such as Oracle, MySQL, and OpenGauss.

The application is an open fan full-range working characteristic expansion method based on skeleton characteristics, belonging to the field of open rotor engine; the method introduces a skeleton characteristic representation system with loss coefficient as the core to solve the problem that the efficiency in the representation of traditional propeller characteristics has singular points in the autorotation state, leading to the collapse of cross-condition numerical simulation calculation. By defining the work coefficient, pressure coefficient and loss coefficient, and based on the blade element theory to determine the ridgeback point and its existence condition, the ridgeback characteristics describing the characteristic change law of ridgeback point and the non-ridgeback characteristics describing the deviation law of non-ridgeback point are constructed. Finally, by combining the ridgeback characteristics and the non-ridgeback characteristics, the known local working condition characteristic data is continuously and smoothly expanded to the full working range including braking, autorotation and windmill state. The application fundamentally eliminates the numerical singular points in the characteristic curve, realizes the stable simulation and high-performance expansion of the open fan in the full working condition.

The invention discloses a deflection-angle-free (111) single crystal diamond substrate and a preparation method thereof. The preparation method comprises the following steps: selecting a (111) single crystal diamond substrate with a deflection angle less than 5 degrees; (111) crystal face atomic steps are formed on the surface of the (111) single crystal diamond substrate with the deflection angle smaller than 5 degrees; growing (111) crystal face diamond on the surface of the substrate with the (111) crystal face atomic step in a self-selection manner to obtain a step-free (111) single crystal diamond substrate; two-dimensional growth is carried out on the step-free single crystal diamond substrate to obtain the deflection-angle-free single crystal diamond substrate, so that the (111) single crystal diamond epitaxial layer with high quality and low surface roughness can be obtained in the subsequent two-dimensional epitaxial growth stage. The preparation method provided by the invention can solve the problems that the current (111) diamond substrate generally has a substrate deflection angle, the substrate cannot realize expanded growth due to (111) homoepitaxy caused by the substrate deflection angle, and the epitaxial layer has fault and cannot realize high-quality growth.

The application relates to a mobile approval system and method supporting multiple languages, which comprises a multilingual resource management center and a language environment interceptor arranged at a rear end, an internationalization processing module, a minority language font library and a font embedding and rendering module arranged at a front end; the internationalization processing module requests a language resource file from the multilingual resource management center according to a language environment of a front-end operating system or user preference setting, then receives the language resource file, intercepts and replaces interface texts; the font embedding and rendering module renders language font files to applications; after the front end initiates an approval request, the language environment interceptor analyzes a language identifier and informs a server; the server requests a language resource file from the multilingual resource management center, then carries out language conversion on business data, so that the font of the business data is consistent with the font of the current server system. The application realizes seamless and efficient support for minority languages in a decoupled and configured manner.

The utility model discloses a modularized gas deodorization device which is characterized by comprising a mounting frame, the mounting frame is in a hollow ring shape, at least one deodorization medium box is arranged in the mounting frame in a sliding mode, the top side and the tail side of the deodorization medium box are open, a deodorization medium is arranged in the deodorization medium box, and the deodorization medium box is provided with a gas inlet and a gas outlet. The deodorization medium is used for filtering peculiar smell molecules in gas, an air inlet channel is formed between the deodorization medium box and the inner wall of the top of the mounting frame, and the gas enters the deodorization medium box through the air inlet channel and flows out from the tail side of the deodorization medium box. Compared with the prior art, the gas deodorization device has the advantages that the expansion cost is lower, the splicing is simple, the use flexibility of the gas deodorization device can be effectively improved, and the use cost is reduced.

The application discloses a function calling method and device, a vehicle and a storage medium, and relates to the technical field of vehicle-mounted system software. The scheme is applied to a target system, the target system comprises a master process and at least one local function process, and the master process is the only process in the system that interacts with a large language model through a model context protocol server. The method comprises the following steps: receiving a target calling instruction based on a model context protocol sent by the large language model based on the master process; analyzing the instruction to obtain a target function identifier and a calling parameter; forwarding the parsed instruction to the corresponding local function process according to the target function identifier, so that the process executes a function according to the calling parameter. Through centralized and standardized calling management, the scheme improves the response efficiency and scalability of function calling, and reduces the function extension and maintenance cost.

This utility model discloses a material transport system with integrated status monitoring and energy-saving functions, belonging to the technical field of material transport equipment. It includes an upper-level terminal, a lower-level controller, and at least three belt conveyors connected in sequence. All belt conveyors are connected to the lower-level controller, which in turn is connected to the upper-level terminal. Each belt conveyor's drive motor is equipped with an integrated temperature and vibration sensor to detect abnormal vibration and temperature. High-definition cameras are installed under the belts to detect belt tears. A material flow detection device is installed above the belts to detect material passing through a designated location and control the start and stop of the next-level belt conveyor accordingly. This utility model effectively solves the problems of high energy consumption during belt conveyor startup, low transport efficiency, and insufficient equipment health status monitoring in existing material transport systems.

The application discloses a kind of text-to-image auxiliary machine translation methods based on large model, comprising: obtaining source text, generating initial image using diffusion model;Text semantic graph of source text and visual semantic graph of initial image are constructed, and the diffusion model is optimized based on consistency reward function using reinforcement learning, to obtain the optimized image;Degenerate text is obtained by degeneration processing to source text, and the text features of source text and degenerate text and the image features of optimized image are extracted, and after fusion, input large model to obtain multi-modal semantic representation;Diffusion model and large model are sequentially trained, and translation result is generated according to the large model after training.The application does not need artificial marking of graph-text data, improves the semantic consistency of generated image and source text through reinforcement learning feedback, realizes the deep cooperation of text and image using multi-modal fusion and sequential training, and improves the performance of multi-modal machine translation.