22results about How to "Improve deployment flexibility" patented technology

The application discloses a raw material storage device for nitrate production in the chemical industry, and belongs to the technical field of raw material storage devices, which comprises a storage tank body, relief ports are connected to the two sides of the top of the storage tank body, a breather valve is installed on the relief port, a main pipeline is connected to the air inlet and outlet of the breather valve, and an air inlet pipe is connected to the main pipeline; tail gas recovery mechanisms are symmetrically arranged on the storage tank body, the tail gas recovery mechanisms comprise recovery boxes which are fixed to the storage tank body through supporting rods, the two recovery boxes are connected to the main pipelines on the two sides respectively, a solution tank is fixed to the top wall of each recovery box, and two liquid outlet hoses are connected to the bottom of the solution tank; the raw material storage device can realize automatic spraying after waste gas accumulation reaches a standard and automatic shutdown after a reaction is completed, does not need uninterrupted spraying operation, greatly saves electric energy consumption and alkali consumption, and reduces invalid volatilization and loss of alkali.

The invention discloses an unmanned aerial vehicle assisted holographic digital-energy integrated transmission performance evaluation method. The method comprises the following steps: constructing an unmanned aerial vehicle assisted holographic IDET system model; determining a data transmission and energy transmission model under the limited code length mechanism; a holographic IDET beamforming design scheme assisted by the unmanned aerial vehicle is provided; analyzing a closed expression of an end point of the Pareto optimal curve; and designing an optimization algorithm to obtain a Pareto optimal boundary, and giving rate-energy domains of the DU and the EU. Aiming at the problem that a traditional fixed base station is difficult to meet the requirements of wide coverage, high energy efficiency and reliable transmission under the limited code length at the same time, the performance is remarkably improved by introducing unmanned aerial vehicle flexible deployment, holographic surface near-field focusing capability and a reliability analysis mechanism under the limited code length; and a rate-energy domain is obtained according to the Pareto optimal boundary curve.

The application belongs to the technical field of wireless communication, and relates to a hardware fast reconfiguration terminal and method. A VPX architecture is used to configure a plurality of slot positions of the terminal to support flexible plugging of a signal processing module and a radio frequency front end. Standardized interaction of data communication between modules, system control monitoring and health management is realized based on PCIe / SIRO, a local area network and an IIC bus. An embedded operating system is deployed in a main control module. A core framework of a domain manager and a device manager and basic equipment and service components corresponding to hardware capabilities / resources are included. The hardware platform is logically abstracted to release deep binding of software and hardware. In addition, terminal functions are disassembled into independent waveform components with standard interfaces. Standardized communication between the components is realized through a soft bus and a hardware abstraction layer interface. Waveform application is dynamically deployed by relying on external waveform deployment management software to complete function switching. In this way, the architecture limitation of strong coupling of existing communication equipment software and hardware is broken, and the technical problem that hardware resources cannot be dynamically reconfigured is solved.

The invention belongs to the technical field of deep sea navigation positioning, and provides an inertial navigation tight coupling correction method and system based on a time-sharing collaborative underwater acoustic beacon, and the technical scheme is as follows: in the navigation process of an AUV (Autonomous Underwater Vehicle), before acoustic auxiliary information arrives, a pure inertial navigation mode is adopted, and based on acquired IMU data and depth sensor data, inertial navigation calculation is carried out; obtaining an AUV navigation state containing an accumulative error; after the acoustic auxiliary information arrives, switching to a combined navigation mode for resolving, utilizing beacons at different positions to cooperatively broadcast acoustic information in a time-sharing manner, resolving the horizontal distance between the AUV and each beacon, and applying a vertical coordinate constraint; based on the measurement distance and the AUV navigation state containing the accumulative error, fusion processing is carried out through a tight coupling filtering algorithm, the accumulative error in the AUV navigation state is corrected, and a corrected AUV positioning result is obtained. And long-endurance and high-precision underwater positioning of the AUV is realized.

The invention discloses an event identification method and system for pig house intelligent monitoring, a camera and a medium. The method comprises the following steps: collecting and sampling pig house video streams; performing target identification on the sampled image to obtain an identification result including categories and position coordinates of pigs, personnel, vehicles and the like; performing label matching on the identification result based on a pre-configured label rule, wherein the matching rule at least comprises a target appearance rule, a region intersection rule, a target continuous appearance rule and a target disappearance rule; matching a pre-configured event rule according to a label matching result, and generating hog house monitoring events such as entry of pigs into a corridor and entry of personnel into an effective area; and intercepting an event associated video clip, monitoring a 4G network signal state, and adaptively selecting to locally store or report the event data and the video clip to a cloud platform. According to the invention, accurate and flexible event identification aiming at a pig house scene is realized, and the reliability of monitoring data in an environment without a wired network or a weak network is guaranteed.

The application belongs to the technical field of deep sea navigation and positioning, and provides an inertial navigation close-coupled correction method and system based on time-sharing cooperative underwater acoustic beacons, which is characterized in that, during the navigation process of an AUV, a pure inertial navigation mode is adopted before the arrival of acoustic auxiliary information, inertial navigation is solved based on collected IMU data and depth sensor data, and the navigation state of the AUV containing accumulated errors is obtained; after the arrival of the acoustic auxiliary information, a combined navigation mode is switched to for solving, acoustic information is broadcasted by time-sharing cooperation of beacons at different positions, the horizontal distance between the AUV and each beacon is solved, and a vertical coordinate constraint is applied; the measured distance and the navigation state of the AUV containing accumulated errors are fused by a close-coupled filtering algorithm, the accumulated errors in the navigation state of the AUV are corrected, and the corrected positioning result of the AUV is obtained. Long-time navigation and high-precision underwater positioning of the AUV are realized.

The invention discloses an iron tower health monitoring method and system based on multi-source data fusion, and belongs to the technical field of iron tower health monitoring, and the method comprises the steps: collecting the multi-source heterogeneous data, including deformation monitoring data and meteorological monitoring data, of an iron tower in real time; performing preprocessing and feature extraction on the multi-source heterogeneous data to obtain multi-source feature data; and performing multi-source data fusion based on the multi-source feature data to obtain feature fusion data. And constructing an iron tower health assessment model, training the model, and outputting an iron tower health index. And carrying out evaluation and early warning on the iron tower structure health condition based on the iron tower health index. According to the invention, space-time alignment and collaborative analysis are carried out on three kinds of heterogeneous data of Beidou positioning, inclination angle vibration and meteorological data, data preprocessing is carried out through an equipment end, and remote data transmission is realized by adopting an MQTT protocol. A multi-source data fusion model is constructed at the cloud, and real-time evaluation and early warning of the health condition of the iron tower structure are realized by combining Kalman filtering and BP neural network algorithms.

The invention provides a vehicle body pose measurement method, an error measurement method, equipment, a medium and a vehicle, and belongs to the technical field of vehicle body positioning. The measurement method comprises the following steps: establishing a regional coordinate system, and arranging at least two reference points on each of two sides of a preset positioning region; after the target vehicle enters the preset positioning area, collecting linear distances from a projection point of each wheel on the ground to two reference points on the same side through a sensor; obtaining projection point coordinates of each wheel in the regional coordinate system based on the linear distance; and based on the projection point coordinates, obtaining a measurement result of the vehicle body pose of the target vehicle in the regional coordinate system. According to the method, the reference point is set, the distance between the wheel projection point and the reference point is collected through the sensor, the vehicle body pose parameters are obtained in combination with geometric calculation, high-cost RTK equipment or laser radar equipment is not needed, and the method has the advantages that the hardware cost is reduced, the implementation process is simplified, and the deployment flexibility is improved.

The invention discloses a breath sound classification detection method based on an improved convolutional neural network, and the method comprises the steps: firstly carrying out the preprocessing and feature extraction of a breath sound signal collected in real time, and constructing a breath sound cepstrum feature matrix; and then the breath sound features are input into an improved convolutional neural network comprising a convolution module, a channel attention mechanism module and a lightweight feature extraction module for training and classification, and accurate recognition of breath sound categories is realized. The method realizes real-time, efficient and high-precision classification detection of breath sound, has the advantages of low calculation complexity, strong real-time performance and high identification accuracy, and is suitable for the field of biomedical signal processing.

The invention provides a mobile oxyhydrogen electricity, heat and cold combined supply system and a capacity configuration method, and relates to the technical field of energy. A photovoltaic module is arranged at the top of the integrated device, and hydrogen storage equipment, oxygen storage equipment, heat storage equipment and cold storage equipment are arranged outside the integrated device; the integrated device comprises an electric energy conversion module, a liquid purification module, an electrolytic cell module, a hydrogen energy fuel cell module, a heat management module, a refrigeration module and a control module. According to the movable hydrogen-oxygen-electricity cold-heat co-production system with multi-energy co-supply, high integration, intensive resources, convenient deployment and intelligent cooperation and the capacity configuration method thereof, hydrogen-oxygen-electricity-heat-cold multi-energy co-supply can be achieved, and the adaptability, flexibility and economical efficiency of the system are remarkably improved. The important defects that an existing mobile energy supply system is single in energy supply form, low in module integration level, low in comprehensive energy utilization rate, short in continuous operation time and the like are well overcome, and the requirements for high efficiency, flexibility and economical efficiency of the mobile energy supply system under various complex scene dynamic working conditions are met.

The present application relates to the cross field of artificial intelligence and computer architecture, in particular to an adaptive KV cache compression method, system and device for an audio-text multi-modal large model. The method comprises opening a full-precision hot table and a low-bit cold table in the video memory; writing the newly generated audio or text Token in the FP16 format into the hot table, and when the number of Tokens in the hot table reaches a preset threshold, gradually migrating the Tokens in the hot table to the cold table; identifying the modal type of the Token during the migration process, selecting a corresponding quantization grouping strategy according to the identified modal type, and performing INT2 compression storage to the cold table after online calculation of the quantization parameters of the Token to be migrated. Through the cooperation of the above steps, the present application realizes the improvement of the model inference throughput and the natural full-duplex human-computer interaction under the premise of ensuring the audio synthesis quality.

The application belongs to the technical field of image compression and transmission, and particularly relates to an image compression and reconstruction method and system based on hidden space discretization and parameter decoupling, which comprises a preprocessing module, an encoder, a light-weight code stream serialization storage module and a decoder; through construction of an asymmetric coding and decoding architecture of hidden vector structured discretization on the encoder side, physical decoupling of coding and decoding parameters, serialization of light-weight code streams, independent deployment of a pre-trained decoder and industrial feature adaptive entropy coding, high compression ratio, low transmission delay and edge lightweight deployment of industrial image compression and reconstruction are achieved.

This invention discloses a miniaturized log-periodic antenna and its assembly and design methods, belonging to the field of communication antenna technology. The antenna includes a dielectric substrate, a semi-steel cable, an adapter plate, and a copper layer; the copper layer is disposed on the front and back sides of the dielectric substrate; an adapter plate is disposed at one end of the dielectric substrate; the outer shell of the semi-steel cable is welded to the copper layer at the center line of the back side of the dielectric substrate, and the inner core of the semi-steel cable is bent upwards through the side of the dielectric substrate without the adapter plate and then welded to the copper layer on the front side of the dielectric substrate; the antenna operates in the frequency band of 150MHz–400MHz, with a VSWR ≤2, and uses linear polarization. This invention addresses the performance shortcomings of traditional antennas in this frequency band by optimizing the structure of the vibrator and improving the feeding method, enabling it to cover the target frequency band while meeting the lightweight and miniaturized usage requirements of different scenarios.

This application discloses an adaptive multi-granularity text embedding representation learning method based on gradient optimization. The method includes: text encoding, multi-granularity embedding generation, difficulty-aware dynamic sampling, sliding window negative sample management, adaptive loss optimization, and model training and validation. It aims to address the performance degradation caused by existing embedding vector dimensionality compression techniques. Through a difficulty-aware dynamic sampling mechanism, a Beta distribution is introduced to achieve progressive learning, significantly improving low-dimensional performance; sliding window negative sample management improves sample quality and training efficiency; and a gradient-oriented update strategy avoids gradient conflicts and optimizes training stability. Furthermore, this invention supports incremental dimensionality expansion, reducing training costs while maintaining consistency across multi-granularity embeddings. This method significantly improves performance in low-dimensional environments, maintains lossless performance in high-dimensional environments, significantly improves training efficiency, greatly reduces storage costs, and exhibits strong generalization ability, making it suitable for data from various domains.

This invention relates to the field of autonomous driving technology and discloses an autonomous driving simulation testing system and equipment. The system allows end-users to configure an interface that sends simulation tasks with target scenarios to a central scheduler and visualizes the process. The central scheduler receives the simulation tasks and, via a message middleware, sends them to distributed simulation computing nodes, building a global task scheduling engine to automate the entire testing process. The message middleware uses an asynchronous communication mechanism to provide message transmission channels. The terminal scene database writes the entire simulation task process data and encrypts scene resources from a cloud data management platform. After confirming receipt of the simulation task via a vehicle-to-everything (V2X) interaction module, the distributed simulation computing nodes retrieve the corresponding target scenario data from the terminal scene database, call the simulation engine to load the target scenario, and execute it. This invention solves the technical problems of insufficient flexibility and scalability, chaotic scene management, low automation, and weak security in existing systems.

The invention discloses a low-delay streaming voice wake-up method and device for an embedded terminal, is applied to the technical field of data processing, and adopts a GRU as a backbone architecture for the requirements of low power consumption and high real-time performance of voice wake-up of the embedded terminal. The original audio is segmented and reconstructed through time domain lightweight preprocessing, and the FFT overhead is eliminated through linear layer mapping; building a multi-layer GRU full-causal streaming coding network, wherein no sliding window cache exists, and the memory bandwidth occupies O (1); outputting Token probability distribution containing Blank through a CTCLoss training module, and directionally extracting a target Token probability; constructing a target index sequence based on the target wake-up word and a CTC protocol, and setting three types of state transition rules to generate anti-interference decoding information; and a non-Top-K pruning decoding mechanism is adopted, an attenuation factor and threshold judgment is combined, a judgment result with constant calculation complexity and low false and missing wakeup is generated, and subsequent interaction is triggered if the judgment result succeeds.

The application discloses a low-resource environment large model deployment reasoning method based on deep learning, comprising the following steps: S1, acquiring hardware physical characteristic parameters and constructing a hardware perception cost model; S2, generating an initial lightweight model by using a neural architecture search algorithm; S3, monitoring the device environment state in real time, calculating the maximum theoretical calculation amount that can be borne, and generating a resource availability vector; S4, calculating semantic entropy values, judging sample complexity and allocating initial calculation resource quotas; S5, introducing an improved CMA-ES algorithm of semantic entropy guided population center initialization, adaptively searching and outputting an execution strategy; S6, performing high or low precision fixed-point operation on key layers and non-key layers respectively; S7, performing decoupling recovery and inverse quantization processing on feature maps, and outputting final results. The application realizes efficient deployment of a large model in a low-resource environment, significantly reduces delay and energy consumption while ensuring reasoning accuracy.

This invention discloses a bearing fault diagnosis method and system based on cross-modal feature alignment, belonging to the field of rotating machinery condition monitoring and fault diagnosis technology. The method includes: constructing sound feature vectors and vibration feature vectors based on the envelope spectrum of sound signals and the envelope spectrum of vibration signals to form a sound-vibration synchronization dataset; training the constructed cross-modal feature alignment diagnostic model using the training and validation sets in the sound-vibration synchronization dataset to obtain a trained cross-modal feature alignment diagnostic model, thereby obtaining a solidified cross-modal feature alignment diagnostic model; using the sound feature vectors from the test set in the sound-vibration synchronization dataset or the sound feature vectors to be diagnosed as input to the solidified cross-modal feature alignment diagnostic model to obtain the bearing diagnosis result. This invention can realize fault diagnosis under the same operating conditions based on sound vibration components, and further realize fault diagnosis across operating conditions.

The utility model relates to the technical field of stackers, and discloses a novel stacker. The novel stacker crane comprises a supporting plate, a supporting beam, a first guide rail, a lifting frame, a supporting arm, a first chain wheel, a first driving motor, a second chain wheel and a chain, according to the novel stacker crane, the longitudinal supporting beam is adopted for replacing a huge gantry frame, the size of the whole stacker crane is greatly reduced, the requirement for the installation space is remarkably reduced, and the novel stacker crane is suitable for the narrow plant environment. The frame structure is simplified, steel consumption is reduced, machining difficulty and transportation cost are reduced, and meanwhile equipment deployment flexibility is improved.

The invention discloses a real-time temperature control type immersion heat dissipation system and method based on the calorific value of computing power equipment, and relates to the field of immersion type liquid cooling heat dissipation. The system comprises a temperature monitoring unit integrated in a machine case unit, the machine case unit with a thermal load partition and an outer loop heat dissipation unit, the interior of the machine case unit is divided into three independent partitions, namely a high-heat partition, a medium-heat partition and a low-heat partition, the three independent partitions are matched with IT equipment with different heating intensities respectively, and each partition is provided with an exclusive temperature sensor, a flow field control device and a drainage pump; the flow field control device comprises an axial vortex generation pump, a Bernoulli principle flow deflector and a wave-shaped flow guide block, the flow field is optimized, and the use amount of cooling liquid is reduced; the temperature monitoring unit synchronously acquires the partition temperature and the chip core temperature, and intelligently switches between a natural circulation mode and a forced external circulation mode; accurate temperature control and quick response are achieved, the using amount of cooling liquid is reduced, and the heat dissipation efficiency and the energy-saving requirement are both considered.

This utility model discloses a material changing mechanism for a central feeding system, belonging to the field of material conveying technology. It aims to solve the problems of difficult movement and positioning, poor height adaptability, and low pipeline connection efficiency of traditional equipment. The mechanism achieves flexible movement and stable positioning through universal wheels with locking function; a servo motor drives a synchronous belt and threaded shaft to precisely raise and lower a sliding block along a sliding groove with an inverted trapezoidal cross-section. Combined with scale indicators and guide pillars, it accurately matches feeding interfaces of different heights. The core connecting unit drives the opening and closing of symmetrical sliding plates through a rotating handle, controlling the hinged clamps to quickly clamp the input / output pipes. The guide protrusions on the pipe connecting plate are used for insertion and positioning with the slots. Finally, the threaded ring on the outer wall of the pipe is locked by a rotating connector, forming a double seal. This invention significantly improves material changing efficiency and connection reliability, and is suitable for high-frequency material switching scenarios in automated production lines.

This utility model relates to an integrated IoT fan coil unit, including a housing and a surface cooler, a surface cooler bracket, a motor, a fan, a condensate drain pan, and a condensate lift pump located within the housing. The surface cooler is mounted on the surface cooler bracket, and the condensate drain pan is located below the surface cooler and fixed to the surface cooler bracket. The fan is connected and fixed to the central shaft of the motor. It also includes a proportional-integral (PI) control valve, a temperature sensor, an edge calculator, and a thermostat. The PI control valve is connected to the outlet of the surface cooler. The temperature sensor is located at the inlet and outlet of the surface cooler to detect the inlet and outlet water temperatures. The edge calculator is located inside the housing and is electrically connected to the PI control valve, the motor, the temperature sensor, and the condensate lift pump. The thermostat and the edge calculator are connected via Bluetooth. This utility model enables intelligent sensing and precise control of the fan coil unit's operating status, while simplifying the installation and commissioning process of the terminal equipment.