3635 results about "Multiplexing" patented technology

The invention discloses a large language model semantic query acceleration method based on a sparse KV Cache index. The method comprises a KV Cache semantic pruning strategy based on an attention mechanism and a set of asynchronous pipeline reasoning architecture with overlapped calculation and I / O. According to the method, the attention sparsity characteristic of the large language model in the reasoning stage and the asynchronous transmission capacity between the Host memory and the GPU video memory are fully utilized, calculation redundancy and video memory occupation in repetitive semantic query are greatly reduced, and high-throughput, low-delay and high-performance batch semantic data processing service is provided. Through a mechanism for mapping a static text into a compressed semantic index and combining a prefix cache technology in a reasoning process to realize state multiplexing, high-performance reasoning acceleration and high-efficiency storage compression are provided for a data-intensive semantic analysis task in a resource-constrained environment.

The invention provides an optical system, a control method, a vehicle optical signal device, a vehicle lamp and a vehicle, the optical system comprises at least one first light source assembly and a second light source assembly, and light emitted by the first light source assembly passes through an optical assembly and is emitted from a first light emitting area of a light emitting surface of the optical assembly; light emitted by the second light source assembly passes through the optical assembly and is emitted from a second light-emitting area of the light-emitting surface of the optical assembly, and at least partial area overlapping exists between the first light-emitting area and the second light-emitting area; wherein light emitted by the first light emitting area serves as a first light function, light emitted by the second light emitting area serves as a second light function, the first light function serves as a marker lamp of the driving automation system, and the second light function serves as other light functions with colors different from those of the first light function. Therefore, the structure and function multiplexing of the automatic system marker lamp and other signal lamps is realized, the cost of independently designing and manufacturing the automatic system marker lamp is effectively reduced, and the occupied space of the whole vehicle lamp is effectively reduced.

The present application belongs to the technical field of wireless communications and radar systems. Disclosed are a dual-band fusion and multiplexing frequency source, chips, a radio-frequency transceiver, a system, and a device. The system comprises: a full-function millimeter-wave CMOS 4+1 phased array chip, which is connected to an antenna end, and is used for realizing signal processing and frequency conversion between a millimeter-wave signal and a C-band intermediate-frequency signal, performing phase control and gain adjustment on the millimeter-wave signal, and providing an internal frequency source and a calibration function; a C-band front end, which is connected to the antenna end, and is used for receiving or transmitting a C-band signal; a C-band / intermediate-frequency transceiver multiplexing chip, which is used for performing frequency conversion between a C-band or intermediate-frequency signal and a baseband signal in a multiplexing manner, and performing signal processing and gain control on the C-band intermediate-frequency signal; a dual-band fusion and multiplexing frequency source, which is used for providing a local oscillator signal; and an all-digital control logic circuit, which is used for providing a control instruction. The present application improves the efficiency and performance of dual-band joint communications.

The invention relates to the technical field of power distribution automation auxiliary power supply control, and discloses an intelligent power distribution circuit system for a digital ring main unit, which comprises a transient voltage gradient induction bridge circuit, a hardware signal processing circuit, a power switch unit and a controlled energy release unit, according to the system, a transient voltage gradient induction bridge circuit is used for collecting a primary side bus voltage change rate signal, a hardware signal processing circuit carries out derivation operation on an induction signal and compares the induction signal with a voltage gradient threshold value, and a power switch unit is blocked at the moment that the voltage abnormally drops so as to cut off a load current path. According to the invention, a hardware parameter direct mapping mechanism is utilized to realize microsecond transient protection response, and the physical decoupling of protection logic and digital software is realized by balancing bus potential pulse caused by sudden unloading of a load through the inherent dehumidification resistor in the multiplexing cabinet.

Aspects of this disclosure relate to a bulk acoustic wave resonator and an integrated capacitor. In certain embodiments, the integrated capacitor can be a metal-insulator-metal capacitor in parallel with the bulk acoustic wave resonator. The metal-insulator-metal capacitor can include a portion of a first electrode and a portion of a second electrode of the bulk acoustic wave resonator. At least part of the metal-insulator-metal capacitor is positioned laterally relative to an acoustic reflector of the bulk acoustic wave resonator. Other embodiments of capacitors integrated with a bulk acoustic wave resonator are disclosed. Related filters, multiplexers, radio frequency modules, radio frequency systems, wireless communication devices, and methods are disclosed.

The invention belongs to the field of PCIe (Peripheral Component Interconnect Express) chips, and relates to a CXL switching chip of a multi-core-grain architecture and a cross-core-grain routing switching method thereof, and a plurality of CXL switching core grains with uniform specifications are packaged and interconnected to form the CXL switching chip; each CXL switching core particle comprises a plurality of ports, each port can be configured to be in a CXL mode and a D2D mode, and the port of each CXL switching core particle at least comprises a physical layer, a CXL link layer, a CXL transaction layer, a D2D routing layer, a bypass and a multipath selection module; the CXL switching chip further comprises a multi-VC on-chip CXL switching network, and the multi-VC on-chip CXL switching network comprises at least one order-preserving switching VC and one out-of-order switching VC. According to the method, CXL switching chips of various specifications can be flexibly constructed, the design complexity and verification difficulty of the large-scale or multi-port CXL switching chips are effectively reduced, and the production yield of the chips is improved.

The invention belongs to the technical field of industrial system integration, and provides a design method for rapid adaptive integration of a cross-domain system. Through the method, developers extract business logic rules of the industry, standardize characteristic mapping and select, develop and integrate functional modules, logic processes and scene components of the cross-domain system, so that the cross-domain system which conforms to industry characteristics and meets business circulation is formed, and meanwhile, the development efficiency of the cross-domain system is improved. The newly added industry logic component can be classified into each industry characteristic model based on the knowledge base, and a multi-dimensional characteristic vector is generated in the platform, so that cross-industry integrated component multiplexing is realized, and the cross-field applicability and flexibility of the platform are further improved. Modularized construction and configuration type development are supported, mapping and matching functions are carried out on various industrial manufacturing industry characteristics such as machining, electronic and electrical appliances and equipment sets, the requirements for subdivision production in the industrial field and rapid adaptation of business processes are met, the system integration cost is reduced, and the application development efficiency is greatly improved.

The invention discloses a construction method and a modulation method of a three-port multi-frequency MMC topology based on bridge arm multiplexing, and relates to the technical field of power electronics. The construction method comprises the steps that sub-modules of a traditional double-bridge-arm MMC are integrated and optimized into a three-bridge-arm structure through a current conversion branch, independent energy interaction of three ports is achieved through an auxiliary bridge arm, multi-frequency operation modes of power frequency, frequency division, direct current and the like are supported at the same time, and therefore the utilization rate of the bridge arm sub-modules is increased, and the energy utilization rate of the MMC is increased. And the overall size and the manufacturing cost of the converter are reduced. Based on three-port multi-frequency coordination control and power balance requirements, a modulation method of the topological structure is provided, a multi-frequency coordination modulation strategy is adopted, frequency offset is performed through an auxiliary bridge arm to reduce cross interference of port frequency, and output waveform quality is improved through multi-objective optimized capacitor voltage balance control. And meanwhile, direct-current side faults are effectively eliminated by means of the auxiliary bridge arm.

The invention relates to a customer service scene-oriented collaboration system and method, and belongs to the technical field of artificial intelligence. In the system, a shared memory storage center stores knowledge entries from a plurality of customer service agents in a predefined data structure; the shared memory storage center is provided with an access control mechanism associated with a visible domain identifier, and an access range is limited during retrieval through a logic partition; the memory management module receives candidate knowledge entries from each customer service agent based on a trigger event, determines visible domain identifiers for the candidate knowledge entries, associates the candidate knowledge entries with the visible domain identifiers, and stores the associated candidate knowledge entries and the visible domain identifiers in the shared memory storage center; the problem processing and decision-making module responds to an access request from any customer service agent and determines a corresponding visible domain identifier according to the identifier of the customer service agent; and executing access based on the visible domain identification and the input question, and retrieving knowledge entries with access from a shared memory storage center to realize cross-agent multiplexing.

The present invention provides assays and assay systems for use in spatially encoded biological assays. The invention provides an assay system comprising an assay capable of high levels of multiplexing where reagents are provided to a biological sample in defined spatial patterns; instrumentation capable of controlled delivery of reagents according to the spatial patterns; and a decoding scheme providing a readout that is digital in nature.

Systems and methods for performing hardware approximation of functions are provided. In one example, a system comprises a controller, a plurality of multiplexors, configurable arithmetic circuits, and a mapping table that stores a set of function parameters. According to a mode of operations, the controller may configure the plurality of multiplexors to forward the set of function parameters or a subset of the function parameters to the arithmetic circuits to compute an approximation result. In a case where the subset of the function parameters is forwarded to the arithmetic circuits, the controller may configure the arithmetic circuits to perform post-processing, such as quantization, of the approximation result.

The embodiment of the invention provides an application starting method and device, and relates to the technical field of terminals. The method comprises the following steps: when the electronic equipment runs a first task of a first application in the background, the electronic equipment receives a first operation of opening a first function by a user; in response to the first operation, the electronic equipment displays a first interface of the second application and runs a first task in a first view of the first interface, a first activity located on the stack top in the first task corresponds to the first function, and the first task is a child node of the first view; and when the electronic equipment displays the first interface, the electronic equipment stops running the first task in the background and runs the first task in the foreground. Therefore, the first application is started in a task multiplexing mode, and the task of the first application is operated in the view of the second application, so that display abnormities such as application flash quit and the like caused by starting of the first application can be reduced.

The invention provides a closed-loop nerve regulation and control system and method based on electroencephalogram detection. The defects that a traditional open-loop method lags in stimulation, equipment is heavy and the like are overcome. The system integrates an electroencephalogram acquisition module, an electroencephalogram analysis module, a remote program control module and a stimulation output module to form a closed-loop link. Through 32-channel 1kHz signal acquisition and 1-100Hz filtering pretreatment, pathological features are extracted in real time by using a deep convolutional neural network, stimulation parameters are modified through a remote program control module, and multi-channel collaborative stimulation and safe electrical stimulation are realized in combination with a time division multiplexing architecture and bipolar optocoupler isolation output. The method has the advantages that the pathological feature recognition accuracy of the deep convolutional neural network is greater than or equal to 95%, and the dynamic response delay is less than 50ms; a time division multiplexing hardware architecture is designed, multi-channel collaborative stimulation is achieved through a single DAC, and the size is reduced by 30%; + / -5V bipolar output is realized by using an optocoupler isolated H bridge driver, and the IEC 60601 medical safety standard is met; wireless regulation and control of stimulation parameters are supported, the long-term implantation requirement is met, and the risk of a secondary operation is reduced.

The invention provides a multi-tenant visual large model reasoning resource dynamic allocation and isolation method. An integrated system of a multi-tenant environment, visual large model reasoning, dynamic resource allocation and isolation guarantee is constructed. The system controls resources such as model copy number, video memory quota, batch size, queue weight, bandwidth and the like at tenant level fine granularity, and adjusts priority and quota based on closed-loop feedback of real-time indexes (such as queuing length, delay and throughput). Interference suppression among tasks is realized through priority grading, a hard / soft isolation strategy and a tenant-model copy mapping mechanism in combination with GPU MIG, video memory partitioning, queue scheduling and other technologies. Aiming at the characteristics of high video memory, large input and the like of a visual large model, model loading, batch processing and video memory multiplexing strategies are optimized, and differential resource allocation of heterogeneous models is supported. According to the overall scheme, on the premise that the service quality and isolation are guaranteed, the GPU resource utilization rate is remarkably increased, and the operation cost is reduced.

The invention belongs to the technical field of hidden danger detection, and more specifically relates to a power transmission hidden danger filtering method and system fusing a visual model and spatial logic correction, and a medium. The method comprises the following steps: hidden danger candidate generation based on a high-performance target detection model: adopting a reference segmentation image dynamic multiplexing strategy to complete background semantic analysis of a power transmission channel inspection image, and generating a final reference segmentation image as a semantic mask; performing spatial mapping and logic correction on hidden danger candidate boxes and semantic masks; performing confidence weighting and soft filtering on boundary cases; identifying the boundary cases, constructing a three-dimensional confidence fusion model to calculate fusion confidence, realizing soft filtering according to the fusion confidence, and outputting a high-credibility hidden danger alarm list. The problems that in the prior art, due to the fact that a background semantic base is unstable and total image re-segmentation is frequently triggered, computing resources are wasted, scenes violating common sense cannot be recognized, and the uncertainty of a boundary area is lack of fine-grained description are solved.

The invention provides a Max plug-in assisted video data transmission method and system, and relates to the technical field of video transmission optimizing.The method comprises the steps that a monitoring video stream is analyzed through a Max plug-in, and video content information containing a plurality of semantic areas, associated position identifiers and importance parameters is obtained; generating multiple paths of coding code streams by using different coding strategies according to the importance parameters, and obtaining a composite code stream through bit field non-orthogonal multiplexing; transmission resources and coding parameters are dynamically adjusted in combination with real-time network states and importance parameters during transmission; and a receiving end decodes and recovers the video data through sequential interference elimination, and finally, the Max plug-in drives the digital twinning scene to realize differentiated three-dimensional rendering presentation according to the recovered data and the associated position identifier and importance parameter, so that efficient transmission of the monitoring video data and differentiated three-dimensional rendering presentation of the digital twinning scene can be realized.

The invention discloses a dynamic mapping virtual mailbox multiplexing control method and device and a storage medium, and the method comprises the steps: creating a plurality of software queues in an RAM, taking each software queue as a software carrier of a logic mailbox, and dividing the logic mailbox according to message priority or CAN ID; when the idle physical mailbox is detected, dynamically selecting a message to be sent from all the software queues, and loading the message to be sent to the idle physical mailbox; and after detecting that the sending of the physical mailbox is completed and the interruption is triggered, selecting the next to-be-sent message from all software queues and loading the next to-be-sent message to the physical mailbox which triggers the interruption, so that the utilization rate of physical mailbox resources of the CAN controller can be effectively improved, the problem of scheduling starvation of low-priority messages caused by accumulation of high-priority messages is avoided, and the service life of the CAN controller is prolonged. The instant multiplexing of the mailbox and the fair and efficient message sending are realized, and the system resource utilization efficiency and the communication reliability are obviously improved.

The invention discloses an embedded system integration test case multiplexing method and system, and the method is characterized in that the multiplexing of a test case is carried out based on the existing software assets: firstly, based on the existing software requirements and the test case, an existing requirement-test case tracking chain is constructed; performing demand change analysis on the new demand activity graph information; performing change influence range analysis based on a change analysis result in combination with a demand-test case tracking chain, and determining a test case reuse and generation scheme; and finally, executing the test case multiplexing and generating scheme. According to the method, the tracking relation between existing requirements and test case assets is established, meanwhile, the execution cost of downstream test tasks caused by requirement change is reduced through multiplexing, and the test efficiency of integration testing is improved.

The invention discloses an LLM-DoS attack protection method based on a multi-level defense strategy and a related device, and belongs to the field of artificial intelligence security. According to the method, denial of service attacks aiming at a large language model are effectively prevented through a three-layer defense strategy: firstly, dynamic frequency control is carried out on an API request, a three-stage rate limiting system is established, a load-aware dynamic adjustment algorithm is introduced, and when the system load is too high, the quota is automatically adjusted; secondly, constructing an input perception classifier by adopting a lightweight Transform model, breaking through context limitation in combination with a random fragment compression and length penalty strategy, and realizing hierarchical interception of attack requests through a harmfulness scoring function; and finally, a request hash mapping and streaming response multiplexing technology is implemented at a server side, and real-time detection and response multiplexing are performed on repeated high-concurrency attack requests, so that consumption of computing resources is reduced. According to the method, the security and availability of the large language model service can be effectively improved, and the influence of DoS attacks on the system is reduced.

The invention provides a distributed large model reasoning method and system based on modular cache, and the method comprises the steps: disassembling a prompt into modules which can be independently reused through modular division driven by Schema, and calculating the attention Key-Value state of each module in advance before reasoning for caching. When a user request arrives, the system firstly performs module analysis and cache assembly on a prompt, loads a cache as required and supplements a missed part, and then injects a complete KV state into a reasoning process. Meanwhile, a layered and distributed scheduling mechanism is designed, a global scheduler selects a target GPU based on a cache hit rate and a node load, and a local scheduler manages cache copying and elimination in a node, so that load balancing and cache multiplexing are both considered. According to the scheme, repeated calculation can be remarkably reduced, the first Token delay is reduced, the system throughput rate is increased, and the method is suitable for high-repetition prompt scenes such as code generation and long text question and answer.

The invention relates to the technical field of meteorological data visualization, in particular to a meteorological data visualization method and device, electronic equipment and a storage medium. According to the meteorological data visualization method and device, firstly, a lightweight multi-resolution data organization unit is established based on a spatial index, a server-side pre-rendering slice is replaced, the storage pressure is reduced, and the data updating flexibility is improved; secondly, a dynamic thread pool and a load balancing mechanism are adopted, data processing tasks are intelligently decomposed and executed in parallel, client computing resources are fully utilized, the processing efficiency is remarkably improved, then coordinate conversion, interpolation and graph drawing are completed in a working thread through an off-screen rendering engine, calculation and rendering decoupling is achieved, and the system performance is improved. And finally, an incremental updating mechanism and a dirty marking technology are introduced, only local redrawing is carried out on a change area, and a cache multiplexing mechanism is combined, so that global rendering overhead is effectively avoided, and real-time, smooth and high-interactivity visualization of mass meteorological data is realized at a Web end.

An integrated circuit includes a first multiplexer, a second multiplexing-input, a first clocked comparator, and a second clocked comparator. A first input node is connected to a first multiplexing-input of the first multiplexer, a first multiplexing-input of the second multiplexer, a first comparing-input of the first clocked comparator, and a first comparing-input of the second clocked comparator. A second input node is connected to a second multiplexing-input of the first multiplexer and a second multiplexing-input of the second multiplexer. An output of the first multiplexer is connected to second comparing-input of the first clocked comparator, and an output of the second multiplexer is connected to a second comparing-input of the second clocked comparator. The integrated circuit also includes an offset control circuit connected to a comparator-output of the first clocked comparator and a comparator-output of the second clocked comparator.

The invention discloses a network configuration automatic recovery method and device, equipment and a medium, and relates to the technical field of computers. Comprising the following steps: acquiring a network configuration request for target network configuration, and calling a system command based on the network configuration request to execute a network configuration operation for the target network configuration; if the network configuration operation is successfully executed, persistently storing the structured data of the current target network configuration into a preset independent storage space; and when an operating system where the gateway software is located is restarted, executing a preset configuration recovery process so as to read the structured data of the stored network configuration from the preset independent storage space, and re-issuing the stored network configuration to the operating system through a preset multiplexing issuing logic so as to complete recovery of the network configuration. Therefore, the network configuration can be automatically recovered after the system is restarted, and the adaptability and manageability for different operating system versions are improved.

The invention discloses a video reasoning training data generation method based on key frame causal chain extraction, and the method comprises the steps: obtaining a to-be-processed video data set, and decoding the to-be-processed video data set to obtain a frame sequence; performing redundancy elimination screening on the frame sequence to obtain a key frame sequence; performing semantic structured representation on the key frame sequence to generate key frame description and entity information, and generating an event candidate set based on the key frame description and the entity information; executing multi-event causal discovery based on the event candidate set to obtain an event-level causal graph, and extracting at least one key frame causal chain from the event-level causal graph; generating a video reasoning training sample based on the key frame causal chain; and performing quality verification and screening on the video reasoning training sample, and outputting a screened training data set. According to the method, through offline key frame index multiplexing and causal structure constraint, the frame-by-frame calculation and labeling cost is reduced, data illusion is inhibited, and the verifiability of reasoning training signals is improved.

The invention relates to the technical field of substation inspection and maintenance intellectualization, in particular to a substation inspection and maintenance multi-expert agent collaborative task generation method based on an Agent2Agent protocol, which comprises the following steps of: receiving a substation inspection and maintenance natural language instruction, analyzing the instruction into a task tree containing a root node and a leaf node based on the physical topology of a substation, and labeling a capability label; calculating the semantic similarity between the current task tree and the historical task, and multiplexing and finely adjusting a historical agent capability chain or calling a capability graph to form a current capability chain; the intelligent agent is controlled to execute subtasks according to a capability chain, detection equipment is deployed in combination with a point location type, the intelligent agent performs standardized interaction through an Agent2Agent protocol, and secondary scheduling is performed when a capability gap appears; and performing streaming processing on task data, completing incremental indexing, performing fine tuning on a correlation model, and endowing high-quality feedback data with an initial weight so as to realize continuous optimization. According to the invention, the generation efficiency and execution accuracy of the substation inspection and maintenance task are improved, and intelligent operation and maintenance requirements are met.

The invention relates to the technical field of satellite navigation and wireless communication, and discloses an indoor positioning method and system based on feeder multiplexing Beidou double-frequency, and the method comprises the steps: receiving outdoor B1I and B2a signals through a high-gain Beidou double-frequency receiving antenna, carrying out the frequency separation and signal conditioning through a double-frequency separation filter, and carrying out the positioning of the B1I and B2a signals; and transmitting the B1I signal to an indoor distributed antenna for broadcasting through the existing communication feeder network, and transmitting the B2a signal to a central directional antenna to establish a time synchronization reference. A mobile terminal receives a dual-frequency signal, capturing and tracking are carried out through digital signal processing, and pseudo-range and carrier phase observation values are extracted. And performing pairing processing on observation data of the same antenna, calculating a dual-frequency pseudo-range difference value, eliminating system errors, and performing position calculation by adopting a weighted least square method and Kalman filtering. The method solves the problem that the traditional GPS signal cannot cover the indoor environment, and has a good application prospect in the scenes of underground commercial complexes, tunnels, parking lots and the like.

The invention discloses a digital acoustic fluidic chip, a preparation method and a multi-dimensional sample processing method, the digital acoustic fluidic chip comprises an upper polar plate and a lower polar plate, and the upper polar plate comprises a glass layer, a coating layer and an upper hydrophobic layer which are arranged in sequence; the lower polar plate comprises a lower hydrophobic layer, a dielectric layer, an electrode layer and a substrate which are arranged in sequence; the electrode layer comprises an EWOD pin electrode, a driving electrode, a sound wave pin electrode and an interdigital transducer; the interdigital transducer comprises a sound field and electric field multiplexing module, a sound standing wave field and thermal field module, a focusing sound field and thermal field module and a sound vortex enrichment module; each driving electrode is connected with one EWOD pin electrode, the two sides of the sound field and electric field multiplexing module, the two sides of the sound standing wave field and thermal field module, the two sides of the focusing sound field and thermal field module and the two sides of the sound vortex enrichment module are all connected with the two sound wave pin electrodes, and one side of the sound field and electric field multiplexing module is further connected with one EWOD pin electrode. The method has the advantages of multi-dimensional sample treatment and high-flexibility treatment.

Wireless circuitry may include a transmission line that carries a signal and a power detector that measures the signal. The detector may include a rectifier coupled to a comparator over a differential path. An offset calibrating digital-to-analog converter (OSDAC), a reference generator, and a multiplexer may be disposed on a negative line of the differential path. The OSDAC may produce an offset-compensated voltage by adding different offset voltages to a voltage on the second line over time. The reference generator may generate a set of threshold voltages by adding different reference voltages to the offset-compensated voltage. A temperature sensor may adjust the reference voltages used to generate the threshold voltages based on a temperature of the power detector. The multiplexer may route different threshold voltages to a negative input of the comparator.

The invention discloses a capacitive flexible sensing array based on an FPC technology and a preparation method thereof, PVDF-TrFE and ionic salt [BMIM / Br] in a specific mass ratio are prepared into an ionic thin film dielectric layer under the action of a cross-linking agent, and a specific rectangular pyramid microstructure is given to the ionic thin film dielectric layer through a microneedle structure mold; and an upper electrode layer and a lower electrode layer for channel multiplexing are prepared in combination with an FPC technology, and finally the sensing array is packaged through vacuum hot pressing bonding. The core innovation of the invention lies in that through the synergistic effect of the ion double electric layer effect, the microstructure stress concentration and the FPC electrode multiplexing design, ultrahigh sensitivity (0-20kPa interval sensitivity gt, 30kPa) and low crosstalk (array crosstalk rate lt, 5%) are realized at the same time. The local deformation characteristic of the microneedle structure effectively restrains edge electric field diffusion, the crosstalk problem of a high-density array is solved from the physical structure, and a key technology is provided for application of flexible electronics in the field of precise tactile perception.