407results about How to "Improve fault tolerance" patented technology

The embodiment of the invention discloses a speech recognition method and a related device, and at least relates to a speech recognition technology in artificial intelligence, speech data to be recognized are used as input data of a time delay neural network in an acoustic model, and an output layer of the time delay neural network comprises acoustic modeling units corresponding to a plurality of syllables respectively, so that the speech recognition efficiency is improved. And the syllable probability distribution corresponding to the voice frames included in the voice data can be obtained by taking the syllables as the recognition granularity through the time delay neural network. When syllable recognition is carried out through the output layer, auxiliary judgment can be carried out on the syllables to which the voice frames belong on the basis of pronunciation rules in combination with front and back syllable information of the voice frames, so that more accurate syllable probability distribution is output. Moreover, since the syllables are generally composed of one or more phonemes, the method has higher fault-tolerant capability, not only can more accurately determine the speech recognition result based on the probability distribution of the syllables, but also has low requirements for the quality of the speech data to be recognized, and effectively expands the application scenarios of the speech recognition technology.

The invention relates to a multi-shaft compressor unit and a control method thereof, in particular to a steam-electricity double-drive multi-shaft compressor unit connected with double clutches and a control method of the steam-electricity double-drive multi-shaft compressor unit. The technical problems that in the prior art, a pure electricity mode, a pure steam mode and a steam-electricity double-drive mode cannot be switched on line, and when steam is sufficient, a motor cannot be disengaged, so that energy is wasted are solved. The system comprises a multi-shaft compressor, a steam turbine, a motor, a first clutch, a second clutch, a steam flow sensor, a controller and a rotating speed sensor. The multi-shaft compressor comprises a driving shaft; the driving shaft is provided with two driving ends which are respectively a first driving end and a second driving end; the output shaft of the steam turbine is coaxially connected with the first driving end through a first clutch; an output shaft of the motor is coaxially connected with the second driving end through the second clutch; and rotating speed sensors are arranged on the first driving end, the second driving end, the steam turbine and the output shaft of the motor.

The invention relates to a physical unclonable function modeling attack resisting method based on a neural network. The method comprises the following steps: generating a multi-bit random bit stream by using a random seed generated by an LFSR module, and simultaneously acting PUF excitation on a plurality of physical unclonable functions to generate a multi-bit physical unclonable function response; the method comprises the following steps: constructing an initial feature matrix by utilizing a Decoder module, and equally dividing position decoding excitation into a plurality of binary position decoding sub-excitation with a certain length; sequentially guiding multi-bit physical unclonable function responses to storage units of corresponding addresses to form a physical unclonable function response feature matrix, and inputting the physical unclonable function response feature matrix into the neural network model to obtain one-bit confusion response; and the other bit of confusion response of the random seed is reutilized until a final confusion response sequence with a preset bit is generated, and the server compares the final confusion response sequence with an expected confusion response to complete identity authentication. By adopting the method, the hardware security can be improved.

This invention discloses a capture device for unmanned underwater vehicles (UUVs), comprising a submersible fixed by a mounting frame. The mounting frame is symmetrically mounted with a grasping mechanism for capturing the UUV via a base. The grasping mechanism includes a drive arm driven by the base, with the other end of the drive arm rotatably connected to four corner support frames. The four corner support frames are driven by connecting arms to contact claws for gripping. A groove is provided on the side of the drive arm near the base, and a servo motor is mounted on the base. The servo motor is connected to a gear at one end of a limiting link, transmitting power through the other end of the limiting link to the groove on the drive arm to achieve the movement of the grasping mechanism. This invention also discloses a method for launching and recovering UUVs in conjunction with the above-mentioned launch and recovery device. This invention simplifies precise hook docking operations in high sea states, reduces the dependence of UUVs on computing power during launch and recovery, and improves the automation level of the launch and recovery process.

The application provides a kind of multi-microgrid collaborative economic dispatching method based on blockchain, comprising: based on microgrid group multi-group architecture model, construct microgrid group dispatching architecture model based on blockchain;Based on objective function submodel, constraint condition submodel and incremental cost submodel, an economic dispatching model is constructed;Wherein, economic dispatching model includes: microgrid layer economic dispatching model and microgrid group layer economic dispatching model;Based on microgrid group dispatching architecture model, intra-group scheduling is carried out on microgrid layer economic dispatching model, and inter-group scheduling is carried out on microgrid group layer economic dispatching model, to obtain multi-microgrid collaborative economic dispatching result.The application constructs microgrid group system distributed dispatching architecture, designs system economic dispatching model with comprehensive economic and environmental protection targets, and proposes a kind of intra-group optimization and inter-group optimization economic dispatching method based on blockchain.In the case of ensuring good expansibility, the economic efficiency and fault tolerance of microgrid group are improved.

The utility model relates to a wall seat structure for installing and fixing a sunshade, which belongs to the technical field of the sunshade and comprises an arm seat frame connecting piece, a transition connecting block is arranged at the front end of the arm seat frame connecting piece, and a bolt through hole connected with the arm seat frame connecting piece through a screw and nut structure is arranged on the transition connecting block. The arm seat frame connecting piece comprises an inserting block, an adjusting connecting block integrated with the inserting block is arranged at the front end of the inserting block, and a U-shaped fixing groove penetrating through the bolt is formed in the adjusting connecting block. The device has the characteristics of simple structure, convenience in installation and high structural strength. The wall body fixing part and the sunshade arm connecting part are connected through the movable transition pivot, so that angle fine adjustment after installation is achieved, and different sunlight irradiation angles or storage requirements are met.

The invention relates to the technical field of vehicle-mounted data processing, in particular to a safety method for instrument alarm icon detection, which comprises the following steps: after a black screen signal is received, an instrument application program receives a fault signal and performs logic judgment on the fault signal to obtain an alarm icon state of the fault signal; the instrument application program sends an alarm icon state of the fault signal to a UI thread and an SOC system adaptation layer; the SOC system adaptation layer forwards the alarm icon state of the fault signal to an SOC MCU, and the SOC MCU forwards the alarm icon state of the fault signal to a screen MCU; the screen MCU receives the alarm icon state of the fault signal and performs state detection so as to determine a mapping decision according to a detection result; and the screen MCU continuously monitors and updates the alarm state, and then resets the monitoring state after the system is recovered. According to the invention, the fault signal processing efficiency is effectively improved.

This invention relates to the field of construction monitoring equipment technology, specifically a settlement observation instrument for roadbed safety construction. The instrument includes a string strain gauge and a differential pressure static level. The differential pressure static level has connecting pipes on both its air and water inlets, which are fixed to a conduit support. The signal line of the string strain gauge is fixed within the conduit support. The two ends of the string strain gauge are fixed to a fixed end and a floating end, respectively. The floating end includes a movable end seat fixedly fitted onto one end of the string strain gauge. This invention allows adjustment of the floating end's height position through a sliding fit between the upper ball joint and the rotating sleeve. The orientation and tilt angle of the floating end can be freely adjusted through the universal joint between the upper ball joint and the upper spherical shell, and between the lower ball joint and the lower spherical shell. During on-site installation, adjusting the floating end position ensures that the string strain gauge's measuring rod remains in a natural, unloaded, and unstressed state, thus avoiding measurement errors introduced by initial stress and effectively improving the accuracy and stability of roadbed settlement and tilt monitoring.

The application provides a game automation system and method based on OCR recognition, comprising a business module, an automation module, an OCR module and a scheduling module; the business module is used for game automation configuration information management, task management and execution device management; the automation module is used for matching automation scripts, controlling a task queue, distributing a task and standardizing an automation execution result; the OCR recognition module is used for receiving a screenshot sample in an automation execution process, judging a current automation execution state, extracting and calling key parameters in an automation scene; and the scheduling module is used for distributing device nodes for task execution and controlling a task execution queue; the application provides a high-stability and high-fault-tolerance game full-automation solution according to the business characteristics of great changes in game automation field elements and difficult positioning.

The utility model discloses an integrated system host for controlling large-screen display, and relates to the technical field of large-screen display control systems. The backboard comprises a backboard module, the inner wall of the backboard module is provided with a PCIE slot, and an interface of the PCIE slot is plugged with a signal processing module. According to the utility model, through cooperation of the signal processing module and the backboard module which are designed in a modularized manner, a hardware architecture assembled according to needs and a PCIE slot interconnection technology are adopted, a user is allowed to flexibly select HDMI input and output, audio processing and network exchange of functional modules according to actual needs, the construction complexity is greatly reduced, redundant wiring is reduced, and the cost is reduced. And meanwhile, primary and secondary hierarchical design among the modules and an independent communication mechanism ensure that operation can be recovered by quickly replacing a secondary module when any module fails, so that the fault-tolerant capability and the maintenance efficiency of the system are remarkably improved, and a reserved expansion space supports subsequent function upgrading or capacity expansion and is adaptive to large-screen display scene requirements of different scales.

The application discloses a crosstalk suppression module and a control method thereof, a display panel and a display device, and relates to the technical field of display. The crosstalk suppression module is provided with a gate drive circuit integrating gate scanning units at all levels, and a replacement circuit connected between a direct-current high-voltage signal line and a first passage end of a pre-charging transistor in the gate scanning unit. When the direct-current high-voltage signal line is normal, the pre-charging transistor is enabled to be connected to a pre-charging voltage provided by the direct-current high-voltage signal line under the drive of a front-stage scanning signal to pre-charge the first node, so that an output transistor is turned on and a current-stage scanning signal is output. When the direct-current high-voltage signal line is invalid, the pre-charging transistor is enabled to provide a pre-charging replacement voltage of the replacement circuit to the first node under the drive of the front-stage scanning signal, so that the output transistor is normally turned on. The application aims to improve the fault tolerance of the gate drive circuit to the invalid direct-current high-voltage signal line to improve the uniformity of picture display.

The application relates to the technical field of building information technology and discloses a building 3D model automatic generation method and system based on a DXF file, which comprises the following steps: extracting floor numbers by analyzing the file name of a multi-floor DXF file, performing concurrent format conversion on multiple files to generate GeoJSON data, identifying building components such as walls, doors, windows and rooms of each floor through geometric semantic analysis to generate a floor building semantic model, assembling and generating a complete 3D building model according to the floor sequence based on the semantic model, and simultaneously realizing asynchronous task scheduling of each link by using a distributed micro-service architecture. The application can realize full-automatic, efficient and high-precision generation of a building 3D model, greatly reduces the modeling cost and technical threshold, and is suitable for application requirements of multiple scenes such as building design, real estate display and city digital twinning.

This application provides a picking method, apparatus, system, and equipment. The method includes: determining picking waves across outbound timeframes based on picking area inventory information, with the goal of freeing up more empty space in the picking area to replenish goods for pending order packages and ensuring timely delivery; determining first replenishment information corresponding to the picking waves across outbound timeframes, and performing replenishment processing on order packages in the picking waves across outbound timeframes based on the first replenishment information; and performing picking processing on order packages in the picking waves across outbound timeframes based on the replenished picking area inventory information. This approach enables dynamic coordination between picking waves and replenishment, creating picking waves that improve inventory turnover in the picking area when picking area capacity is limited and order timeframes are critical. This solves the problem of outbound delivery due to insufficient replenishment in the picking area, thereby ensuring normal order fulfillment.

The application provides a kind of design method of token codebook based on deep hash, it is related to intelligent communication technical field, including: extracting the high-dimensional semantic embedding vector set of target token list, and constructing similarity matrix;With similarity matrix as guidance, construct triplet data set for training deep hash model;Deep hash model is trained using triplet data set;The target token list is input into the trained deep hash model, to obtain the binary code word corresponding to each token, form the preliminary token codebook;Identify the code word that occurs hash collision in preliminary token codebook, reassign code word for the token corresponding to these code words, obtain the final token codebook.The application can better preserve the semantic association between tokens to improve the fault tolerance and expression coherence of semantic communication system.

This invention discloses a topology-variable bias control method and system for a five-axis magnetic levitation bearing, belonging to the field of magnetic levitation bearing control technology. The method includes: calculating the target instantaneous levitation force required for each winding based on rotor displacement; using threshold limits for winding and common arm currents and the constraint that the electromagnetic force of each winding equals the target instantaneous levitation force; and optimizing to minimize copper losses in the five-axis magnetic levitation bearing system. The method then solves for a dynamic allocation scheme of bias current and differential current combinations for each radial winding pair; dynamically adjusting the bias current of each radial winding pair to reduce the effective value of the winding current while ensuring levitation force requirements, thereby significantly reducing losses. By globally coordinating the bias current allocation of all radial winding pairs, the common arm current is always kept within the device's safe capacity, and optimized allocation and fault-tolerant operation are achieved using the control degrees of freedom between multiple winding pairs. This method effectively improves system efficiency and reliability without increasing hardware costs.

The application discloses a top-opening drum washing machine inner drum positioning method and a washing machine. The method comprises the following steps: after setting a washing program, the inner drum is controlled to rotate at a preset low speed; the rotation angle of the inner drum is collected and the deviation of the rotation angle from a target angle is calculated; the motor speed is controlled by PID closed-loop regulation with feedforward compensation based on the deviation, wherein the feedforward compensation amount is calculated according to the load weight, the speed fluctuation and / or the deviation change rate; when the deviation is less than a preset threshold, segmented braking is performed until the inner drum is locked at the target angle position; if the locking fails, the inner drum is reversely rotated to a preset angle, and the segmented braking is retried with corrected braking parameters. The application combines PID regulation with feedforward compensation and segmented braking, introduces a failure retry and parameter self-adaptive correction mechanism, and realizes high-precision, impact-free and high-reliability positioning of the inner drum.

This invention discloses a requirement tracking method, apparatus, device, and medium based on a large language model, relating to the field of artificial intelligence technology. The method includes: performing hierarchical structure tracking and nested table linearization on the document to be processed, and semantically segmenting it; submitting each text block in parallel to a large language model to extract structured requirement items according to preset fields; calculating the cosine similarity between requirement items using an embedding model to recall a candidate matching set; then using the large language model to perform semantic association matching and coverage determination to obtain association mapping data; using the large language model to perform multi-level semantic matching on the requirement items, constructing a multi-level tracking link that supports cross-level matching between non-adjacent documents; and parsing the test item result description text and mapping it to a completed state when the link ends and contains a test document. This invention can improve the automation and accuracy of requirement item extraction, association matching, coverage determination, and full-link tracking, and enhance the fault tolerance of the tracking link.

This invention discloses a sensing spring adjustment mechanism for mask mounting, belonging to the field of semiconductor lithography equipment technology. It includes a mounting base, a rotating shaft rotatably mounted on the mounting base, and a detection lever fixedly connected to the rotating shaft. Double torsion springs are sleeved at both ends of the rotating shaft, with the middle connecting end of the double torsion spring abutting against the bottom of the mounting base. A translational limiting mechanism is provided at the bottom of the mounting base, cooperating with the connecting end of the double torsion spring to guide and radially limit the movement of the connecting end. This invention uses double torsion springs instead of the traditional single-sided spring structure, providing a more balanced and stable reset force for the detection lever, ensuring the consistency and accuracy of the detection lever's reset action. Simultaneously, the double-sided support structure significantly reduces the risk of large deformation of the spring due to external force interference, fundamentally avoiding the problem of the spring detaching from the preset mounting slot.

The application discloses a method for automatically detecting the light column inclination angle of light column laser paper, and relates to the technical field of testing material performance by measuring material physical properties, which comprises the following steps: fixing the light column laser paper on a rotating platform; irradiating the surface of the light column laser paper with a linear light source at a set angle γ, and collecting the reflection image of the surface at the same time; controlling the rotating platform to rotate the light column laser paper and collecting images; calculating the evaluation value of each collected image and the corresponding rotation angle; comparing each evaluation value, setting the collected image corresponding to the evaluation value with the largest value as the target collected image; and obtaining the light column inclination angle α1 according to the set angle γ and the rotation angle β1 corresponding to the target collected image. The method can detect the light column inclination angle of the light column laser paper with any initial deviation, solves the problem of the maximum allowable inclination angle limitation in the prior art, and greatly enhances the fault tolerance of the initial placement position of the light column laser paper.

The invention discloses a seismic source positioning method considering the influence of an abnormal value and a dynamic wave velocity, and the method comprises the steps: arranging sensors to detect sound waves emitted by a seismic source, and observing the time when the sound waves reach each sensor; subtracting the control equation established based on the reference sensor from the control equation established based on the measurement sensor, and carrying out linearization and matrix representation on a nonlinear equation line obtained by subtraction; the matrix is solved, an analytic solution enabling the sum of squares of the residual errors to be minimum is found, the obtained analytic solution is substituted into an overdetermined matrix equation to obtain a fitting model, and weight estimation is conducted according to the relation between the residual errors and weights; iteration is carried out between weight estimation and overdetermined matrix solving, after optimal weight estimation is obtained, overdetermined matrix solving is completed according to the optimal weight, and a final seismic source positioning result is obtained based on an analytical solution. According to the method, the technical problem that the influence of abnormal values and wave velocity measurement errors on the seismic source positioning precision in a complex engineering environment is large is solved, and the fault tolerance and the real-time performance of seismic source positioning are improved.

The invention particularly relates to a cable line fault detecting and positioning system for an intelligent power distribution network ring main unit, and relates to the technical field of intelligent power distribution network fault detection. A multi-dimensional feature extraction and intelligent diagnosis module; a health state evaluation module; and a self-adaptive calibration compensation and control module. According to the invention, a passive acquisition mode is combined with steady-state conventional monitoring and transient double-trigger high-speed recording, so that hard fault related data can be accurately captured; the active test mode automatically injects a standard signal at a power grid load trough, flexibly captures soft degradation characteristics such as element aging and parameter drift, and the magnetic isolation design avoids interference to the power grid.

The application discloses a kind of substation wireless meter reading method, system, electronic device and storage medium, belong to electronic communication technical field, solve the technical problem of insufficient accuracy of existing wireless meter data acquisition.It includes using multiple wireless acquisition nodes deployed in substation to collect multi-modal raw data and pre-process, obtain the multi-modal data set of each node after pre-processing;Semantic feature extraction is carried out on the multi-modal data set after pre-processing, to generate the unified multi-modal feature representation of each node;Using lightweight Transform model to process the unified multi-modal feature representation, to obtain the meter reading prediction value of the node and the corresponding reliability evaluation result;Based on the reliability evaluation result, the meter reading prediction value of each node is weighted and fused to obtain the global meter reading;Based on the error between global meter reading and reference value, the parameters of lightweight Transform model are optimized.The reliability and fault tolerance of data reading are realized.

The utility model belongs to gas detection technical field, concretely is small flow gas detection device based on valve opening degree regulation, including valve body, still include: gas detection subassembly, set up at valve body top, flow detection subassembly, set up in valve body outer wall and with its inside intercommunication, baffle, set up in valve body inner wall middle segment, and its top is provided with ladder hole, first limit frame, fixed setting in valve body inside bottom, sealing assembly, through setting on first limit frame, and its top cooperates with ladder hole, first sealing plate, set up in valve body one end, the utility model provides small flow gas detection device based on valve opening degree regulation, can carry out quick sealing to valve body when needing closing valve body in the emergency, relative to the speed of screw rod closed mode is faster, adapts the gas flow demand of quick change, especially in small flow gas detection, can realize more accurate control, simple structure simultaneously.

This invention provides a remote smart lock control method based on a cloud platform, relating to the field of artificial intelligence technology. The method includes: the cloud platform sending a sound wave signal transmission command to the target smart lock, triggering the smart lock to send multi-band sound wave signals to preset auxiliary signal enhancement device deployment locations A and B; the cloud platform receiving sound wave feedback signals from locations A and B, calculating the final communication path based on signal propagation delay, frequency band interference intensity, and attenuation parameters, combined with real-time signal quality indicators reported by the remote communication module; and dynamically encrypting the command and sending it to the remote communication module through the final communication path. This invention dynamically calculates the optimal transmission path, breaking through the limitations of traditional single-channel transmission. Even in complex electromagnetic environments or obstacle-blocked scenarios, it can still ensure reliable transmission of control commands and reduce communication interruption rates.

The invention discloses a small-size high-efficiency large-current lithium battery active equalization circuit and a control method thereof, the small-size high-efficiency large-current lithium battery active equalization circuit comprises active equalization modules which are connected in parallel between two connected battery cells, and each active equalization module is composed of four switching tubes and a group of resonance branch circuit which is formed by connecting an inductor and a capacitor in series; n-1 active equalization modules are needed for a battery pack composed of N battery cells. The LC series branch serves as an energy temporary storage unit to replace a traditional inductance type single inductor or capacitance type multiple capacitors, the number of elements is reduced, meanwhile, low-loss energy transfer is achieved through resonance characteristics, and the equalization scheme has the advantages of being small in size, large in equalization current, small in equalization energy loss between battery cells, high in efficiency and the like.

The present application relates to the technical field of image state recognition, in particular to a kind of based on pattern recognition's picking and replacing hook AI accurate identification and capture system and method, in the present application, the outline change of hook assembly in image sequence frame, boundary difference and gray dynamic are carried out node construction, and combined with edge displacement accumulation analysis, while through graph neural network, cosine value between nodes and coordinate difference are compared jointly, construct path jump sequence, to enhance the response sensitivity of state mutation, under the condition of complex background interference or local occlusion, still can stably extract the key path of morphological evolution, effectively improve the anti-interference and fault tolerance of spatial path recognition, further statistical modeling is carried out on state rate mutation point by hidden Markov model, by matching standard state mode, paragraph merging and invalid segment rejection operation are carried out to abnormal point section, construct state label sequence, realize the accurate division of high confidence, multi-section continuous state.

The application provides a bolt recognition method and system and an electronic device, and relates to the technical field of computer-aided engineering. The application determines the connection relationship features between the candidate sub-regions and adjacent virtual components by acquiring the geometric features of the candidate regions in the assembly model, combining the connection topological information of the candidate sub-regions in the assembly model, and inputting the connection relationship features into a multi-modal learning model for bolt recognition, so that efficient and accurate recognition of the bolt is realized, and the fault tolerance and recognition accuracy are higher in complex assembly structures.

This invention discloses a shift speed control method and vehicle for an AMT (Automated Manual Transmission) gearbox, relating to the field of vehicle transmission technology. The shift speed control method for the AMT gearbox includes the following steps: When the AMT gearbox shifts and adjusts speed, the motor executes a speed response mode, calculates the real-time speed difference at the sliding sleeve end, and determines whether the real-time speed difference at the sliding sleeve end is within the speed difference setting range. If yes, the speed adjustment is considered complete; otherwise, the speed adjustment is considered a failure, and the number of speed adjustment failures is accumulated. It is then determined whether the accumulated number of speed adjustment failures exceeds a failure limit. If so, the initial value of the speed difference setting range is adjusted based on the accumulated number of speed adjustment failures, and the adjusted speed difference setting range is the adjusted value of the speed difference setting range. If the adjusted value of the speed difference setting range exceeds the speed difference limit, the motor's speed response mode is switched to a torque response mode. This shift speed control method for the AMT gearbox improves the success rate of shift speed adjustment.

The utility model provides a kind of wastewater treatment system of reinforced AOA-MBBR process stable operation, including secondary sedimentation tank, the periphery of secondary sedimentation tank is sequentially provided with anaerobic tank, first MBBR aerobic tank, second MBBR aerobic tank, first anoxic tank, second anoxic tank, filter tank;First and second MBBR aerobic tank are equipped with several combined suspended fillers, the outlet of two MBBR aerobic tanks is installed with cylindrical screen, and float ball liquid level meter, aerator are installed in the pool, online dissolved oxygen instrument and ammonia nitrogen on-line monitor are installed in second MBBR aerobic tank, float ball liquid level meter, online dissolved oxygen instrument and ammonia nitrogen on-line monitor are respectively covered with a protective cover outside.AOA-MBBR process is formed by coupling MBBR process and AOA process, wastewater is treated by anaerobic, aerobic, anoxic alternation state, improves total nitrogen and total phosphorus removal rate biological treatment, maximally utilizes raw water organic matter and internal carbon source to achieve the purpose of strengthening biological denitrification and phosphorus removal.