91results about How to "Save bandwidth" patented technology

The invention discloses a cabinet type server and a communication method, and relates to the field of computers. The cabinet type server comprises a plurality of fusion nodes and a plurality of switching nodes, a first fusion node in the plurality of fusion nodes comprises a plurality of computing chips and at least one first switching chip, the at least one first switching chip is used for realizing communication connection among the plurality of computing chips, and a second fusion node in the plurality of fusion nodes is used for realizing communication connection among the plurality of computing chips. The first switching node in the plurality of switching nodes is coupled with the first fusion node through the connector, and the first switching node is used for realizing communication connection between the first fusion node and other fusion nodes in the plurality of fusion nodes. A switching chip and a computing chip are deployed nearby, a plurality of fusion nodes and a switching node adopt a mode of orthogonal connection and data two-stage switching, and a high-density system with high performance, large bandwidth and low time delay is constructed.

The invention provides a tensor core system for a large language model and a hardware chip, and relates to the technical field of artificial intelligence chips, and the tensor core system comprises an activation matrix processing part which is used for caching and preprocessing activation matrix data read from the interior of a memory to obtain a target activation matrix; the weight matrix processing component is used for caching and preprocessing the weight matrix data read from the inside of the memory to obtain a target weight matrix; the tensor kernel operation component is used for performing matrix floating-point number multiply-add operation on the target activation matrix and the target weight matrix to obtain a floating-point number multiply-add operation result, and accumulating the accumulation matrix and the floating-point number multiply-add operation result based on a preset floating-point number system to obtain a target result matrix; and the accumulation matrix and result processing component is used for performing floating-point number system conversion processing on the target result matrix to obtain a to-be-written result matrix, and writing the to-be-written result matrix into the memory. The method is more beneficial to improving the computing power utilization rate of the chip.

The invention provides a large model-based text generation and information recovery method and system. The method comprises the following steps of: converting complex situation information into simple and accurate text description by utilizing a text generation model; converting the text description of the situation information into a semantic vector by using a text encoder of a pre-trained CLIP model, and capturing semantic information of the text; encoding the noise image into a hidden space representation by using an encoder of a variational auto-encoder, and learning a compressed representation of the image; fusing the semantic vector generated by the pre-trained CLIP model with the image hidden space representation generated by the encoder of the variational auto-encoder, and optimizing the fused hidden space representation by using the trained de-noising Unet network; the optimized hidden space representation is decoded into a high quality situational information image using a decoder of a variational auto-encoder. By applying the technical scheme of the invention, the technical problem of incomplete received information caused by inevitable information loss in the data compression process in the prior art is solved.

The embodiment of the invention provides a data packet transmission control method and device, a computer readable medium and equipment. The transmission control method comprises the following steps: receiving data packets respectively sent by a data source through a plurality of transmission paths; calculating a hash value corresponding to the data packet according to the packet information of the received data packet; if the Hash value corresponding to the data packet is found in a Hash value storage space, the data packet is discarded, and the Hash value storage space is used for storing the Hash value corresponding to the received data packet; and if the hash value corresponding to the data packet is not found in the hash value storage space, storing the hash value corresponding to the data packet in the hash value storage space, and transmitting the data packet to a downstream service. According to the technical scheme provided by the embodiment of the invention, on the premise of discarding repeated data packets and reducing bandwidth occupied by network transmission, the memory space occupied by storing the information of the received data packets can be reduced, and the data transmission efficiency can be improved.

The invention discloses an H-infinity intermittent sampling load frequency safety control method for a multi-region power system under spoofing attack, and belongs to the technical field of power system safety control. According to the method, the intermittent control signal is only used in the working interval, the control cost is effectively saved, the limited network bandwidth can be saved through sampling control, and the influence of spoofing attacks on the multi-area power system is effectively resisted through safety control. Firstly, according to the characteristics of intermittent control and sampling control, the influence of spoofing attack is considered at the same time, and a multi-region power system intermittent sampling load frequency safety control dynamic model under the spoofing attack is established. Secondly, a switching Lyapunov functional is provided, meanwhile, under the framework of an LMI, sufficient conditions for existence of an intermittent sampling load frequency safety controller are given, and the conditions guarantee that the index of a multi-area power system is stable and meet the H infinity performance; and finally, verifying the effectiveness of the H-infinity intermittent sampling load frequency safety controller through three-region power system simulation under spoofing attack.

The invention relates to the technical field of picture rendering, in particular to a media picture rendering optimization method and system based on artificial intelligence. The method comprises the following steps: acquiring a terminal video stream; performing RGB format conversion on the terminal video stream to generate an RGB type frame image set; pixel partitioning is carried out on the RGB type frame image set, and RGB pixel matrix blocks are generated; performing color space acceleration conversion on the RGB pixel matrix block to generate a YUV cache data matrix; the YUV storage data matrix is sent to a DSP internal processing unit for key frame selection, and YUV key video frame data and YUV non-key video frame data are obtained; and performing intra-frame coding on the YUV key video frame data so as to generate YUV key frame coded data. According to the method, the media picture rendering efficiency is improved by optimizing data processing, transmission and rendering and combining user visual focus analysis.

The invention provides a power line carrier and wireless communication combined power data transmission method and system, and the method comprises the steps: collecting the operation data of terminal power equipment in real time through terminal collection equipment in a power distribution area, and adding a timestamp for the data; and the hybrid communication sub-equipment transmits the operation data with the timestamp to the hybrid communication main equipment of the substation by using the power line carrier. And after receiving the data, the main device calculates the communication distance between the power distribution area and the transformer substation according to the timestamp information, and selects a proper communication mode. And the master device continues to monitor and calculate the bit error rate, obtains parameters such as the signal-to-noise ratio, the attenuation value and the interference intensity of the communication channel, and inputs the parameters into the pre-trained support vector machine to predict the current bit error rate. And based on a prediction result, a communication distance, a load and other factors, a pre-trained reinforcement learning model is adopted to adaptively select an optimal communication mode, and the communication mode is dynamically adjusted to optimize a data transmission effect.

The embodiment of the invention provides a data transmission method and device, equipment and a storage medium. A receiving end receives target data sent by a sending end; verifying a plurality of to-be-verified data packets in the target data, and determining the to-be-verified data packets passing the verification as first data packets; combining the plurality of first data packets into verification data; and comparing the verification data with the target data, determining whether the verification data is consistent with the target data, and further judging whether the target data is successfully transmitted. The receiving end in the embodiment of the invention ensures data consistency, reduces data loss or damage and improves transmission reliability by verifying the data packets to be verified, timely finds errors by comparing the verification data with the target data, avoids further processing of error data, improves transmission accuracy, and improves the transmission efficiency by subpackage verification and combination and parallel processing of a plurality of data packets. The verification time is reduced, and the transmission efficiency is improved.

The invention relates to an industrial quality inspection data cooperative transmission method driven by image recognition, in particular to the field of artificial intelligence, which analyzes video frames in real time through a lightweight neural network, accurately locates and extracts key areas and features in industrial product images, constructs a dynamic semantic scene graph to understand contents and divide priorities, and improves the quality inspection efficiency. The core of the method is that non-uniform intelligent coding is carried out on video streams according to semantic importance of contents, high-fidelity transmission of key information such as defects is ensured, non-key areas are greatly compressed to save bandwidth, scheduling is carried out in the transmission process according to priorities, confidence feedback based on a cloud recognition result is introduced, closed-loop control is formed, and high-fidelity transmission is realized. According to the method, the front-end analysis model, the coding strategy and the network path are adaptively optimized, so that low-delay and high-reliability visual quality inspection data transmission and accurate identification can be continuously and stably realized in a complex industrial network environment, and the overall efficiency and the intelligent level of online quality inspection are remarkably improved.

According to the distributed photovoltaic ring communication group key generation method provided by the invention, safe and efficient group key management is realized by combining a group number derivation mechanism, environment feature binding and a quantum key enhancement technology. The method comprises the steps that a preset broadcast root key is called from a concentrator, and the broadcast root key is used for deriving group keys of members in each group; acquiring environment characteristics of members in each group in a communication group corresponding to the concentrator, and recording an acquisition timestamp; summarizing the environmental characteristics of all members in the communication group, and performing hash processing on the summarized environmental characteristics to obtain an initial environmental characteristic vector; and generating a first group key with a preset length through a key derivation function based on the broadcast root key, the initial environment feature vector, the timestamp and a preset group identifier.

The invention relates to the field of gas pipe network leakage early warning, in particular to a gas pipe network leakage early warning method based on edge calculation, which is applied to an equipment terminal, and comprises the following steps: responding to a leakage monitoring request, and receiving pipe network current data acquired by an acquisition end, the pipe network current data comprising pressure current data and leakage current data; calling a preset screening model to process the pipe network current data to obtain abnormal deviation values, and obtaining effective pipe network data based on the abnormal deviation values and the corresponding data fluctuation thresholds; receiving pipe network annotation data, obtaining hidden layer output values of the pipe network annotation data according to the initial evaluation model, and determining a state output weight based on each hidden layer output value to obtain a pipe network state evaluation model; and inputting the effective pipe network data into the pipe network state evaluation model for evaluation, generating a state evaluation result, processing an abnormal state evaluation result based on a preset early warning strategy, generating leakage evaluation data, and sending the leakage evaluation data to a data center.

The invention provides a live broadcast delay determination method and apparatus, and a terminal device. The method comprises the steps of determining a user type corresponding to a user watching a live broadcast; acquiring network parameters when the user watches the live broadcast; based on the user type and the network parameter, determining a delay parameter when the user watches the live broadcast, the delay parameter being associated with the delay when the user watches the live broadcast; and sending the delay parameter to a multimedia server, and receiving live broadcast data sent by the multimedia server based on the delay parameter.

The invention discloses a cross-platform real-time central control screen projection instrument display method, which relates to the technical field of vehicle-mounted systems, and is characterized in that real-time transmission of high-definition video data between a central control system and an instrument system is realized through an intelligent transmission technology and a special hardware component, the system integrates a serializer and a deserializer, and the real-time transmission of the high-definition video data is realized. The method supports automatic identification and conversion of various video formats, combines an AI enhancement algorithm, dynamically adjusts a compression mode, resolution and refresh rate according to factors such as content complexity and bandwidth, and ensures an efficient and clear display effect. According to the method, the high-performance shielding twisted pair is adopted, the good anti-interference capability is achieved, and the stability of data transmission is guaranteed. According to the invention, the limitation of inter-core communication which depends on the same SOC traditionally is overcome, the separated deployment of the central control system and the instrument system is supported, the synchronous management of data resources is simplified, the adaptability is high, the real-time performance is high, and the flexibility and the user experience of the central control projection screen in vehicle-mounted and industrial applications are remarkably improved.

The invention discloses a flood situation detection method and device based on big data hyper-fusion and a medium. The method comprises the steps of generating a corresponding ponding area list and a monitoring area list; performing spatial association on the ponding area list, the camera list and the water level gauge sensor list so as to determine the camera and the water level gauge sensor which have a spatial association relationship with the ponding area; for each ponding area, calculating the water depth of the ponding area corresponding to the ponding area according to the water depth data collected by the camera and the water level gauge sensor which have the spatial association relationship with the ponding area, and predicting the predicted ponding depth of the ponding area in the future according to the water depth of the ponding area; dividing camera monitoring grading lists of different grades according to the first camera list, the second camera list and the fourth camera list; and feeding back the gathered flood situation information and the map capturing frequency to a control center, so that the control center carries out rescue scheduling according to the flood situation information, and carries out dynamic scheduling on cameras in the ponding area and the monitoring area.

The invention discloses a cloud side end cross-domain training efficient communication technology based on partition sparse communication, and aims to research how a cross-domain distributed computing framework efficiently realizes quick communication of cloud side end cross-domain nodes, research a back propagation sparsification mechanism, determine the influence of sparsification of a model layer on the model accuracy and communication through a sparsity heuristic algorithm, and improve the communication efficiency. Therefore, the communication scale is greatly reduced; a partition and grading communication scheduling mechanism is researched, and mainly comprises a computing intra-cluster communication scheduling mechanism and a computing inter-cluster communication scheduling mechanism. Wherein the computing intra-cluster communication scheduling mechanism maximizes computing-communication overlapping of the computing cluster through an activation value and weight separation mechanism; and calculating an inter-cluster communication scheduling mechanism, decoupling forward propagation and back propagation, and obtaining a most suitable micro-batch scheduling target through a custom optimization target. Through the above mechanism, the communication scale is effectively reduced, the calculation-communication overlapping rate is improved, efficient and rapid cross-domain node communication is realized, and the efficiency of cross-domain distributed training is improved.

The invention provides an industrial application configuration data transmission method and system, computer equipment and a medium, and the method comprises the steps: carrying out the plug-in installation through a software development tool, then selecting a fragmentation strategy and a compression strategy according to a plug-in installation result and industrial application configuration data parameters, and carrying out the fragmentation of industrial application configuration data based on the fragmentation strategy; performing multi-thread concurrent compression on each piece of fragmented data based on a compression strategy, and uploading all compressed data; and finally, decompressing all the compressed data according to the fragmentation strategy and the compression strategy, and outputting a corresponding decompression result. According to the method, rapid and high-compression-ratio compression of the industrial configuration data can be realized, the fragmentation and compression strategies can be self-adapted according to the configuration data condition, concurrent compression and decompression can be performed on different fragments, the bandwidth of configuration data transmission and the access traffic overhead of a mobile terminal are reduced on the premise of almost not influencing the original performance and process, and the data transmission efficiency is improved. Therefore, the application development and use experience is improved.

The invention discloses an unmanned aerial vehicle actuator dynamic characteristic test and modeling method based on a real installation state, belongs to the technical field of unmanned aerial vehicles, and comprises a test system constructed based on an unmanned aerial vehicle actuator and a control surface in an in-situ installation state. The test system comprises an MEMS inertial navigation device which is fixed on the surface of the control surface and is closest to a control surface rotating shaft, and the MEMS inertial navigation device is connected with a test computer and a test power supply through cables respectively; a suction cup located on the surface of the control surface is installed at the position L away from the MEMS inertial navigation device, the suction cup is connected with a spring through a rope, and the other end of the spring is connected with a fixed frame; and when the control surface deflects by 0 degree, the rope at one end of the spring is just in a straightened state. According to the dynamic characteristic testing and modeling method for the unmanned aerial vehicle actuator based on the real installation state, the modeling precision of the actuator can be improved, and the reliability of flight control law design and the confidence coefficient of a flight simulation result are improved.

The application discloses an OTA cross-version incremental synthesis upgrading method and system based on a version graph, and belongs to the field of vehicle OTA upgrading. The method comprises the following steps: step S1, acquiring upgrade package metadata of all versions to construct a version dependency graph; step S2, synthesizing an incremental package according to a current vehicle version, a target version and the version dependency graph; and step S3, performing OTA upgrading on the vehicle through the incremental package. The application solves the problem of intermediate change loss during cross-version upgrading by constructing a version dependency graph and dynamically synthesizing a differential package.

The application relates to a pre-warning processing method, device, equipment, storage medium and computer program product. The method comprises the following steps: acquiring current inspection data corresponding to a target inspection point on an inspection route; in the case that it is determined according to the current inspection data that an inspection report needs to be generated, generating the inspection report according to the current inspection data and historical inspection data corresponding to the target inspection point; and sending the inspection report to a server, so that the server performs pre-warning processing based on the inspection report. By adopting the method, the inspection data collected can be analyzed in time while saving the time and bandwidth of data transmission, and the inspection point with a safety hidden danger can be pre-warned.

The invention relates to an image super-resolution method, system and equipment based on low-resolution increment, and the method comprises the steps: carrying out the initial down-sampling and multi-round increment reconstruction, obtaining information increment through the additional down-sampling of each round of increment reconstruction, and carrying out the up-sampling and super-resolution processing based on the information increment until a reconstructed image meeting the requirements is obtained. Through multi-round on-demand additional down-sampling and focusing on an interested area (which can be a whole image or one or more local image blocks) of a user, incremental optimization is realized, transmission of redundant data is effectively reduced, unnecessary resolution overhead is avoided, dependence on low-resolution image size is reduced, data transmission and processing are enabled to be more simplified, and user experience is improved. The system efficiency is improved, and the bandwidth is saved.

The present application relates to the field of data processing, and provides a method and device for asynchronously updating multiple channels of image data. The method includes: creating multiple graphics buffer memory blocks and multiple external extension textures; based on the application extension interface, establishing a binding relationship between each external extension texture and each graphics buffer memory block, and storing the binding relationship in a relationship cache set; if it is detected that the non-shared memory of the CPU writes at least one channel of newly added image data, then determining the target graphics buffer memory block and target external extension texture corresponding to each channel of newly added image data; for each channel of newly added image data, running a data writing thread, and writing the newly added image data into the corresponding target graphics buffer memory block. The present application improves the overall synchronization efficiency of multiple channels of image data, shortens the time consumed for overall synchronization update, and greatly reduces the CPU load.

The present application provides a firmware sending method, a firmware receiving method and a firmware upgrade system, and relates to the technical field of wireless earphones. The firmware upgrade system includes: a terminal device, which is used to divide the upgrade firmware corresponding to the earphone device into multiple target data blocks; send the multiple target data blocks to the first earphone and the second earphone in the earphone device; perform an XOR operation based on the first receiving information fed back by the first earphone and the second receiving information fed back by the second earphone to obtain an XOR data packet; send the XOR data packet to the first earphone and the second earphone; the earphone device includes a first earphone and a second earphone, which are used to receive the multiple target data blocks sent by the terminal device respectively; send the first receiving information and the second receiving information to the terminal device according to the reception status of the multiple target data blocks; perform XOR analysis based on the received XOR data packet sent by the terminal device to obtain the first data block and the second data block.

The application relates to a fast mirror system based on a transmissive application of an eddy current sensor, which comprises a detector seat, a working mirror body, a mirror seat, a flexible hinge, a voice coil motor, an inner ring seat, an outer ring seat, a rotating ring, an eddy current sensor and a controller; the working mirror body is fixedly connected with the mirror seat; the mirror seat is fixedly connected with the inner ring seat through the flexible hinge; the inner ring seat is fixedly connected with the outer ring seat through the flexible hinge; the outer ring seat is fixedly connected with the rotating ring; the rotating ring is fixedly connected with the inner ring seat and the outer ring seat through the flexible hinge; the outer ring seat is fixedly connected with the detector seat; four eddy current sensors are assembled in the detector seat to collect angle information when the mirror seat rotates; the application has the characteristics of high bandwidth and high precision, can realize the ability of reflecting and transmitting light paths of the working mirror body, provides more possibilities and operability for application environments of a light path stabilization system, and provides a good selection scheme in the directions of light path stabilization and system light path design.

The invention provides an Internet of Vehicles security batch authentication method oriented to a cloud service scene, and relates to the field of Internet of Vehicles and post quantum security. The batch authentication method comprises the following steps: a trusted mechanism generates a public parameter and a public and private key pair of the trusted mechanism; the vehicle and the road test unit apply for registration from the trusted mechanism, and the trusted mechanism generates an identifier and a private key for the vehicle and the road test unit according to an access control strategy; when the vehicle and the road test unit communicate with the cloud server, the vehicle and the road test unit perform digital signature on the sent message by using the private keys thereof; when the cloud server receives massive messages, the cloud server carries out batch authentication on the messages with the same attribute to verify the authenticity and integrity of the messages and accelerate the authentication process at the same time; according to the batch authentication method provided by the invention, a safe and efficient data authentication service can be provided for the cloud server, the consumption of storage and computing resources is reduced, and meanwhile, the response delay and the average message packet loss rate are reduced; and the method can realize fine-grained access control and post-quantum security.

The invention provides a map updating method, device and system. The method comprises the following steps: acquiring updated map data; determining a first area to which an updating area corresponding to the updated map data belongs; determining a first terminal device meeting a map updating condition according to the obtained position relationship between the position of the terminal device and the first district; and pushing a map updating notification to the first terminal device to control the first terminal device to complete map updating. According to the method and the device, the problems of discontinuous map display, white screen and the like of the terminal equipment are solved, the network bandwidth is saved, and the map updating efficiency is improved.

The embodiment of the present application provides a CCM message counting method, device, electronic device and storage medium, including: when it is detected that a continuity detection CCM message is received, a hash table lookup operation is performed according to the CCM session identifier to obtain a table lookup result; the table lookup result includes the local identifier of the detection chip, the identifier of the target detection board and the target detection chip, and the counting index; if the local identifier of the detection chip indicates that it is not the target detection chip, the first message header is encapsulated in the CCM message and transparently transmitted to the target detection chip; instead of using the multicast replication mechanism, the message transparent transmission method is adopted to transparently transmit the CCM message to the target detection chip. It ensures that there are no large numbers of redundant duplicate messages between each line card board and each chip, and avoids the impact of attack messages from the source. It saves the bandwidth of cross-chip transmission within the device, reduces the impact of attack flows on normal services, and improves the anti-attack and overall stability of the device.

The application relates to the technical field of semiconductor chip packaging, and discloses a lead frame structure and a packaging structure for SOW10 packaging of a Hall current sensor chip, which comprises a lead frame base island, the thickness of the lead frame base island is greater than that of a conventional SOW16 lead frame base island, so as to compensate for the loss of the cross-sectional area of the original side current path caused by the reduction in the number of pins in a first pin group; the lead frame pins comprise a first pin group and a second pin group; the first pin group is directly arranged on the lead frame base island and is electrically connected to the lead frame base island, is located at the opposite side of the side where the protruding part is located, and is used for leading in and out the original side current; the second pin group is connected to a die through a lead wire; an isolation device is arranged between the die and the lead frame base island, and the isolation device completely covers the area where the lead wire passes through the lead frame base island on a projection plane. The application provides the lead frame structure with lower conduction resistance, and the structure has lower conduction impedance.

The application discloses an interactive form design method and system based on a language large model, and belongs to the technical field of human-computer interaction and form automatic generation. In view of the problems that a traditional form design tool is tedious to operate, highly depends on manual experience, and an existing method based on a large model has high training cost, weak interaction ability and difficulty in real-time collaboration, the following scheme is used for improvement: defining a JSON Schema specification of a form and constructing a visual form designer, integrating a language large model interaction entrance in an interface, dynamically assembling a request containing user instructions, a JSON Schema specification and existing form data according to whether the form is empty, and submitting the request to the language large model; receiving a response in a streaming manner, and real-time analyzing and progressively updating a form designer preview interface, so that a user can directly generate or modify a form through natural language instructions, and multi-round interaction and manual subsequent adjustment are supported. The method is mainly used for reducing a form design threshold, improving design efficiency and intelligent collaboration level.

The application relates to the technical field of computers, in particular to the field of artificial intelligence, and provides a message forwarding method and device, electronic equipment and a storage medium. The method comprises the following steps: based on forwarding information of a to-be-processed message, querying a matched address table entry in a first address table of an initial board card in a network device; in the case that no address table entry is queried in the first address table, based on the forwarding information, querying a matched address table entry in a second address table of a switching network board in the network device; in the case that the matched address table entry is queried in the second address table, forwarding the to-be-processed message through a target forwarding port in the matched address table entry. Since the switching network board of the application can also be used for address table entry query, flooding forwarding is reduced, system bandwidth is further saved, and the hardware resource utilization rate is improved.

A method and apparatus for determining an initial bandwidth portion (BWP), the method comprising: determining a frequency resource of a first initial BWP, the frequency resource of the first initial BWP being included in a frequency resource of a second initial BWP, the frequency resource of the first initial BWP corresponding to a first type terminal device, the frequency resource of the second initial BWP corresponding to a second type terminal device, and the frequency resource of the first initial BWP being included in a frequency resource of the second initial BWP; the two types of terminal devices have different capabilities. According to the scheme provided by the embodiment of the invention, a new method for determining the BWP is provided, so that the required system bandwidth is saved while the system supports different types of terminal equipment.