145results about How to "Reduce communication costs" patented technology

The invention discloses a detection laboratory total factor digital twinning model construction method and an intelligent operation and maintenance system, and relates to the field of digital twinning, and a total factor digital twinning model comprises a detection laboratory physical object model, a detection laboratory virtual model, a total factor data model, a data interaction model, a learning decision model and an operation and maintenance management model. Real-time mapping is kept between the physical object model and the virtual model of the detection laboratory; the total element data model is used for collecting, inputting, processing and updating multi-source heterogeneous data in the digital twinborn model of the detection laboratory; the data interaction model carries out fusion transmission on real-time data among the models; and meanwhile, the learning decision-making model performs autonomous decision making and diagnosis, and functions are integrated and displayed through the operation and maintenance management model, so that the intelligent operation and maintenance level of the detection laboratory is further improved. The whole detection process of the detection laboratory is digitalized, the communication cost of the detection test process is saved, the manpower and time cost is saved, and the detection efficiency is improved.

The invention discloses a code review method, system and device, a storage medium and a program product, and the method comprises the steps: cleaning a to-be-reviewed code snippet, and obtaining an incremental code snippet and position information; vectorizing the incremental code snippets; semantic query is conducted on the vectorized incremental code snippet data based on an ES database according to the cosine law, development rules of similar semantics are obtained and ranked from high to low according to similarity values, and a preset number of first several snippets are obtained; recombining cue words, wherein the recombining cue words comprise incremental code snippets and a preset number of development rules; and sending the recombined cue word to a general generative large model to obtain an examination suggestion. The code snippets to be examined are pushed to the gilab warehouse by a developer, and are cleaned by the gilab warehouse through webhook triggering. According to the review method, the whole process is automatic, the review efficiency is high, review can be conducted according to the development rules in the team, knowledge sharing is achieved, the review result is objective, and the problem that the text exceeds the upper limit is avoided.

The invention discloses a federated learning communication method, system and device, and the method comprises the steps: receiving a quantization model parameter and an inverse quantizer sent by each first client participating in the current round of federated aggregation in any round of federated aggregation of transverse federated learning; adopting an inverse quantizer corresponding to the target client to inversely quantize the quantization model parameter corresponding to the target client to obtain a reduced model parameter; and after the restoration model parameters corresponding to the first clients are obtained, aggregating the restoration model parameters of the first clients to obtain aggregated model parameters, and feeding back the aggregated model parameters to the first clients. Therefore, the federal learning communication efficiency is improved.

The invention relates to a dual-controller communication system and method, a computer product, equipment and a storage medium. The dual-controller communication system comprises a first controller and a second controller, wherein the first controller comprises a first processor and a first non-transparent bridge; the second controller comprises a second processor and a signal relay device; the input end of the first non-transparent bridge is connected with the output end of the first processor, the output end of the first non-transparent bridge is connected with the input end of the signal relay device, and the output end of the signal relay device is connected with the input end of the second processor. By adopting the system, the communication cost of the first controller and the second controller can be reduced, and the reliability of the system is improved.

The invention discloses a model parameter updating method and system based on graph federated learning, and relates to the field of federated learning and graph neural networks, and the method comprises the steps: completing the training of a graph convolutional neural network through employing a federated learning frame, and enabling the training to be capable of training a model on a graph distributed on a plurality of clients; a neighbor feature aggregation function is used, so that the clients do not need to communicate in each training iteration, and the communication burden is remarkably reduced; an aggregation mechanism based on an attention mechanism is used, a server is allowed to dynamically allocate weights for each local model participating in training, and the robustness and accuracy of the model are improved.

The invention discloses a DAG-based cross-fragmentation point-to-point energy transaction optimization method, and belongs to the technical field of cross-fragmentation transaction processing. Comprising the following steps: defining a cross-fragment transaction; describing a dependency relationship among a plurality of sub-transactions based on a directed graph (DAG) structure; constructing a coordination layer based on a DAG account book, and selecting a node with the highest reputation value to serve as a DAG committee member; and carrying out transaction verification, cross-fragment transaction forwarding and verification during forwarding on the nodes serving as DAG committee members. According to the method, the optimistic strategy is adopted to process the cross-fragmentation transaction in parallel, and the cross-fragmentation transaction is decomposed into a plurality of sub-transactions in combination with a DAG structure, so that the atomicity and consistency of the transactions are ensured; in a high-frequency transaction scene, the throughput and the processing rate of the system are remarkably improved, and the transaction delay is effectively reduced; a DAG committee is formed in cross-fragment internal nodes, a DAG account book is maintained, efficient cross-fragment transaction processing is achieved, dependence on a trusted third party is eliminated, and communication and verification cost is remarkably reduced.

The invention discloses a traffic police dynamic gesture recognition method and system based on vehicle-mounted federal learning, and the method enables a key feature extraction model and a preset global model to be combined as a traffic police gesture recognition model shared by vehicle-mounted terminals of a whole network, and facilitates the improvement of the gesture recognition precision of the traffic police. Different vehicle-mounted terminals are organized to train the models based on the collected traffic police gesture image data through vehicle-mounted federal learning, and then each vehicle-mounted terminal only shares model update data instead of original training data, so that the data privacy of vehicle users is protected, the communication cost for uploading mass data to a central server is saved, and the user experience is improved. Meanwhile, real-time traffic police gesture image data dispersed in different vehicle-mounted devices are fully utilized in vehicle-mounted federation learning, training data from vehicle-mounted images have different observation angles, different environment conditions and the like, the diversity of model training data is enriched, and the training efficiency is improved. And thus, the adaptive capacity of the trained traffic police gesture recognition model in a variable traffic environment is enhanced.

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 invention discloses a flashover active feature recognition method for a power transmission line image, and the method comprises the steps: obtaining a historical picture, carrying out the preprocessing of the historical picture, and obtaining an image matrix, and the preprocessing comprises the first preprocessing and the second preprocessing; a sliding window technology is applied to the image matrix, the value of each window is calculated, the maximum value of the mean value of the windows is obtained, first judgment is conducted on the maximum value of the mean value of the windows, and a lightning flashover track is obtained; according to the method, the average brightness value and distribution in the lightning flashover track are extracted, the lightning flashover track is subjected to second judgment, lightning flashover is obtained, operation and maintenance personnel can understand the monitoring result more easily, and therefore operation and maintenance work can be guided more effectively.

The invention discloses a federated learning-based power load prediction system and a training method thereof, the system comprises a server, a client, an encryption and decryption module and a storage and processing module, and the method adopts a federated learning framework to realize distributed training and optimization of a power load prediction model. According to the method, a plurality of clients are allowed to locally train the power load prediction model, the model parameters are sent to the server for aggregation and updating, and the data set is distributed at each client locally, so that the risk that the data is illegally accessed or leaked is reduced, and the user experience is improved. According to the method, the LSTM neural network model is adopted to predict the power load, the characteristics of the time sequence data can be fully utilized, the prediction accuracy is improved, the regularization hyper-parameters are optimized by utilizing the Bayesian optimization algorithm in the prediction process, the performance of the model can be further improved, and the prediction result is more accurate.

The invention belongs to the technical field of federated learning, and more specifically relates to a privacy protection personalized federated learning method based on model hierarchical optimization. The method comprises the following steps: each client obtains a global model of a server side, updates a local model by using a training data set of a self-side, then performs hierarchical optimization on the local model, and determines partial model parameters to be uploaded of each client; partial model parameters to be uploaded by each client are cut, differential privacy noise is introduced, and then the model parameters are sent to a server side; the server side allocates a weight to each client according to the size of the data volume of each client, then aggregates the model parameters uploaded by each client to obtain a new global model, and sends the new global model to each client; and repeating the steps S1 to S3 until a set training round or convergence is achieved. According to the method, the problems of model performance influence and communication efficiency caused by heterogeneity in federal learning are solved while user privacy is guaranteed.

The invention discloses a rapid batch switching authentication method for predictability and aggregation promise proof, which realizes batch authentication of a plurality of user terminals by generating system parameters and secret keys through a trust mechanism and combining predictability promise and aggregation signature technologies. In the authentication process, the edge base station to be switched exchanges with the current communication edge base station through an encrypted message, verification information such as a pre-authentication certificate and a signature is generated through the user terminal, and identity authentication is completed step by step. The method comprises the following specific steps: a trust mechanism generates a private key, a public key, a pseudo identity and an expiration timestamp for each device; through multiple rounds of message exchange, the edge base station and the user terminal perform identity verification by using a shared key and a hash function; and finally carrying out efficient verification on a plurality of users through batch signature and aggregation commitment. According to the method, on the basis of ensuring the security, the speed of switching the authentication process is greatly improved, and the method is suitable for a large-scale network environment, especially in edge computing and mobile communication networks.

The invention discloses a parameter adaptive control method and device for frequency modulation participation of a network construction type wind power plant. The method comprises the following steps: respectively establishing an inertia adaptive curve model and a droop adaptive curve model of each wind turbine generator set; distributing a corresponding cooperative control coefficient for each droop adaptive curve model to obtain a droop adaptive curve cooperative control model of each wind turbine generator set; in the frequency modulation process from power disturbance to maximum frequency deviation of the system, the inertia coefficient and the droop coefficient of each wind turbine generator set are respectively changed according to the corresponding inertia adaptive curve model and the corresponding droop adaptive curve cooperative control model; in the frequency modulation recovery process after the system has the maximum frequency deviation, the inertia coefficient of each wind turbine generator set is maintained at the set minimum inertia coefficient, and the droop coefficient of each wind turbine generator set continues to change according to the corresponding droop adaptive curve cooperative control model; therefore, cooperative control of each wind power set in the wind power plant participating in system frequency modulation is completed.

The invention discloses a one-way data communication method, device and equipment of a micro front end and a storage medium, and belongs to the technical field of research and development frameworks. The method comprises the following steps: firstly, configuring and generating an adaptive sub-application according to a front-end service demand, and packaging a one-way data communication rule for the sub-application; then the sub-applications are loaded into the micro-front-end framework, and it is ensured that the sub-applications can cooperatively work with the main application and other sub-applications. Next, when the sub-applications in the micro-front-end framework need to communicate, the system stores a communication instruction in an asynchronous message queue, and takes out the instruction from the queue in an asynchronous loading mode; and the system identifies the state parameters of the target sub-application in the state pool, analyzes the transmission parameters in the instruction, and updates the state of the target sub-application based on the parameters. The invention further relates to the technical field of block chains, and the state parameters are stored in a block chain network. According to the invention, the pollution problem of the state pool is effectively avoided, the maintainability of the system is enhanced, and the system is convenient to expand and update.

The invention discloses a device and a method for adaptively adapting a cloud platform to a domestic AI accelerator card, belongs to the technical field of cloud computing and artificial intelligence, and aims to solve the technical problems of how to realize rapid integration and deployment of the domestic AI accelerator card by the cloud platform, improve the flexibility and expandability of the cloud platform and further better meet diversified AI computing requirements. According to the technical scheme, the device comprises a script tool set, a resource management module, an adaptation task management module, a data storage module and an adaptation analysis module; wherein the script tool set integrates a CPU and a script tool for identifying, sensing, driving and monitoring physical machine accessories of a domestic AI acceleration card, and is used for automatically realizing hardware identification, driving installation and hardware state monitoring functions according to the characteristics of hardware; and the resource management module is used for realizing platform access of the AI server and management of the domestic AI acceleration card, and monitoring states and loads of the nodes and the AI acceleration card in real time.

The utility model discloses a low-power-consumption intelligent necklace for shepherds based on dual-mode energy supply. The low-power-consumption intelligent necklace for the shepherds based on dual-mode energy supply comprises a necklace body, the necklace body is provided with a detachable core unit and a dual-mode energy supply unit, and the core unit is connected with the dual-mode energy supply unit; the core unit is a TPU shell in which a positioning chip, a movable sensor, a microprocessor and a communication module which are connected with one another are packaged, and the dual-mode energy supply unit comprises a flexible solar cell, a ceramic micro generator and a micro battery.

The invention provides a source code compiling method and device, computer equipment and a readable storage medium, and aims to solve the problems that in components in an existing code bin, due to the fact that modification does not have an independent submission record, information of all component versions is lost, it cannot be determined that each modification is compatible with current modification, and the compatibility between the current modification and the current modification is poor. And the version management of the code library is very disordered. The method comprises the following steps: acquiring identification information corresponding to a plurality of code libraries to be managed; code storage bins and modification information storage areas corresponding to a plurality of code libraries are generated according to the identification information, each code library runs in the code storage bin, and the modification information storage areas are stored in the code storage bin; when first modification information of at least one code library in the code storage bin is received, generating a corresponding binary component according to the first modification information; and adding the binary component to the modification information storage area according to the identification information of the code library corresponding to the first modification information to complete compiling of the source code in the code storage bin.

The invention relates to a communication fault monitoring and passive meter reading system and method based on an electric meter communication module. The method comprises the following steps: step 1, inserting a pin header at an interface of an intelligent electric meter and a communication module, and leading out weak current terminals of the intelligent electric meter and the communication module to an adapter plate; 2, connecting the isolation boosting device 1 to the adapter plate, and supplying power by using an isolation power supply; step 3, switching communication signals RXD and TXD led out from a carrier line communication port to a voltage boosting chip; 4, connecting the boosted communication port to an HPLC carrier communication port and a TTL communication port in parallel; step 5, supplying nominal voltage to the intelligent electric meter, enabling the simulation master station to communicate with the intelligent electric meter, monitoring communication data by a TTL module through a common serial port, and recording the data to a memory; and step 6, performing continuous communication on the carrier module and the TTL module through the communication upper computer, observing a communication result, and performing judgment and data analysis.

The invention discloses a federal learning model fine tuning method and system for heterogeneous data, and belongs to the technical field of artificial intelligence, and the fine tuning method specifically comprises the following steps: S101, each client carries out the local training of a deep learning model of a random initialization parameter until a convergence standard is reached on a local data set; the method can reduce the number of iteration rounds of federal training, reduce the communication cost, enable the local model to be aligned to the fine-tuned global model in the early stage of training, reduce the weight divergence, reduce the constraint on the parameter update in the later stage of training, facilitate jumping out of the local optimal solution, reduce the risk of privacy leakage to a great extent, and improve the privacy protection efficiency. The decision-making layer is usually easily influenced by data isomerism in federal training, and the global decision-making layer is finely adjusted by using feature representatives of each category, so that the generalization performance of a global model is enhanced at very low calculation cost.

The invention discloses a vehicle OTA upgrading method and device, computer equipment and a storage medium. Belongs to the technical field of network resource planning. The method specifically comprises the following steps: acquiring upgrade demand information of a current vehicle over-the-air (OTA) upgrade event initiated by a user side; according to the upgrade demand information, network resource information needing to be occupied by the vehicle OTA upgrade event is determined; according to the network resource information and the historical load condition of the associated network element, the upgrade start time of the current vehicle OTA upgrade event is predicted; and the upgrade starting time is fed back to the user side, so that the user side executes the vehicle OTA upgrade event based on the upgrade starting time. On the basis of the method and the device, the communication cost and the human input between a vehicle enterprise and an operator are effectively reduced, the existing situation that due to the fact that accurate upgrading starting time is lacked, OTA upgrading is mostly selected to be carried out late night, and some vehicles are parked in a basement, upgrading fails is avoided, and the vehicle OTA upgrading efficiency and the service experience of a user are improved.

The invention belongs to the field of vehicle-mounted tire pressure monitoring, and discloses a vehicle-mounted tire pressure monitoring platform and method based on the Internet of Things. Comprising the following steps: collecting comprehensive state data of each tire, and constructing a joint structure chart of the tires; calculating the difference data of the joint structure chart at the current time point and the previous time point to form a difference packet, and transmitting the difference packet by using an improved difference technology; building and training a monitoring and early warning model, and importing the difference packet into the trained monitoring and early warning model to obtain early warning information; and the early warning information is displayed through a vehicle-mounted interface, and the vehicle is adjusted, so that intelligent, high-efficiency and high-quality monitoring of the vehicle-mounted tire pressure based on the Internet of Things is realized.

The application provides a traffic classification method and system based on federal semi-supervised learning. The method comprises the following steps: constructing an unlabeled traffic dataset and a labeled traffic dataset; a center server decomposes a global model into supervised learning parameters and unsupervised learning parameters and initializes the parameters; the parameters and an auxiliary agent are sent to each client; each client performs unsupervised training based on the supervised learning parameters, the unsupervised learning parameters and the auxiliary agent by using a local unlabeled traffic dataset, uploads the difference between the unsupervised learning parameters to the center server; the center server aggregates and updates each unsupervised learning parameter; supervised training is performed by using a local labeled traffic dataset, and the difference between the supervised learning parameters and the difference between the unsupervised learning parameters are sent to each client; a new auxiliary agent is obtained based on nearest neighbor search, and the new auxiliary agent is sent to each client when a set sending condition is met; the two steps are iteratively executed until a stop condition is met.

The application provides a JPA-based web code automatic generation system, comprising a JPA client, a code generation server, a Java backend service module and a Vue frontend service module; the system combines JPA technology with template technology, and can generate database tables and the front end and the back end of a web application system in batches according to JPA classes at one time, thereby greatly improving development efficiency. When a user adjusts requirements, the properties of the classes are modified, and one-key generation is very convenient.

This application relates to a device control method, apparatus, digital signal processor, and electronic device. By acquiring a wake-up signal, the digital signal processor is controlled to exit a sleep state; a Bluetooth connection is established with an external terminal device via a Bluetooth module within the digital signal processor; a control signal is acquired through the Bluetooth connection, and a target functional module within the digital signal processor or electronic device is controlled to respond based on the control signal. This technical solution integrates Bluetooth functionality into the digital signal processor of the electronic device, enabling communication between the electronic device and the terminal device without requiring an additional Bluetooth chip, thereby reducing the communication cost between the electronic device and the terminal device.

The invention discloses quantum encryption communication of a low-voltage transformer area based on a quantum security service platform. The quantum encryption communication comprises the steps that S1, a local communication bin of side equipment C sends network access authentication data to a main control board; s2, the end equipment S carries out network access authentication; s3, the edge device C obtains an initial secret key K; s4, the end equipment S obtains an initial secret key K; s5, session key distribution is carried out between the side equipment C and the end equipment S; s6, the side equipment C sends downlink data to the end equipment S; and S7, the end equipment S sends uplink data to the side equipment C. According to the invention, the problems of poor communication security and low key use efficiency of the low-voltage transformer area in the prior art can be solved.

The present invention relates to a method and device for predicting customer repurchase based on an RF-LightGBM fusion model in a non-contract scenario. The method includes: obtaining historical data of users, preprocessing and feature engineering the data; using the preprocessed data as samples, balancing the sample set using the SMOTE-ENN method; optimizing the random forest algorithm and the LightGBM algorithm through the TPE optimization algorithm to perform hyperparameter optimization, and constructing a weak classifier; using each weak classifier to perform ensemble learning on the training samples to obtain a strong classifier, and obtaining the final result of repurchase prediction. The present invention analyzes the consumption data of the company's purchased customers, accurately predicts the repurchase behavior of existing customers, guides customer relationship management decisions and precision marketing strategies based on the repurchase behavior, improves marketing conversion rates, and reduces corporate operating costs.

The invention discloses a power system economic dispatching method based on a distributed hybrid differential evolution algorithm, and belongs to the technical field of power systems. In order to solve the problems that a distributed gradient descent algorithm is easy to fall into a local optimal solution and a centralized network framework can cause the problem that the computing power is limited when a large-scale optimization problem is processed, an initial population is randomly generated and is decomposed according to the dimension of the problem, and each sub-population comprises generator information of a single dimension. The distributed gradient descent algorithm is introduced before mutation operation is executed, so that the power generation cost of population individuals is reduced, the search efficiency of the algorithm is effectively improved, and meanwhile, the problem that the distributed gradient descent algorithm is prone to falling into a local optimal solution is solved. A preset time average consistency algorithm is introduced in a selection link, so that the fitness of different generators is averagely consistent within preset time. The method is distributed, the communication cost is saved, the robustness of the network is improved, and the search efficiency of the differential evolution algorithm is improved.

The invention provides a reliability test system and test method of a cabin controller and the cabin controller, the reliability test system comprises a vibration platform, a host and the cabin controller, the cabin controller comprises a board-to-board connecting piece, a signal generation module, a signal receiving module and a judgment processing module; after the cabin controller is fixed on the vibration platform, the host sends a detection instruction to the cabin controller and sends a starting instruction to the vibration platform; in response to a starting instruction, the vibration platform starts to vibrate, the signal generation module generates an initial voltage signal within a preset time period and sends the initial voltage signal to the input end of the board-to-board connecting piece, and the signal receiving module obtains a real-time voltage signal from the output end of the board-to-board connecting piece; the judgment processing module generates a detection result according to the initial voltage signal and the real-time voltage signal and stores the detection result; and the host accesses the cabin controller after a preset time period to obtain a detection result, and judges whether the cabin controller passes the reliability test according to the detection result.

A virtual power plant distributed regulation and control method based on communication review is characterized by comprising the following steps: step 1, generating a distributed regulation and control model of node power in a virtual power plant and between virtual power plants by using a direct current power flow constraint problem; 2, solving the distributed regulation and control model based on an alternating direction multiplier method, and introducing a local variable of the virtual power plant, a local intermediate variable of the virtual power plant and a global variable of the virtual power plant; 3, presetting a communication review mechanism, updating the variables in the iteration process, and judging whether the iteration meets the communication requirement or not according to the communication review mechanism; and step 4, performing communication on the variables meeting the communication requirements, and performing distributed regulation and control of the virtual power plant based on a communication result. According to the method, the operation cost of virtual power plant communication scheduling is minimum, and the economic benefit is maximized.

This specification provides a method and system for multi-party collaborative data filtering based on privacy protection. In the filtering method, collaborating parties decompose a total filtering condition into m sub-filtering conditions, where each sub-filtering condition defines the value of a feature. The collaborating parties then send each of the m sub-filtering conditions to the participating parties that possess the feature defined by that sub-filtering condition. Next, each participating party receiving a sub-filtering condition determines its corresponding logical result set for the shared sample set and provides this logical result set to the collaborating parties. Finally, the collaborating parties aggregate the logical result sets returned by the participating parties and, based on the aggregated result, determine the filtering result for the shared sample set and provide this filtering result to the n participating parties.