152 results about "Network level" patented technology

The invention discloses an object model access parameter dynamic optimization method and system based on reinforcement learning. According to the method, a multi-agent reinforcement learning model is constructed, a device-level state and a network-level state of the Internet of Things terminal device are used as environment state input, access parameters of the terminal device are modeled as an action decision of a reinforcement learning decision model, and the reinforcement learning decision model outputs optimized access parameters according to the current environment state. And continuously updating the strategy network through a reward function. And finally, the control platform deploys the trained strategy model, automatically selects the optimal access parameter configuration when the terminal equipment accesses, and adjusts the access strategy in real time according to the state change of the equipment and the network. Compared with an existing static strategy, the method has the advantages that efficient, low-time-delay and low-power-consumption access of the power internet-of-things terminal equipment can be achieved, and the system resource utilization rate and the communication reliability are remarkably improved.

The invention provides a low-power-consumption cooperative sewer water body integrated monitoring method. The method comprises the following steps: carrying out model selection and deployment on sensing node hardware of an adaptive scene; a low-power-consumption internet of things network topology and power supply design; a time-sharing triggering acquisition strategy of the multi-modal data is carried out; node-level environment self-adaptive data calibration is carried out; node-level flow velocity-stink association anomaly recognition is carried out; carrying out network-level data fusion and validity verification; collaborative traceability, diffusion simulation and linkage control are carried out. According to the invention, the problems of data distortion, difficulty in source tracing of pollution sources, short equipment endurance and response lag caused by independent collection of flow and stink and environmental interference in traditional sewer monitoring are solved; the integrated effect of ultrasonic Doppler and electronic nose cooperative monitoring, environment self-adaption accurate calibration, odor source rapid positioning and diffusion simulation and low-power-consumption and long-endurance linkage control is achieved.

A Smart Service Analyzer obtains Call Direct Record (CDR) Data from Probing Devices of a mobile network. The CDR Data is processed at a Key Performance Indicators (KPI) Generator to generate User Level KPIs. The User Level KPIs are provided to a Customer Experience Index (CEI) Estimator. Machine Learning is applied to the User Level KPIs at the CEI Estimator to generate Generalized User Level CEI Estimates. The Generalized User Level CEI Estimates are provided to a Service Quality Index (SQI) Estimator. Machine Learning is applied to the Generalized User Level CEI Estimates at the SQI Estimator to generate Network Level SQI Estimates.

The invention provides a video high-definition coding and low-delay transmission method for an IPTV (Internet Protocol Television) large-screen short play, which relates to the technical field of low-delay transmission, and comprises the following steps: defining an end-to-end measurement boundary and a time reference, configuring an acquisition mechanism of a network environment real-time state index, extracting high-definition coding metadata by adopting a coding end, and transmitting the extracted high-definition coding metadata to a transmission end; and timestamp information is marked and a unified structured operation state object is configured. A short window prediction path and a middle window prediction path are configured, confidence interval correction and lower limit fusion are combined to form a conservative available bandwidth estimation value, and when a network level label is judged to be a weak network level, a protection buffer mode and a low code rate pre-pulling strategy are triggered. The mechanism effectively solves the problems of frequent playing lagging and overhigh switching delay in a weak network environment in the prior art, so that a short play can still keep stable and continuous playing experience in a complex network environment.

The invention relates to the technical field of industrial internet network security, and discloses an industrial internet network security situation awareness analysis method and system, and the method comprises the steps: carrying out the deep comparison analysis of a network instruction, an actual production process state, equipment operation conditions, material characteristics, final product quality requirements, and other information; instructions which seem legal in the network level but have potential destructiveness in the process level can be identified. For example, for a hidden attack which slowly destroys the production process through small parameter tampering or suboptimal path selection, the method can find and give an alarm in time through comparison of a simulation result and a safety threshold, thereby avoiding false alarm and a safety blind area caused by a fixed reference behavior mode of a traditional system. Therefore, the accuracy and effectiveness of industrial internet network security situation awareness are remarkably improved, and the method is particularly suitable for complex industrial environments such as high-precision aviation component manufacturing and the like which have extremely high requirements on the production process and the product quality.

Systems and methods are disclosed that relate to Key Performance Indicator (KPI)-driven alarm threshold optimization using machine learning. In one embodiment, a computer-implemented method comprises obtaining network-level KPI data for a network and radio log data for one or more radio systems in the network. The KPI data comprises KPI values for one or more network-level KPIs, and the radio log data comprises a number of faulty or uncalibrated antenna branches in the radio system and cell or user beamforming weights. The method further comprises pre-processing the KPI data and the radio log data, labeling the pre-processed KPI data as degraded or non-degraded, and training, with the labeled KPI data and the pre-processed radio log data, a fault classifier Machine Learning (ML) model to output a value(s) that represent a probability that the KPI(s) will be degraded for a given input feature set.

A method and system are provided at a local network level and a remote cloud level. The local network level includes compressors and a controller configured to create a short-term schedule having a short-term switching sequence for operation of the compressors, perform a validation assessment on the short-term schedule, and send the short-term schedule to a platform embedded with the main controller of the system. A prediction model, a long-term scheduling approach, and cloud storage that stores measurements from the compressors and executable instructions are provided at the remote cloud level. The long-term schedule is transferred to a safety check module at the local network level. Following validation, the embedded platform refines the long-term switching sequence and the safety check module allows the system to implement the long-term schedule, having the short-term schedule as a backup option.

Approaches are described for providing quality of experience (QoE) driven in-transport communications services. For example, in-transit communications services are provided over a network to content consumption devices on one or more transport craft. During a measurement window, delivery of the in-transit communications services can be measured to obtain network-level delivery (NLD) measurements, and consumption of the in-transit communications services by one or more of the content consumption devices can be measured to obtain customer-level consumption (CLC) measurements. The NLD and CLC measurements can be used to compute a QoE score that indicates a delivered quality of service as perceived by one or more passengers of the one or more transport craft. The QoE score can be used to update stored service level data. The update can cause generation of one or more service level triggers, which can trigger automatic remedial action to address QoE-related conditions.

The invention belongs to the technical field of computer vision and artificial intelligence, and particularly relates to a semantic image segmentation method based on hierarchical neural architecture search, which is characterized in that efficient architecture optimization is realized by adopting a differentiable search algorithm by constructing a double-layer search space comprising a unit level and a network level. 12 kinds of heterogeneous operations are designed in a unit-level search space, including a multi-scale context extractor and boundary sensing convolution; and a resolution self-adaptive adjustment mechanism is realized in a network-level search space. Through a two-stage gradient optimization strategy, search is completed within three GPU days, the decoded model reaches 80.4% mIoU on a Cityscapes data set, and the parameter quantity is only 10.15 M. Compared with the prior art, the method has the advantages that the segmentation precision and the model efficiency are remarkably improved while efficient search is kept, and the method can be widely applied to the fields of automatic driving, medical image analysis and the like.

The invention discloses an urban rail transit network level signal professional data management and interface method, and relates to the technical field of urban rail transit signal professional data management. A data resource pool platform is established between an online network signal professional communication front-end processor and an external interface; the data resource pool platform comprises a data access module, a data transmission module, a data storage module, a data subscription module and a data service module; the data access module is used for acquiring signal data from a network signal professional communication front-end processor, the data transmission module establishes labels for the signal data and performs classified storage, the data storage module performs standardized processing on the signal data and then stores the signal data, and the data subscription module manages an external interface and performs external interface subscription content configuration. And the data service module performs identity verification on the external interface and provides data. According to the method, the intelligence level and decision-making efficiency of line network operation are remarkably improved, and a core technical guarantee is provided for safe and efficient operation of urban rail transit.

Implementations for chiplet fault management through hierarchical error reporting and coordinated recovery in an accelerator interconnect. In some implementations, a system for a UALink-based network comprises a chiplet die coupled to an accelerator die, a management agent on the accelerator die, and a centralized controller. Upon fault detection, the chiplet die transmits a fault notification to the accelerator die, the management agent reports the fault to the centralized controller, and the centralized controller determines a recovery action based on the fault identifier and severity. In some implementations, the accelerator die reports the fault through dual paths, to a host processor via CXL or PCIe error reporting and to the centralized controller via the management agent, and the host processor and the centralized controller coordinate recovery, with the host managing workload-level recovery and the centralized controller managing network-level recovery.

The application discloses an APT flow detection method based on a knowledge graph, which comprises the following steps: analyzing network data flow characteristics in network flow, establishing a knowledge graph of network data flow, and classifying the established network flow knowledge graph by using a graph neural network algorithm to detect APT attack flow in the network, so that the attack behavior of an APT organization can be detected from the network level. The application can detect unknown APT attack network data flow, and is helpful for flexibly and accurately detecting APT flow.

The invention discloses a distributed system and method for providing end-to-end deterministic time delay guarantee, a terminal and a storage medium. The system comprises a super object model time sequence extension module, a global time synchronization module and a distributed deterministic scheduling module. The super object model time sequence extension module is used for defining distributed tasks or equipment capabilities in the system, and the global time synchronization module is used for synchronizing clocks of whole network equipment by using a high-precision synchronization protocol to obtain a unified time reference; the distributed deterministic scheduling module is used for planning a scheduling period of a hard real-time task based on an end-to-end time sequence requirement and a time reference, calculating an end-to-end communication path meeting the time sequence requirement of the hard real-time task according to a real-time network state and the scheduling period, and scheduling the hard real-time task based on the time reference. According to the invention, resource reservation is carried out on an end-to-end communication path on a network level and an operating system kernel level, a service layer time sequence demand is associated with bottom layer network scheduling, and end-to-end microsecond-level deterministic guarantee is realized.

The invention provides an energy consumption data intelligent calling method and device, and the method comprises the following steps: S1, initializing a collection period parameter, and obtaining the collection tasks of all devices from a database; s2, initializing an acquisition data link according to the acquisition task; s3, analyzing the received data file, and updating the data file to the collection data chain according to the data content; and S4, regularly counting the data of the collection data chain to generate a call item, updating the call item to a call data chain, further generating a call message, and sending the call message to the front collection equipment for data call collection. According to the invention, the problems that the urban rail transit network level energy management system is lack of data due to unstable communication, equipment failure and the like and cannot carry out data compensation and calling in time are solved, 24-hour intelligent compensation and calling of energy consumption data all day long are realized, the data integrity of the system is guaranteed, and the working efficiency of operators is improved.

The present application relates to a kind of metro network level engineering vehicle protection system data acquisition method, equipment and medium, the method includes: signal interlocking system acquisition equipment information and send to ATS system, signal vehicle-mounted system sends train operation information to ATS system;ATS system sends equipment information and train operation information to network command center, while uploading the line number of this line;The interface server of network command center divides the received information to the processing thread of each line, and then the information output to network gate server by processing thread upload;Network gate server packs the information received to data pool and ferry to management network domain;Request information from network gate server, according to line, information is distinguished different thread processing, according to line, port number is sent to engineering vehicle interface front-end computer;Engineering vehicle interface front-end computer obtains information, and forwards information to engineering vehicle system.Compared with prior art, the present application has improved engineering vehicle operation efficiency and other advantages.

The application relates to a light wavelet enhancement fusion medical image segmentation method, and belongs to the technical field of medical image processing and computer-aided diagnosis. The core of the method is to construct a light WEF-Net network, which comprises an encoder, a bottleneck layer and a decoder. The encoder adopts a dual-domain perception module, extracts complementary features from the frequency domain and the time domain through wavelet transformation and convolution, and simultaneously extracts complementary features from the frequency domain and the time domain. The bottleneck layer is designed with a feature aggregation Kolmogorov-Arnold transformation module, which is used for efficiently fusing multi-scale semantic information and enhancing the nonlinear modeling capability. The network level also integrates a sawtooth rolling feature fusion module, which enhances the global continuity of the features through the channel rolling and spatial scanning mechanism to improve the integrity of the boundary segmentation. Experiments show that the method realizes excellent performance on multiple public medical image datasets, while maintaining low parameter quantity and low computational quantity, and significantly improves the segmentation precision.

The invention relates to the technical field of industrial Internet of Things communication, in particular to an AUTBUS node device and system with a CN-TN main and standby redundancy function. The device comprises a device main framework, a main CN module, a standby CN module, a high-speed internal communication bus and a local redundancy management module, the main CN module and the standby CN module are independent pluggable hardware units, and the main CN module synchronizes core state data to the standby CN module through a high-speed internal communication bus; the local redundancy management module is integrated on the device main framework, is logically independent of an AUTBUS bus protocol stack, and is used for monitoring an internal fault of the main CN module and triggering a local takeover process when monitoring that the main CN module has the internal fault; according to the method and the device, rapid seamless switching is realized on the premise of not triggering network-level re-election, so that the reliability of the AUTBUS node device is improved, and the continuity and the certainty of the time sensitive characteristic of the AUTBUS system are guaranteed.

The application discloses an adaptive industrial surface defect segmentation network establishment method and application thereof, belongs to the field of defect segmentation, and comprises the following steps: based on neural architecture search, combining prior knowledge of an existing manually designed defect detection architecture, designing a potential segmentation network search space which is simple but suitable for an industrial scene, and completing network-level search space design; summarizing common operations in the industrial detection architecture, additionally adding channel attention operations and spatial attention operations as candidate operations, constructing basic units in the network, and completing unit-level search space design; since the overall network architecture is relatively fixed, the final search range is simplified to the structure of the unit instead of the entire network, so that the architecture construction with good performance is completed in a shorter time under the condition of limited time resources, and the network design cost is greatly saved. On this basis, the search strategy is further optimized, and the search efficiency can be further improved.

The present application discloses a method and device for responding to voice commands, a storage medium, and an electronic device, and relates to the field of smart home technology. The method for responding to voice commands includes: when receiving voice control information, sending the voice control information to a set voice server for voice recognition, and determining a waiting time according to a network level of current network quality; if a first voice recognition control instruction fed back by the set voice server is received within the waiting time, responding to the first voice recognition control instruction; if the first voice recognition control instruction fed back by the set voice server is not received within the waiting time, recognizing the voice control information based on a preset voice recognition model, obtaining and responding to a second voice recognition control instruction.

Various embodiments of the present disclosure provide a network linking technique that improves the functionality of a computer in various aspects. The techniques comprise receiving a member-specific assertion request encapsulating an assertion request from a user of the first member platform; determining a local construct of a cross-platform data object; determining, using the local construct of the cross-platform data object, a second member platform; providing an internal member-specific assertion request to an internal service associated with the second member platform; providing, via the internal service, an incoming member-specific assertion request to the second member platform; responsive to receiving a user input at the second member platform, receiving an assertion response at the internal service and from the second member platform; and providing the assertion response to the first member platform.

A network level adaptive based lightweight feature map compression method, in the forward inference process of a convolutional neural network, when the intermediate feature map output by the activation function needs to be exchanged between the calculation unit and the off-chip memory due to the limited on-chip storage resource, the intermediate feature map is first compressed, that is, the network shallow layer feature map is subjected to transform domain processing by using integer discrete cosine transform, the network deep layer feature map is subjected to block maximum value extraction processing, then unified quantization processing and entropy coding processing are performed, the compressed feature map is written into the off-chip memory, so as to reduce the data size, thereby reducing the off-chip storage access bandwidth requirement and data carrying energy consumption. The present application is a feature map compression method which does not depend on neural network retraining, can fully utilize the spatial correlation of shallow layer feature map, has low calculation complexity and good hardware implementation characteristics, so as to significantly reduce the storage and energy consumption overhead in the inference process of the convolutional neural network under the premise of ensuring the inference accuracy.

Various embodiments of the present disclosure provide a network linking technique that improves the functionality of a computer in various aspects. The techniques comprise responsive to an exchange request, receiving, originating from a member platform within the member computing ecosystem, a linking code request comprising a member-specific identifier corresponding to the member platform; providing, to an internal service, the linking code request to generate an ephemeral transport object storing a linking code and a link reference corresponding to the member-specific identifier and the linking code; providing the linking code to the member platform; receiving, originating from the another member platform within the member computing ecosystem, a linking request comprising the linking code; and authorizing, using the second member platform, the exchange request based on the linking code.

The invention discloses an AI service strategy updating method and device, a functional entity, a medium and a product, and relates to the technical field of communication, and the method applied to a first functional entity specifically comprises the steps: determining the execution capability information of the first functional entity based on the network state information of the first functional entity, when it is detected that execution capability information does not meet a preset capability index, a first service strategy adopted for executing the first AI task is updated, and the execution capability information is used for representing the capability of the first functional entity for executing the first AI task; and / or, receiving a first policy adjustment request sent by a second functional entity, and updating the first service policy based on the first policy adjustment request, the second functional entity and the first functional entity being functional entities deployed at the same network level; the first service strategy comprises at least one of a connection control strategy, an AI model selection strategy, a radio bearer DBR configuration strategy and a resource configuration strategy of the first functional entity.

The application belongs to the technical field of launch vehicle electronic information system, and particularly relates to a launch vehicle communication network system based on Powerlink. The system comprises a two-stage star network of a launch vehicle, an inter-stage star network, a first-stage tail star network and a main bus adopting a Powerlink bus; a plurality of terminal slave nodes in the two-stage section access a two-stage network repeater through a star topology communication link, a master node and the two-stage network repeater are connected in series to the main bus through a two-stage interface; a plurality of terminal slave nodes in the inter-stage section access an inter-stage network repeater through a star topology communication link, a redundant master node and the inter-stage network repeater are connected in series to the main bus through an inter-stage interface; a plurality of terminal slave nodes in the first-stage tail section access a first-stage tail network repeater through a star topology communication link, and are connected to the main bus through a first-stage tail interface. The system has high real-time performance and high reliability.

The embodiment of the invention provides a network link reliability monitoring method and device, a medium and a program product, and relates to the technical field of network monitoring. The method comprises the following steps: acquiring network index sampling data of a plurality of network levels of a target network system in real time; determining a self-adaptively adjusted sliding window according to the network entropy calculated in real time; converting the network index sampling data into corresponding network index sequence data based on a sliding window; processing the network index sequence data by using a pre-trained anomaly score model to obtain a corresponding anomaly score; and determining an anomaly monitoring result of the target network system based on the anomaly score and a preset anomaly score threshold. According to the embodiment of the invention, by collecting the index data of the plurality of network levels and converting the index data into the sequence data based on the self-adaptive adjustment sliding window, the cross-layer association analysis of the network link can be realized, so that the accuracy and timeliness of network anomaly monitoring are effectively improved.

The invention provides a rail transit heterogeneous network seamless roaming method and system based on multi-level switching, and relates to the technical field of rail transit network communication, and the method comprises the steps: constructing a space-time transfer matrix to predict a motion track, carrying out the level division of a candidate access network, and calculating an optimal switching path; and calculating a resource reservation time length and reserving resources in advance in combination with the prediction trajectory and the network coverage range, and setting a sampling point to determine a switching trigger point to complete network switching. According to the invention, smooth transition between network hierarchies can be realized, the seamless roaming success rate is improved, and switching delay and service interruption are reduced.

Various embodiments of the present disclosure provide techniques for modifying a value of a credential-less exchange in near real time using a plurality of identifier mappings, member interfaces, and exchange modifiers. The techniques may include receiving an exchange request for executing a value-based exchange that is indicative of a universally unique ephemeral key (UUEK) and one or more request attributes. The techniques include identifying an exchange data object based on the UUEK and identifying a user exchange modifier for the exchange request based on the exchange data object and the request attributes. The techniques include generating a modified exchange value for the exchange request based on the user exchange modifier and providing an exchange authorization request to a member platform with the modified exchange value.

Various embodiments of the present disclosure provide a network linking technique that improves the functionality of a computer in various aspects. The techniques comprise receiving an exchange request comprising a universally unique ephemeral key (UUEK); identifying a local data construct of a cross-platform data object based on the UUEK; providing a value request to a internal service corresponding to a member platform using the cross-platform data object; receiving, based on a contribution ruleset corresponding to the member platform, an updated exchange request from the internal service; providing an adjusted authorization request to another member platform based on the updated exchange request; receiving, from the other member platform, an authorization response based on the adjusted authorization request; and providing the authorization response to the terminal.

The embodiment of the invention provides a task scheduling method and device, management equipment and a computing equipment cluster, and relates to the technical field of cloud computing, the task method is applied to the computing cluster, the computing cluster comprises different network levels, and each network level comprises one or more network performance domains. Computing resources with the same network performance are distributed in the same network performance domain, and the method comprises the steps that the computing resource demand quantity of each task in a plurality of tasks and computing resource position constraints of one or more tasks in the plurality of tasks are obtained; according to the computing resource demand quantity of each task and the computing resource position constraint of one or more tasks, respectively dividing one or more affinity groups from computing resources of each network performance domain under each network level; scheduling of a plurality of tasks is performed in one or more affinity groups of each network performance domain under each network level. The method and the device can be compatible with task scheduling of computing clusters in different network forms.

Various embodiments of the present disclosure provide a network linking technique that improves the functionality of a computer in various aspects. The techniques comprise receiving an exchange request comprising a universally unique ephemeral key (UUEK); identifying a local data construct of a cross-platform data object based on the UUEK; providing a value request to a internal service corresponding to a member platform using the cross-platform data object; receiving, based on a contribution ruleset corresponding to the member platform, an updated exchange request from the internal service; providing an adjusted authorization request to another member platform based on the updated exchange request; receiving, from the other member platform, an authorization response based on the adjusted authorization request; and providing the authorization response to the terminal.