55 results about "Global system" patented technology

The invention discloses a kernel-level heterogeneous collaborative AI reasoning method and device, equipment and a storage medium, and relates to the field of AI reasoning acceleration. When the KDS receives a reasoning request submitted by a user space, generating a cross-device plan list according to a global system resource state and issuing the cross-device plan list to the HCEE; the HCEE drives the corresponding target execution equipment according to the task execution sequence, reads the data information from the UCMP, or restores intermediate / result data reasoned and output by the target execution equipment; after reasoning of the plan list is completed, the kernel space triggers an event notification mechanism, and a user process of the user space is awakened; and the user process directly reads the reasoning result from the UCMP. According to the scheme, deep collaboration of task scheduling, memory management and equipment execution is realized on a Linux kernel level, the problems of high overhead, resource islands, memory redundant copy, extensive scheduling granularity and the like of a user mode reasoning framework system are solved, and end-to-end low-delay, high-throughput and high-energy-efficiency AI reasoning acceleration is realized.

The application relates to the field of computing models and industrial safety data processing under new-generation information technology, and discloses a chemical fire and explosion coupling consequence prediction method and system based on a large model, which has the technical scheme as follows: physical mapping of a multi-source heterogeneous environment and construction of a graph node digital twin, evolution calculation of a fluid mechanics source term and triggering of a disaster critical point, deduction of a physical large model thermal radiation field in a restricted environment, analysis and quantification of a "thermal-power" dual coupling tensor of a target device and buckling failure, deduction of a space-time graph attention network with strong intervention of a physical prior probability, tracing of global system vulnerability and closed-loop blocking control of a physical execution end; by using intelligent calculation technologies such as a physical information neural network and a space-time graph attention network, a thermodynamic partial differential equation and a material thermal-power coupling failure criterion are forcedly embedded into a graph deduction framework as a physical prior bias, so that complex disaster consequence prediction is accelerated from "hour level" to "second level" in the traditional CFD.

The application relates to the technical field of digital security management of assets, and discloses a digital security management system and method, which extracts to-be-scanned asset identification and scanning strategies by analyzing a digital security scanning task, considers network bandwidth and scanning result data warehousing queues of a scanning engine performing scanning, combines asset risk attributes, scanning features and system resource constraints, generates a digital security management strategy, and further controls scanning and data warehousing actions performed by the scanning engine. Thus, the application realizes accurate matching of scanning frequency and asset risk through a dynamic risk assessment mechanism, avoids the complexity and subjective bias of traditional manual strategy setting, comprehensively considers global system resource restrictions such as scanning engine load, network bandwidth and data warehousing queues, optimizes scanning task scheduling, completely solves the resource competition problem that is prone to occurring in traditional scanning, avoids system performance bottlenecks, and significantly improves scanning efficiency and the timeliness of security assessment.

The invention provides a dynamic self-adaptive DNN collaborative reasoning method supporting a backspacing mechanism, relates to the technical field of image processing, and designs an elastic two-way task flow from equipment to edge to equipment, so that the elastic two-way task flow becomes an intelligent scheduling decision point. And when the edge server is overloaded, the task can be actively returned to the source mobile equipment to be executed, so that the bottleneck of queuing delay caused by edge resource competition is fundamentally avoided. A rollback decision is introduced, so that an unloading strategy is upgraded from'static 'or'semi-static' to real'dynamic self-adaption ', the change of edge resources can be directly responded, and a decision target is expanded from optimizing a single task to optimizing a global system load. Three key dimensions, namely a backspacing decision, resource allocation and a model segmentation point, are arranged in a unified framework for joint optimization. Through an integrated reinforcement learning model, the three decisions are output at the same time, so that the three decisions can cooperate with each other and have a synergistic effect.

The invention discloses a power grid hybrid network performance optimization model considering construction cost, and belongs to the technical field of intelligent power grid communication. According to the model, a parameter system including communication equipment, terrain and a global system is defined, a function with the purpose of minimizing the total cost of a life cycle and multivariate constraint conditions of transmission distance, retransmission times, delay, reliability and flow velocity are constructed, and optimal equipment and transfer device configuration is solved by adopting a node traversal and iterative optimization algorithm. Cost and performance collaborative optimization of a honeycomb, Wi-Fi and ZigBee heterogeneous hybrid communication network under a complex terrain is realized, planning accuracy and reliability are remarkably improved, and construction, operation and maintenance cost is reduced.

The objective of the proposed invention is to make it possible to mount the multifunctional mobile module onto the roof of any vehicle to turn it into an autonomous vehicle by providing all the functions of an autopilot system within a single module, ensuring the operability of the module within a wide range of ambient temperatures (−40° C. to +50° C.), as well as within a wide range of weather conditions (rain, downpour, snow) by means of a thermal management system (TMS) and sensor cleaning systems that are integrated into the body of the module.The multifunctional mobile module for automatic vehicle control comprises an underframe made of a metal profile and equipped with metal brackets that hold screw-connected sensors (cameras, a lidar, an inertial navigation system, a Global Navigation Satellite System (GNSS) receiver), equipment (calculators, fans, heaters, network switches, a peripherals controller, GSM (Global System for Mobile Communications) antennas, sensor cleaning system components), and an outer plating that is also mounted onto the underframe.

The invention relates to the technical field of railway wireless communication, in particular to a GSM-R (Global System for Mobile Communications-Railway) vehicle-mounted front-end wireless interference test system and a method thereof. Comprising a high-resolution spectrum acquisition unit; an integrated multi-dimensional test unit; a dynamic scene adaptation unit; and an intelligent interference diagnosis unit. According to the invention, multipath channel modeling and Doppler frequency offset estimation in a vehicle-mounted high-speed mobile environment are completed through a dynamic scene adaptation unit, and scene prior information is generated; and the intelligent interference diagnosis unit constructs a vehicle-mounted adaptive multipath frequency offset estimation and interference diagnosis model according to the multi-path frequency offset estimation and interference diagnosis model, and realizes full-dimensional interference feature extraction. According to the method, sudden interference is detected by adopting double modes of known interference accurate identification and unknown interference clustering, internal and external source interference is decoupled and traced through feature analysis and hypothesis testing, the criticality is quantified, and a disposal instruction is generated. Meanwhile, based on a diagnosis-collection closed-loop scheduling mechanism, collection parameters are adjusted in a targeted mode, an interference model is iteratively optimized, and self-adaptive adaptation of the system to a vehicle-mounted dynamic scene is achieved.

The utility model discloses a touch and press dual-purpose type GSM digital mobile phone modularization detachable button comprising a GSM digital mobile phone and a keyboard assembly, a touch screen is arranged above the front end of the GSM digital mobile phone, the top of the GSM digital mobile phone is provided with a loudspeaker, the bottom of the front end of the GSM digital mobile phone is provided with a keyboard groove, and the keyboard groove is provided with a touch screen. A keyboard assembly used for the replaceable numeric keyboard is installed in a keyboard groove, the keyboard assembly comprises a maintaining shell, fixing screws, buttons, a circuit board and metal contacts, the four corners of the maintaining shell are in threaded connection with the fixing screws, the buttons are arranged in the maintaining shell, the tail ends of the buttons are electrically connected with the circuit board, and the metal contacts are electrically connected with the circuit board. A metal contact is arranged on the left side of the circuit board, and the circuit board is electrically connected with a GSM digital mobile phone through the circuit board. According to the modularized detachable keys of the touch-press dual-purpose GSM digital mobile phone, the key replacement time can be shortened, the key replacement difficulty can be reduced, and meanwhile the key protection capability can be improved.

The invention discloses a global digital authentication terminal device supporting an eSIM, and the device comprises a password security chip which is used for executing global digital identity authentication, credit evidence storage and secure communication control, and a security module, an SM9 identification password algorithm module, an electronic seal engine, a multi-platform digital certificate management unit and a trusted execution environment are built in the password security chip; the eSIM module conforms to the standard of the global system for mobile communication (SGP). 22 and supports the switching of configuration files of a multi-operator subscriber identity module; wherein the eSIM module and the password security chip cooperatively operate, when a user initiates a security service, the password security chip generates a dynamic digital identity identifier based on an SM9 identifier cryptographic algorithm, and binds the dynamic digital identity identifier with an international mobile subscriber identity code in an eSIM configuration file; therefore, two-factor authentication of the communication identity and the digital identity is realized. The terminal device has the advantages of strong authentication and high compliance.

The present application belongs to the technical field of exploration and engineering surveying, and specifically discloses a tunneling machine vibration coupling navigation self-calibration method, which comprises: based on a global system reference beat, feeding forward compensating the physical transmission delay of a heterogeneous bus and the inherent hardware latch delay of an analog-to-digital converter, controlling a micro-electro-mechanical system accelerometer to collect and output a clean vibration acceleration sequence; performing window segmentation on the clean vibration acceleration sequence to extract time-frequency domain features, and synchronously extracting the thermal flow gradient change rate features of the inertial measurement component kernel, and cascading and combining them into a holographic intrinsic state descriptor; the present application aims to solve the problems in the prior art, such as low positioning accuracy and poor stability of the combined navigation system, caused by the problems of the tunneling machine under vibration and impact working conditions, such as non-synchronization of heterogeneous sensor data, vibration rectification error and thermal hysteresis effect caused thereby, model misalignment caused by sudden geological changes, integral error accumulation, and decline of long-period operation model generalization ability.

The invention provides an industrial Internet of Things system based on a vehicle lamp mold and an execution method, and the system comprises a data transmission layer, an equipment layer, and a system application layer, and the data transmission layer is used for achieving data collection, data transmission and data safety; the equipment layer is used for monitoring the mold and interacting with the data transmission layer; and the system application layer is used for realizing full-life-cycle management of a message queue center and equipment. And full-link data closed-loop management is realized through cooperation of the Internet of Things equipment and the cloud. According to the invention, the overall system crash is avoided through modularized independent upgrading and expansion capability; heterogeneous technology stacks are compatible and flexibly adapted; a decentralized high-availability architecture reduces the failover time; and data security and privacy isolation are enhanced, so that the risk of data leakage in supply chain collaboration is reduced.

The invention discloses a two-for-one twister spindle speed anomaly detection method based on health layering and a hybrid auto-encoder model, relates to the technical field of yarn twisting, and solves the problem of training data pollution caused by scarcity of industrial field labeled data. Health layering is carried out by mining long-time-history statistical characteristics of a spindle group, and a health core data subset representing an ideal state is automatically screened out for model training. On this basis, a hybrid auto-encoder model fusing a one-dimensional convolutional neural network and a long-short-term memory network is constructed, the hybrid auto-encoder model only performs unsupervised training on purified health data, and real single-spindle local anomaly is stripped from complex global system fluctuation, so that the accuracy of the single-spindle local anomaly is improved, and the accuracy of the single-spindle local anomaly is improved. Therefore, sensitive and accurate detection of various unknown anomalies is realized.

The invention discloses a high-performance time sequence parallel simulation system and method based on a time acceleration and synchronous control mechanism, and belongs to the technical field of high-performance parallel calculation. The system comprises a timing trigger module, a time mapping engine module, a task generation and distribution module, a parallel processing module, a synchronous blocking module, a state aggregation module and a visual output module. By establishing a double-layer time system of physical time takt-simulation time window, a time mapping engine dynamically calculates the number of virtual time steps needing to be propelled according to acceleration multiplied speed, and acceleration mapping of physical time and simulation time is achieved. The parallel processing module executes tasks in a thread pool with a fixed size, and the synchronous blocking module performs unified time sequence control on execution of each thread through a counting synchronizer, so that state consistency and calculation reproducibility under a high-speed operation condition are ensured. And the state aggregation module is used for integrating the synchronized local calculation results and updating the global system state. According to the method, high-speed simulation and high-fidelity playback of large-scale time-dependent data are realized on the premise of not damaging sequential logic, and the modeling verification efficiency and stability of a complex system can be remarkably improved.

The invention provides a GSM-R (Global System for Mobile Communications-Railway) cross-domain coordinated multipoint cluster wireless network simulation method and system, and the method comprises the steps: receiving a call request from at least one vehicle-mounted device for each domain, determining the call request as a circuit domain request or a packet domain request based on the call request, the calling request carries address information and identification information of the vehicle-mounted equipment, local domain address information and identification information of called target ground control equipment, or carries request content and address information of the vehicle-mounted equipment and a target domain name; if the call request is determined to be a circuit domain request, determining whether the call request is cross-domain communication based on identification information of the called target ground control equipment and address identification information stored in the domain; and after the address information of the local domain is reconstructed and the address information of the target ground control equipment of the target domain is determined through the central server, the reconstructed call request is forwarded to the target ground control equipment based on the address information of the target ground control equipment.

The invention discloses a micro-grid group electric energy-warrant hybrid transaction optimization method, which comprises the following steps of: establishing a multi-micro-grid system architecture comprising a virtual operator decision-making layer and a micro-grid optimization layer, and constructing an electricity utilization right and power generation right transaction mechanism based on the architecture; constructing a double-layer game optimization model which comprises an upper-layer virtual operator price decision model and a lower-layer multi-microgrid resource scheduling model; solving the double-layer game optimization model by adopting a mixed solving strategy combining an improved particle swarm optimization algorithm and an fmincon function; and fairly distributing the income of the virtual operator among the multiple micro-grids based on an improved Shapley value method. According to the method, the virtual operator is introduced as a coordination center while the operation autonomy of each micro-grid is guaranteed, the mechanism can effectively aggregate the system flexibility and guide the multiple micro-grids to cooperatively participate in main grid interaction, organic unification of local scheduling autonomy and system global optimization is realized, and the overall operation efficiency and the renewable energy consumption capability are improved.

The invention discloses a deep learning-based surveying and mapping point cloud data intelligent correction method and system, and relates to the technical field of data processing, and the method comprises the steps: obtaining original point cloud data of surveying and mapping equipment, carrying out the denoising and adaptive density normalization processing, constructing standardized point cloud data, and extracting a multi-scale feature set; inputting a deep learning model to perform feature learning, extracting different scale spatial features containing global surveying and mapping errors and local detail distortion in parallel, performing interactive fusion, and decoupling and identifying parameterized global errors and local distortion displacement fields; and fusing scene physical constraints, carrying out iterative adaptive optimization until a convergence condition is met, and outputting final corrected point cloud data. The technical problem of poor geometric fidelity and reliability of the point cloud data caused by difficulty in cooperative processing of global system errors and local complex distortion with huge scale difference in the prior art is solved, and the technical effect of improving the geometric fidelity and reliability of massive point cloud data is achieved.

The invention provides a financial risk perception-oriented distributed artificial intelligence system and method, and the method comprises the steps: carrying out the dynamic compensation of missing data in a financial Internet through the static features of static business data and the time-space correlation features of dynamic time series data in the financial Internet, and obtaining the reconstruction data in the financial Internet; deducing a risk evolution state of the financial internet by using a risk conduction mechanism in the financial risk perception model; when the risk conduction intensity of the risk evolution state meets a preset event driving condition, carrying out combined driving constraint on the risk propagation constraint in the risk evolution state based on the information gain entropy of a local node and the topology collaboration of a global system, and generating a distributed training constraint containing an adaptive gain parameter and a calculation communication resource ratio; and completing distributed training and risk optimization of the financial risk perception model according to the distributed training constraint. Based on the scheme, distributed collaborative modeling of the risk perception model in the Internet financial environment can be realized.

The invention discloses a multi-target dynamic tracking and encircling method based on multi-agent cooperative control, and the method comprises the steps: S1, obtaining the position and speed information of multiple agents and scattering and escaping targets after the initialization of a system; s2, according to the relative position and relative speed information of the intelligent agent and the target, carrying out primary encircling task distribution, and evaluating the influence condition of the intelligent agent on encircling task distribution; s3, realizing global distributed task allocation based on an improved particle swarm optimization algorithm in combination with actual application scene constraints to achieve the purpose of optimal global system performance; and S4, establishing a gravitational force / repulsive force cooperative control model based on a potential energy field function, and realizing tracking and encircling control of multiple agents on the target. According to the method, the multi-agent system is dispersed to form a plurality of sub-groups with the number equal to that of the targets, a dynamic multi-target encircling problem is converted into a multi-dynamic single-target encircling problem, and finally, tracking and encircling of the dynamic multiple targets are realized by establishing a self-organizing collaborative motion model of the multi-agent system.

The application discloses a project whole-process penetration control method and system based on multi-source data fusion, and relates to the technical field of engineering project risk control and time sequence logic formal verification. The method comprises the following steps: acquiring a control instruction generated by a multi-agent cooperation system; extracting an attention weight distribution in an inference process to lock a target subgraph in a fusion knowledge graph, and cutting a global compliance rule to obtain a local formal context according to the target subgraph; mapping numerical variables in the local formal context into interval logic operators and constructing a local linear time sequence logic formula; and performing safety verification on the formula, and executing the control instruction when the formula passes the safety verification. The application is used to solve the problems of instruction illusion risk caused by the probabilistic decision of the generative model and the calculation space explosion caused by the global system state verification.

The invention discloses a fully distributed path planning method and device based on correlation equilibrium and a medium, belongs to the technical field of traffic system path planning, and solves the problems that the Nash equilibrium solution is not unique and the centralization strategy distribution dependence is high in the existing traffic path planning. Comprising the steps that a distributed optimization model is established, the objective function of the model is to minimize global system cost, and related equilibrium constraints are applied; solving the distributed optimization model based on a fully distributed algorithm, updating the vehicle path selection preference during each iteration, and taking the path selection preference corresponding to each vehicle as a final driving path when the algorithm converges; according to the method, a mapping relation between suggested path preference and system travel cost is established, an objective function is constructed according to the mapping relation, and a mathematical model of a traffic jam problem is established in combination with simplex constraints and related equilibrium conditions which are required to be met by the path preference; and distributed correlation equilibrium calculation and path selection are realized on the premise that a central coordination node is not needed.

The invention discloses an SSD multi-queue timeout low-cost control method and device and a medium, and relates to the technical field of SSDs. The method comprises the steps that global system master time is provided; when a command operation occurs, taking the system main time as a timestamp of the command operation, and storing the timestamp to a storage position corresponding to a command identifier of the command operation; reading a timestamp corresponding to the uncompleted command operation from the storage position in a polling mode; according to the current system main time and the read timestamp, the executed duration of the command operation is calculated; and comparing the executed duration with a preset overtime threshold to judge whether the command operation is overtime or not. Under a multi-queue concurrent command processing scene, the area cost and the dynamic power consumption of hardware implementation can be effectively reduced, the expandability of the system is improved, and the problem that an existing scheme is difficult to expand due to too large resource occupation is solved.

The invention provides a distributed predictive control method for an unmanned aerial vehicle cluster system under hybrid network attack. The method comprises the following steps: firstly, decoupling a global cost function, and approximately converting a distributed prediction control problem of an unmanned aerial vehicle cluster into a centralized prediction control architecture of independent operation of each unmanned aerial vehicle; on the basis, a state observer is designed in an off-line mode to estimate the system state, a multi-step control invariant set is introduced based on the estimated state, and a distributed predictive control strategy is constructed. Furthermore, a health judgment mechanism is designed, the neighbor relation between the unmanned aerial vehicles is dynamically updated, and nodes suffering from network attacks are isolated from the communication topology in a fault isolation mode. Meanwhile, through introduction of compatibility constraints, asymptotic stability of a global system is ensured. And finally, the feasibility and superiority of the proposed method in a hybrid network attack environment are verified in a mode of combining simulation with a physical experiment.

The present application relates to the technical field of fatigue detection, in particular to a hybrid tractor fatigue spectrum mutation detection method for multi-dimensional channel synchronous positioning, comprising the following steps: synchronously collecting and aligning three-channel vibration signals of the hybrid tractor to generate a synchronous multi-dimensional signal; extracting a dominant coupling strength characteristic quantity of a coupling relationship matrix in a preset characteristic frequency band in real time from the synchronous multi-dimensional signal to form a dominant coupling strength time sequence; determining a spectrum characteristic mutation point representing the fatigue state evolution of a key component and outputting an alarm signal. The present application can not only reflect the strongest mode of global coupling relationship, but also be highly sensitive to instantaneous energy concentration and mode enhancement. Compared with traditional amplitude correlation coefficient methods, the present application retains phase coupling information and can capture the global system coupling mode mutation caused by fatigue cracks, gear tooth breakage, bearing jamming and the like to realize rapid identification of potential faults of key components in an early stage.

The utility model relates to an alcohol heater device based on double-chamber heat exchange, and belongs to the technical field of waste heat recycling. The methanol heater device comprises a shell, a methanol fuel heater, a heat exchanger and a GSM (Global System for Mobile Communications) communication module; wherein the methanol fuel heater comprises a combustion chamber and a heat exchanger; the combustion chamber is arranged in the heat exchanger, the heat exchanger comprises a heat exchange shell, a tail gas inlet and a tail gas outlet, the tail gas inlet and the tail gas outlet are formed in the heat exchange shell, and a cooling liquid heat exchange unit is arranged in the heat exchange shell; the heat exchanger comprises an air side heat exchange unit and a cooling liquid side heat exchange unit; a tail gas inlet of the heat exchanger is connected with a tail gas outlet of the combustion chamber. A tail gas outlet is connected with an air inlet of an air-side heat exchange unit of the heat exchanger. And a cooling liquid outlet of the cooling liquid heat exchange unit is connected with a cooling liquid inlet of the cooling liquid side heat exchange unit. According to the alcohol heater device, the tail gas waste heat utilization rate and the energy conversion efficiency can be effectively improved through a double-cavity heat exchange structure and multi-stage heat recovery.

The invention relates to a multi-target multi-constraint hybrid decision-making agent planning method for a coal mining process, and belongs to the technical field of coal mine intellectualization. Comprising the following steps that: a global system-level planning organization Agent plans a target yield of a current production cycle; each local equipment level planning type organization Agent obtains a four-dimensional target optimization function and a reward function of the corresponding scene task; planning by using an MAPPO algorithm to obtain a plurality of multi-target connection task strategies; and after the local device level planning type arbitration Agent obtains the conflict-free multi-target connection task strategy, the local device level planning type organization Agent calculates the multi-target connection task strategy with the optimal value according to the corresponding reward function based on the arbitration strategy of the cooperative game, and issues the multi-target connection task strategy with the optimal value to the Agent alliance. The intelligent degree of decision making can be improved, the labor cost can be reduced, the decision making efficiency can be improved, and potential safety hazards can be reduced.

The invention provides an encrypted traffic monitoring method and system based on multi-modal perception and confidence evaluation, and belongs to the technical field of network data security. The system comprises a data preprocessing engine used for preprocessing and feature extraction of encrypted traffic data to obtain session feature vectors; the feature sample dynamic compressor is used for adaptively switching a compression strategy according to a data flow destination and a resource occupancy rate, and compressing a session feature vector according to the switched compression strategy; the global system coordinated scheduling controller is used for scheduling the session feature vector to the processing module group; the processing module group comprises an asymmetric confrontation defense perception module, a multi-granularity flow interaction graph construction module and a dynamic confidence fitting and topology perception analyzer; and the task polling type model training engine forms a model evolution feedback cycle through the model files stored in the model warehouse and the knowledge base and the structured knowledge and processing module group. According to the invention, the problems of reduced detection capability, unreliable decision and poor adaptability in the prior art can be solved.

The invention discloses a data center observability data processing method. The method comprises the steps of collecting multi-source observability data, performing association and standardization to obtain standardized time sequence data, and constructing a global system topological graph based on an asset relationship; sliding window sampling is carried out on the standardized time series data, forward reasoning is carried out based on a long short-term memory (LSTM) network model, and predicted values of corresponding performance indexes are obtained; comparing the predicted value with a dynamic threshold interval to generate a predictive early warning event, and associating the event with nodes with an asset relationship in the global system topological graph; when an early warning event is received, extracting observability data corresponding to the associated node, and obtaining a fault type through a fault classification model; taking a corresponding node as a starting point, performing reasoning on a global system topological graph by using a graph neural network GNN model, and positioning a root cause node; and based on the fault type and / or the root cause node positioning result, calling an optimization algorithm to obtain an optimization decision suggestion for execution.

The invention discloses a self-intelligent-driving thermal power plant unattended implementation method, and particularly relates to the technical field of thermal power plants, which comprises the following steps: S1, collecting multi-modal heterogeneous data in the operation process of a thermal power plant, S2, carrying out feature extraction and fusion on the multi-modal heterogeneous data by using a multi-modal fusion encoder, and S3, carrying out feature extraction and fusion on the multi-modal heterogeneous data; the method comprises the steps of S1, generating a global system state vector, S3, jointly inputting the global system state vector and preset operation target text description into an industrial field large language model, S4, inputting numerical time series data into a time series prediction model based on a ProbSparse self-attention mechanism, S5, generating a cooperative control instruction set by using a center decision-making device based on a space-time Transform architecture, and S5, generating a cooperative control instruction set by using a central decision-making device based on the space-time Transform architecture. S6, issuing the cooperative control instruction set to each subsystem execution mechanism; and S7, carrying out online incremental learning on the time sequence prediction model and the central decision maker. According to the invention, the fundamental transformation of the thermal power plant from local automation to global autonomous intelligence is realized, and finally, the real unattended operation with safety, high efficiency and self-optimization capability is achieved.

The application provides an encrypted traffic monitoring method and system based on multi-modal perception and confidence evaluation, belonging to the technical field of network data security. The system comprises: a data preprocessing engine for preprocessing and feature extraction of encrypted traffic data to obtain a session feature vector; a feature sample dynamic compressor for adaptively switching compression strategies according to data flow destination and resource occupancy rate, and compressing the session feature vector according to the switched compression strategy; a global system coordination and scheduling controller for scheduling the session feature vector to a processing module group; the processing module group comprises an asymmetric adversarial defense perception module, a multi-granularity flow interaction graph construction module and a dynamic confidence fitting and topology perception analyzer; and a task polling type model training engine forms a model evolution feedback cycle with model files and structured knowledge stored in a model warehouse and a knowledge base and the processing module group. The application can solve the problems of reduced detection capability, unreliable decision and poor adaptability in the prior art.