635 results about "Plug and play" patented technology

The invention discloses a laboratory analytical instrument automatic calibration system based on artificial intelligence, and belongs to the technical field of instrument automatic calibration. Comprising the following modules: an intelligent hardware interaction module for realizing plug-and-play of different instruments and accurate conveying of standard substances; the instrument digital twinning module automatically identifies the characteristics of the instrument and simulates the behaviors of the instrument; the multi-mode AI calibration module is used for optimizing calibration parameters and adjusting the calibration process in real time; the environment intelligent compensation module calculates and generates a compensation coefficient in real time, and eliminates the influence of environmental factors on a calibration result; the instrument health management module is used for evaluating the health state of the instrument and predicting the service life of a core component, and realizing conversion from passive maintenance to predictive maintenance; the calibration knowledge graph module is used for realizing automatic extraction, reasoning and application of calibration knowledge through a graph neural network and continuously optimizing a calibration strategy; and the intelligent scheduling and management module is used for generating an optimal calibration plan, coordinating the work of each module and realizing reasonable distribution and efficient utilization of calibration resources.

The invention relates to the technical field of data acquisition, in particular to a multi-source heterogeneous bus data acquisition protocol configuration method, which realizes protocol identification and structure analysis of CAN (Controller Area Network), Modbus and Ethernet / IP (Internet Protocol) bus data by dynamically compiling protocol source codes and registering analysis classes, and comprises the following steps of: extracting protocol feature information from a data stream; the method comprises the following steps of: calling a corresponding protocol analysis class to perform recursive analysis, and executing parameter analysis and mapping, in addition, defining a new protocol semantic feature and dynamically generating a protocol analysis plug-in through an XML (Extensible Markup Language) configuration file to realize protocol semantic-level custom extension, adopting source code level dynamic compilation and plug-in design, supporting a new protocol plug-and-play mode, and being high in practicability. System codes do not need to be modified and redeployed, and the flexibility and expandability of the system are effectively improved.

The invention relates to the technical field of power distribution automation of a power system, and particularly discloses a general operation and maintenance method and system supporting multiple types of power distribution terminals. According to the method, millisecond-level automatic accurate identification of the power distribution terminal is realized through a multi-layer protocol feature fingerprint extraction technology and a weighted dynamic similarity matching algorithm. On the basis of a double-library cooperation mechanism, after equipment identification succeeds, communication parameters, point table mapping relations and operation and maintenance instructions which conform to international / industrial general standards are automatically triggered, loaded and issued. The plug-and-play and zero-configuration deployment of the power distribution terminal is realized, and the point table configuration error rate is reduced to below 1%. And an unidentified equipment self-learning closed-loop process is designed, so that the system can quickly adapt to a power distribution terminal of a new manufacturer, a new number or a new firmware version. And an automatic parameter read-back and verification comparison mechanism after configuration execution forms an operation and maintenance operation closed loop, so that the accuracy of configuration issuing each time and the controllability of the equipment operation state are ensured.

The invention discloses a multispectral image panchromatic sharpening method based on plug-and-play gradient feature guidance fusion, through effective extraction and integration of gradient guidance features (GRAD), excellent recovery results are obtained in the aspects of structural fidelity, detail retention and spectral consistency, and in addition, the method has the advantages of being simple in structure and convenient to use. Gradient feature guidance is fused and designed into a highly flexible plug-and-play method, and the existing architecture can be seamlessly enhanced. In particular, the present invention explicitly models a GRAD by integrating fine spatial details of a high resolution PAN with intrinsic spectral information of an MSI. Meanwhile, an attention mechanism and learnable weighting and residual connection are adopted, and selective aggregation and adaptive fusion of image information are realized, so that the MSI quality is remarkably improved. In addition, the GRAD is systematically analyzed from the aspects of structure, detail and spectrum, and the result is optimized through a multi-objective loss function. In this way, gradient information is fully mined, high-frequency details are captured, cross-modal structure guide alignment is achieved, panchromatic sharpening of the multispectral image is flexibly and efficiently achieved, and the panchromatic sharpening performance is improved.

The invention discloses a front-end multi-module development method and system based on module federation. The method comprises the following steps: extracting necessary examples, public components and public methods in a front-end project to a public basic module, and constructing a module federated product; decoupling other functional modules, and configuring a module code and a module federation construction entry in the construction configuration file; remote module lists and configuration information are managed in a centralized mode through the project configuration module; executing a construction command to call the project construction module to independently package each module and update the project configuration module; and when the application is started, the project loading module loads the public basic module and the remote function module according to the project configuration module. Compared with an existing Module Federation scheme with application as granularity, independent construction, configuration management and flexible loading during operation are achieved on the module level, dependence on repeated and redundant logic is avoided, module plug and play and version switching are supported, and therefore development efficiency, construction performance and expansibility are improved.

The invention provides a target detector construction method based on transfer learning in a foggy scene, and belongs to the technical field of deep learning. According to the method, two lightweight neural network modules, namely an image adaptive module and a feature attention enhancement module, are designed according to the requirements of an unmanned aerial vehicle target detector in a foggy scene, and the two lightweight neural network modules are integrated into a plug-and-play image defogging and enhancement network; and then designing a combined target detector network according to the characteristics of the image defogging and enhancing network and various widely used target detectors. According to the invention, the target detector can improve the target detection performance in a foggy scene under the condition that model parameters and computing resources are hardly increased, technical support is provided for reliable visual perception under a severe weather condition, and the method has a wide application prospect.

The invention provides a multi-modal ecological environment monitoring element plug-in integration system and method based on unified objectification modeling. Firstly, an object model and an inheritance system of monitoring equipment, data and an algorithm are constructed; on the basis, a unified logic storage and metadata cataloguing mechanism oriented to heterogeneous databases such as relational databases, non-relational databases and file object databases is designed and realized; the plug-and-play and full-life-cycle management of equipment, data and algorithm models is further realized through a dynamic registration module; common APIs (Application Program Interface) such as object-level addition, deletion, modification and query, space-time association query, concurrent task scheduling and aggregation and the like are externally provided by relying on a unified service bus. And the hybrid storage engine cooperatively manages structured and unstructured data and maintains a topological relation between objects. According to the scheme, object dynamic integration and expansion are completed on the premise that operation is not interrupted, the multi-source heterogeneous data management and cross-modal analysis efficiency is remarkably improved, and the method is suitable for various ecological environment monitoring scenes such as natural reserve, drainage basin pollution supervision and disaster emergency response.

The invention discloses a multi-modal sensor abstract description and packaging method and system based on software definition. Firstly, a unified sensor abstract description model based on software definition is introduced, so that hardware independence is realized, and the differences of a bottom-layer multi-mode sensor in the aspects of interfaces, protocols and data formats can be effectively shielded; secondly, abstract packaging of data is achieved through an autonomous information processing module, and high-bandwidth and high-redundancy original data is converted into low-bandwidth and high-value structured information; in addition, by means of automatic registration, discovery, scheduling and control mechanisms of the sensor management control module on sensor resources, the system realizes a real plug-and-play function. According to the invention, by deploying the autonomous information processing modules at various sensor nodes, the original sensing data is preprocessed, and standardized data packaging with semantic consistency is output, so that the operation load of the central controller is reduced, and the information processing efficiency and the system response capability are also improved.

The invention discloses an island micro-grid distributed cooperative control method based on local information and a storage medium, and belongs to the technical field of micro-grids, and an island micro-grid system comprises a converter of network construction type control and network following type control. And according to local information, the network tracking type control adopts a maximum power point tracking point tracking (MPPT) algorithm to control power output. The reactive power of the network construction type control parallel converter is output and distributed according to the fixed voltage regulation droop coefficient; and the active power of the network construction type control parallel converter is output and distributed according to the frequency modulation droop coefficient adaptively updated based on local information. According to the method, power instruction distribution of island micro-grid and network type control and network construction type control is generated by local information, instruction distribution of upper-layer centralized control does not need to be carried out, plug and play, flexible reconfiguration and decentralized control of an island micro-grid inverter can be realized, and the method has relatively good practicability.

The invention relates to the technical field of industrial communication and Internet of Things, in particular to an IO-Link gateway system and method for heterogeneous protocol unification and edge enabling. Comprising a hardware layer, a protocol adaptation layer, a data processing layer, an intelligent computing layer and a communication interaction layer which are connected in sequence to achieve standardized conversion and edge intelligent processing of heterogeneous industrial data, and data time sequence consistency and system operation stability are guaranteed in the whole process. According to the invention, a plurality of heterogeneous industrial protocols are downwards and uniformly mapped to a technical path of an IO-Link international standard, and heterogeneous data are converted into a standard IO-Link object dictionary at a gateway level, so that field devices of different manufacturers and different protocols can carry out data interaction with a uniform and semantic-rich information model; therefore, plug-and-play of data is realized in the equipment access layer, and a solid foundation is laid for constructing an integrated industrial Internet of Things system.

The invention discloses a pluggable Kubernetes cluster deployment method, a medium and equipment, and solves the problems of rigid deployment process, serious component coupling and low fault recovery efficiency in the prior art. The method comprises the following steps: receiving user module label combination and node resource information through a visual interface, and carrying out structured storage as a deployment context object; after the object is analyzed, a dependent support module is automatically identified and added, a cluster architecture mode is dynamically determined based on the number of Master nodes, and a DAG of a deployment task is constructed; module deployment is executed according to the DAG sequence, and only the fault module and downstream dependency thereof are retrying when failure occurs. According to the method, plug and play of the component is realized through label driving, a deployment process is dynamically arranged based on DAG, intelligent framework cutting is supported, accurate fault isolation and local recovery are realized, time consumption of component switching and resource occupation are greatly reduced, and deployment flexibility and operation and maintenance reliability are remarkably improved.

The invention discloses a hand-eye cooperative robot control system and method based on dynamic operator arrangement, an upper computer planning layer operates a master control computer to periodically trigger task scheduling, a joint controller of a lower computer execution layer receives an instruction through a redundant bus to perform servo control, hand-eye camera data triggers a visual assembly line through an interrupt event, and a visual assembly line is controlled through a control interface. According to the method, pressure is calculated through visual processing, motion planning and joint control in a load sharing mode, a double-buffering mechanism is adopted to enable current frame visual processing and previous frame motion control to be executed in an overlapping mode, real-time scheduling and parallel computing of tasks are completed, hot data are cached in a memory database, cold data are archived to an HDFS and migrated through an LRU strategy, and the real-time scheduling and parallel computing of the tasks are completed. Data type conversion is automatically derived and performed based on a feature type registry, and conditional branch execution is triggered according to real-time sensor data. According to the control system, cross-hardware plug and play, algorithm flow configurability and data flow real-time sharing are achieved, and the requirement for improving the efficiency of complex operation tasks is met.

The invention discloses a multi-brand and multi-model automatic identification control method and system for an intelligent fixed-line telephone, and belongs to the technical field of plug and play of communication peripherals. The method comprises the following steps: after a USB plugging event is detected, extracting multi-dimensional features such as a manufacturer ID, a product ID, a USB serial number and a user-defined identifier of a fixed-line telephone in sequence to form a multi-dimensional feature vector with clear priority; then, through a multi-stage progressive matching strategy of first-stage (VI + PID accurate matching)-second-stage (self-defined identifier supplementary matching)-third-stage (serial number + descriptor fuzzy matching)-second-stage (self-defined identifier supplementary matching)-third-stage (serial number + descriptor fuzzy matching)-second-stage (self-defined identifier supplementary matching)-second-stage (self-defined identifier fuzzy matching)-second-stage (self-defined identifier fuzzy matching)-second-stage (self-defined identifier fuzzy matching)-second-stage And aiming at the scene of feature missing or VAD / PID conflict, a fault-tolerant mechanism is built in the system: when the scene is missing, the universal DLL is used for reverse query confirmation and identification, and when the scene is conflict, candidate DLLs are cyclically loaded according to a brand market share sequence until the communication verification succeeds. The method is high in accuracy, multi-stage progressive matching and confidence evaluation are combined, and the error recognition rate is lower than 0.1%; zero sample online: the new type number can be connected for the first time to complete the adaptation in the second level, and manual maintenance is not needed.

The invention provides a simulation tool for a self-adaptive joint debugging device of a power distribution terminal. The simulation tool comprises a tool main body, a first connecting unit, a second connecting unit and a third connecting unit, the tool main body panel is provided with a power supply signal input interface, a power supply signal output interface and an interface combination corresponding to the simulation tool adapter ring main unit. The tool main body realizes that power supply signal analog quantity generated by the joint debugging device is input to the power supply signal input interface through the first connecting unit; the simulation tool adapts to an interface combination corresponding to the ring main unit to obtain a power supply signal analog quantity through internal wiring, and then the power supply signal analog quantity is sequentially transferred to a ring main unit interval through a second connecting unit; and the power supply signal output interface is used for outputting part of voltage signal analog quantity generated by the joint debugging device to the ring main unit through the third connecting unit. According to the invention, plug-and-play of the simulation tool on the ring main unit and the existing self-adaptive joint debugging device can be realized, and the convenience and reliability of the test are improved.

The invention discloses an electric power inspection exoskeleton modular structure, and belongs to the technical field of human body exoskeleton machine assistance, the electric power inspection exoskeleton modular structure comprises a supporting system, the supporting system comprises a back supporting plate, a waist supporting frame and thigh supporting rods, and a main body supporting frame is formed; the modular interface system comprises at least one group of mechanical interfaces, electrical interfaces and data interfaces which are uniformly distributed on a force transmission path of the support system; the functional module comprises a plurality of interchangeable power assisting modules, operation modules and protection modules, and the functional module is detachably connected with the supporting system through the modular interface system; a hybrid drive system includes a rigid driver and a flexible driver for providing position control and compliance force control, respectively. By means of the modular interface system, plug-and-play of functional modules such as power assisting, operation and protection is achieved, needed exoskeleton configuration can be rapidly assembled, multiple sets of exoskeleton equipment do not need to be arranged, and the equipment utilization rate and task adaptability are greatly improved.

The present application relates to the technical field of software engineering, and provides a device management method and apparatus, a device, and a storage medium, which are used for performing automatic identification when a device is connected, and enabling a corresponding function, thereby achieving plug-and-play of the connected device. The method comprises: when detecting that a device is connected to a port, an inverter identifies a device type of the connected device, and establishes a communication connection with the connected device if the device type of the connected device is successfully identified; furthermore, the inverter enables a function corresponding to the device type of the access device.

The application discloses a video question and answer modeling method based on a hierarchical selection framework, characterized in that: firstly, video data and text data are preprocessed to obtain video segment features, object features, question features, question word features and candidate answer features as inputs of a video question and answer model; video segment features, video frame features and object features are selected hierarchically from large to small according to time span and space span, and then are fused with question features to realize prediction of answers. The method can be combined with different pre-training models for training and fine-tuning, that is, plug and play; the video segment features, video frame features and object features related to questions can be selected hierarchically to realize extraction of question-related key features and filtering of irrelevant redundant information.

The invention discloses a multi-interface fingerprint identification card type USB flash disk, which belongs to the technical field of USB flash disks, and comprises a hardware core module, a software interaction module and a software-hardware combination module, the hardware core module is a physical basis of the USB flash disk, and is responsible for interface adaptation, fingerprint acquisition, data storage and state indication; the hardware core module comprises a multi-interface protocol module, a dual-master control cooperation module, a multi-specification plug-in card adaptation module and a fingerprint identification module; and the software interaction module is used for realizing fingerprint management, authority configuration and log recording functions through a four-layer architecture. In the use process, the plug-in type USB flash disk with the multi-interface fingerprint identification function is compatible with Lightning, TYPE-C and USB-A interfaces, adapts to multiple devices such as a computer, a mobile phone, a tablet computer and an industrial controller, does not need an additional adapter, is plug-and-play, improves the convenience of a use scene, is matched with a high-speed protocol of USB3.2 and above, gives consideration to the transmission efficiency and the device compatibility, has multiple permissions, and is convenient to use. And efficient protection of the information is realized through three-level permission classification.

The invention discloses a distributed cooperative control method and system for cluster temperature controllers, and relates to the technical field of temperature controller control. Comprising the following steps: establishing a temperature control consensus: establishing a temperature controller alliance chain, and adopting a practical Byzantine fault-tolerant algorithm to achieve a temperature control target consensus; sharing a prediction model: based on a federated learning architecture, each node locally trains a temperature prediction model and uploads a model parameter gradient to a cloud server for aggregation updating; global energy consumption optimization: adopting a multi-agent depth deterministic strategy gradient algorithm to carry out minimization optimization on the global energy consumption; and dynamic topology recombination: supporting plug and play of the equipment and automatically recombining the cluster topology according to the state of the equipment. According to the method, the information island problem is solved, sharing and privacy protection of the temperature prediction model are realized, global energy consumption is optimized, cluster temperature control consistency can be improved, energy utilization efficiency can be optimized, user comfort experience can be improved, hundred-level equipment expansion is supported, and complex scene requirements are met.

The invention discloses a plug-and-play power distribution terminal topology self-identification method and system, and relates to the technical field of power system power distribution network operation and maintenance, and the method comprises the steps: employing a modular hardware architecture and automatic identification configuration software, and constructing a plug-and-play intelligent terminal; through topological information acquisition, electrical parameter analysis and topological structure construction, the topological structure of the power distribution network is automatically identified and updated; and operation and maintenance decision support is provided for a master station system based on a power distribution network topological structure. According to the method, plug and play of the intelligent power distribution terminal can be realized, the intelligent power distribution terminal can quickly adapt to a complex and changeable network topology structure, and meanwhile, the beneficial effect of quick self-healing of a power distribution network fault is provided.

The invention discloses a spatial self-adaptive plug-and-play watermarking method and device for image editing traceability, and the method comprises the steps: obtaining the hidden space feature representation of an edited image through an image editing model according to an input original image and an editing condition; constructing a structured watermark grayscale image used for bearing image editing traceability core information, and encoding the structured watermark grayscale image into watermark potential features through a watermark encoder; based on the watermark potential features, watermark embedding of content and structure perception is realized in the submerged space, and the submerged space feature representation of the edited image with the watermark is acquired; and inputting the hidden space feature representation of the edited image into an image decoder, decoding the edited image into an edited image with a watermark, inputting the edited image with the watermark into a watermark decoder, extracting an embedded watermark grayscale image, and restoring complete traceability metadata information based on the watermark grayscale image. The device comprises a processor and a memory. According to the invention, high visual quality of the image is maintained while reliable traceability is realized.

The invention relates to a modular intelligent power distribution cabinet safety management method based on the Internet of Things, particularly discloses a modular intelligent power distribution cabinet safety management method based on the Internet of Things, and aims to solve the problems that a traditional power distribution cabinet monitoring system is poor in expansibility, weak in fault recognition capability and rigid in protection strategy. The method comprises the following steps: deploying a modularized sensing unit with a standardized interface to realize plug and play; multi-dimensional data of current, voltage, temperature and arc light are synchronously acquired at a high sampling rate; data are fused on the edge side through Kalman filtering, and key features such as harmonic distortion rate, voltage sag depth, temperature gradient change rate and electric arc energy accumulation value are extracted; performing dynamic risk assessment by using a five-layer deep belief network model, and triggering early warning when the high risk probability exceeds a threshold value; adaptively adjusting an overcurrent protection constant value, opening delay and isolation logic according to an evaluation result; and the edge-cloud coevolution is realized by continuously optimizing the model and the strategy through cloud reinforcement learning.

The invention relates to electroencephalogram collection, in particular to a stackable self-reconfiguration electroencephalogram collection device and a self-reconfiguration algorithm. The stackable self-reconfigurable electroencephalogram acquisition device comprises a plurality of stacking plates and a stacking interaction interface unit, the stacking plates are used for filtering, amplifying, acquiring, transmitting and instruction communication of 32-channel electroencephalogram signals, and the stacking plates are connected through the stacking interaction interface unit. According to the stackable self-reconfiguration electroencephalogram acquisition device provided by the invention, through the synergistic effect of the modular stacking design and the self-reconfiguration algorithm, dynamic lossless expansion and plug-and-play of an electroencephalogram channel are realized, and the expansibility and flexibility of a system are remarkably improved; and meanwhile, the synchronization and integrity of signals are guaranteed by adopting cascade clock synchronization and global isolation design, automatic identification and function distribution among multiple stacked plates are realized by access detection and a hot plug consensus algorithm, and hot plug and topology reconstruction of the system can be completed without manual intervention.

The invention discloses a design method of a novel backplane bus. The design method comprises the following steps: step 1, constructing a basic backplane bus architecture; 2, implementing a module physical isolation design; step 3, integrating a fault detection mechanism; 4, designing a slot number dynamic transmission mechanism; step 5, establishing a hybrid communication architecture; 6, deploying a redundant power supply and communication link; step 7, realizing state self-diagnosis and recovery; 8, designing a remote monitoring interface; and step 9, formulating a modular extension specification. Through modularized physical isolation and dynamic slot position transmission, electrical fault diffusion is effectively avoided, the maintenance efficiency is remarkably improved, the hybrid communication architecture integrates the advantages of a real-time Ethernet and a parallel bus, high-speed real-time performance and low-speed compatibility are both considered, redundant power supply and ring topology communication ensure high-reliability operation, and the service life of the system is prolonged. Remote monitoring and standardized interfaces are combined, and plug-and-play expansion is supported.

The invention relates to the technical field of intelligent warehouse logistics, and particularly discloses a warehouse logistics AI path planning system and method based on multi-modal dynamic optimization. The system comprises a multi-source data layer which aligns inventory data, traffic flow data and meteorological data of a warehouse management system (WMS) in real time through a sliding window synchronization mechanism, and ensures time-space consistency of multi-modal input; the dynamic fusion engine responds to a traffic interruption event within 3 seconds based on a space-time alternative path algorithm and generates an anti-interference optimized path scheme; according to the path decision core, a two-stage reinforcement learning architecture formed by a graph convolutional network (GCN) and a deep Q network (DQN) is adopted, static topology analysis and dynamic real-time characteristics are fused, and a final transportation path is output; the decision auditing module is used for recording Q value distribution of decision nodes, constructing a traceable path decision tree and automatically detecting regional distribution deviation through a quantization formula; and the zero-code compatible layer is used for presetting a standardized API template library and a rule compiler so as to realize plug-and-play integration with an enterprise resource plan (ERP) or a warehouse management system (WMS). Through multi-source data collaborative optimization, dynamic anti-interference decision and whole-process audit tracing, the real-time performance, robustness and transparency of warehouse logistics path planning are remarkably improved.

The invention provides an operation and maintenance system and method based on an MCP protocol and electronic equipment, the system comprises a standardized interface packaging module and a large model context protocol layer service module, interfaces of various operation and maintenance tools are packaged in a large model context protocol layer through the standardized interface packaging module according to a unified packaging standard, and the packaged interfaces are stored in the large model context protocol layer service module; and generating adapters corresponding to the operation and maintenance tools, recording the adapters through a large model context protocol layer service module, obtaining the corresponding adapters in response to an operation and maintenance request of a client, and operating the adapters to execute an operation and maintenance task carried by the operation and maintenance request. Various operation and maintenance tools can be packaged according to a unified packaging standard and then integrated into the operation and maintenance system provided by the application, and when a client initiates an operation and maintenance request, a corresponding adapter is quickly obtained according to the operation and maintenance request, so that plug and play of the operation and maintenance tools is realized, the maintenance cost of an operation and maintenance scheme is effectively reduced, and the operation and maintenance efficiency is improved. And the expansibility of the operation and maintenance system is improved.

The invention discloses a modular sensor adaptive integration and data processing method and system in a complex terrain environment, and the method comprises the steps: constructing an extensible sensor sensing layer through a standardized interface module, and achieving the plug and play of various environment monitoring sensors; performing an inspection task in a complex terrain area based on a mobile device, and performing fusion processing on real-time data acquired by multiple sensors through an edge calculation layer; according to the environment data characteristics, the terrain complexity and the running state of the mobile device, the inspection behavior strategy of the mobile device is dynamically adjusted, and follow-up data monitoring and processing are carried out; the processed real-time data is uploaded to a cloud collaboration layer for collaborative analysis in a self-adaptive selection communication mode, and fundamental conversion from a static and fixed monitoring mode to a dynamic and adaptive mode is realized through modular sensor integration, mobile inspection, edge intelligent processing and self-adaptive communication; and the coverage rate, the efficiency and the intelligent level of environment monitoring are remarkably improved.

The invention discloses a peripheral interface module of computer equipment, which comprises a module side expansion interface, and the module side expansion interface is connected with a host side expansion interface of the computer equipment in a physical plugging mode; each of the host side expansion interface and the module side expansion interface comprises a module function signal and is used for identifying a protocol type supported by the peripheral interface module; and the protocol self-matching system is used for automatically identifying the protocol type of the peripheral interface module according to the module function signal and switching the corresponding controller configuration to realize a plug-and-play function. After the interface is replaced, the interface output from the processor is lower in delay, no external extra equipment exists, carrying is more convenient, management and maintenance are convenient, meanwhile, the interface is used for self-defining, secondary development and transformation of a notebook computer interface according to self requirements, the interface adapts to application scenarios of non-standard peripherals, different function protocol modules are automatically recognized, and the interface switching efficiency is improved. The switching of the corresponding controllers is completed, and the plug and play of the module is realized.

The invention discloses a bidirectional information retrieval enhancement generation method for a large language model, and belongs to the technical field of artificial intelligence. In order to overcome the defects that noise is introduced and key evidences are omitted due to the fact that traditional RAG only executes'query-document 'one-way retrieval, a two-way semantic perception retrieval enhancement generation model and a two-stage training framework are constructed, wherein in the first stage, the positive / negative example distance is increased in an embedded space in a contrast learning self-supervision mode; in the second stage, fine-grained correlation discrimination is carried out on query-document bidirectional sentences through supervised dichotomy, and probabilistic correlation scores are output; in the reasoning stage, the bidirectional probabilities are fused according to Bayesian to obtain final relevancy, document reordering is carried out, and plug and play can be achieved without fine adjustment of LLM in the whole process. According to the method, the accuracy and consistency of single-hop and multi-hop questions and answers and fact checking tasks are remarkably improved, and the method has the advantages of light weight and low deployment cost.

The invention provides a split type robot dynamic task flow engine system. Comprising a basic actuator used for generating a task instance according to an externally input task request; the operator set comprises at least one operator, and each operator is used for calling hardware resources of the corresponding function cabin to execute actions corresponding to the operator types when receiving an execution instruction of the basic actuator; the task model is used for maintaining identification, types, priorities and state information of task instances between the basic actuator and the operator set and supporting persistent storage and recovery of the task instances; and the communication cooperation module is used for providing a unified message transmission interface among the chassis, the functional cabin and the edge computing node of the split type robot so as to realize the transmission of instruction and state data between the basic actuator and the operator set. According to the invention, plug-and-play and dynamic adaptation of the cabin body, task scheduling based on capability, unified communication and state persistence can be realized, and task continuity is guaranteed.