156results about How to "Eliminate dependencies" patented technology

The invention discloses a mechanical arm operation method based on a multi-mode large visual language model, which comprises the following steps of: firstly, acquiring image data of a mechanical arm operation scene through a visual sensor, and acquiring a user instruction in combination with a voice recognition module or a text input module; and inputting the natural language instruction and the visual scene data into a multi-modal large visual language model for multi-modal task analysis. The model identifies the category, position and attribute of a target object in a scene, and generates a specific operation sequence in combination with an input instruction. And then, according to the operation sequence generated through analysis, the system directly drives the mechanical arm to execute corresponding tasks including specific actions such as grabbing, moving and placing of the target object, and in the task execution process, the mechanical arm sequentially completes the actions according to the operation sequence generated in advance. The problems that in the prior art, a mechanical arm is inaccurate in complex task instruction analysis, poor in operation flexibility and adaptability and poor in real-time response capacity in a dynamic environment are effectively solved.

The invention provides a learning-based map-free autonomous navigation obstacle avoidance method, a medium and equipment. The method comprises the following steps: carrying out map-free autonomous navigation obstacle avoidance on the mobile robot by adopting a reusable planning capability library; the multiplexing planning capability library comprises a global planning module and a local planning module; the global planning module is used for guiding a global navigation direction and generating ordered and discrete path points; the local planning module generates real-time actions based on a learned local navigation obstacle avoidance model; and controlling the mobile robot to move according to the action generated by the local planning module so as to sequentially navigate the mobile robot to each path point generated by global planning. According to the method, through a global and local combined planning architecture, an algorithm is modularized, so that the method adapts to different navigation tasks, a large amount of retraining based on a deep reinforcement learning navigation method is avoided, and the universality and interpretability of a model are improved.

The invention discloses a GNSS interference identification method based on spatial transformation, multi-scale feature fusion and KAN. The method comprises the steps of forming an interference signal data set; obtaining a time-frequency image training set and a time-frequency image test set; constructing a GNSS interference identification classifier composed of a spatial transformation module, a multi-scale feature fusion module and a KAN module; a trained GNSS interference identification classifier is obtained; and obtaining a final GNSS interference identification classifier. And finally, outputting the category of the GNSS interference signal acquired in real time by the GNSS interference identification classifier. According to the invention, the spatial transformation module is used to carry out translation, scaling and other operations on the interference signal time-frequency image, and the number of training samples is increased. And the multi-scale feature fusion module adopts parallel convolution kernels of different scales to extract multi-scale spatial features of the time-frequency graph, so that the richness of feature representation is increased, and the recognition accuracy is further improved. The KAN module uses a KAN to replace a traditional multi-layer perceptron to execute a final classification task, and a more complex nonlinear function can be fitted by using fewer parameters.

The invention discloses a remote sensing image ship segmentation method based on a U-KAN network model. The method comprises the following steps: constructing the U-KAN network model; the U-KAN network model is a hybrid network architecture fusing a U-Net network and a KAN network; wherein the U-Net network comprises an encoder and a decoder which are in jump connection; a plurality of KAN layers are nested in the KAN network; u-KAN network model training is carried out in combination with the remote sensing image data set, and an optimized U-KAN network model is obtained; and inputting a remote sensing image to be processed into the optimized U-KAN network model to obtain a corresponding ship segmentation result. The method not only can effectively capture ship targets with different scales and complexities in the remote sensing image, but also can realize effective fusion between shallow detail features and deep semantic features, thereby overcoming the limitation in processing complex backgrounds and diversified ship forms in the prior art.

The invention discloses a lightweight vehicle intranet identity authentication and key negotiation method based on a PUF (Physical Unclonable Function). The method comprises a system initialization stage, a PUF registration stage, an identity authentication and key negotiation stage and a block chain synchronization stage. The method is oriented to an automobile domain architecture, direct low-cost identity authentication is carried out on the electronic control unit by utilizing the PUF (namely a physical unclonable function), the necessity of installing safety hardware on the electronic control unit can be eliminated, and secret key leakage is prevented; meanwhile, symmetric cryptography is adopted, a hash function, a pseudo-random function and a Lagrange interpolation formula are combined to balance the speed and safety of the key negotiation process of the electronic control unit, and compared with an existing vehicle intranet identity authentication and key negotiation method, the method has remarkable performance advantages.

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

The invention discloses a digital pattern reference image library construction method for multi-style fusion building restoration. The method comprises the following steps: constructing a reference image library consisting of a plurality of single-style building component reference images; acquiring a plurality of images of the same building component, performing preprocessing to extract contour information, selecting one of the images with the most complete contour information, calculating the similarity between the selected image and the corresponding reference image in the reference image library, and matching the reference image with the similarity higher than a similarity threshold value; comparing the contour information of the to-be-repaired building component image with the reference image to obtain a defect position; fusing and reconstructing the plurality of to-be-repaired building component images with different defect positions to obtain a fused image; the pictures meeting the requirements are added into a reference picture library through manual approval; and if the defect exceeds the defect threshold, matching the reference images with the same age, the same region and the same main style to perform fusion reconstruction, performing manual approval again, and adding the fused image meeting the requirement into the reference image library.

A magnet array holder configured to hold a magnet piece and / or a non-magnet piece to form a magnet array, the magnet array configured to manipulate one or more electron beams in a vacuum electronic device when assembled, the magnet array holder comprising: a set of slots configured to receive the magnet piece and / or the non-magnet piece; a set of pockets for accommodating the magnetic and / or non-magnetic parts; and one or more attachment interfaces configured to couple the magnet array holder to the vacuum electronic device.

A batch process model-free output feedback control method with non-repetitive interference belongs to the technical field of industrial process control, and specifically comprises the following steps: step 1, giving a mathematical description for an optimal tracking control problem of a non-repetitive batch process; 2, designing an optimal control law based on output feedback; 3, a game Q function construction method based on two-dimensional output feedback in the intermittent process; 4, solving a two-dimensional output feedback problem of batch operation by utilizing reinforcement learning; 5, solving an optimal control strategy based on output feedback through iteration to enable the optimal control strategy to be finally converged to an ideal optimal value; according to the method, the problem that a traditional method for processing the complex characteristics of a two-dimensional system is too tedious is solved, through the output feedback control method based on reinforcement learning, the method does not depend on a system model, the influence on the control performance is reduced, and the control effect is greatly improved.

This invention discloses a diagnostic method for abnormal fluctuations in quality characteristics based on heterogeneous data optimization, applied to the field of abnormal fluctuation analysis of automotive crankshafts. Addressing the problem that existing technologies are ineffective in handling the ambiguity of multi-layered, multi-dimensional, and multi-source data, this invention constructs a dimensionality reduction and optimization method for multi-dimensional, multi-source quality characteristic data. This method eliminates the correlation between variables and sampling dimensions, extracts more principal features, and optimizes the dimensionality-reduced data. Then, a diagnostic model for abnormal fluctuations in quality characteristics based on the principal feature dimensionality reduction and optimization method is constructed, thereby obtaining the location results of abnormal quality in automotive crankshafts.

The invention discloses an industrial control private protocol vulnerability mining method based on protocol analysis. The method comprises the steps of 1, preprocessing industrial control protocol flow data; and 2, analyzing the target protocol by using the byte change characteristics and the heuristic rule. And 3, generating a test case of the target protocol by using the analysis specification. And 4, reconstructing a protocol state machine by using the state conversion track of the flow, and guiding the protocol to reach a target state for testing. And 5, tracking the state of the detected target and monitoring and positioning the triggered vulnerability through a network. According to the method, the analysis accuracy of the private protocol specification can be effectively improved, the test case acceptance rate, the test target anomaly rate and the state coverage rate in the fuzzy test process are remarkably improved, and meanwhile, more protocol states can be efficiently tested under the condition of the same number of test cases so as to mine more potential vulnerabilities in the protocol.

The application relates to the field of flight control technology, in particular to a hovering control method and device for a short take-off and vertical landing aircraft of a tilt-rotor type, wherein the method comprises the following steps: constructing a hovering nonlinear dynamics model of the short take-off and vertical landing aircraft containing a tilt-rotor power system; converting the hovering nonlinear dynamics model of the short take-off and vertical landing aircraft into a high-order full-drive system model to design a state tracking controller; designing an adaptive law based on a neural network to estimate nonlinear terms in the hovering nonlinear dynamics model of the short take-off and vertical landing aircraft; and feeding back the nonlinear terms to the state tracking controller to perform hovering mode stable control. Thus, the problems that the hovering control method based on dynamic inversion needs a large amount of linearization processing, the actuator dynamic characteristics are significantly different, the controller needs to verify the closed-loop performance through repeated simulation and experiment, and the multi-actuator decoupling efficiency of the aircraft is low and the robust control performance is poor are solved.

The application provides a polarization / inertial navigation combined navigation method based on a deep Kalman filtering network, belongs to the field of bionic polarization navigation, and comprises the following steps: taking intermediate characteristics in Kalman filtering, a polarization sensor and historical data of inertial navigation as inputs of the network, building a priori prediction error covariance matrix estimation network and a measurement residual covariance matrix estimation network in polarization / inertial filtering, simultaneously training the priori prediction error covariance matrix estimation network and the measurement residual covariance matrix estimation network by adopting a two-stage training strategy, predicting states and measurements, calculating optimal gain in polarization / inertial filtering, and fusing the prediction of states and new information according to the calculated optimal gain to obtain estimation of the states at the current time. The application realizes high-precision navigation of the polarization / inertial navigation combined navigation system and improves the environmental adaptability of the system.

The present application relates to the technical field of computer vision, deep learning and graphics rendering, in particular to a 3D portrait reconstruction method and system based on cascade diffusion prior and UV space physical constraint, which adopts a parameter efficient adaptation paradigm of freezing a unified DiT backbone and a cascade LoRA adapter, and sequentially completes UV space texture completion, illumination homogenization and intrinsic material decomposition; introduces a cross-intrinsic attention mechanism to realize multi-material pixel-level spatial alignment, and executes differential physical rendering supervision to constrain material physical consistency in the UV domain. The present application realizes 100,000-level field image generalization under less than 100 3D scanning samples, and the generated 4K PBR material assets have no baking light and shadow residues, and the physical behavior is consistent under re-illumination, which is suitable for high-fidelity 3D portrait asset production in the fields of digital human creation, film special effects and virtual reality.

The invention discloses a vital sign monitoring system based on a WiFi signal, which comprises two WiFi CSI data acquisition platforms, a transmitting end platform and a receiving end platform, the transmitting end platform and the receiving end platform are respectively connected with a transmitting end antenna and a receiving end antenna through radio frequency lines, the transmitting end antenna and the receiving end antenna are directional antennas, and the transmitting end antenna and the receiving end antenna are both directional antennas. The beam of the directional antenna points to the chest and abdomen of the human body, the height of the directional antenna is flush with the chest position of the human body, and wave-absorbing materials are arranged on the sight distance paths of the transmitting end antenna and the receiving end antenna. According to the invention, vital signs can be reliably extracted, and separation of breathing and heartbeat signals can be realized. The invention further discloses a vital sign monitoring method based on the WiFi signal.

This invention discloses a method for autonomously constructing an evolvable knowledge graph by integrating a multimodal large-scale model, relating to the fields of artificial intelligence and knowledge graph technology. The method includes: collecting raw multimodal data from text, images, audio, and structured tables; using a pre-trained multimodal understanding large-scale model to perform unified semantic encoding on the raw multimodal data, generating modality-independent deep feature vectors to form an initial multimodal feature pool; performing cross-modal clustering analysis on the vectors in the feature pool, grouping semantically similar vectors into the same feature cluster, with each feature cluster defined as a candidate knowledge concept node; constructing an initial concept relationship network based on the co-occurrence relationship and feature similarity between candidate concept nodes; iteratively optimizing and evolving the network according to preset graph quality evaluation indicators, outputting the final evolvable knowledge graph. This invention achieves the autonomous and unified construction and evolution of a knowledge graph from multimodal data.

The embodiment of the invention provides a Guest OS starting exception detection method and system, and the method comprises the steps: recording the starting moment of a virtual machine instance in response to a detected starting event of the virtual machine instance; under the condition that the virtual machine instance is in the running state, determining whether the virtual machine instance successfully responds to a detection request sent by a host machine where the virtual machine instance is located or not; if it is confirmed that the predefined starting exception judgment condition is met, it is judged that starting of a guest operating system Guest OS on the virtual machine instance is abnormal; wherein the startup exception judgment condition comprises that the virtual machine instance does not successfully respond to the detection request within a preset duration after the startup moment.

The invention relates to the technical field of mobile robot navigation, and discloses a robot adaptive motion control method and system, and the system comprises a step test module which is used for recording odometer data and control quantity data with a timestamp; the parameter fitting module is used for analyzing the odometer data and the control quantity data and establishing a mapping function of the control quantity and the movement speed; the delay compensation module is used for maintaining a historical control quantity queue and predicting the expected pose of the robot when the control instruction sent at the moment t is executed based on the control delay tau; the multi-objective optimization module is used for solving the optimal control quantity; and the online updating module is used for continuously collecting motion data in the motion process of the robot and dynamically updating the mapping function and the key parameters. According to the method, actual motion parameters are automatically obtained through step testing, different robot individuals and performance attenuation scenes are adapted, the problem of model mismatching is solved, and time delay compensation is controlled to reduce positioning deviation.

This application provides a network device management method and apparatus, and a system, and belongs to the communications field. The method includes: obtaining, by a first network device, management information of a service flow, where the management information includes a type of a management operation and an identifier of the service flow; determining a second network device on a forwarding path of the service flow based on the identifier of the service flow, where the second network device is a next-hop network device of the first network device; sending a control packet to the second network device, where the control packet includes the management information, and the control packet is used to indicate the second network device to perform a management operation and obtain information associated with the management operation; and receiving a response packet from the second network device, where the response packet includes the information that is obtained by the second network device and that is associated with the management operation. The network device management method provided in this application is less costly and has relatively high flexibility.

This utility model relates to the field of pipeline inspection technology, specifically to a drone inspection device for oil and gas pipelines. It includes an inspection instrument with an infrared camera fixedly mounted on one side. A removable battery is detachably mounted on the adjacent side via a clamping structure, and the removable battery has a charging slot for independent charging and maintenance. A sliding frame is fixedly mounted at the bottom of the inspection instrument, with hooks slidably embedded within it, allowing the device to be hung on the operator's clothing or mounted as a whole on a drone bracket. A display screen is embedded on the side of the inspection instrument opposite the removable battery. A microcontroller fixedly mounted inside connects to the removable battery via preset contacts to obtain power, and simultaneously connects to the infrared camera via flexible wires to transmit thermal signals and to the display screen to output detection results. This utility model solves the problems of shortened battery life due to reliance on drone power supply and the inconvenience of battery removal affecting continuous operation, achieving seamless switching between autonomous power supply and manual / drone dual-mode operation.

The utility model relates to the technical field of medical instruments, in particular to a hanging frame and a therapeutic apparatus, the hanging frame comprises a handle, a rod frame and a locking mechanism, the handle is provided with a fixed channel, and the side wall of the fixed channel is provided with a first limiting hole; the rod frame is provided with a second limiting hole, and a part of the structure of the rod frame is inserted into the fixing channel. The locking mechanism comprises a driving piece, a transmission piece and a locking piece, the locking piece is arranged in the rod frame in a sliding mode, the driving piece and the transmission piece are both movably arranged on the rod frame, the driving piece is movably matched with the transmission piece, the transmission piece abuts against the locking piece, and when the transmission piece moves relative to the driving piece, the locking piece is locked by the locking piece. And the locking piece can move towards and extend into the first limiting hole or be far away from and separated from the first limiting hole, so that the rod frame is detachably connected with the handle. The utility model mainly aims to provide a hanger, and aims to improve the disassembly and assembly convenience of the hanger.

The application relates to the field of mechanical manufacturing and automatic assembly, in particular to a fan support bearing head device and a method thereof; the device comprises a rack base, a composite spindle, a micro boring cutter, a glue injection head and a control and detection unit; the device receives detection signals through the control unit and drives each component, so that the outer pipe of the composite spindle drives the micro boring cutter to cut grooves under the driving of a fixed cam, and the inner pipe synchronously cooperates with the glue injection head to inject glue; the core is that the cooperative operation of material removal processing and fluid filling is realized in the same device, so that the solidified glue and the micro groove form mechanical interlocking; the device replaces the traditional physical extrusion mode, eliminates the assembly stress and bearing deformation caused by radial extrusion from the physical mechanism, effectively improves the fixing reliability of the axial thrust, and effectively captures the debris by using the negative pressure cavity to suck air flow, and completely solves the pollution problem of the assembly surface.

This invention belongs to the fields of medical image processing and deep learning technology. It discloses a pathological image classification method and system combining stable learning and hybrid enhancement. The method involves acquiring a pathological image dataset, dividing it into a training set, a validation set, a test set, and an external validation set, and preprocessing the dataset. A deep learning network combining stable learning and hybrid enhancement is constructed and trained using the training set. The optimal deep learning network model is obtained using the validation set. The test set and the external validation set are input into the optimal deep learning network model to output the pathological image classification results. This invention utilizes a well-fitting pathological image classification model to effectively improve the overfitting problem and weak recognition ability of traditional models for domain-biased data, enhancing the recognition accuracy of independent and identically distributed data, improving the robustness and generalization ability of the pathological image classification model, and increasing the diagnostic accuracy of pathological images.

The application discloses a kind of multi-agent system output compensation control methods under non-periodic DoS attack, first construct follower fuzzy approximation system model and non-periodic DoS attack model;When attack activation causes output loss, trigger attack compensation mechanism to keep last output to provide continuous compensation signal;Switch fuzzy state observer is constructed to estimate agent state, and adaptive law in attack activation and dormancy stage is designed using Lyapunov function;Global preset performance function is constructed to establish dynamic performance boundary, and consistency tracking error and normalized conversion error are calculated to construct barrier function;Distributed adaptive security control law is obtained by combining backstepping method.The application overcomes the defect that observation performance decreases due to output loss in attack stage and traditional performance function depends on error initial value, guarantees global preset performance and high-precision state estimation, solves the cooperative control problem under non-periodic DoS attack.

An upturning type door body hinge with a random stop function comprises a driving arm fixedly connected to a door body and a fixing frame fixedly installed on a box body, one end of the fixing frame is rotationally provided with an output shaft, the outer cylindrical surfaces of the two ends of the output shaft are sunken to form an outer connecting table and an inner connecting table respectively, and the outer connecting table and the inner connecting table are connected with the driving arm. One end of the connecting rod is clamped and mounted on the inner connecting table, and the driving arm is clamped and connected to the outer connecting table; the other end of the connecting rod is hinged to a pull arm, the other end of the pull arm is connected with a sliding rod, and the sliding rod is installed in a movable groove formed in the fixing frame in a sliding mode. A spring is arranged between the sliding rod and the fixed frame, when the door body is turned upwards, the connecting rod is driven to swing through the output shaft, the pull arm stretches the spring, and the tension of the spring is balanced with the gravity of the door body, so that the door body can be paused at any position.

The utility model provides a kind of machining center contact type mechanical tool changing structure, it is related to mechanical tool changing structure technical field, including stand mechanism, the front of the stand mechanism is equipped with main shaft box, the front of the stand mechanism is close to top end place, and punch arm mechanism is installed in the inside of main shaft box, the stand mechanism includes stand main body, the top end of the stand main body is equipped with motor with output end towards bottom, the front of the stand main body is close to top end place and is equipped with punch block, the inside of main shaft box is close to the front place and is equipped with main shaft body, the top end of the main shaft box is close to left side place and is screw-connected with first fastening screw, the periphery of the first fastening screw is equipped with tension spring, the top end of the main shaft box is close to the front place and is screw-connected with second fastening screw, mechanical quick tool changing can be realized, additional transmission device is not needed, with the advantages, such as quick tool changing efficiency, small size, cost saving.

The invention discloses a construction method and system of a medical dialogue thinking chain data set. The method comprises the following steps: firstly, preprocessing original medical session data, and screening a target session round to obtain structured data; then, in combination with downstream tasks such as medical dialogue generation and disease diagnosis assistance, a forward reasoning Prompt template is constructed, and the problems that an existing method depends on the response of a real doctor, and a thinking chain lacks clinical authenticity and depth are solved. And through a forward reasoning-review-reflection mechanism, on the basis of structured data and a Prompt template, generating a thinking chain and reaction data which conform to real clinical reasoning logic, and finally forming a medical dialogue thinking chain data set. The data set is consistent with clinical diagnosis logic, and does not need to depend on the reaction of an original doctor. The diagnosis accuracy, the response pertinence and the interpretability of the medical dialogue system can be effectively improved, and powerful support is provided for constructing the reliable and interpretable medical dialogue system.

The invention relates to the technical field of vibration measurement, and discloses a wireless energy transmission vibration sensor for vibration monitoring of a generator runner, which comprises a heat insulation and conduction protection mechanism and a transmitting circuit module, an acquisition mechanism, a wireless module, a receiving circuit module, a signal processing unit and a magnetostriction actuator are arranged in the heat insulation and conduction protection mechanism, the wireless module is arranged at the output end of the acquisition mechanism, and the receiving circuit module is arranged at the input end of the wireless module and the input end of the signal processing unit. According to the wireless energy transmission vibration sensor for generator runner vibration monitoring, the resonant frequency can be controlled through the magnetostriction actuator when the sensor is used, that is, the coupling capacitance is adjusted through the magnetostriction actuator according to the characteristic that the magnetostriction rod is stretched or shortened in the magnetization direction when being magnetized, and then the resonant frequency is adjusted; therefore, self-tuning is realized, and high-efficiency transmission is ensured.

This invention discloses an intelligent adaptive dose verification calculation method and system. The intelligent adaptive dose verification calculation method includes: obtaining preliminary dose distribution data for a patient based on a deep learning-based dose prediction model; dynamically assigning variance reduction parameters to particles in a Monte Carlo simulation based on the high and low dose distribution information in the preliminary dose distribution data, wherein the variance reduction parameters include simulation weights, with simulation weights assigned to particles predicted to be in high-dose regions being less than those assigned to particles predicted to be in low-dose regions; and using a Monte Carlo dose calculation engine, simulating the planned beam based on the dynamically assigned variance reduction parameters to obtain high-precision verification dose distribution data. This invention's intelligent adaptive dose verification calculation method ensures the physical authenticity of the final verification dose distribution and the credibility of the "gold standard."

This invention relates to an automatic attitude calibration method, apparatus, device, and medium for a TMR current sensor. The method includes the following steps: acquiring reference attitude information detected by the attitude sensor in the TMR current sensor during initial installation; acquiring the magnetic signal sensed and output by the TMR sensing element in the TMR current sensor in real time during use; detecting whether the TMR current sensor undergoes positional change during use; when a positional change is detected, acquiring the current attitude information detected by the attitude sensor at the current position in real time, calculating and outputting the measured current after the position change based on the magnetic signal, the current attitude information, and the reference attitude information; when no positional change is detected, calculating and outputting the measured current based on the magnetic signal and the reference attitude information. Compared with the prior art, this invention has the advantages of improving the measurement accuracy of the current sensor and having low deployment cost.