29 results about "Active sensing" patented technology

The application provides a detection device and method for fasteners and a fastener, and belongs to the technical fields of precision manufacturing and industrial visual detection. The method comprises the following steps: performing multi-spectrum active illumination configuration on a fastener production line to generate multi-spectrum illumination parameter data; deploying a multi-spectrum light source array according to the multi-spectrum illumination parameter data to construct a multi-spectrum active sensing network; and through the fusion application of multi-spectrum active illumination and neural radiation field three-dimensional reconstruction technology, the recognition accuracy of the surface defects of the fastener is significantly improved, the geometric size measurement is realized to the level of ±0.003 mm, and the detection rate of the micro defects is improved to 99.5%.

This invention discloses a compliant connection platform, comprising a first cross spring group, a frame unit, and a second cross spring group sequentially connected along the axial direction. The frame unit includes a support component and a first connecting portion and a second connecting portion respectively disposed at both ends of the support component. The first cross spring group is fixedly connected to the first connecting portion, and the second cross spring group is fixedly connected to the second connecting portion. The support component can deform in a plane perpendicular to the axial direction. The first and second cross spring groups are orthogonally arranged in a plane perpendicular to the axial direction, enabling the compliant connection platform to achieve five degrees of freedom passive compliance compensation. This invention achieves automatic alignment of docking points without active sensing control, has a simplified and compact structure, significantly reduces the accuracy requirements for docking and assembly, and is suitable for the rapid docking and reconfiguration needs of modular vehicles under complex operating conditions.

The application discloses an AI teaching feedback and self-evolution method and system based on active sensing and mixed thinking mechanism. The method comprises the following steps: generating student behavior characteristics, emotion characteristics and teacher-student voice characteristics; generating student state-teaching content structured information through a semantic analysis model; obtaining behavior emotion analysis information through a thinking scheduling module core model; generating visual perception module optimization parameters, semantic analysis module optimization parameters and thinking scheduling module optimization parameters according to all historical high-quality classroom video annotated efficient interaction segments and behavior emotion analysis information of each period; and updating the visual perception type model, the semantic analysis model and the thinking scheduling module core model. The application constructs a self-evolution system, which can continuously improve the sensing accuracy, analysis accuracy and scene adaptability without manual intervention, and completely solves the problems of static and no evolution ability of the traditional teaching analysis system.

The present application relates to the technical field of pipeline hydrostatic test, and especially relates to a pressure water supplementing and metering device for municipal pipeline hydrostatic test, which comprises a water tank, a booster pump, a flowmeter and a control cabinet, a program controller is arranged in the control cabinet, further comprising a pressure sensor and a vibration sensor arranged along the axial direction of the pipeline to be tested, and an active tunable damper arranged between the pump outlet and the pipeline; the program controller synchronously collects pressure and vibration data, carries out correlation analysis on the data, generates a pressure wave space-time propagation graph, calculates a water hammer risk index and divides a risk level through a nonlinear fusion model according to the dominant frequency, amplitude, propagation speed change rate and vibration energy of the pressure wave, selects an inhibition strategy according to the risk level, calculates a pressure peak value reduction rate and a wave fluctuation decay time according to the data after inhibition, and generates an effect evaluation value. The present application realizes active sensing, graded inhibition and effect evaluation of water hammer risk, and improves the safety and intelligent level of the hydrostatic test.

The application discloses a monocular depth estimation method based on active sensing, comprising: acquiring a monocular RGB image of a surgical area; acquiring an initial sparse depth measurement set; inputting the RGB image and the sparse depth measurement set into a depth estimation integrated model to output an integrated dense depth prediction map and a prediction uncertainty map; calculating a gradient of the prediction uncertainty map to a measurement value in the current sparse depth measurement set; selecting a next group of points to be probed by using a Stein variational gradient descent algorithm based on the gradient map; controlling a tactile sensing end of a surgical robot to move to an anatomical position corresponding to the point to be probed to perform contact measurement; updating the newly acquired depth measurement value to the sparse depth measurement set; iteratively optimizing the depth estimation and uncertainty evaluation, and outputting a final integrated dense depth prediction map as a surgical navigation three-dimensional scene model. The application can enhance the accuracy of monocular depth estimation by sparse proprioceptive depth measurement.

The application provides an intelligent pollution cleaning method based on multi-modal deep learning, real-time monitoring of a pollution accumulation area is performed, and a real-time picture of a trash rack area is transmitted, a pollution detection data set is generated, and a pollution recognition algorithm based on deep learning is trained, the pollution recognition algorithm is used to recognize pollution in the real-time picture of the trash rack area and mark position information of the pollution, a multi-modal deep learning model is established to determine whether a pollution cleaning device is started, and factors such as an area of the pollution accumulation area, a water level difference of an inlet, weather conditions and rainfall are comprehensively considered, when an output result of the model reaches a preset value, it is determined that the pollution cleaning device starting requirement is met, through intelligent identification, automatic operation and informationized collaboration of the whole process of the hydropower station pollution cleaning system, the application improves the pollution cleaning efficiency of the trash rack of the hydropower station inlet and the reliability of the pollution cleaning device, and promotes the intelligent leap of the pollution cleaning system from passive response to active sensing.

This invention belongs to the field of road engineering technology, specifically relating to an integrated self-adhesive rubber asphalt asphalt crack repair and monitoring system. The system includes: a self-adhesive rubber asphalt composite material layer for filling cracks and absorbing stress; a flexible sensing network unit embedded within the layer to sense strain and temperature; a data acquisition and signal conditioning module; a self-powered energy capture system that utilizes ambient energy to power the system; an edge computing terminal for localized damage assessment; and a remote cloud management platform. This system improves bonding strength and low-temperature flexibility through material modification, and combines embedded sensing and self-powered technology to achieve a leap from passive filling to active sensing of the repair body. It solves the problems of difficult power supply in the field and lack of coordination in monitoring, enabling real-time feedback of the interface mechanical state, significantly improving the scientific nature and predictability of highway maintenance, and reducing overall operation and maintenance costs.

The application relates to the field of artificial intelligence, and in particular discloses a digital human intelligent service system for human-computer interaction; the system comprises a sensing end, an inference end and an interaction end; the sensing end is arranged in a physical space of a service site, collects behavior data of a user in a service scene and environment-related information in real time through a multi-modal sensor array, and sends original sensing data collected to the inference end; the inference end is used for receiving the original sensing data, performing identity recognition and intention inference processing on the original sensing data, generating a service instruction sequence matched with a current service context of the user according to an inference result, and sending the service instruction sequence to the interaction end; and the interaction end is used for receiving the service instruction sequence, driving a digital human virtual image to perform corresponding service actions. The application replaces the dependence on explicit operation of a user in a traditional instruction driving mode through active sensing of a user identity and behavior context by the sensing end.

This invention discloses a method, device, and medium for dynamic sensing across the entire distribution network, belonging to the field of power system operation and control technology. The method includes constructing an active sensing layer to collect transient waveform data; fusing the transient waveform data with distribution network operation data to generate a global data correlation cloud map; performing reverse backtracking based on the global data correlation cloud map when a tripping event occurs to form an event chain; constructing an operational stress accumulation model based on the event chain for latent defect early warning; sensing the defect source based on the latent defect early warning results; and generating a dynamic sensing map of the distribution network by integrating the event chain, early warning results, and defect source. This invention achieves end-to-end connectivity and closed-loop optimization from bottom-level data sensing to top-level decision support.

The application discloses a fruit autonomous picking method and system based on active perception and visual arm cooperative optimization, and relates to the technical field of robot control. The method comprises the following steps: acquiring a color image and a depth image of a working environment, and performing semantic segmentation on the color image by using a preset zero-sample learning model; performing semantic point cloud extraction on the depth image based on a mask corresponding to a target category, and constructing a fruit tree environment map containing spatial information and semantic information; performing candidate target screening and extraction according to the fruit tree environment map, and generating a candidate grasping pose according to the screened candidate target; performing kinematic reachability analysis on all candidate targets and candidate grasping poses, and calculating a joint score; determining a next grasping object according to the joint score; and picking the next grasping object according to a preset trajectory. The application deeply couples active visual perception and mechanical arm grasping action, and realizes efficient fruit autonomous picking of the robot.

This invention discloses a novel method and system for actively sensing the operating status of a power distribution system. The method includes: identifying multiple monitoring points in the power distribution system; actively injecting a power frequency signal into the neutral point and simultaneously acquiring the three-phase and zero-sequence electrical quantities at each monitoring point; then decomposing the quantities into positive-sequence, negative-sequence, and composite zero-sequence components using the symmetrical component method; and comparing the composite zero-sequence quantity with the measured zero-sequence quantity to accurately determine whether the monitoring terminal is functioning correctly, thereby identifying reliable diagnostic sections and avoiding misjudgments caused by measurement anomalies. Based on this, the load imbalance is quantified by the ratio of negative-sequence to positive-sequence current at normal monitoring points, providing a direct assessment of the economic efficiency of the power grid operation. Simultaneously, the zero-sequence impedance is calculated using the zero-sequence voltage and current under normal conditions to effectively verify the protection settings of the switching equipment. Finally, a state evaluation equation integrating multiple characteristic quantities is established to output the normal or fault state of the power grid in real time. This method can actively sense the operating status of the power distribution system during normal operation.

The present application belongs to the technical field of mine hoist, and proposes a cage control system and control method based on visual feature extraction and deformation compensation, which comprises a cage integrated terminal installed on the cage of mine, a wireless relay network arranged along the shaft, and a ground control center arranged on the ground; the cage integrated terminal comprises an explosion-proof box body; the present application sets the integrated terminal on the cage, and configures a weighing sensor, an acceleration sensor, and an image acquisition and processing module, so that the cage has real-time sensing capability for its own load state and running state; meanwhile, the image feature extraction module is combined to dataize the internal situation of the cage, realizing the change from traditional indirect detection relying on the hoist side to active sensing at the cage end; and in the case of abnormal wireless bridge communication, the key state data can still be sent through feature extraction and LORA communication module, thereby ensuring the continuous availability of the internal personnel and material state of the cage.

A LIDAR system that includes optics configured to (a) transmit, using a scanner, a transmitted signal, and (b) receive, using the scanner, reflections from objects, during a scan segment time window that corresponds to a scan segment of a field of view (FOV) of the LIDAR system; a sensing unit that includes multiple sensing elements; and a controller that is arranged to (a) activate the multiple sensing elements before starting to receive the reflections, and (b) selectively deactivate at least some of the multiple sensing elements, based on a scan direction of the scan segment, thereby reducing a number of active sensing elements during the scan segment time window.

This invention provides a calibration-free path reconstruction method for buried cables based on Topo-NeRF and active sensing. Addressing the issues of traditional magnetic field inversion relying on calibration and being prone to mismatch in environments with strong interference such as reinforced concrete, the method preprocesses raw data to obtain a spatiotemporal feature matrix. This matrix undergoes feature extraction, spatial alignment, and drift compensation to eliminate extrinsic parameter calibration, resulting in multi-source fused data. This multi-source fused data is combined to characterize the cable's magnetic field using a continuous three-dimensional topological manifold. An implicit neural radiation field is constructed and coupled with differentiable electromagnetic rendering. The cable body and branch nodes are separated within a persistently cohomologically constrained topological bottleneck layer. The initial reconstruction results are combined with conditional diffusion and ground-penetrating radar dielectric priors to enhance low signal-to-noise ratio weak fields, ensuring consistency between the reconstruction results and the underground physical structure. Based on reconstruction uncertainties, topological entropy is calculated, and online acquisition trajectories are planned to form a closed-loop active sensing system. Finally, a three-dimensional path model is output, enabling rapid detection of underground cables.

This invention relates to the field of server fault detection technology, providing a method and system for proactive fault detection in computer servers. The invention updates the phased detection data of the server cluster; performs dynamic time planning and detection control for proactive fault detection based on the previous detection time and data; extracts non-standard detection data from the phased detection data; performs fault risk analysis on the non-standard detection data; and identifies risky servers and provides off-peak warnings when there are hidden fault risks. It can extract previous detection data from the phased detection data for dynamic time planning and detection control, achieving adaptive adjustment of the detection frequency. This avoids generating a large amount of redundant detection data, reduces computational burden, improves operating efficiency, and ensures timely detection of server faults, enhancing operational stability and reliability.

Example embodiments relate to active sensing in joint communication and sensing. An apparatus can transmit, to a sensing management entity, an active sensing request for sensing with respect to a terminal device; receive, from the sensing management entity, a sensing beam configuration indicating a beam group for the sensing; detect, based on the sensing beam configuration, at least one beam in the beam group; transmit, to the sensing management entity, information of the at least one detected beam; and receive, from the sensing management entity, an active sensing response including a sensing result.

The application discloses an AI native intelligent interaction terminal with active sensing and emergency response capability, and belongs to the technical field of artificial intelligence terminals and man-machine interaction. The terminal integrates a multi-modal interaction module, a local AI inference engine capable of offline voice recognition, negotiation strategy generation and multi-language real-time translation, and an end-cloud collaborative scheduling unit. The terminal adopts an independent transfer box form with integrated charging function or an integrated terminal form with integrated charging bin, and can provide charging and power supply functions for matched earphones, glasses and other peripherals. The core feature of the terminal is the built-in motion state sensor and environmental sound recognition model, which can actively sense user falls, impacts or dangerous acoustic events, and automatically trigger an emergency response program. The terminal also supports a covert real-time word prompting function based on voiceprint recognition and cognitive state detection, provides long-term personalized services through a localized user cognitive model, and supports seamless task relay with a local AI host. The application deeply integrates intelligent interaction, active safety guardianship and privacy protection, and can be widely applied to various independent AI interaction scenes.

The application relates to the technical field of tunnel monitoring, and relates to a tunnel surrounding rock state monitoring device and method based on a piezoelectric intelligent anchor rod. The device comprises an anchor rod unit and a sensing unit. The anchor rod unit is arranged on the tunnel surrounding rock in the form of a rectangular array. The anchor rod unit comprises four wave-emitting anchor rods and one wave-receiving anchor rod. The wave-receiving anchor rod is arranged at the center of the rectangle, and the four wave-emitting anchor rods are arranged at the four vertices of the rectangle. The sensing unit is arranged inside the wave-receiving anchor rod and the wave-emitting anchor rod. The sensing unit is used for monitoring the surrounding rock structure damage. The application can expand the monitoring range from a single point to an area. The active sensing principle is used to sensitively capture the wave propagation characteristic change caused by the early damage such as micro-cracks and looseness in the surrounding rock, so that early warning can be realized before macroscopic deformation occurs.

The application belongs to the technical field of satellite-ground communication and physical layer security, and particularly relates to a satellite sensing assisted secure communication method based on dynamic uncertainty modeling. The method comprises the following steps: constructing a sensing-integrated transmission architecture, using a dual-function beam to actively detect potential threat areas while transmitting data to downlink users; performing primary positioning by capturing target echo signals, and dynamically calculating the uncertainty boundary of positioning in combination with echo signal-to-noise ratio; establishing a multi-dimensional channel uncertainty model, including a user-side angle uncertainty model established for the high-speed movement of low-orbit satellites, and a norm-bounded uncertainty region constructed for eavesdroppers based on the above boundary; taking the maximization of the worst-case secrecy rate as a global optimization goal, jointly designing optimal beamforming and radar receiving filters under the constraints of sensing, power and security; and using semi-positive relaxation and convex-concave process algorithms to convert the non-convex problem into a convex optimization sub-problem for iterative solution. The application realizes a closed loop of active sensing and mathematical modeling, and significantly improves the dynamic adaptability and robustness of satellite secure communication.

This application relates to the field of intelligent storage technology, and provides an automatic fault repair method and related equipment based on storage cell architecture. By collecting defect-sensitive electrical parameters of the storage cells, the dynamic defect entropy of the accumulated defect level within the storage medium is calculated. When the dynamic defect entropy exceeds a preset repair threshold, the voltage amplitude and pulse width of a repair pulse are adaptively generated. Based on the voltage amplitude and pulse width of the repair pulse, a repair excitation matching the storage cell type is applied to the storage cell to drive the rearrangement of atoms, dipoles, or magnetic moments in the storage medium. After repair, the defect-sensitive electrical parameters of the same address cell in the storage cell are collected, and the dynamic defect entropy after repair is recalculated. If the repair is ineffective, the repair parameters are strengthened and the repair is repeated. This achieves active sensing and material-level lattice reconstruction repair of metastable defects within the storage medium, thereby significantly extending the effective service life of the storage cells under radiation accumulation conditions.

The application discloses an Ethernet power supply switch and a control method, equipment, storage medium and program product thereof, relates to the technical field of intelligent switch management, and the Ethernet power supply switch comprises an Ethernet intelligent control module, an artificial intelligence management and control module and a sensing module. Since the multi-dimensional environment data and the service traffic data are acquired through the sensing module, active sensing of various types of service data is realized, the Ethernet power supply switch actively determines the service priority, and the terminal negotiation result is no longer relied on, thereby providing a basis for subsequent scheduling. The artificial intelligence management and control module generates a bandwidth allocation strategy and a power supply priority adjustment instruction according to the multi-dimensional data, realizes intelligent decision of resource scheduling, makes the power supply priority and the bandwidth allocation cooperate, guarantees the end-to-end quality of core services, can realize rapid response to emergencies, and avoids the problem of unbalanced bandwidth allocation during peak periods.

The application provides a magnetic field data monitoring and analyzing method of a passive tag, the passive tag, equipment and products, and relates to the field of logistics. Through a radio frequency energy power supply mechanism, a threshold voltage starting monitoring mechanism and a magnetic field threshold value triggering sampling mechanism, the passive tag can intelligently perceive magnetic field abnormal events without a battery, has a long service life and no maintenance cost, and overcomes the defects of high cost, battery replacement, complex maintenance and inadaptability to complex environments of an active sensing tag. The passive tag can still maintain listening in a low energy environment, consumes energy for collection when an abnormal event occurs, and ensures that the energy is used for valuable event sampling. The method enables the passive tag to not only provide static information such as cargo information and shelf information, but also lays a data foundation for background intelligent analysis of environmental abnormal events such as illegal interference, dumping, collision and illegal opening during the logistics transportation or storage process.

The application discloses a vehicle anti-collision early warning method and system based on active sensing control, and comprises the following steps: S1, scanning a target vehicle to obtain the relative position between the target vehicle and the ego vehicle; S2, estimating the motion state of the target vehicle based on the relative position as the initial condition; S3, delimiting a target area, and searching and tracking the target vehicle in the target area based on the motion state of the target vehicle; S4, obtaining the position and speed of the target vehicle in real time based on the tracking of the target vehicle; and S5, determining the collision risk between the target vehicle and the ego vehicle based on the position and speed of the target vehicle, and performing anti-collision early warning according to the collision risk. The application can timely early warn the collision risk of the vehicle behind and the vehicle at the side and rear of the ego vehicle, achieves the effect of preventing vehicle collision, and guarantees the driving safety of the vehicle.

A complementary metal-oxide-semiconductor sensor array includes an active sensing area of pixels arranged in an array with a pitch, each pixel including an exposed surface electrode alongside switches and logic gates, and non-overlapping clocks configured to rapidly charge and discharge the exposed surface electrode, wherein control signals steer a switched output current between shared column outputs.

This invention discloses a novel active sensing method for grid-connected entities based on deep learning. The method includes collecting operational data from new grid-connected entities to obtain raw power signals; extracting feature values ​​using signal processing techniques to construct an active sensing dataset, which is then divided into training and testing sets; constructing a deep learning network model and training it using the training set to obtain the active sensing model; and inputting the testing set into the model to achieve active sensing of new grid-connected entities. This invention provides an effective data source through data acquisition and feature extraction. The constructed model can intelligently identify new grid-connected entities such as unreported distributed photovoltaic systems, energy storage systems, and electric vehicle charging piles, effectively solving the problem of difficult-to-detect implicit connections. This achieves accurate identification and supervision of new grid-connected entities on the user side, significantly improving identification accuracy and supervision efficiency.

A depth map of a scene can be generated by: projecting a lighting pattern onto the scene using a light projector; capturing a camera image of the scene using a camera; generating multiple reference true depth values ​​for sample pixels of the camera image using a computing device, based at least in part on the lighting pattern; and estimating the depth map of the scene based at least in part on the camera image and the reference true depth values ​​of the sample pixels.

This invention relates to the field of intelligent monitoring technology for power systems, and discloses a control device and method for a power grid edge monitoring terminal combined with active energy sensing. The device includes: a micro-meteorological sensing module for real-time acquisition of meteorological parameters of the environment where the monitoring terminal is located; an energy prediction module for predicting expected energy gains within a future time window using a time-series prediction model; an energy assessment module for calculating an energy availability score; a mode decision module for dynamically switching the operating mode of the monitoring terminal based on a comparison of the energy availability score with a preset threshold; a defect detection module for real-time defect detection of acquired images or video streams; and a verification decision module for determining whether to trigger a second power consumption mode for defect verification when the defect detection module detects a suspected defect in a first power consumption mode. This invention can significantly reduce the false negative rate of dynamic defects without increasing hardware costs or battery capacity.

This invention discloses a proactive perception closed-loop inspection method, device, system, and medium based on a risk field. The method employs a unified voxelized multimodal risk field data structure, a reproducible update mechanism for the multimodal risk field, a unified proactive perception scoring mechanism, a quantifiable and reproducible coverage quality model, and a coverage evaluation-driven supplementary observation generation and minimum increment incorporation mechanism. This enables a quadruped robot and its onboard sensors to automatically construct a multimodal risk field in multiple scenarios and conduct proactive perception adaptive inspections under the constraints of this risk field. Supplementary observation actions are automatically generated when there is insufficient coverage, high uncertainty, unconfirmed anomalies, path obstruction, or sudden risk changes, thereby improving inspection coverage, the reliability of anomaly detection, and the stability of task execution.