137 results about "Sensor observation" patented technology

The invention provides a multi-sensor fusion robot grabbing path planning method and device, and relates to the field of automation. The method comprises the following steps: constructing a first time correction relation based on observation parameters of a multi-source sensor; correcting the relation based on the first time; inputting the hardware timestamp synchronized to the master clock source into the extended state estimation model; iteratively updating the observation parameters of the multi-source sensor based on the extended state estimation model, and constructing a second time correction relation according to the updated observation parameters of the multi-source sensor; constructing a pose continuous time trajectory function corresponding to the target robot based on the second time correction relation; the continuous time trajectory function and the multi-source sensor observation parameters serve as input, and a grabbing path corresponding to the target robot is output through a time sequence consistency constrained joint optimization model; and controlling the target robot to execute grabbing operation based on the grabbing path. The problem that in the prior art, the precision of a robot grabbing path is low is solved.

The invention relates to the technical field of path planning, in particular to a pesticide spraying robot positioning method and system based on graph neural network multi-source fusion. Performing data preprocessing based on the acquired multi-source sensor information; graph node multi-modal data embedding is carried out on the preprocessed information based on perception information mapping; a multi-modal space-time diagram structure is built based on graph node multi-modal data embedding; performing graph feature propagation and fusion based on space-time attention weighting; and carrying out positioning calculation and path decision on the fused features based on spatial-temporal feature constraints. According to the method, a sensor observation graph structure is constructed, multi-source information such as GPS, IMU, vision, radar and the like is modeled in a node form, and nonlinear feature fusion is realized through graph convolution and an attention mechanism.

A system captures via one or more sensors of a computing device, data of an environment observed by the one or more sensors at a first timeslot, and stores the data in a data store as a first portion of a timelapse memory experience. The system also captures, via the one or more sensors of a computing device, data of the environment observed by the one or more sensors at a second timeslot, and stores the data in a data store as a second portion of the timelapse memory experience. The system additionally associates the timelapse memory experience with a memory experience trigger, wherein the memory experience trigger can initiate a presentation of the timelapse memory experience.

The invention belongs to the technical field of distributed multi-sensor passive detection multi-target tracking. The invention provides a factor graph optimization passive detection multi-target tracking method based on a Gaussian mixture model. According to the embodiment of the invention, the multi-target batch number is distributed by constructing the distributed passive sensor cooperative coordinate system and combining the multi-target identity judgment result between the two sensors; calculating direction finding lines based on two-dimensional observation of a sensor, combining the direction finding lines of the same batch number, obtaining a multi-target position estimation point set through a least square method, and obtaining a multi-target coarse positioning point through weighted fusion; modeling by adopting a Gaussian mixture model, fusing measurement distribution characteristics, solving parameters through an expectation maximization algorithm, and completing solvable conversion of an optimization problem; a factor graph optimization model containing multiple factors is constructed, state estimation is achieved through sliding window optimization, and track association and state updating are completed in combination with the Mahalanobis distance and the Hungary algorithm; and the passive detection multi-target tracking performance of the distributed sensor is effectively improved.

The invention discloses an illegal unmanned aerial vehicle source tracing method and system based on visible light, thermal imaging and radio frequency fingerprint fusion. The method comprises the following steps: synchronously acquiring target visible light, a thermal infrared image and a radio frequency signal; extracting appearance, thermal radiation and radio frequency fingerprint features in parallel; multi-modal features are dynamically weighted and fused through an adaptive cross-modal attention fusion network (ACMAF-Net), and unmanned aerial vehicle model identification and individual association are completed; reconstructing a three-dimensional track of the unmanned aerial vehicle by using multi-sensor observation data; the possible positions of a take-off point and an operator are deduced by combining track features and geographic information. The system comprises a multi-mode sensing subsystem, a data processing and fusion subsystem, a traceability analysis subsystem, a display control storage subsystem and the like. According to the invention, through three-mode complementary fusion, the limitation of a single technology is overcome, all-weather and high-precision detection identification and effective traceability of the illegal unmanned aerial vehicle are realized, and the low-altitude security and protection capability is remarkably improved.

The invention provides an agricultural robot-oriented state machine modeling and exception response method, which comprises the following steps of: modularly dividing a main state space, introducing an external driving event set and an exception state set, and constructing a multi-layer state transition function comprising main state transition, exception triggering and exception recovery; and in combination with a multi-sensor observation vector configured by the agricultural robot, a conditional probability distribution model between a state and an observed quantity is established, and parallel reasoning and dynamic switching of a main state and an abnormal state are realized. In addition, an abnormal state evaluation function is designed to carry out quantitative analysis on various abnormities, and a grading response strategy is set. The method is suitable for various agricultural operation scenes, and has strong behavior modeling expression capability and abnormal response capability.

The invention belongs to the technical field of unmanned aerial vehicle surveying and mapping, and particularly relates to an unmanned aerial vehicle cluster dynamic measurement method based on cooperative positioning and multi-source fusion, and the method comprises the steps: constructing a cooperative operation mechanism of a dynamic reference station network and an observation unmanned aerial vehicle cluster, and combining global positioning data provided by RTK and relative distance measurement constraints provided by UWB, and multi-sensor observation information is fused to realize displacement calculation with centimeter-level precision. In addition, the flight path of the unmanned aerial vehicle cluster is optimized through a clustering algorithm and a heuristic optimization algorithm, and the measurement efficiency of the unmanned aerial vehicle cluster is further improved.

Provided are methods, systems, devices, and tangible non-transitory computer readable media for providing data including vehicle map service data. The disclosed technology can perform operations including accessing vehicle map service data including information associated with a geographic area and sensor observations of a vehicle. The vehicle map service data is based on a vehicle map service protocol specifying layers associated with portions of the vehicle map service data to which each client system is subscribed. Further, the layers to which the client systems are subscribed can be determined and access to the plurality of layers to which each of the vehicle client systems is subscribed can be provided to each of the client systems. Access to the layers can include authorization to send or receive one or more portions of the vehicle map service data associated with a corresponding layer.

The invention discloses a multi-sensor data processing method suitable for high-altitude meteorological detection, and the method comprises the steps: collecting meteorological data from a plurality of different types of sensors, and obtaining initial data; missing values and abnormal values in the initial data are detected and marked; carrying out data interpolation by adopting a lightweight missing detection perception type neural network to obtain a complete data sequence; performing point estimation and uncertainty joint inference on the complete data sequence to obtain a corresponding point predicted value and uncertainty; data conflicts of different sensor data in overlapping measurement intervals are solved, and a fused single observation value and estimation uncertainty are obtained; and adopting a multi-expert hybrid model to dynamically select or perform weighted fusion on outputs of different expert models based on the fused single observation value and the estimated uncertainty to obtain a fused meteorological parameter set. According to the invention, real-time processing and smooth switching of observation data of multiple sensors are realized, and continuity, stability and precision of detection data are effectively improved.

The invention discloses a space-based distributed multi-sensor multi-target tracking and positioning method. The method comprises the following steps: firstly, acquiring a grayscale image of each sensor, and extracting coordinates of a target in the image; and matching a serial number for the extracted target, and taking a first frame coordinate and an image sequence of the target as tracking input. The tracking process is based on a discriminant correlation filter, an energy concentration threshold value of small target features is added, a double-constraint trajectory correlation model of target spacing and speed similarity between continuous frames is constructed, and a Kalman filtering algorithm is used to predict a target position. An adaptive covariance cross fusion positioning algorithm is designed among multiple sensors, a geometrical relationship between multiple detection satellites and a target is established, and the three-dimensional position of the target is calculated by using sensor observation parameters as a basis for adjusting fusion weights. The method can significantly improve the target tracking precision and suppress the positioning offset problem caused by instability of part of sensors, can adapt to the dynamic change of a distributed detection scene, and ensures the accurate target positioning.

The invention relates to the technical field of limited space fire hazard flashover prediction. The invention provides a limited space fire hazard flashover prediction method and system based on a physical information neural network. The method comprises the following steps: acquiring sensor observation data and initial fuel distribution data; inputting the space-time coordinate vector and the initial fuel distribution data into a Mamba neural network to obtain predicted values of temperature, incident radiation and fuel mass fraction; calculating a space-time derivative of the predicted value, and constructing a physical residual term containing energy conservation, radiation transfer and fuel consumption equation residual, a boundary residual term and an observation residual term; weighted combination is carried out on the three types of residual errors to construct a loss function, and network parameters and to-be-learned physical parameters are updated through an optimization algorithm; and predicting a temperature field based on the updated network, and judging the time of detonation when the temperature reaches a detonation critical value. The problems of poor reliability and prediction deviation caused by lack of physical rule constraints of an existing data-driven fire prediction model are solved.

The present invention provides a method for mining the observation capabilities of physical sensors and social media for disaster missions. It addresses the current situation in which traditional disaster perception research ignores the observation needs of social factors, lacks the construction of comprehensive disaster observation tasks, and the observation capabilities of physical sensors and social media. A technical process for mining the observation capabilities of physical sensors and social media for disaster observation tasks is formed to meet comprehensive disaster observation needs. Compared with existing sensor capability mining methods, this method can cover more observation factors of disaster phenomena, can effectively enhance the management of sensor observation capabilities for disaster observation tasks, provide sensor observation capabilities for observation tasks in different fields and at different granularities, and provide a basis for disaster sensor planning and selection and improving disaster observation plans. It has been proven that this method is a practical and reliable method that is conducive to increasing observable factors and enhancing sensor observation capabilities in disaster observation tasks.

The invention belongs to the technical field of network security, and discloses an optimization filter-based cyber-physical system cascading failure security assessment detection method, which comprises the following steps of: performing data acquisition on a power system by analyzing a CPS (Cyber-Physical System) architecture, and respectively constructing a data uploading model and a data issuing model; then, learning an optimal filter over time by using the sensor observations to filter out malicious sensor observations while retaining other sensor measurements during data stream transmission; and finally, whether the data stream is attacked in the transmission process is judged by analyzing whether the data transmission quantity is suddenly and abnormally increased or decreased, a detector is optimized, and the false alarm rate is reduced. According to the method, cascading failure analysis is carried out on the CPS through the dynamic complex network, the analysis accuracy is enhanced, and therefore the reliability of the CPS is improved.

The present application relates to a kind of infrared scene quantitative radiance calculation methods based on light ray tracking module, comprising: obtaining target mask information image, the radiance image of infrared scene, depth information image and position information image, sensor position information, sensor field of view and sensor observation direction;Infrared scene is divided into several view cones according to the resolution of target radiance area image, wherein several view cones and the object surface in infrared scene intersect to form several face elements;The radiance of each face element at the sensor is calculated by the theorem of solid angle projection;The radiance of several face elements at the sensor is accumulated, and the radiance of infrared scene to the sensor is obtained.The method supports full scene, single target, multiple targets as radiation energy source, and can be applied to high complex infrared scene simulation work;Without additional data support, the radiance at the sensor can be calculated without performing redundant work.

The present invention corrects the axial offset of a sensor. The present invention is a sensor alignment device comprising: a target position relationship processing unit that outputs positional relationship information of a first target and a second target; a sensor observation information processing unit that transforms the observation results of the first target and the second target into a predetermined unified coordinate system based on coordinate transformation parameters, synchronizes the time to a predetermined timing, and extracts first target information indicating the position of the first target and second target information indicating the position of the second target; a position estimation unit that uses the first target information, the second target information, and the positional relationship information to estimate the position of the second target; and a sensor correction amount estimation unit that uses the second target information and the estimated position of the second target to calculate the offset of the second sensor and estimate the correction amount, wherein the sensor alignment device changes the coordinate transformation parameters based on the correction amount.

The invention belongs to the technical field of marine environment monitoring and signal processing, and relates to a marine dissolved gas sensor data denoising and time-delay correction real-time processing method, which comprises the following steps: identifying a steady-state interval of historical observation data of a sensor, and fitting steady-state jump by using a Sigmoid function to generate a gas concentration virtual true value; a neural network model with observation data as input and the gas concentration preliminary correction value as output is constructed, and the model is trained with historical observation data as input and the gas concentration virtual true value as a training target; intercepting observation data fragments of the sensor at current and historical moments by adopting a sliding window mechanism, and inputting the observation data fragments into the model to output a gas concentration preliminary correction value at the current moment; and performing post-processing by using the logic state machine to obtain a final correction value of the gas concentration at the current moment. According to the invention, a real-time compensation function considering high dynamic response and zero steady-state drift can be provided for the sensor, so that in-situ acquisition and real-time return of high-precision observation data are realized.

The invention discloses a state estimation and health management method and system for a main fuel regulating system of an aero-engine. The method comprises the following steps: constructing a state space model of a main fuel regulating system; estimating the fuel flow and the engine speed based on the state space model so as to obtain optimal estimation values of the fuel flow and the engine speed; optimizing a system parameter matrix of a state space model according to the optimal estimation value; the system parameter matrix comprises a state transition matrix and an input control matrix; constructing an innovation sequence based on the optimal estimation value and the sensor observation value, and calculating a residual error between the optimal estimation value and the sensor observation value; and performing fault diagnosis on the state of the main fuel regulating system and performing trend analysis on the health state according to the innovation sequence and the residual error. According to the method, the state estimation precision of the full service cycle of the engine is guaranteed, and the fault diagnosis reliability is greatly improved.

The present application relates to the technical field of navigation positioning, and in particular to a multi-source sensor adaptive weight fusion positioning method and related equipment, which comprises the following steps: collecting multi-source positioning sensor observation data, calculating corresponding health degree factors according to the data characteristics of each sensor, which are used to represent the data quality at the current time; inputting the health degree factors into a pre-trained reinforcement learning model, the model taking the maximization of the reward function constructed by the positioning error and the sensor health degree as the goal, and outputting the fusion weight of each sensor. According to the output weight, the multi-source observation data is fused and solved through factor graph optimization, and the final positioning result is output. This method can dynamically adapt to the differences in sensor working conditions and environmental disturbances, suppress the influence of abnormal data, realize the collaborative improvement of positioning accuracy and system reliability, and is suitable for multi-source fusion positioning scenes such as automatic driving, robots, intelligent navigation, etc.

The invention discloses a method for predicting spatial distribution of multiple components in a transformer, and relates to the technical field of power equipment operation state monitoring, and the method comprises the steps: firstly constructing a multi-gas component transport physical simulation model of transformer oil based on fluid simulation software, and configuring an observation system mapped with the position of an actual sensor; constructing a time sequence error matrix of actual observation and simulation prediction, and dividing the time sequence error matrix into a training set and a verification set; on this basis, a dynamic B matrix is generated by using an LSTM network, simulation data and real-time sensor observation data are fused through an EnKF algorithm, and dynamic updating and error correction of a simulation model are realized; the data assimilation process is assisted by a self-adaptive step length regulation and control and exception handling mechanism; and finally, dynamically displaying the spatial distribution of each gas component in the oil tank in various forms. According to the method, the problems of space limitation of traditional monitoring, error accumulation of static simulation prediction and the like are effectively solved, and accurate and dynamic prediction of spatial distribution of multiple gas components in the transformer is realized.

A processor-implemented method for multimodal beam management, implemented by a network device, includes: receiving an input stream from one or more sensors by the network device; generating a digital twin modeling the environment of an area observed by the one or more sensors; the digital twin including one or more objects detected based on the input stream; and managing wireless communication signal beams, at least in part, based on the digital twin for communication with at least one user equipment (UE) in the area observed by the one or more sensors.

The invention discloses a coal mine catastrophe scene deduction method, which belongs to mine disaster emergency rescue, and comprises the following steps: collecting environmental data of a coal mine stope; performing data screening according to the environment data and the key elements, and extracting time sequence features; deducing time sequence characteristics by using an energy-based dynamic time sequence model to obtain probability energy distribution of disaster development; according to the probability energy distribution, constructing a Boltzmann machine model with a limited time sequence for predicting a mine catastrophe development trend; wherein the explicit layer unit corresponds to a sensor observation value at a current time point, the hidden layer unit captures potential time sequence characteristics and modes, correlation between a hidden layer at the current moment and hidden layers at the previous N moments is established through time sequence connection, and model training is conducted through a time-dependent conditional contrast divergence algorithm; aiming at the problem that a traditional Boltzmann machine model is mainly used for processing static distribution, so that the processing precision of a coal mine catastrophe scene with obvious time sequence evolution characteristics is low, the coal mine catastrophe scene deduction precision is improved.

The present application discloses a method and apparatus for intelligently predicting the remaining life of equipment, which relates to the field of reliability engineering. The method comprises collecting sensor observation data, preprocessing the sensor observation data, and extracting data fluctuation characteristics; processing the sensor observation data and the data fluctuation characteristics based on a time window input strategy to determine input characteristics; training a global-local feature extraction network according to the input characteristics to determine a trained global-local feature extraction network; the global-local feature extraction network comprises a multi-layer temporal convolutional network, a causal convolution gating unit, a two-dimensional compressed excitation fusion module, and a Transformer encoder connected in sequence; performing performance verification on the trained global-local feature extraction network to determine a life prediction model; and predicting the remaining life of the equipment to be tested according to the life prediction model. The present application can comprehensively improve the life prediction performance.

The application discloses a polarization differential reference-based navigation orientation method and device, equipment and a medium, which are applied to a polarization differential reference station and a user station, and the user station is located within the effective radius of the reference station. The polarization angle non-perpendicular error of each pixel point observation area is taken as a state vector to be estimated, a state equation of a polarization angle non-perpendicular error time domain interference observer is established; a polarization vector is constructed based on a polarization angle measurement value; a measurement equation and a system model of the interference observer are established; a corresponding relationship between the polarization angle non-perpendicular error to be compensated and a sensor observation vector is established; the polarization angle non-perpendicular error value is determined according to the corresponding relationship and a real-time sensor observation vector obtained through a user station navigation system; the polarization angle is compensated, and polarization heading calculation is carried out by using the compensated polarization angle, so that real-time heading information of a carrier is obtained. According to the embodiment of the application, the navigation precision and environmental adaptability of the user end combined navigation system can be improved.

The application discloses a kind of based on digital twinning unmanned mine car reliability rapid early warning method, system, it is related to unmanned mine car technical field, including the following steps: the digital twin simulation model of unmanned mine car is constructed;Based on the digital twin simulation model, the digital twin reliability model that the reliability state change of each module of the unmanned mine car is characterized is constructed;Based on the measured data of the unmanned mine car, the consistency of the digital twin simulation model and the digital twin reliability model is judged, and after judging consistent, reliability state database is constructed;Based on the reliability state database, digital twin reliability rapid early warning model is obtained by training, and the early warning model is used to identify and output the current reliability state of the unmanned mine car according to real-time acquisition vehicle-mounted sensor observation data, to realize early warning.The reliability state of each subsystem and module of unmanned mine car is rapidly and accurately warned.

Techniques for performing object detection in a vehicle environment using sensor data captured by one or more sensors of the vehicle are described herein. In some cases, an object in a vehicle environment can be detected based on at least one of (i) a first similarity matrix that represents a first similarity value for two sensor observations associated with the vehicle environment, or (ii) a second similarity matrix that represents a second similarity value for a sensor observation and a track associated with the vehicle environment.

The invention discloses a marine test parameter identification method for a high-speed moving ship, which comprises the following steps of: 1, integrating a plurality of inertial sensors through an attitude sensing module, and establishing a kinematics relation equation set among an angle, an angular velocity and an angular acceleration corresponding to three-degree-of-freedom rotating motion; 2, constructing a state estimation framework with fault detection capability, analyzing an actual observation value of the sensor and a Kalman filtering predicted value in real time based on a Kalman filtering algorithm to generate a residual sequence, and realizing diagnosis and identification of a sensor fault through statistical test and logical judgment; step 3, carrying out smoothing processing on the sea test data after fault cleaning by adopting a spectral analysis method, and eliminating measurement noise and data delay; and 4, based on the processed high-quality data, establishing a motion equation of the high-speed moving ship, performing linearization processing, converting the motion equation into a discrete ARMAX model structure, and executing ship system model parameter identification by using a generalized least square method.

Disclosed are systems and methods to control a robotic device using natural language commands. A natural language command may be received by a system. The system may convert the command into low-level machine controls and logic for implementation by a robotic device to achieve a desired action. In some instances, an API module may include mapping data to associate high-level commands with low-level machine controls. A language model may process the natural language command (input), high-level commands, and / or other information, such as system state, sensor observation data, parameters, etc., to determine one or more commands to execute by a robotic device and possibly logic for execution by the robotic device. The robotic device may receive the low-level machine controls and logic to cause the robotic device to perform the requested actions.

The invention discloses a time-varying parameter identification method of a wind power mixed tower structure, a storage medium and equipment. The invention belongs to the field of new energy engineering structure health monitoring, and aims to solve the problem that time-varying parameters of a wind power mixed tower under the action of dynamic wind load are difficult to track in real time in an existing method. The method comprises the following steps: firstly, determining a covariance matrix of an initial state quantity, carrying out preliminary parameter identification based on a UKF algorithm, obtaining a sensitive parameter eta k corresponding to each step, drawing a time history curve, switching to an MAF-UKF algorithm for identification when peak pulses appear in the time history curve, dynamically adjusting a forgetting factor alpha k when eta k is greater than eta 0 in the process, and finally, carrying out identification according to the MAF-UKF algorithm. And the measurement prediction covariance, the cross covariance and the state quantity covariance are corrected according to the alpha k, a gain matrix is updated, iterative filtering is carried out until circulation is finished, a time-varying parameter identification result of the wind power mixed tower structure is obtained, parameters are updated through a sensor observation value in combination with a finite element model, a time history curve is output, and abnormal early warning is triggered.

The invention relates to a lamp ring-free handle positioning method and system based on Lie group state estimation and dynamic offset calibration, and the method comprises the steps: defining the pose state of a handle based on Lie group SE (3), defining an error state in a corresponding Lie algebra se3 space, and building a system kinetic equation based on error state Kalman filtering; representing a multi-source sensor observation model as an observation function on the Lie group; modeling dynamic offset transformation between the handle and the hand into a time-varying random process, and expanding the state of the time-varying random process into a system state vector for online joint estimation; updating and optimizing the system state vector by using a nonlinear optimization method based on multi-sensor observation to obtain an optimized handle positioning result; the system is implemented based on the method. According to the method, the consistency between the positioning output and the actual physical position of the handle is remarkably improved, the pose jumping phenomenon is effectively eliminated, the immersion and continuity of user experience are improved, and the continuous track which is consistent in time and accurate in space is finally output.

The present invention discloses a track association method based on track similarity segmentation, which mainly solves the problems existing in existing track association algorithms in actual engineering applications, such as the difficulty in obtaining prior information, the difficulty in determining the association threshold, and the long association time. The method utilizes deep learning technology, starting from the association matrix, and using a neural network to deeply explore the tracks of multiple sensors. All track information of each sensor is integrated to obtain a track matrix. By performing similarity segmentation on the track matrix, the association relationship between different tracks is directly obtained, thereby completely avoiding the traversal calculation of a large number of tracks, reducing the manual modification and debugging of models and parameters, and greatly improving the association efficiency. The method is suitable for the track association between multiple sensors in areas with rigid or non-rigid deformation during the sensor observation process, and can achieve rapid and accurate association of multiple sensors and multiple targets in actual engineering applications.