105 results about "Data fusion algorithms" patented technology

Each of multiple probability distribution outputting units computes a probability distribution of a data value detected by the corresponding sensor or algorithm in an image recognition processing or the like. The respective probability distributions of the multiple probability distribution outputting units are given as outputs to a synthetic determination processing unit. Data formats of the outputs to the synthetic determination processing unit can be thereby standardized. Hence, the synthetic determination processing unit is exempted from considering which type of sensor or algorithm each of the outputs is based upon. Even when a sensor or algorithm is added or changed, the same data-fusing algorithm in the synthetic determination processing unit can be uniformly used.

A system and method for monitoring and controlling an aquatic environment thus regulating the aquatic environment and maximizing the stability of the aquatic ecosystem. This system uses sensor and data fusion algorithms to perform system anomaly analysis and predictive failure diagnostics based on the output of sensors associated with the aquatic environment control devices. Based on the system anomaly analysis and/or predictive failure diagnostics, the system may continuously adjust the system parameters to maintain an efficient and stable aquatic environment, alert local or remote monitors of failed or impending failure of control devices. In addition, the system may present the system anomaly analysis and predictive failure diagnostics information at a local or remote location to enable the monitor to address any problems presented in situ.

Provided is a human motion tracking system based on Zigbee/ institute of electrical and electronic engineers (IEEE) 802.15.4. A wireless wearable sensing network is covered on nine independent limbs of a person, and a wireless wearable sensing node is adhered onto each independent limb for monitoring free motion of the limb in the three-dimensional space. In each wireless wearable sensing node, an inertia measuring unit which comprises a three-axis acceleration sensor, a three-axis gyroscope and a three-axis magnetic force sensor is used for calculating the position of the independent limb in the three-dimensional space through an on-line data infusion calculation method. In the wearable wireless sensing network, real-time location data can be transmitted to a substation through the low-consumption wireless communication technology Zigbee/IEEE 802.15.4. The substation forwards the obtained wireless wearable sensing network to a computer in real time. Human motion can be reconstructed in the computer according to a forward motion model. The reconstructed human motion can be transmitted to a remote user through the internet for tememedicine. The human motion tracking system totally eliminates ligature between sensor nodes, do not need a light source for assisting and facilitates system expansion and application development by means of the low-consumption wireless operation system.

The invention provides a ship power station fault diagnosis method based on data fusion. The method comprises the steps of respectively adopting different data fusion algorithms on different levels of a fault diagnosis, using a plurality of sensors for detecting fault messages in multiple aspects, conducting stage treatment on multi-source messages, accurately and timely judging the state of a system, and giving the correct judgment on whether the system fails or not and the correct judgment on the fault mode. According to the ship power station fault diagnosis method based on the data fusion, intelligent monitor of a ship power station unit is effectively achieved, the reliability and safety of the operation of the unit are improved, and the phenomena of false alarms, misinformation and information missing are reduced. The sensors respectively monitor parameters of the fault messages and integrate the monitored data to conduct a first-stage fusion to detect layer data, a second-stage fusion is carried out on the data passing through the first-stage fusion, a third-stage fusion is carried out on the data passing through the second-stage fusion, the data after the three-stage fusion are matched with data inside a fault diagnosis knowledge database, and the fault diagnosis result is output.

The invention discloses a ZigeBee-based industrial environment dangerous gas leakage detection system. The invention can be mainly divided into three parts: a wireless sensor network composed of sensor nodes, a data processing center of an ARM platform, and a data storage and monitoring center based on a cloud server. In order to improve the installation flexibility and detection efficiency of the system, and reduce deployment costs and power consumption, ZigBee technology is used to build a wireless sensor network, and the solar charging function based on Texas Instruments' single battery management chip is added to the sensor nodes. The data processing center uses the ARM processor as the control core to realize data reception, leakage judgment, data upload to the cloud server, early warning SMS sending, and data display synchronously. Combined with the homogeneous sensor data fusion algorithm and fragment anomaly detection algorithm, the industrial Fast and accurate alarm of environmental hazardous gas leakage.

The invention provides a full-automatic parking method and system. The full-automatic parking method comprises the steps of acquiring space parking space data and obstacle data recognized by an ultrasonic radar and look-around parking space data and obstacle data recognized by a fisheye camera; performing fusion based on a data fusion algorithm to obtain parking space fusion data and obstacle space fusion data; according to a free track planning algorithm, carrying out planning in combination with the parking space fusion data and the obstacle space fusion data to obtain a parking track; and acquiring moving position information of a current vehicle, and making the vehicle run along the parking trajectory in real time based on a high-precision positioning algorithm and a high-precision track tracking algorithm. In the scheme, the parking track is determined by calculating the parking space fusion data and the obstacle space fusion data obtained by fusion by using the free track planning algorithm. And the vehicle is controlled to run along the parking track by using the high-precision positioning algorithm and the high-precision track tracking algorithm. Precision of parking spacerecognition and automatic parking control is improved so that the method and the system adapt to more parking scenes.

The invention relates to a multi-individual multi-parameter multi-sensing intelligent fireman physical ability early-warning and monitoring device, which mainly comprises heat-insulation fireproof clothes, a monitoring system network, a human body information collecting module, an upper computer display module, a phone APP, and an EnOcean energy collecting unit, wherein the monitoring system network comprises multiple wireless transmission nodes, a relay and a monitoring display terminal; the human body information collecting module is internally provided with a central processing unit, a human body temperature detection unit, a heart rate detection unit, a human body blood pressure detection unit, an ambient gas detection unit, and a falling detection unit. The device has the advantages that the environments and positions of the fire places where multiple firemen are as well as the human body key physiological parameters of the firemen can be monitored in real time, and early warning is sent, so that the harm to the firemen can be avoided furthest; meanwhile, a proper data fusion algorithm is adopted, so that the accuracy and reliability of collecting the human body data of the firemen are improved, and the condition that the best rescue time is lost because of delayed decision is avoided.

The invention discloses a wearable Tai Chi movement gait evaluating and training system based on a cloud platform. The system comprises a gait parameter collecting module, a gait parameter processingmodule, a gait movement function evaluating and training module and a cloud platform module. An attitude sensor, a plantar pressure sensor and a surface electromyography sensor are adopted to acquiregait data in Tai Chi movement of a human body in real time, and transverse and longitudinal comparison of user Tai Chi movement gait training is realized based on a cloud platform data fusion algorithm. Meanwhile, a big data deep mining algorithm is adopted to realize reverse analysis from electromyographic signals to gait specifications. Based on the human-computer interaction function of multiple users, clear digital visual guidance can be provided for Tai Chi teaching, and objective quantitative reference can also be provided for scoring of competitive Tai Chi competition judges. Through the packaging integration design of the devices, the system is portable, wearable, easy to use and reliable, and has good market value.

Provided are an earth satellite attitude data fusion system and method, applicable to an earth satellite space environment to estimate attitude data of a satellite. When the earth satellite attitude data fusion system of the present invention is used to perform the earth satellite attitude data fusion method, the first step is to perform a body rates/quaternion attitude data processing operation. Then, the next step is to perform an attitude/rates data fusion processing operation, wherein an attitude data fusion algorithm module receives a first IAE result data from a first EKF, and a second IAE result data from a second EKF, and performs an attitude/rates data fusion algorithm in a subsystem level to evaluate an attitude estimation IAE performance based on the first IAE result data, and the second IAE result data.

A satellite attitude data fusion system and method is disclosed, applicable to the earth satellite environment to estimate attitude data of the satellite. When the satellite attitude data fusion system of the present invention is used to perform the satellite attitude data fusion method, the first step is to perform a body rates quaternion attitude data processing operation. Then, the next step is to perform an attitude/rates data fusion processing operation, wherein an attitude data fusion algorithm module receives the first IAE result data from the first EKF, and the second JAE result data from the second EKF, and performs an attitude/rates data fusion algorithm in a subsystem level to evaluate an attitude estimation JAE performance.

The invention relates to a multi-mode intelligent inertial navigation sensing system and a data processing method thereof. The multi-mode intelligent inertial navigation sensing system comprises a sensor module, a processing module and an interface module, wherein the sensor module is connected with the processing module, and the processing module is connected with the interface module; the processing module comprises an engine kernel module and a selector, the engine in the engine kernel module comprises two or more engine modes, every engine mode corresponds to one data fusion algorithm; thesensor module is used for measuring acceleration, angular speed, axial geomagnetic component, external temperature, external air pressure and/or positioning information; the processing module is usedfor receiving measured data, and according to user or system configuration, performing data fusion on a part of or all the measured data from the sensor module according to the data fusion algorithmcorresponding to one of the engine modes; the selector selects and sends one data steam from data steams fused according to different engine modes to the interface module; the interface module is usedfor outputting processed data streams to an external module or a host; the processing module is embedded with a real-time operation system, in which a plurality of threads are arranged, and the plurality of threads are rounded at a preset time slice interval and used for controlling execution of different tasks.

The invention discloses a data fusion algorithm and system for multi-band radar target recognition. The system comprises: a data acquisition module used for acquiring electromagnetic echoes of an identified target body; a feature extraction module used for extracting echo data of the electromagnetic echoes; a data fusion module used for fusing the plurality of extracted feature data into feature data F by using a specific fusion algorithm; a category identification module used for classifying the identified target body according to the fused feature data F; and an image processing module usedfor converting the fused feature data F into a visual image by utilizing a technology and an image conversion principle. A radar system is used for emitting multiband electromagnetic waves, then linear discriminant analysis is carried out on multiband electromagnetic echoes of the identified target body to extract the features of the target, the features of the multiband electromagnetic echoes arefused, the fused features are classified through adoption of a classification algorithm, and therefore, the categories of the different identified target bodies can be obtained.

The invention discloses an Internet of Things multi-dimensional data fusion system and method. The method comprises the steps that a data supply direction data fusion sharing center provides Internetof Things data or publishes a data service, a data demand direction data fusion sharing center submits a data demand, and the data fusion sharing center integrates an Internet of Things spatio-temporal data description model and a data fusion model; the Internet of Things spatio-temporal data description model takes a Beidou grid code as a spatial domain identifier, and performs mapping processingon Internet of Things data of different data standards to complete data resource normalization; and the data fusion model uniformly describes detailed features of data required by a data demander, extracts feature vectors based on a geographic space information domain according to the space-time data demand description of the Internet of Things, and performs data fusion according to the data features by taking Beidou grid codes as space domain identifiers and adopting a data fusion algorithm. According to the invention, the fusion efficiency of the Internet of Things spatio-temporal data in the geographic space information domain is improved.

The invention discloses a water surface unmanned surface ship path tracking method based on intelligent predictive control, aiming at the influence of under-actuated performance and time-varying wind wave flow disturbance of an unmanned surface ship on the maneuverability of the unmanned surface ship. The method aims at solving the problems that during the power-on period, data deviation is large, and during high-speed navigation, data drifting occurs to the gyroscope. A data fusion algorithm is provided, a complementary filtering algorithm is improved by setting threshold increment constraint, and the navigation data precision is improved. The problems that an unmanned ship is prone to being interfered in a severe environment, a ship body shakes greatly, and steering is advanced or lagged when the unmanned ship enters a receiving circle are solved. The invention provides a self-adaptive sight line method, and the radius of a receiving circle is automatically adjusted, so that the unmanned ship has sufficient time to adjust the course, steers gently, and is prevented from deviating from a target path. The method aims at solving the problem that the solving precision is not high when a traditional objective function is used in a model prediction controller. The invention provides an improved artificial fish swarm algorithm for global optimization. And the anti-interference capability of the control system and the convergence speed of the algorithm are improved.

The invention relates to a fault diagnosis method for a horizontal attitude angle of an unmanned aerial vehicle, and the method comprises the steps: obtaining two groups of unmanned aerial vehicle horizontal attitude angle sensor information from a horizontal attitude angle hardware dual-redundancy sensor, carrying out the cross combination of the two groups of sensor information, and generating four-path gyroscope and accelerometer combination information; respectively resolving the four paths of combined information into four-redundancy horizontal attitude angle information through a data fusion algorithm; and carrying out information diagnosis on the four-redundancy horizontal attitude angle through motion over-limit judgment and redundancy voting, and carrying out fault diagnosis on the horizontal attitude angle sensor according to a four-redundancy horizontal attitude angle information diagnosis result. According to the method, a voting fault diagnosis mechanism which can be carried out only by existing redundancy hardware redundancy can be completed, wider sensor faults can be effectively diagnosed, the hardware structure of the system is simplified, the requirements for theinstallation space and the loading capacity are reduced, the system cost is reduced, and the reliability of the flight control system is improved.

The invention discloses a head posture recognition method based on data fusion. The method comprises the steps of firstly, collecting a precise head X-axis rotation angle and a precise head Y-axis rotation angle during the head rotation process; acquiring a Z-axis rotation angle during the head rotation process, wherein the Z-axis rotation angle comprises a gyroscope Z-axis rotation angle and an infrared feature point Z-axis rotation angle, and the gyroscope Z-axis rotation angle and the infrared feature point Z-axis rotation angle are used as the input quantities of a data fusion algorithm; obtaining the accurate head Z-axis rotation angle through data fusion, and finally outputting the head posture parameters composed of the accurate head X-axis rotation angle, the accurate head Y-axis rotation angle and the accurate head Z-axis rotation angle. According to the method, the measurement cost is low, the drift distance of the measurement result of a gyroscope can be corrected by adopting the infrared scanning measurement result through the fusion algorithm, and the purpose of high-precision measurement is achieved on the basis of keeping the original advantages of the two measurement methods.

The invention discloses a software platform applying a CART algorithm to realize electric energy meter detection data management. The invention relates to the technical field of electric energy metering detection. A cloud big data management platform is adopted, and information in mass data is fully mined to realize big data analysis, mining and processing, a cloud computing technology is adoptedto process hundreds of millions of data types within a few seconds, the classification capability and calculation precision of different data are improved through a CART algorithm and an improved CARTalgorithm, the construction of a distributed hierarchical model of big data can be realized through the CART algorithm, and the data identification capability is simplified. According to the invention, a data fusion algorithm is adopted to reflect various data of the electric energy meter with diversified information representation forms, huge information quantity and complicated information relationship timely, accurately and reliably, through a BP neural network algorithm model, the error of a direct leading layer of an output layer is estimated by using an output error, and the learning algorithm precision is improved.

The invention discloses a seismic attribute multi-scale fusion method and device, electronic equipment and a medium. The method comprises the following steps: performing spectral decomposition processing on three-dimensional superposition seismic data through generalized S transformation to obtain a multi-scale seismic data volume; calculating multi-scale attribute data according to the multi-scale seismic data volume; comparing the geological information with the multi-scale attribute data, and determining a source attribute of data fusion; and performing seismic attribute image fusion through a visual saliency image fusion algorithm to obtain a fused attribute image. A data fusion algorithm and a generalized S-transform spectrum decomposition method are combined to perform seismic attribute slice image fusion, detail information of source attribute slices can be reserved, and abnormal signals in the source attribute slices can be highlighted.

The invention discloses an MARG attitude calculation method with motion acceleration compensation, and belongs to the field of navigation. The method comprises the following steps of: firstly, carrying out attitude one-step prediction by utilizing a gyroscope, then, detecting an accelerometer and a magnetometer by utilizing a one-step prediction value, and finally, executing a corresponding data fusion algorithm according to different detection results. By using the method to estimate the attitude of the MARG sensor, the motion acceleration and the magnetic field interference of the carrier are effectively detected; directed at the dynamic characteristics of the motion carrier, sliding window data is introduced to fit a motion acceleration change rule and carry out corresponding model compensation; and an analytical solution of a gravity-magnetic (acceleration and magnetic induction intensity measured values) vector pair under the Wahba problem is derived so as to accelerate the attitude solving speed and remarkably improve the MARG attitude estimation efficiency.

The invention relates to the technical field of data processing, in particular to a database fusion association system, which comprises a plurality of sensors, a real-time database, a historical database, a fusion management module and a filtering tracking module; the fusion management module comprises a calibration module, a correlation module, an identification module and an estimation module; the real-time database is used for data entry and maintenance, real-time data updating, historical database communication and application program communication. The fusion model and the database are combined into a unified whole, the database is designed from top to bottom, an integrally optimized database strategy is achieved, great superiority is achieved, the performance and the reliability of afusion system can be guaranteed, the working efficiency of the database can be greatly improved, and starting from the requirement of data fusion for the database, a database model supporting a datafusion algorithm is established, and a special database of a data fusion system is developed on the basis of the model, so that the real-time performance is met, and the storage amount of data in a memory is reduced.

The invention provides a meadow environment monitoring system based on the Internet of Things, and belongs to the technical field of ecological environment monitoring. According to the meadow environment monitoring management information system based on the Internet of Things, the Internet of Things technology is applied to the field of grassland meadow environment monitoring, the NB-IoT wireless communication technology and the image processing technology are combined and applied to the system, and management of meadow environment monitoring information is scientifically achieved. According to different weights of various environmental factors in grassland information for herdsman grazing, characteristics of the factors are analyzed and compared, a multi-sensor data fusion algorithm is adopted, grassland environment data are analyzed and processed, an evaluation matrix is constructed, and grassland environment grades are scientifically evaluated.

The invention discloses an antenna monopulse performance evaluation method under a complete machine condition, and the method comprises the following steps: firstly, positioning the position of a target relative to a radar, then correcting the deviation between the zero depth of a difference channel of an antenna of the radar and the main beam of a sum channel, obtaining the normalization information of the difference channel and the sum channel of the antenna at different pointing angles, analyzing the normalized information to obtain an evaluation result of the monopulse performance of the antenna. According to the method, the antenna is controlled to move through a control mechanism of the radar according to the condition of the whole machine, normalized information at different angles is obtained by adopting a signal power estimation algorithm and a data fusion algorithm, and the monopulse performance of the antenna is rapidly evaluated.