20results about How to "Strong perception" patented technology

The invention discloses a dynamic threat-based LGVF-APF-SAC spacecraft path planning method, a storage medium and an electronic device, and the method specifically comprises the following steps: firstly, building a real-time risk function according to a dynamic threat (radar detection exposure risk), and further generating an early warning region according to the risk function (radar radiation); based on the early warning area, an improved dynamic time-varying LGVF is used to construct a guide potential field based on the dynamic time-varying LGVF, so that the spacecraft is guided to avoid the early warning area to reach a target point. Then, an end-to-end reinforcement learning framework is developed; according to the framework, a maximum entropy RL (SAC) algorithm is selected, and a constructed guide potential field is integrated into a learning process in a reward function form so as to evaluate and optimize decision actions of a spacecraft. According to the algorithm provided by the invention, the dynamic adaptive capacity and reliability of path planning can be improved while the dynamic overload condition of the spacecraft is ensured to be met.

The invention provides an interconnected electronic flight bag system and an interconnection method thereof. The system comprises portable equipment, airborne equipment and a ground server, the portable equipment comprises a portable electronic flight bag and a multi-screen collaborative plug-in; the airborne equipment comprises an installation type electronic flight bag, multi-screen collaborative software and an aircraft information system; the installation type electronic flight bag comprises an integrated user interface and an information system data relay platform. The information system data relay platform of the mounted electronic flight bag performs data information interaction with an aircraft information system; and the information system data relay platform of the mounted electronic flight bag performs data information interaction with the ground server. According to the invention, flight route operation functions such as PEFB and MEFB multi-screen collaboration, data synchronization, video monitoring, chart display, flight manual management, performance calculation and WEB browser display are realized, good interoperability and real-time performance are realized, and use habits of users are developed to adapt to operation of airplanes of novel models.

The invention discloses a dexterous hand distal finger force transducer for a robot and a calibration method. The dexterous hand distal finger force transducer comprises a phalanx, a base plate, an elastic body, a cover plate and a protective layer which are sequentially arranged from bottom to top, the base plate is connected with the top of the phalanx, the top of the base plate is provided with a supporting boss, the geometric center of the elastic body is connected with the supporting boss, the elastic body extends from the geometric center to form a plurality of blades, the blades are provided with sensors, the bottoms of the blades and the base plate are arranged at intervals, and the geometric center of the elastic body coincides with the geometric center of the upper cover plate. A contact is arranged at the tail end of the blade at the bottom of the cover plate and makes contact with the top of the tail end of the blade, and the top of the cover plate is connected with the protective layer. The position and size information of the applied force value can be obtained, and the sensitivity detection of the detection assembly on the force is realized.

The invention discloses a screening method for a riboswitch capable of specifically recognizing doxycycline as well as rational transformation and application of the riboswitch, and belongs to the field of gene expression regulation and control. The invention provides a doxycycline specific riboswitch, establishes a high-throughput screening method based on a flow cytometry, and obtains the doxycycline riboswitch with further improved activation multiple by means of computer-aided prediction of binding sites of the riboswitch and calculation of mutation sites. According to a dose reaction curve, the riboswitch shows linear correlation at 40-100 [mu] g / L, and meanwhile, the constructed whole-cell sensor is low in price, high in stability and simple to operate, and can be used for on-site rapid detection.

The invention discloses an intelligent fireproof rescue robot and a rescue system, and belongs to the technical field of fire rescue equipment. The rescue robot comprises a main body part, and a driving mechanism, a moving mechanism and a fire extinguishing mechanism are connected to the main body part; the intelligent reconnaissance mechanism comprises a reconnaissance device and a storage cabin; the reconnaissance device is connected with the rescue robot in a split manner and can be independently and movably controlled; the storage cabin is used for placing the reconnaissance device; and the rescue mechanism comprises a rescue mechanical arm and a rescue cabin, and the end of the rescue mechanical arm is connected with a rescue manipulator. Under the cooperative control effect of the control system, the rescue system is connected and communicated through a 5G network, when a mine fire occurs and after the ground command center receives an alarm, the site of the mine fire can be determined in time and monitored in real time, and an instruction whether to start or not is given to the rescue robot through the communication base station; the robot is started to carry out rapid intelligent rescue work, rescue is efficient, and the safety coefficient is high.

The invention relates to the field of computer vision, in particular to an improved high-resolution remote sensing image road extraction method based on a CATransFuse model. The method comprises the following steps of: 1, acquiring a remote sensing image, and slicing the remote sensing image according to a fixed size to obtain remote sensing image slices; and a second step, inputting the remote sensing image slices into a trained network CATransFuse for forward propagation operation, obtaining road features in the network CATransFuse, and outputting the road features. According to the method provided by the invention, by optimizing the structure of the network CATransFuse and training the parameters, the network CATransFuse capable of accurately identifying the road features is obtained. In the training process, the ability of identifying various road features is achieved by optimizing the sample quality and the number. Therefore, the identified road features are closer to the actual situation, and the accuracy is higher. According to the method, the CBAM and ASPP modules are introduced, so that feature relevance, context information and road target modeling can be better processed, and better performance can be obtained in a road extraction task.

The invention provides a missing modal data processing method based on mutual information constraint and adversarial learning, and belongs to the technical field of artificial intelligence. The method comprises the following steps: firstly, collecting full-modal data and corresponding missing modal data; then preprocessing the collected full-modal data and the corresponding missing modal data to improve the quality of the full-modal data and the corresponding missing modal data; secondly, after the preprocessed full-modal data and the corresponding missing modal data are obtained, a missing modal processing model based on cross-modal mutual information and adversarial learning is constructed, and meanwhile, a loss function of the missing modal processing model is constructed to constrain the missing modal processing model; continuously training and optimizing the missing modal processing model to obtain a trained missing modal processing model; and finally, inputting the data with the missing modality facing an actual task into the trained missing modality processing model to obtain a segmentation graph.

The invention discloses a coronavirus appearance imitating hyperelastic memory alloy rod triggering type spherical robot. The robot is mainly composed of a rod end contact, a hyperelastic memory alloy rod, a tough wire, a rod sleeve, a spherical shell, a micro camera, a hexagonal connecting prism, a straight-through connector, a take-up and pay-off device, a dodecahedron inner core, a control panel and a nine-axis attitude sensor. The take-up and pay-off device drives a winding reel to rotate through a motor so as to take up the tough wire, an elastic rod on the outer surface of the spherical robot deforms, then the tough wire is rapidly released, the bent elastic rod rebounds to provide power for movement of the robot, meanwhile, the take-up and pay-off device and the elastic rod reset, and continuous movement of the robot is guaranteed. The bending degree of the elastic rod can be controlled by changing the position of the knot on the toughness line, and the ejection period is controlled by the control system. The robot is unique in structure, novel in principle, composite in function, high in man-machine interaction and sensing capacity and high in movement flexibility and stability, and has important theoretical significance and practical value for research of spherical robots.

The application provides an industrial product quality prediction method, device and medium, the method comprises the following steps: obtaining multivariate time series data of an industrial process, and dividing the multivariate time series data into a training set and a test set; obtaining enhanced frequency domain features corresponding to the training set through discrete Fourier transform; performing topological preservation and low-dimensional projection processing on the enhanced frequency domain features through a complex frequency domain self-organizing map (SOM) network to obtain SOM projected frequency domain features; designing a time-frequency attention fusion mechanism containing time domain features and SOM projected frequency domain features, and constructing a time-frequency attention fusion network model based on the time-frequency attention fusion mechanism and a time convolution network to capture global-local time patterns; and inputting the test set into the time-frequency attention fusion network model to obtain a predicted value of the industrial product quality. The method, device and medium can solve the problem that existing industrial product quality prediction methods cannot adaptively fuse time domain and frequency domain features, resulting in low accuracy of industrial product quality prediction.

The embodiment of the invention provides a claim clause text generation method and device, electronic equipment and a storage medium, and belongs to the technical field of text generation. The method comprises the following steps: acquiring sample claim clauses and sample score data; screening the sample claim clauses to obtain reference claim clauses, and obtaining reference score data of the reference claim clauses; according to the reference claim clauses and the reference score data, training a preset clause score model to obtain a target clause score model; training a preset claim clause generation model through the target clause scoring model and the initial clause generation prompt word to obtain a target claim clause generation model; obtaining a target claim term generation task, and generating through the target claim term generation model and the target claim term generation task to obtain a target claim term; and splicing the target claim clauses and the preset general claim clauses to obtain a claim clause text. According to the embodiment of the invention, the efficiency of generating claim clauses can be improved.

The invention discloses an intelligent stereoscopic warehouse system and method based on a laser demarcation device, and belongs to the technical field of intelligent warehousing. The system establishes a high-precision space coordinate system for a warehouse through a space reference network formed by fixed laser demarcation devices; the mobile device visually perceives the network to realize submillimeter-level accurate positioning, scans the goods allocation by using a laser demarcation device carried by the mobile device, and perceives the state of the goods allocation in a non-contact manner; the central processing system is integrated with a digital twinning technology, inventory can be managed, and warehouse structure deformation can be monitored by analyzing long-term change of reference laser. The problems that a traditional stereoscopic warehouse is large in positioning accumulative error, goods location sensing depends on a large number of sensors, and the self-calibration capability is lacked are solved, and the precision, reliability and intelligent level of the system are remarkably improved.

The invention relates to the technical field of anti-riot equipment, and discloses an attack behavior grade identification method based on an improved reinforcement learning intelligent algorithm and a folding shield thereof. Multi-modal sensor fusion, reinforcement learning and a real-time feedback control algorithm are applied to defense equipment, a protection strategy is optimized, the algorithm adapts to different working conditions and environment changes, prediction and early warning are performed on a future threat state level, and an intelligent algorithm is used for predicting and early warning the threat state level through combination of real-time data and historical data. And the prediction and planning capabilities of the system are improved. The system not only can quickly respond when a threat comes, but also can plan a protection strategy in advance to ensure that a protection action is ready before the threat occurs. The advanced planning reduces the response time, and ensures that the shield holder can cope with various sudden threats.

The invention discloses an electricity market development potential analysis and evaluation method and system. The analysis and evaluation method comprises the following steps: evaluating the marketization development degree of a power generation side according to member participation and electric quantity transaction conditions of the power generation side of a power market; evaluating the marketization development degree of power transmission according to the electric quantity transaction condition of the power grid side of the power market; the marketization development degree of the electricity utilization side is evaluated according to member participation and electricity utilization conditions of the electricity utilization side of the electricity market; and determining the weight coefficient of each marketization development degree based on the power transaction characteristics, and comprehensively evaluating the development potential of the power market. According to the method, a multi-dimensional power market power generation potential measurement and calculation index system is constructed from the three aspects of a power generation side, a power transmission link and a power utilization side, main power generation links, power transmission links and power utilization links in the marketization process are comprehensively considered, and the method has the advantages of being comprehensive in evaluation and high in practicability.

The embodiment of the invention provides a user tag prediction method and device, electronic equipment and a storage medium, and belongs to the technical field of text processing. The method comprises the following steps: acquiring a sample dialogue text of a sample object, and acquiring a sample object label of the sample object; based on a preset dialogue label relevance evaluation model, performing relevance dialogue evaluation on the sample dialogue text to obtain dialogue relevance evaluation data; performing text screening on the sample dialogue text according to the dialogue tag evaluation data to obtain an associated dialogue text; performing model training on a preset tag classification prediction model according to the associated dialogue text and the sample object tag of the sample object to obtain a target tag classification prediction model; and obtaining a target dialogue text of the target object, and performing label classification prediction on the target object based on the preset dialogue label relevance evaluation model, the target label classification prediction model and the target dialogue text to obtain a target object label. According to the embodiment of the invention, the efficiency of carrying out label classification on users can be improved.

The present invention belongs to the technical field of reservoir safety detection, and provides a system and method for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals. The system includes a fixing device, a signal acquisition device, and a cloud data analysis system; the fixing device is used to install the signal acquisition device on the dam body; the signal acquisition device is used to collect microseismic signals of rockfill material rupture, and send the microseismic signals to the cloud data analysis system; the cloud data analysis system is used to obtain whether the rockfill dam body is ruptured based on the received microseismic signals and the information vector machine model for identifying the type of rockfill dam rupture; based on the received microseismic signals and the spatial distribution map of the source position, the distribution of the source position in the spatial distribution map of the source position is analyzed to determine whether there is a potential seepage channel in the local area. The present invention can track the rupture of rockfill materials and its evolution process caused by seepage inside the rockfill dam body of the upper reservoir of the pumped storage power station.

The application discloses a SAM2 small sample segmentation method based on semantic-visual double memory fusion, constructs semantic query memory, visual query memory and query related support visual memory, fuses the semantic query memory and the visual query memory through a memory refinement module guided by the query related support visual memory, and combines SAM2 dense matching and decoding module end-to-end training. The method solves the single memory and foreground-background confusion problems of the existing method, enhances target semantic consistency and fine-grained modeling capability, significantly improves segmentation precision and generalization capability in a complex scene, and is suitable for medical image analysis, automatic driving and the like.

The utility model provides hand-wearable blind guiding equipment. The hand-wearable blind guiding equipment comprises a packaging shell; the control panel is arranged in the packaging shell; the upper fixing plate is fixedly connected with the packaging shell, and one face of the upper fixing plate is a continuous arc-shaped curved face and used for being attached to one side of an arm; the lower fixing plate is detachably connected with the upper fixing plate, and the face, facing the upper fixing plate, of the lower fixing plate is a continuous arc-shaped curved face and used for being attached to the other side of the arm; the power supply module is arranged in the packaging shell, is electrically connected with the control panel and is used for supplying power to the control panel; the holder is rotatably mounted on one side of the packaging shell; the depth camera is rotatably connected with the holder, and the holder is used for keeping the stability of the depth camera; the earphone is in communication connection with the control panel and is used for playing voice prompts; and the more than two vibration fingerstalls are in communication connection with the control panel, and the vibration fingerstalls are used for sending vibration signals. The problems that in the prior art, visually impaired people are inconvenient to go out, and an ultrasonic sensor scheme is not high in reliability are solved.

The present invention discloses a complex electromagnetic environment spectrum sensing method and system based on a multi-dimensional feature domain, wherein the method includes the steps of: dividing the azimuth dimension 360° airspace into 24 regions, respectively receiving electromagnetic wave signals in different directions; calculating the time-frequency instantaneous energy of the electromagnetic wave signal in each region; detecting effective signals in each airspace and calculating the time-frequency amplitude; screening out all signals whose time-frequency amplitude exceeds a preset threshold, and using the cumulative phase difference method to calculate their instantaneous frequencies, obtaining a spatial multi-angle time-frequency two-dimensional feature matrix, identifying unoccupied spectrum ranges, i.e., spectrum holes; planning communication links according to the number of spectrum holes and their available spectrum ranges. The present invention can effectively improve the real-time and accuracy of electromagnetic spectrum cognitive equipment in terms of electromagnetic spectrum situational awareness through multi-dimensional domain feature analysis of spectrum in complex electromagnetic environments, thereby guiding the frequency utilization strategy of frequency utilization equipment.

The invention provides a feature extraction and fusion reconstruction method, an electronic device, a medium and a product, the feature extraction and fusion reconstruction method comprises the following steps: obtaining channel time sequence data, and carrying out dimension standardization and time sequence arrangement on the channel time sequence data to obtain preprocessed data; pCA dimension reduction processing is carried out on the preprocessed data, and static features are extracted; sampling the preprocessed data by adopting a sliding window, and constructing a time sequence sample; extracting dynamic characteristics of the time sequence samples; performing full-connection mapping on the static features; performing feature splicing and regression output on the static features and the dynamic features; and training, testing, evaluating and visualizing the output result. According to the feature extraction and fusion reconstruction method, the static space feature and the dynamic time-varying feature of the channel can be precisely restored, high-precision channel modeling and reconstruction of a millimeter wave communication system in a dynamic environment are achieved, robustness, stability and channel estimation performance are good, delay is short, and generalization ability is high.

This disclosure provides a flexible tactile sensing system and a wearable device. The flexible tactile sensing system includes electronic skin and a driving device. The electronic skin includes a piezoresistive functional layer and an electrode layer. The piezoresistive functional layer contacts an object to be grasped and detects the pressure of the grasped object, while the electrode layer contacts the skin of the hand. The driving device includes a control module and a current source module. The control module is electrically connected to the piezoresistive functional layer and the current source module. The control module receives the pressure signal from the piezoresistive functional layer and controls the current source module to output current to the electrode layer. The electronic skin also includes an insulating substrate layer. The piezoresistive functional layer includes multiple spaced pressure detection patches, and the electrode layer includes multiple spaced electrode patches. The pressure detection patches and electrode patches are arranged in pairs and are correspondingly distributed on two opposing surfaces of the insulating substrate layer. By amplifying and converting minute contact forces into multi-dimensional electrical stimulation signals, the patient's tactile feedback loop can be effectively reconstructed.