85results about How to "Improve perception accuracy" patented technology

The invention discloses a short-wave spectrum sensing method and a short-wave spectrum sensing system. An RF (radio-frequency) front-end unit dynamically adjusts the center frequency of a narrow-band electric tuning filter for multi-band short-waves at an RF analog front-end, and the Q value of a tuned loop is improved through an effective impedance matching network, thereby controlling the bandwidth of a filter to be fixed at a communication channel bandwidth or slightly greater than the communication channel bandwidth; after carrying out sampling on a signal, an analog-to-digital conversion module converts an analog short-wave signal in a pass band into a digital short-wave signal, and a digital signal processing unit realizes the sensing on the power spectrum of a received signal by using an energy detection algorithm. The method and system disclosed by the invention have the characteristics that the frequency band covering of short-waves is comprehensive, no image interference is introduced, and the method and the system are easy to implement.

The invention provides a deep learning cognitive radio network (CRN) multi-user cooperative spectrum sensing method, in which correlations of adjacent secondary users and adjacent frequency channels in spaces and spectrum domains are utilized, and fine granularity characteristic distinguishing is carried out on spectrum sensing data through a convolutional neural network (CNN) to improve the sensing precision. The method specifically comprises the following steps: a constructed multi-layer CNN model is firstly trained through the sensing data of the various users; current spectrum sensing results of the secondary users are then input into the trained CNN model; and the model automatically extracts characteristics of the sensing data, and carries out classification and identification on theextracted characteristics, so that cooperative spectrum sensing results of the secondary users for a current frequency channel of a primary user are obtained.

The invention discloses a hierarchical matching method based on channel quality prediction in a cognitive wireless network. The hierarchical matching method comprises the following steps that: S1, multiple cognitive users form a collaborative perception group; and each cognitive user performs perception by selecting to add the group having the maximum benefit; S2, a group leader trains and constructs a probabilistic suffix tree by utilizing a perception result and predicts the idle probability of the next time slot spectrum through the generated tree; S3, the group leader constructs a channel quality evaluation factor by utilizing the idle probability and sorts the reliabilities of multiple authorized channels in a descending manner; and S4, a dispatching unit distributes proper channels to coded hierarchical data according to a priority-reliability dispatching criterion. By means of the hierarchical matching method, the channel quality can be dynamically estimated in the time-varying cognitive wireless network; the perception delay is reduced; reliable channels can be effectively matched for hierarchical data; and the hierarchical matching method can be used for improving the link adaptation of hierarchical multimedia data in the cognitive wireless network.

The invention discloses a coalitional game based multi-user collaborative spectrum sensing method. The method comprises steps as follows: each secondary user performs independent spectrum sensing at an initial stage; coalitions are formed among the secondary users; a coalition head in each coalition is selected; utility values of the secondary users are computed; whether the coalitions are formed is judged; coalition partitions are updated; whether the coalitions are in a stable state is judged; data fusion is performed in each coalition; coalition collaborative spectrum sensing is performed finally. Meanwhile, environment change such as master user mobility and the like is considered, and the steps can be periodically repeated. On the basis of the cognitive radio network, the problem about the multi-user collaborative sensing is analyzed with the coalitional game theory; the adaptability to the dynamic change of a wireless environment is improved; the influence of the shadow effect and hidden nodes on the sensing performance is reduced; the inspection probability of spectrum sensing is improved, the interference to a master user is reduced, and the protection for the master user is improved.

The invention discloses a rehabilitation robot sensing system which comprises an initialization module, an environment sensing module, a motion intention sensing module and a motion execution module. The initialization module is used for initializing the rehabilitation robot sensing system and starting the rehabilitation training mode; the environment sensing module is used for conducting analytical judgment on the environment where a rehabilitation robot is located and selecting an appropriate walking mode; the motion intention sensing module is used for collecting sensing data through a sensor after the appropriate walking mode is determined and generating motion intention commands of a patient according to the sensing data; the motion execution module is used for controlling the rehabilitation robot to execute the rehabilitation training motion according to the motion intention commands. By means of a rehabilitation robot sensing method, the patient can control and use the rehabilitation robot with the good environmental suitability and the high motion intention sensing accuracy. The invention further discloses the rehabilitation robot sensing method.

The invention discloses an unmanned driving technology perception fusion method based on a mining area environment, and the method employs a plurality of sensors as data sources, and guarantees the stability and reliability of perception information through employing a multi-sensor fusion technology for a special driving environment of a mining area. Various sensors are used for collecting vehiclesurrounding environment information, complementation is conducted according to different sensor characteristics based on the multi-sensor fusion principle, more accurate sensing information can be obtained, the sensing accuracy is improved, and therefore the unmanned mine card vehicle can adapt to the complex mining area environment in the using process, and work is completed safely and efficiently. Obstacle information is obtained by using different sensors; the feature information weight is obtained by calculating the information entropy, the matching matrix is optimized, and the matching precision is improved, so obstacle information and tracking target information extracted in a severe mining area environment are more accurately matched, more accurate driving environment information is obtained, and the safety and fluency in the whole driving process are guaranteed.

The invention provides a power line network topology sensing method based on power line communication. The method comprises the following steps: S1, under the coordination of a main node, transmitting a signal frame m from a node A of a power line network to a node B, which receives a signal frame m'; S2, acquiring channel frequency response estimation Hest between the node A and the node B through channel estimation; S3, calculating an estimated value t of the arrival time of the signal frame m from the node A to the node B; S4, calculating the distance between the node A and the node B; S5, performing the steps S1 to S4 on any two nodes in the power line network to obtain distance information between any two nodes; and S6, acquiring the topological structure of the power line network according to a tree estimation method. The method provided by the invention has the advantages of high sensing accuracy and low complexity.

The invention belongs to the technical field of aerospace total pressure probes and specifically relates to a probe used for total pressure measurement of a rotorcraft. The probe solves the problem that total pressure sensing accuracy of a total pressure probe of the rotorcraft is poor due to influences of a downwash airflow component and also solves the problem that a total atmosphere pressure cannot be detected under a low forward speed. The two probes are installed symmetrically on the left and the right of the rotorcraft, installation positions are located on a flow line of a downwash airflow of a helicopter, the probe has a non-convolution body shape, a front end face of a pressure sensing hole in the front end is designed to form an inclined angle relative to an incoming flow direction, the pressure sensing hole in the front end turns the back to the downwash airflow direction, and an inner conical angle of the pressure sensing hole in the front end is smaller than or equal to 30 DEG. In comparison with the prior art, the probe used for the total pressure measurement of the rotorcraft provided by the invention effectively increases total pressure sensing accuracy during utilization of the total pressure probe in the rotorcraft and increases a low-speed applicable scope of the total pressure probe.

An NMW-OFDM system for terahertz communication and perception integration comprises a terahertz communication and perception integration transmitter located at a transmitting end and a multi-stage perception module located at a receiving end, the terahertz communication and perception integration transmitter modulates communication data into communication and perception integration wireless radio frequency signals and transmits the communication and perception integration wireless radio frequency signals, and the multi-stage perception module estimates K speed candidate values in parallel for the received target reflection signal, and performs serial estimation to obtain a final target estimation distance. According to the invention, the characteristics of terahertz perception detection are fully utilized, the distance estimation precision can be improved by three orders of magnitude, and the speed estimation precision can be improved by one order of magnitude while the maximum detectable distance is not reduced. And meanwhile, an ultrahigh data rate as high as 100Gbps can be realized, which is more than 100 times of the data rate obtained by a millimeter wave system.

The invention discloses an identity authentication method fusing user multi-source sound production characteristics, a storage medium and equipment. The method comprises the following steps: constructing a sensing terminal; designing a transceiving protection mechanism, adding dynamic random noise during signal transmission in a label communication process of an RFID reader-writer and the sensingterminal, eliminating the noise and recovering a label EPC during signal receiving, and judging whether a sensing signal is true or false; identifying a user voice command, and if the user voice command contains a keyword, determining to enter a next step or reject the request and end according to whether the sensing terminal tag EPC can be correctly recovered; performing noise reduction and voiced part detection on the voice signal, converting the voice signal into a time-frequency spectrogram, extracting radio frequency signal phase information and preprocessing the radio frequency signal phase information; and simultaneously inputting the processed voice and radio frequency sensing data into a neural network through two branches, and carrying out user identity authentication. Based on the realization of voice and radio frequency technologies, the system can be ensured to be safe and credible, and relatively high application value, social significance and development potential are achieved.

The invention discloses a sensor scheme determination method, device and equipment and a storage medium, and the method comprises the steps: building a simulation unmanned vehicle and a simulation scene, and enabling the simulation scene to be used for the simulation unmanned vehicle to carry out the simulation driving; determining a first sensor scheme according to the initialization parameters,and determining simulation data generated by simulation driving of the simulation unmanned vehicle in the simulation scene according to the first sensor scheme; and determining a first sensing parameter of the first sensor scheme according to the simulation data, and correcting the first sensor scheme according to the first sensing parameter to obtain a sensor scheme applied to the unmanned vehicle. According to the method, the physical parameters of the sensor, the obstacle form and motion characteristics, the vehicle model size and appearance of the unmanned vehicle and the traffic environment of vehicle driving are fully considered and combined, the determined sensor scheme applied to the unmanned vehicle is high in sensing capacity and high in sensing precision, and safe operation of the unmanned vehicle can be guaranteed.

The invention discloses an ultrasonic array interaction method and system fusing visual tactile perception. The method comprises the steps: collecting hand posture information, and calculating fingertip coordinates of a finger to form a position array; for the palm, calculating palm center coordinates and palm surface normal vector parameters to form a vector array; if the finger is detected, selecting to activate the ultrasonic array based on the distance according to the position array; if the palm is detected, starting a motion platform based on directivity selection according to the fingerposition array and the palm vector array, activating a corresponding ultrasonic array after the platform is parallel to the palm, and returning angle information; adopting focus information after homogeneous coordinate transformation to drive a transducer to emit ultrasonic waves to generate a focus point to form tactile feedback, and displaying the focus point, a virtual hand and surrounding scene information in real time, and completing visual tactile fusion perception. Multi-angle different forms of tactile feedback are provided for fingers and palms, the feeling is finer and more vivid, visual perception is fused, the interaction experience is enriched, and the interaction accuracy is improved.

The embodiment of the invention discloses a model optimization method and device and a computer program product, and the method comprises the steps: obtaining an image sample set, calling a to-be-optimized target perception model to carry out the image perception of each image sample in the image sample set according to a target image perception task, and obtaining a target perception result of each image sample; calling a reference perception model to perform image perception on each image sample according to the target image perception task to obtain a reference perception result of each image sample; performing differential processing on the target sensing result of each image sample and the corresponding reference sensing result to obtain a differential result of each image sample; according to the difference result of each image sample, carrying out difficult sample mining in the image sample set to obtain one or more difficult samples; and updating the target perception model according to optimization parameters of the target perception model determined through the one or more difficult sample. According to the embodiment of the invention, the sensing capability of the target sensing model can be improved.

The invention belongs to the technical field of Internet of Things, and particularly relates to a dynamic monitoring system and method based on a multi-perception Internet of Things. The system comprises a unit sensor group which is configured on each piece of Internet of Things equipment or node needing to be monitored and is used for acquiring operation data and environment data of the Internet of Things equipment or node; the unit sensor group comprises a plurality of types of sensors with different functions, and at least three sensors are arranged in each type; only one sensor of each type of sensors is in a working state at the same moment, the other sensors are in a dormant state, and each sensor stores sensed environment data and operation data. According to the method, dynamic monitoring of the Internet of Things is carried out by utilizing data changes sensed by the multiple sensors in different time periods and data changes sensed by the sensors in different time periods, whether the operation state of the Internet of Things is abnormal or not is judged, and the method has the advantage of high sensing accuracy.

The invention discloses a ship target 6D pose estimation method based on point cloud data, and the method comprises the steps: 1, obtaining a ship point cloud data set of an offshore scene, and enabling a data set label to comprise a target type, the three-dimensional coordinates of a target, the three-dimensional size of the target, and the three-dimensional pose of the target; 2, constructing aneural network, and performing point-by-point point cloud feature extraction by adopting Point Net + + to obtain point-by-point high-dimensional features; 3, generating a 3D bounding box proposal frombottom to top, a real segmentation mask is generated based on the 3D bounding box, foreground points are segmented, and meanwhile the bounding box proposal with angle information is generated from the segmentation points and used for RCNN input; and 4, performing proposal optimization based on the proposal obtained in the step 3, the foreground segmentation features and the spatial features so asto output final classification, a 3D frame and an attitude angle. According to the method, the pose estimation effect of the three-dimensional target is achieved in an end-to-end learning mode, and the real-time performance of pose estimation is improved.

The embodiment of the invention discloses a method and a device for determining a field angle of a camera, and relates to the technical field of data processing. According to the main technical scheme, the method comprises the steps that under different to-be-selected field angles, a simulation test system is controlled to use a plurality of simulation road scenes to test an intelligent driving system of a test vehicle, and a fixed driving distance exists between a traffic vehicle and the test vehicle in each simulation road scene; based on the fixed driving distance corresponding to each simulation road scene and the measured driving distance between the traffic vehicle and the test vehicle in each simulation road scene corresponding to the same to-be-selected field angle in the test result of the intelligent driving system, determining a target field angle from each to-be-selected field angle, wherein, the measured driving distance is determined based on a simulated road image which is shot by a camera of the simulation test system at a corresponding field angle to be selected and corresponds to the simulated road scene.

The invention discloses a surgical instrument clamping force sensing method based on deep learning, and relates to the field of medical robots. The method aims to solve the problems that an existing minimally invasive surgery robot system cannot accurately obtain the clamping force of a surgical instrument and is low in precision. During application, the clamping force can be sensed only by acquiring the current of an existing motor, the angular position of a driving motor and the angular speed of the driving motor of the surgical robot system, extra hardware cost is avoided, and high-temperature disinfection is not affected; Based on a convolutional neural network, the surgical instrument clamping force sensing model is constructed in combination with the attention mechanism module and the feedback of the current of the driving motor, and compared with an existing model, the clamping force sensing precision is improved, so that the surgical instrument clamping force is accurately obtained in a minimally invasive surgical robot system, and the obtaining precision is improved. The invention is applicable to the field of medical robots.

The invention provides a spectrum sensing method for a cognitive anti-interference communication system. The method comprises the following steps: an average energy of M received signal samples withina frequency band is firstly calculated; the calculated energy is then compared with an initial threshold, and a new judgment threshold is reset according to a comparison result; if the energies of received signals are larger than the new threshold, interference on a channel is determined; and if the energies of the received signals are not larger than the new threshold, the interference on the channel is denied. The method provided by the invention has the advantages that through setting of the adaptive judgment threshold, the problem that accurate spectrum sensing cannot be achieved throughfixed-threshold energy detection in a time-varying channel is solved, and the sensing precision is effectively improved.

The embodiment of the invention provides autonomous mobile equipment. According to the autonomous mobile equipment, an inclined installation mode of planar array laser sensors is proposed, that is tosay, the planar array laser sensors are obliquely installed on an equipment body of the autonomous mobile equipment in the direction perpendicular to the field angle, so that the observation range ofa non-rich information area can be reduced, and the observation range of a rich information area is increased; richer external environment information can be acquired in the vertical field angle, andthe quality of the acquired external environment information is improved; and the perception capability of the autonomous mobile equipment to the external environment is further improved, and the perception accuracy of the autonomous mobile equipment to the external environment is improved.