273results about How to "Reduce cumulative error" patented technology

The invention belongs to the technical field of data processing, and particularly relates to a well deviation azimuth measurement system based on a high-precision compensation algorithm. The system comprises a reference vector construction part, a compensation part and a Kalman filtering measurement part. The benchmark vector construction part is used for determining a reference benchmark of a geomagnetic field and a gravitational field in the underground environment; the compensation part is used for compensating original sensor data measured by the underground sensor group, and estimating an underground initial attitude angle according to compensated accelerometer data and compensated magnetometer data in the compensated data; calculating attitude change data; according to the attitude change data, constructing a local interference field model to obtain corrected magnetometer data; and the Kalman filtering measurement part is used for modeling all error sources into state vectors, correcting measurement errors in real time through a Kalman filtering algorithm and outputting a final well deviation direction. According to the invention, the precision, the stability and the robustness of well deviation azimuth measurement are obviously improved.

The invention discloses a coastal port pile foundation structure load calculation method based on a coupling mode, and the method comprises the steps: a hydrodynamic module which evaluates the load and mechanical properties of a port pile foundation structure under different environment conditions; the water power module at least comprises an atmospheric mode module, a sea wave mode module, an ocean mode module and a coupler module; wherein the coupler module exchanges and synchronizes data among the atmosphere mode module, the sea wave mode module and the ocean mode module; the data conversion module is used for converting the wind field data, the wave field data and the flow field data obtained by the hydrodynamic force module into aerodynamic force and hydrodynamic force input files of the pile foundation structure; the finite element analysis module is used for receiving the aerodynamic force and hydrodynamic force input files, establishing a port pile foundation structure model and calculating stress distribution and displacement response of a pile foundation structure, and simulation results output by the finite element analysis module include but not limited to a stress nephogram, displacement field distribution and a buckling phenomenon of a pile foundation.

The invention discloses a three-degree-of-freedom series-parallel hybrid humanoid robot waist structure, and relates to the technical field of robots. The defects that an existing multi-degree-of-freedom waist structure is not compact enough, limited in precision, low in rigidity and poor in stability are overcome. The robot comprises a yaw joint mechanism and a pitching-sidesway joint mechanism which are connected in series, upper limbs and legs of the robot are connected through the yaw joint mechanism and the pitching-sidesway joint mechanism, and the yaw joint mechanism can drive the upper limbs of the robot to do yaw motion; the pitching-sidesway joint mechanism comprises a movable platform, a static platform, a joint universal joint and two joint driving assemblies, the movable platform is fixedly connected with the upper limbs of the robot, the static platform is fixedly connected with the legs of the robot, the joint universal joint is connected with the movable platform and the static platform, and the two joint driving assemblies are symmetrically arranged on the left side and the right side of the joint universal joint; and the two joint driving assemblies are connected with the movable platform and the static platform correspondingly, and when the two joint driving assemblies drive the movable platform, the movable platform can do pitching motion and / or side-sway motion around the joint universal joint.

The invention discloses a wave-current coupling simulation-based pile foundation wharf sediment erosion and deposition prediction system and method. The system comprises a preprocessing module, a wave-current coupling simulation-based pile foundation wharf sediment erosion and deposition prediction module, a wave-current coupling simulation-based pile foundation wharf sediment erosion and deposition prediction module and a wave-current coupling simulation-based pile foundation wharf sediment erosion and deposition prediction module, the coupler module is used for realizing data synchronization and exchange among various modes through an MCT coupler by adopting a coupling numerical mode technology; the data conversion module is used for converting wave field data and flow field data into a hydrodynamic input file of a pile foundation structure; the FlOW3D module is used for inputting the pile foundation three-dimensional model at the current moment, the hydrodynamic conditions generated through coupling and the sediment parameters obtained after data preprocessing into a configured FLOW3D module program to simulate erosion and deposition of sediment around the pile foundation under the real sea area condition, and the erosion and deposition form of the sediment around the pile at the current moment is obtained; and the visualization module is used for displaying the three-dimensional model of the pile foundation at each moment, the erosion and deposition form of the sediment around the pile and the early warning result in a visualization interface.

The invention relates to multi-mode signal processing, and provides a signal processing method of a computer system. The method comprises the following steps: acquiring different modal signal data by using a plurality of sensors and performing timestamp synchronization; a depth time sequence alignment network and a multi-layer dynamic time warping algorithm are utilized to ensure signal time consistency; multi-modal signal features are extracted and fused through a multi-stream depth feature extraction network and a cross-modal attention mechanism; the system adopts a dynamic convolution kernel generation and meta-learning technology to carry out adaptive signal processing; noise suppression and signal optimization are carried out by using a multi-mode noise reduction auto-encoder and a generative adversarial network, and parameters are optimized and processed through closed-loop feedback control and reinforcement learning strategies. According to the method, synchronization, fusion, adaptive processing, noise suppression and real-time feedback control of multi-mode signals are realized, and the method has the advantages of high precision, high robustness and real-time performance.

The invention provides an aircraft flight path optimization method and system based on inertial navigation, and relates to the technical field of flight control, and the method comprises the steps: generating an initial path according to a flight task and a limiting condition; acquiring original acceleration and angular velocity data; the attitude estimation precision is improved based on a quaternion updating algorithm in combination with Kalman filtering and an adaptive noise model; position updating is carried out by combining the acceleration and attitude information; dynamically adjusting and correcting parameters by using a machine learning algorithm; path optimization: defining an objective function which comprehensively considers path length, time, fuel consumption and safety, and adopting a particle swarm optimization algorithm and a simulated annealing algorithm to carry out optimization; evaluating and correcting the path, evaluating the quality of the path in real time, and correcting according to the track deviation degree, the safety and the fuel efficiency. According to the method, through combination of various advanced algorithms and models, the optimization effect of the flight path is remarkably improved, and efficient, safe and economical flight of the aircraft in a complex and dynamic environment is ensured.

The invention belongs to the technical field of power systems, and discloses a method for maintaining electric energy metering, which comprises the following steps of: acquiring electric energy metering related data; preprocessing the electric energy metering related data to obtain an electric energy metering comprehensive characteristic data set; according to the electric energy metering comprehensive feature data set, training to obtain an electric energy metering error prediction model, and based on the electric energy metering error prediction model, predicting to obtain an electric energy metering error; judging whether error correction is needed or not according to the predicted electric energy metering error; if error correction needs to be carried out, early warning information is generated, and electric energy metering error data is automatically collected; training to obtain an error type identification model according to the electric energy metering error data, and predicting to obtain electric energy metering error type data based on the error type identification model; based on the electric energy metering error type data, utilizing a distribution calibration algorithm to correct the electric energy metering error; the requirement for manual intervention is reduced, and the working efficiency of the system is improved.

The invention discloses a wind power generation power prediction method and system, and belongs to the technical field of wind power generation processing, and the method comprises the steps: building a multivariate time series prediction data set based on the historical data of a to-be-predicted target, and training a multivariate time series prediction model based on the multivariate time series prediction data set to obtain a power prediction pre-training model; the power prediction pre-training model is adopted to predict the current parameter data of the to-be-predicted target to obtain the power generation power of the to-be-predicted target in the future set time, the method is excellent in performance in time series prediction tasks of multiple fields based on the multivariate time series prediction data set, and experimental results show that the method is high in prediction efficiency. According to the method provided by the invention, the long and short term dependency relationship in the multivariate time sequence can be effectively captured, and the accuracy and stability of wind power generation power prediction are improved by utilizing physical constraints.

The invention relates to the technical field of sheet metal part bending detection, in particular to a method and system for judging sheet metal part bending accuracy based on a three-dimensional model, and the method comprises the steps: determining an error compensation parameter through historical data of a target production line, and calculating a bending simulation adjustment parameter through combining a sheet metal part design drawing, an expected process and a bending simulation parameter; according to the design drawing, the expected process and the bending simulation adjustment parameters, generating a sheet metal part finished product three-dimensional model by utilizing three-dimensional modeling software, and extracting a target expansion view and a target panoramic reference view of the sheet metal part finished product; respectively carrying out bending accuracy detection on the images to obtain an expanded image deviation result and a panoramic reference image deviation result; if the deviation results are qualified, the bending accuracy is qualified; otherwise, the bending simulation parameters and the design drawing are adjusted, and bending simulation adjustment parameter calculation, three-dimensional modeling and bending accuracy detection are repeated until all deviation results are qualified. Therefore, actual production conditions are met, and the bending precision of the sheet metal part is improved.

The invention discloses an intelligent wind field forecasting method and an intelligent wind field forecasting system based on combined application of ST-LSTM (Long Short Term Memory) and Swin-Transform. The method belongs to the technical field of data processing, and comprises the following steps: constructing a time sequence prediction model based on ST-LSTM; inputting the historical data of the wind field into a time sequence prediction model based on ST-LSTM to obtain future wind field prediction data; the method comprises the following steps: constructing a wind field space downscaling model based on a Swin Transform architecture; and inputting the wind field forecast data into the wind field space downscaling model based on the Swin Transform architecture to obtain wind field refined forecast data. According to the method, the STM-LSTM is adopted as the extension of the LSTM, the long-term and short-term dependency relationship in time and space can be better captured when time sequence data is processed, and the Swindow-Transform can extract features in a global range through a self-attention mechanism of the Swindow-Transform, effectively capture the mutual influence between different positions in a wind field and further improve the accuracy of forecasting.

The invention relates to a liver tumor focal lesion prediction method based on iterative learning, and belongs to the field of computer models, and the method comprises the steps: firstly, carrying out the manual marking of a lesion region in part of medical image data, obtaining ROI training data, and constructing an automatic lesion detection model based on YOLOv8 through the data; in the ViT classification model, the cut image is divided into image blocks with a fixed size, and the image blocks are embedded into a high-dimensional space. Then, location information is added to each image block using location coding to preserve their spatial relationship. And inputting the embedded sequence into a Transform module, carrying out classification processing through a full connection layer, and finally outputting a predicted category. According to the method, a dynamic learning strategy is introduced, sample weights in a loss function are dynamically adjusted in different training stages, accumulative errors of the model can be reduced, and uncertain samples can be effectively processed at the same time.

The invention relates to the technical field of robot control, and discloses a robot continuous control-oriented triple optimization SAC reinforcement learning method, which comprises the following steps that a robot executes an initial strategy in a real environment to collect empirical data; initializing a strategy network, a value network and a target network, and performing experience playback and training updating on the strategy network, the value network and the target network through a triple optimization SAC reinforcement learning method; and continuously circulating the network updating process to obtain an optimal control sequence of task execution of the robot. The problem of underestimation is relieved by improving a strategy updating mechanism and stable value function learning, the problem that the robot cannot execute the optimal strategy is solved, underestimation deviation in the strategy is effectively relieved, the accuracy of robot action selection is improved, and more accurate guidance is provided for the robot to adopt correct actions.

The invention discloses an air conditioner evaporator fin optimization design method which comprises the steps that S1, air conditioner fin multi-dimensional data are collected and preprocessed, and preprocessed air conditioner fin multi-dimensional data are obtained; s2, processing the preprocessed multi-dimensional data of the air conditioner fin into an image, extracting first and second image features, and complementing the first and second image features with the data volume to obtain complemented multi-dimensional data of the air conditioner fin; s3, extracting first, second and third characteristic relations according to the heat efficiency and the gas resistance, and calculating the bending times; s4, multi-dimensional data of the air conditioner fins are screened according to the bending times, the characteristic relation between the notch position and the heat efficiency and the characteristic relation between the notch position and the gas resistance are extracted, and the notch length and the notch depth are calculated; and S5, according to the bending times, the notch positions, the notch lengths and the notch depths obtained through calculation, an air conditioner fin drawing is drawn, and notch machining is conducted. The problems that a traditional air conditioner fin design method excessively depends on artificial experience and data is insufficient can be solved.

The invention provides a battery charge state prediction method comprising the following steps: obtaining to-be-predicted data of a target battery, the to-be-predicted data comprising a terminal voltage, a terminal discharge current and a terminal voltage change rate; inputting the to-be-predicted data into a pre-trained battery state-of-charge prediction model based on transfer learning, and obtaining a state-of-charge prediction result corresponding to the target battery output by the battery state-of-charge prediction model; wherein the battery state-of-charge prediction model comprises a transfer learning model and a Transformer model which are obtained through training of training data and are based on time distribution search and matching. Therefore, the state of charge of the battery can be accurately and quickly predicted.

The invention belongs to the technical field of industrial measurement, and particularly discloses a multi-frame three-dimensional point cloud splicing measurement method and device for high-temperature environment profile steel production and a medium. The method disclosed by the invention comprises the following steps: firstly, acquiring three-dimensional point cloud data of the profile steel surface through a line laser scanner; then, denoising and normalization preprocessing operation is carried out on the three-dimensional point cloud data; thirdly, splicing the three-dimensional point cloud data by using an improved ICP (Inductively Coupled Plasma) algorithm, and carrying out Laplace smoothing and normal vector homogenization processing on the spliced three-dimensional point cloud data; thirdly, performing profile steel edge fitting on the spliced data; and finally, calculating the geometric dimension of the profile steel after edge fitting. According to the invention, the precision of point cloud splicing and the stability of a measurement result are greatly improved. The method can be widely applied to dimension measurement of the profile steel in the high-temperature environment.

The invention discloses a device and a method for manufacturing a laser three-dimensional body with a complex structure, and solves the problems of low processing precision and low processing efficiency of the existing femtosecond laser processing technology. The light beam position presetting unit, the dispersion compensation unit, the acousto-optic modulation unit, the optical relay compensation unit and the microscopic focusing unit are sequentially arranged on a light path where the processing laser is located in the emitting direction of the processing laser. The light beam position presetting unit is used for adjusting the optical axis space position of the processing laser; the dispersion compensation unit is used for performing dispersion compensation on the processing laser after the optical axis spatial position adjustment; the acousto-optic modulation unit is used for dividing the processing laser subjected to dispersion compensation into a plurality of sub-beams through amplitude modulation of electric signals with different and / or the same frequency, and adjusting the spatial position of each sub-beam; the optical relay compensation unit is used for adjusting the optical path of each sub-beam; and the microscopic focusing unit is used for focusing the sub-beams after the optical path adjustment.

The invention discloses a tunnel central axis rapid extraction method based on point cloud data, and belongs to the technical field of tunnel engineering. According to the method, technologies such as global principal component analysis, local segmentation processing, RANSAC robust fitting and spline interpolation are innovatively combined, comprehensive analysis of the overall structure and local details of the tunnel is achieved, an efficient local splicing strategy is innovatively provided, the data processing amount is effectively reduced, the algorithm efficiency is improved, and the method is suitable for large-scale popularization and application. Segmentation processing and elliptic cylindrical surface model fitting are innovatively adopted, and adaptability to complex tunnel shapes is improved. According to the method, many limitations in the prior art are overcome, local features and global features are concerned at the same time, and a more efficient, more accurate and more reliable solution is provided for extraction of the central axis in tunnel engineering.

The invention relates to the technical field of semiconductor detection, and discloses an ATE equipment high-speed AC signal calibration system and method, and the system comprises at least one FPGA module which is provided with a plurality of channels for bidirectional signal transmission; each PE module is correspondingly positioned on one channel of the FPGA and is used for receiving the driving signal output by the FPGA and an external input signal and sending a feedback signal to the FPGA module according to the driving signal; the signal calibration module is located in the FPGA and used for receiving the feedback signals and comparing time delays between the feedback signals input from different channels; the delay compensation module is positioned in the FPGA and is configured to adjust the time delay of each channel based on the comparison result of the signal calibration module so as to realize the synchronization of the signals of the plurality of channels; and when a plurality of FPGAs exist, the system also comprises a PLL module, and the PLL module is configured to provide synchronous clock signals with the same frequency and fixed phase to the plurality of FPGAs at the same time, so as to ensure that the internal clocks of the FPGAs are synchronous.

The invention discloses an Internet of Things machine room control method and device.An Internet of Things machine room is provided with a plurality of temperature sensors and air conditioner assemblys.The method comprises the steps that the real-time detection values of the temperature sensors in the current detection period are obtained, and the temperature average value is calculated; based on the temperature average value of the current detection period, calculating a temperature prediction value of the next detection period by combining the temperature average values of the plurality of historical detection periods; and based on the temperature prediction value, an air conditioner assembly is controlled to adjust the indoor temperature of the Internet of Things machine room according to the indoor constant temperature value of the Internet of Things machine room, and the indoor temperature average value of the next detection period is adjusted to the indoor constant temperature value. The temperature average value of the current period is monitored and calculated in real time, the temperature change of the next period is predicted through historical data, then the air conditioning system is intelligently adjusted, it is ensured that the temperature in the machine room is maintained at the preset constant value, efficient and accurate temperature control is achieved, and the stability of the machine room environment and the reliability of equipment operation are improved.

The invention discloses a brake pad cutting device and a cutting method thereof, and relates to the technical field of brake pad cutting, the brake pad cutting device comprises a machine tool, a conveying assembly, a cutting assembly and a fixed-point cutting assembly.By flexibly adjusting the position of a trepanning laser nozzle, the trepanning requirements of different positions on a brake pad are met, and the size of a hole is flexibly adjusted; the electromagnet is started to generate magnetic force with the same polarity as the magnetic block, the magnetic block is pushed to rotate through repulsive force, rotation and positioning of the magnetic block are achieved under cooperation of the check block, the position of the rotating disc is adjusted, and cutting at different positions is achieved; the number of the magnetic blocks can be adjusted according to the number of actually-needed holes by controlling a fifth electric push rod to drive the magnetic blocks with the needed number to move outwards, the positions of an outer ring laser spray head and an inner ring laser spray head are adjusted, and an outer contour and an inner contour are cut on a brake pad raw material. All cutting operations can be completed through one-time positioning, and accumulative errors caused by multiple times of positioning are reduced.

The invention discloses an inspection robot obstacle avoidance algorithm based on imitation learning. The algorithm comprises sensor data acquisition, picture depth estimation, obstacle detection, multi-modal data fusion and action sequence generation. In the model generated by the action sequence, task actions are divided by adopting an action partitioning technology; the action blocking technology comprises the steps that the size of an action block is set to be k, every k time steps, an action sequence generation model receives a set of multi-modal data, k actions are generated, and the actions are executed in sequence. According to the invention, through multi-modal input data and a generation model, autonomous inspection and obstacle avoidance of the robot can be completed under the conditions of a complex external environment and a dynamic obstacle.

The invention provides a new energy automobile endurance prediction method based on working condition identification and prediction, and belongs to the technical field of new energy automobile endurance mileage estimation. According to the method, sliding window statistical characteristics are constructed through speed, current, voltage and other signals, typical working conditions of cities / suburbs / highways and the like are identified, the working condition of a next time window is predicted, and energy consumption priori which can be updated in real time along with changes of road conditions is formed. Secondly, a rolling capacity estimation model is provided, the available capacity of the battery is continuously evaluated based on historical charging and discharging data, and dynamic changes of the SOH of the battery along with time and use conditions are recorded; according to the method, working condition prior, rolling SOH and time sequence energy consumption characteristics are fused, a TCN-BiLSTM multi-model cooperation framework is constructed, short-term dynamic and long-term attenuation trends are considered, high-precision and generalizable endurance mileage prediction under complex working conditions and battery state fluctuation is realized, reliable support is provided for intelligent energy consumption management and journey planning of vehicles, and the method is suitable for popularization and application. The method can also be applied to an intelligent power distribution system with a high reliability requirement, and provides a core solution for technology promotion service in related fields.

The invention belongs to the technical field of geographic data registration, and discloses an urban map batch geographic registration method and system based on map feature classification. The method comprises the following steps: accurately classifying collected target maps based on map features, wherein the classification comprises dividing drawings with high repetition rate of drawing contents into the same class; unified batch processing of similar modern city maps is completed for the classified same-group maps by means of picture editor drawing software, and external reference grid cross points are loaded; and completing batch geographical registration of the same group of maps by utilizing the established external reference grid intersection points. The method is different from a thought starting from internal elements of a map, the thought of adding an external reference control point to the map is changed to carry out geographic registration, a batch high-efficiency geographic registration mode can be considered at the same time, the learning threshold is low, and the method has wide popularization and promotion value in the field of modern city maps.

The utility model discloses a differential housing inner spherical surface diameter measuring device, which comprises a main body, a fixed end ball measuring head is connected to the lower part of the right side of the main body, the left side of the main body is provided with an up-and-down through groove hole, a horizontally arranged sliding shaft is installed in the groove hole, and two ends of the sliding shaft are connected with two side surfaces of the main body. The sliding shaft is sleeved with a sliding seat, the bottom of the sliding seat is connected with a movable end ball measuring head, the top of the sliding seat is provided with a small handle, the top of the main body is provided with a large handle, the position of the large handle corresponds to the position of the fixed end ball measuring head, and the side face, away from the fixed end ball measuring head, of the sliding seat is provided with a flat measuring head. The technical problems that in the prior art, manual detection efficiency is low, data are not accurate, large equipment detection procedures are complex, and detection cost is high are solved.

The invention relates to the technical field of tunnel engineering, engineering geophysical prospecting, geophysics, seismic exploration and oil-gas exploration, and provides an intelligent radial interpolation seismic data processing method based on a multi-target mayfly naiad algorithm, which comprises the following steps: collecting logging data and seismic data; when radial interpolation seismic trace gather processing is carried out, a wide-azimuth trace gather is converted into a conventional trace gather, and conventional trace gather optimization processing is carried out on each azimuth; initializing parameters of the multi-target mayfly naiad algorithm, and setting a superposition weight range of each direction; constructing a target function of the multi-target mayfly naiad algorithm, and completing initialization of the multi-target mayfly naiad algorithm; carrying out iteration by adopting a multi-target mayfly naiad algorithm, and dynamically updating the superposition weight; and extracting the mayfly naiad position of the middle part in the Pareto set, and taking the superposition weight in the mayfly naiad position as the optimal superposition weight of each direction. According to the method, the multi-target mayfly naiad algorithm is adopted to determine the superposition weight of each direction of the radial interpolation seismic gather data, and high-fidelity processing of the radial interpolation seismic gather data is realized.

The invention relates to the technical field of electromagnetic inverse scattering, in particular to a three-dimensional electromagnetic inverse scattering method based on a diffusion model of physical constraints and BCGS-FFT (Binary Common Gateway Standard-Fast Fourier Transform), which combines the BCGS-FFT, physical models (CF-BPA and CC-SOM) and the diffusion model. The BCGS-FFT physical constraint consistent with the BCGS-FFT physical constraint in a data generation stage is introduced in a back diffusion stage, so that the physical consistency of a three-dimensional reconstruction process is ensured, and the problem of physical distortion is effectively reduced; the step-by-step denoising process of the diffusion model significantly improves the noise suppression capability in the three-dimensional reconstruction, and the preliminary inversion result of the physical model is combined, so that the error accumulation in the whole three-dimensional reconstruction process is reduced, the precision and stability are higher, and the adaptability is also greatly improved.

The invention discloses a clock data recovery circuit suitable for high-speed SerDes and a recovery method thereof, and the circuit comprises an analog front end which inputs a receiving signal and outputs a phase discriminator result; the phase discrimination accumulator is used for accumulating results of the phase discriminator; the gain control unit is used for adjusting the gain of the trunk signal; the first branch comprises a proportional decision device and a phase gain unit; the second branch comprises a weight decision device and a frequency integration unit; the summator is used for adding the output results of the phase gain unit and the frequency integration unit, and the output of the summator sequentially passes through the phase accumulator and the phase interpolator; and the phase interpolator is used for outputting a sampling clock for the analog front end to use. According to the method, timely response can be realized under the condition of small errors, the CDR locking speed is high, and the consumed time is short; the accumulated error of the system is reduced, the clock jitter tracking effect is enhanced, and the overall performance of the system is improved.

The invention relates to a non-visual clue brain-like SLAM (Simultaneous Localization and Mapping) method based on grid cells, which refers to a neural mechanism of the grid cells in a mammalian brain olfactory cortex, tracks and encodes the position of a robot only by utilizing self-motion information of the robot, and constructs a cognitive map of the robot and an interaction environment of the robot. The method comprises the following steps: firstly, constructing a multi-level grid cell continuous attractor network module, and encoding self-motion information of a mobile robot into cognitive representation of a grid cell module through a cognitive space transformation module; and then constructing a cognitive map in a navigation environment without visual clues by utilizing the cognitive representation of the grid cell modules. The invention provides a robust and efficient brain-like SLAM navigation method without visual clues, which can be used for intelligent autonomous positioning and navigation of the mobile robot in satellite denial and visual clue-free environments, and has important significance for expanding the application scene of the mobile robot.

The invention discloses an uplink and downlink combined wireless communication method and system based on channel semantic feature feedback. The method constructs an uplink and downlink closed-loop semantic communication architecture: a terminal side obtains a downlink channel matrix, generates a channel semantic feature vector through a channel semantic feature extraction network, inputs the channel semantic feature vector into a compression coding network to obtain a low-dimensional feedback code word, and sends the low-dimensional feedback code word to a base station side through an uplink limited feedback link; the base station side decodes and restores the low-dimensional feedback code word to obtain a reconstructed channel semantic feature vector, inputs the reconstructed channel semantic feature vector as prior information into a channel sensing sequence coding and modulation network, and generates a complex modulation signal under the condition of meeting the transmission power constraint; and the terminal side is combined with local channel information, and original data is recovered through a channel sensing demodulation network. Through joint optimization of the channel semantic feedback and the transceiver end-to-end transmission process, the channel state information feedback overhead is reduced under the constraint of limited feedback resources, and the transmission reliability and robustness in a complex channel environment are improved.

The invention relates to the technical field of motion posture sensing, in particular to a motion posture sensing method and system of a motion body. The method comprises the following steps: carrying out real-time shooting on a motion image in a motion process corresponding to a motion body to obtain real-time motion change image data of the motion body; performing real-time monitoring on motion state parameters of the motion process corresponding to the motion body to obtain real-time acceleration data and real-time angular velocity data of the motion body; performing time sequence alignment synchronization processing, and performing motion coordinate positioning and long and short term complementary filtering fusion analysis of the motion body to obtain a motion posture change fusion feature sequence of the motion body; and performing motion posture track recognition analysis on the motion process based on the motion posture change fusion feature sequence of the motion body, and performing posture change mode recognition analysis and motion posture sensing classification at the same time to obtain corresponding motion posture category information of the motion body in the current motion process. According to the invention, high-precision sensing of the moving posture of the moving body can be realized.