1125 results about "Data segment" patented technology

The invention discloses a resting electroencephalogram quality evaluation method and system based on double-branch comparative learning, and the method comprises the steps: collecting an original EEG signal X, carrying out the data preprocessing and data enhancement, generating two different enhanced views, transmitting the two different enhanced views to a double-branch encoder in parallel, respectively extracting a time domain waveform and a time-frequency domain rhythm feature, and carrying out the deep fusion, the fused features X1 and X2 are sent to a projection head, and through a self-supervised contrast learning mechanism, the network weight is optimized by using contrast loss; a small amount of labeled fine-tuning data sets including X and corresponding labels y are adopted, after flowing through a pre-trained double-branch encoder, the fine-tuning data sets are directly sent to a classification head connected with the back of the double-branch encoder so as to output a prediction result of an input data segment, and a mixed loss function including classification loss and comparison loss is adopted for training in the optimization process. According to the invention, a complete online real-time quality control system is established, and end-to-end real-time closed loop from data acquisition to quality evaluation is realized.

The invention relates to an aeration fan predictive maintenance method, system and equipment based on multi-modal perception and adaptive learning and a medium. The method comprises the following steps: generating a time sequence data set through synchronous acquisition and combined noise reduction processing of a sensor group; generating a multi-dimensional feature vector through time-frequency feature spectrum characterization and interpretability contribution analysis in combination with dynamic weight distribution coupled by environmental factors; on the basis of the multi-dimensional feature vectors, real-time anomaly detection is carried out at the edge end through a lightweight model, and abnormal data fragments are uploaded to the cloud end; and performing cross-sensor bidirectional reasoning on abnormal data fragments through a reasoning model deployed at the cloud, reconstructing a sensor topological graph, intelligently triggering elastic incremental learning, cooperatively processing equipment degradation trend analysis, multi-source evidence fusion and space calibration, and outputting a life prediction result and a fault thermodynamic diagram. According to the method, the core pain points of high early fault omission ratio, insufficient model robustness and the like are solved, and cost reduction, efficiency improvement and equipment life prolonging are realized while the diagnosis precision is maintained.

The invention discloses a data quality treatment method and system based on an AI Agent, and belongs to the technical field of artificial intelligence and data treatment. An initial data semantic distribution map is constructed, cross-dimension correlation feature factors are extracted, a multi-scale quality anomaly sensitive factor matrix is constructed, time sequence evolution weights are embedded, and a dynamic feature evolution trajectory is formed; performing perception modeling on the evolution trajectory by using an AI Agent, generating a multi-level quality risk thermodynamic diagram, extracting a deviation dense region and constructing an anomaly propagation path set; in combination with upstream and downstream data links and task flow information, calculating a potential impact factor weight, constructing a causal traceability map, and injecting a correction strategy label for a key field node; the AI Agent autonomously selects an adaptive strategy combination according to the target data segment, and performs online intervention on the target data segment; according to the method, closed-loop treatment of the data quality problem from perception and judgment to intervention and feedback is realized, and the method has the advantages of self-adaption, high interpretability and the like.

Systems and methods for generating generative AI based solution are disclosed. A system receives a request for generating the generative AI (GenAI) based solution. The system classifies the received request into solution patterns and performs actions corresponding to at least one of the solution patterns. The system extracts a metadata from the received request and the actions based on the type of GenAI based solution to be generated. Further, the system segments the extracted metadata into data segments and generates a vector representation of the data segments. The system generates the GenAI based solution corresponding to the received request based on the generated vector representation. The system validates the GenAI based solution large language model (LLM). The system continuously updates the LLM and the vector-based machine learning model with the validated GenAI based solution and user feedback. The system outputs the GenAI based solution on a user interface.

Backup and recovery of multi-party computation (MPC) security data utilized to secure digital assets and transactions thereof can enhance user confidence and user experience in digital asset transactions. Example MPC security data can include cryptographic keys and key shares utilized with a N×M MPC signature and validation framework. A computing device participating in generation of MPC secure data can retain a segment of the MPC secure data and can encrypt and store an encrypted segment at a second device. A recovery service or recovery application at the second device can facilitate recovery of the MPC secure data segment at the computing device, or at an additional device not involved in generation of the MPC secure data. The recovery service or application can facilitate recovery of the MPC secure data segment even in the event the computing device is lost.

A correlation-aware adaptive codebook compaction system for multi-modal data compression that preserves cross-modal relationships while providing enhanced reconstruction quality. The system analyzes temporal and spatial relationships between different data modalities to generate correlation maps that guide compression decisions. A virtual management layer performs stream characterization and adaptive routing, while a processing pipeline implements primary codebook compression with mismatch handling for novel data blocks. High-entropy data segments receive pre-compression processing before codebook compression. Sequential registration data is processed through matrix factorization and dedicated matrix codebooks. The system continuously monitors data distribution characteristics and automatically retrains codebooks when drift thresholds are exceeded. A neural upsampling subsystem uses correlation information to guide cross-modal enhancement processes through modality-specific networks and attention mechanisms. The unified output includes compressed data streams, correlation maps, synchronization metadata, neural model parameters, and updated codebooks, enabling synchronized reconstruction with preserved cross-modal relationships and enhanced quality through correlation-guided neural upsampling.

The invention discloses a non-intrusive household load identification method based on CUSUM event detection and a convolutional neural network, which is applied to total load data output by an intelligent electric meter installed at a user total incoming line, and comprises the following steps: data preprocessing; carrying out load event detection based on a sliding window bilateral CUSUM algorithm so as to determine an occurrence moment and a continuous interval of a load event; in a continuous interval corresponding to each load event, calculating an active power variation delta P, a reactive power variation delta Q, a current harmonic distortion variation delta Ihd and a current span Itc in steady-state data segments before and after the load event, and combining to form a multi-dimensional event feature vector; the method comprises the following steps: sequentially executing the steps on operation data of each electric device in a family in different operation modes to obtain a multi-dimensional event feature vector marked with a device type and an operation state label, and constructing a standardized device load feature database; and training a convolutional neural network classification model and realizing load identification.

The invention relates to the technical field of sleep environment regulation and control, and discloses a sleep environment self-adjustment system based on multi-source heterogeneous data. An environmental parameter acquisition module and a physiological feature acquisition module of the system respectively acquire sleep space environmental parameters and user physiological feature data in real time; after the master control system receives the two types of original data, a fusion calculation unit cleans original environment parameters to generate a standardized environment data stream, and extracts features from the original physiological feature data to generate a physiological feature time sequence; the dynamic evaluation module divides the two types of data into a plurality of sleep stage data segments according to a preset rule, and calculates an environmental parameter fluctuation index and a physiological feature deviation index; the label distribution mechanism combines the two types of indexes to generate a comprehensive comfort label of each sleep stage; an adjustment decision engine generates a global sleep environment adjustment strategy according to all labels, and an actuator control module drives environment adjustment equipment to execute; and the feedback learning unit receives the adjusted data and updates the calculation benchmark of the dynamic evaluation module.

The embodiment of the invention provides a cross-network data exchange method, electronic equipment, a storage medium and a program product, and the method comprises the steps: receiving transmission data, determining the identification information of the transmission data according to the service identification, timestamp and segment sequence number of the transmission data, and determining the verification value according to the performance requirement, the security requirement and the data integrity; the hash algorithm is obtained by performing hash operation on the segmented data; according to the identification information, judging whether the transmission data is duplicated data or not; judging whether the transmission data is complete data or not according to the verification value; whether the transmission data is lost data or not is judged according to the segment serial number, if yes, a retransmission instruction is sent to the sending end equipment, the data segments which are repeatedly transmitted can be accurately recognized through the unique identification information, redundant storage and processing are avoided, the data repetition problem is solved from the source, then the integrity of the data is verified according to the verification value, and the data transmission efficiency is improved. And the lost data end is judged according to the segment sequence number, so that the problem of data loss is solved.

The invention relates to the technical field of cable fault monitoring, in particular to a power cable fault monitoring method and system based on big data, and the method comprises the following steps: carrying out the optimization and integration of spatial layout and sampling time based on a cable monitoring node, recognizing a data segment with the reverse trend of a current peak and a voltage phase angle, and carrying out the recognition of a current peak; and further analyzing the synchronism of the current-carrying density and the temperature, extracting a periodic intersection distribution characteristic, screening a continuously progressive trend track, and forming a trend directivity factor. According to the method, the linkage accuracy among multi-point signals is guaranteed by combining synchronous acquisition of node data with time sequence calibration, the trend discrimination method performs step-by-step screening for collaborative fluctuations of different characteristic parameters, and a continuous evolution sequence of various types of tiny disturbances under a complex working condition is dynamically extracted; a temperature and current-carrying density correlation mode is adopted to highlight the progressiveness and the positioning capability of a fault precursor, and dynamic risk signals which are difficult to identify at the initial stage of change are effectively captured.

The invention relates to the technical field of industrial big data storage and management, and discloses an industrial data full life cycle management method and system.The method comprises the steps that original time sequence data of industrial equipment is obtained, Shannon entropy calculation is carried out, and a time sequence importance score sequence is obtained; performing threshold comparison according to the time sequence importance score sequence, and determining a key transient window and a stable normal window; performing lossless differential coding on the key transient window to obtain a high-fidelity data segment; performing feature extraction and sparse processing on the stationary normal window, and performing time domain splicing and index construction to obtain a primary hybrid storage structure; and continuously monitoring the query access path to generate a popularity distribution graph to obtain a hierarchical archiving data set. According to the method, value-aware hierarchical storage and dynamic full-life-cycle management of massive industrial data can be realized.

The invention discloses an electric vehicle charging load data virtual acquisition method and system based on a conditional diffusion model, and a storage medium, and relates to the technical field of power system operation optimization and artificial intelligence, and the method comprises the steps: constructing a dual-channel feature extraction network, and outputting conditional information; injecting conditional information into the conditional diffusion model by using a cross attention mechanism for de-noising; introducing a self-supervised alignment loss function, and constructing a self-supervised regularization mechanism with a stable interpolation result; and training the conditional diffusion model according to the historical charging load data set, and performing reverse denoising on the target missing data fragment. The method has the advantages that data features are comprehensively captured, accurate guidance of condition information is achieved, output stability is effectively improved, and a high-fidelity interpolation result highly consistent with real data distribution is generated through reverse denoising.

The invention discloses an Internet of Things main service platform data analysis cooperation method, and relates to the technical field of data analysis, and the method comprises the following steps: obtaining a to-be-analyzed terminal uploading data set, and detecting the real-time transmission time delay and data encryption level of each data segment in the terminal uploading data set; based on the historical data analysis collaborative log, extracting a tolerance loss coefficient between the real-time transmission time delay and the analysis fault-tolerant rate, and obtaining an analysis fault-tolerant deviation value caused by transmission time delay change according to the tolerance loss coefficient and the real-time transmission time delay; if the real-time transmission time delay of the current Internet of Things main service platform does not cause interference on the data encryption analysis algorithm, processing the data encryption grade and the analysis calculation power consumption to obtain a calculation loss coefficient, and performing calculation power compensation on the initial analysis calculation power distribution quantity according to the calculation loss coefficient, the data encryption grade and the analysis fault-tolerant deviation value; obtaining a first actual analytical calculation force; the method has the effect of improving the collaboration efficiency and stability of data analysis of the whole platform.

The invention discloses a child activity data desensitization processing method and system based on differential privacy, and relates to the technical field of privacy data protection, and the method comprises the steps: carrying out the time synchronization and unit normalization of multi-source data from a wearing device and an environment sensor, and building a unified multi-dimensional time sequence matrix; then, introducing a change point detection method to dynamically segment the time series data, so that the data is divided into event segments with clear behavior semantics; a sensitivity scoring mechanism is designed, factors such as activity types, time, places and behavior frequencies are comprehensively considered, privacy risk quantification is performed on each event segment, adaptive privacy budget allocation is implemented based on a scoring result, and key protection of sensitive data segments and efficient utilization of conventional data segments are realized; after differential privacy noise is injected, time sequence consistency correction and cross-feature correlation correction are combined. Therefore, differentiated and refined protection of the child activity data is realized, and privacy protection and data availability are effectively balanced.

The invention relates to the technical field of embedded data storage, and discloses a data storage method and system for an embedded storage chip. The method comprises the following steps: acquiring a to-be-stored original data stream through a data acquisition interface and performing segmentation marking; performing hierarchical decomposition on the data segments by adopting a dynamic decomposition algorithm to generate a plurality of data components; calculating an information density index of each data component, and screening key data components and non-key data components according to the information density index; key data components are stored in a high-speed cache region, and non-key data components are stored in a conventional storage region; performing redundancy check on the key data components in the cache region and generating a check result; after correcting the key data component based on the verification result, recombining the key data component with the non-key data component to generate an optimized data stream; writing the optimized data stream into a physical storage medium and recording a storage position; and monitoring the storage load state, and dynamically adjusting the distribution proportion of the two storage areas.

The invention discloses a smart city data capitalization-oriented data right confirmation management method and system, and belongs to the technical field of data resource management. The method comprises the following steps: receiving data assets uploaded by a user and carrying out multi-modal preprocessing on data contents to generate a data fingerprint with a unique identifier; counting the proportion of the unconfirmed right data fragments; judging an infringement risk level of the data assets in combination with a risk early warning model; if the unconfirmed right proportion exceeds the dynamically calculated threshold value range or the infringement risk level reaches a preset threshold value, executing an audit interception operation and returning a right confirmation abnormity prompt to the user, and requesting to supplement ownership proof; and embedding a multi-modal fusion watermark into the data assets passing the auditing, and generating a cross-domain verifiable right confirmation voucher. According to the method, the data features are extracted through multi-modal preprocessing to generate the unique data fingerprint, the ownership registration state of the data fingerprint is inquired in combination with the block chain technology, the ownership-confirmed and unconfirmed data fragments are accurately identified, and the problem that the data ownership is difficult to accurately and efficiently identify is solved.

The present application provides a video segmentation method, a server, a storage medium, and a program product. In the method of the present application, video data to be segmented is segmented into multiple data segments, unimodal features of the data segments, including text features of a text modality and visual features of a visual modality, are respectively extracted by means of a video topics segmentation model, and then the text features and visual features of the data segments are fused, so that the fusion of multimodal information can be performed at the intermediate representation level, the relationship and interaction between different modalities can be better captured, and higher-quality multimodal fusion features of the data segments are obtained. Furthermore, on the basis of the multimodal fusion features of the data segments, whether the data segments are topic boundaries is predicted, so that the topic boundaries of the video data can be accurately predicted, improving the accuracy of topic boundary recognition, thereby improving the accuracy and quality of video topics segmentation results.

The invention discloses an access instruction sending method and device, a storage medium, electronic equipment and a computer program product, and relates to the field of computers.The method comprises the steps that under the condition that a first notification message sent by a network card is received, control segment data of a target message is read to a first-level cache and a second-level cache of a central processing unit, wherein the first notification message is used for indicating that the network card has received the control segment data; and determining an access instruction according to the control segment data in the central processing unit, and sending the access instruction to a target hard disk to indicate the target hard disk to read the data segment data of the target message stored in the network card to the target hard disk through a direct connection communication channel between the network card and the target hard disk.

The application belongs to the technical field of medical signal processing, and discloses a sleep apnea detection method and system, an electronic device and a storage medium. The method comprises the following steps: acquiring a synchronous ECG signal and a breathing signal, and processing the synchronous ECG signal and the breathing signal into a bimodal data segment containing multiple time scales; constructing a parallel neural network architecture, extracting ECG features and breathing signal features from each scale data segment respectively, and forming a cross-modal feature pair; designing a cross-scale dynamic weight correction attention mechanism, introducing a correction factor based on the Euclidean distance between feature vectors, dynamically weighting and information fusion on the cross-modal feature pair between different scales, so as to enhance the relevance of local details and global context; inputting the fused multi-scale features into a classifier, and outputting the probability of a sleep apnea event. Through multi-scale collaborative analysis and a dynamic weight correction attention mechanism, the application effectively fuses physiological information of different time dimensions, and improves the detection accuracy.

The invention relates to a sensitive information protection method and device, a storage medium and electronic equipment. The method comprises the following steps: executing structured field identification and general sensitive data identification processing on original input data to identify all sensitive data fragments; performing identification calculation and integrity signature processing on each sensitive data segment to generate a corresponding placeholder, and storing a mapping relationship between each sensitive data segment and the placeholder thereof in a mapping table; according to the mapping relation, replacing each sensitive data segment in the original input data with a corresponding placeholder to obtain replacement input data, and inputting the replacement input data and the output rule into the target large model to enable the target large model to output the original output data; and according to the mapping relationship, replacing each placeholder in the original output data with the corresponding sensitive data fragment to obtain final output data. The technical problem that sensitive data desensitization is prone to misoperation is solved.

The invention relates to the technical field of aircrafts, in particular to an aircraft key component fault monitoring method and system based on wireless WQAR data, and the method comprises the following steps: S1, data acquisition and preprocessing: receiving wirelessly transmitted fast access recorder WQAR data from an aircraft; the WQAR data and the maintenance record data are associated and fused; performing cleaning, abnormal value processing and working condition division on the fused data to extract an effective data segment corresponding to a specific flight stage; s2, establishing a statistical baseline model for one or more monitoring parameters of at least one key component on the basis of the effective data segments in historical normal operation; according to the method, the accurate health baseline model and the real-time anomaly detection mechanism are established, so that the performance degradation and potential faults of the parts can be recognized in an early stage, the maintenance mode is changed from post-processing to pre-warning, and serious faults are effectively avoided.

The invention belongs to the technical field of vehicle networking safety, and relates to a lightweight in-vehicle network intrusion detection system and method based on an ADGRU-Net hybrid model, and the system comprises a data loading and cleaning module, a time sequence and feature correlation module, and a space-time hybrid deep learning model. The data loading and cleaning module adopts a block-by-block loading mode to convert multi-source heterogeneous original time sequence data into a standard format; the time sequence and feature association module adopts a sliding window mechanism to intercept a data segment containing L continuous time steps with a preset length from multi-source time sequence data, each time step contains N features, and the data segment is transposed and then converted into a two-dimensional single-channel grayscale image through linear scaling; the space-time hybrid deep learning model comprises a spatial feature extraction module, a remodeling module, a time sequence feature extraction module, a time sequence branch decoder, a spatial branch decoder and a feature fusion module; the problem that an existing model is poor in performance in a complex and changeable real attack scene is solved.

The invention relates to an acupuncture rehabilitation training data processing method based on the combination of VR and digital therapy, and the method comprises the steps: obtaining an acupuncture stimulation signal and a VR training interaction signal, carrying out the synchronous association of the two signals based on a unified time reference, and dynamically determining a signal processing priority according to the state of a patient when the acupuncture operation and a virtual training task conflict, carrying out backtracking calibration on the offset signal; the task action rhythm, force and completion condition of a patient are monitored in real time in VR rehabilitation training, data of abnormal behaviors are collected and recorded, whether the abnormal behaviors are caused by acupuncture stimulation or not is analyzed by combining acupuncture stimulation parameters and time sequence information, and data confirmed to belong to acupuncture effect interference are marked as independent samples; real-time subjective feedback information of a patient is collected in a VR scene, subjective indexes of comfort, aching feeling and fatigue are associated with physiological and behavior data, analysis weights of corresponding data segments are adjusted according to feedback, and an individualized rehabilitation data processing strategy of the patient is formed based on historical feedback data.

The invention provides an intelligent inspection management system in a hybrid expert mode under a cloud edge collaborative architecture, and relates to the technical field of data processing, the system is used for cutting basic inspection data into a plurality of section units according to a business time window, and combining record elements and node sequences in each section unit to obtain a combined data segment; constructing a track series sequence by taking the occurrence sequence of each record element as a reference, and converting the track series sequence into track matrix data; constructing node load data and calculating a process aggregation factor; screening out candidate sections in section combination data and constructing a section traffic matrix; calculating a process shunting factor; and determining process node data, and constructing an inspection service link according to the process node data. According to the invention, the problem of multiple circulation of error events in the process can be solved, and the uniqueness of the inspection task is ensured.

The invention provides a real-time audio feature extraction method and system based on a sliding window, and belongs to the technical field of audio feature extraction. The method comprises the steps of determining a sliding step length of a current feature extraction period based on audio content characteristics of latest frame data in an audio buffer area; cache index alignment operation is executed, and whether reusable features corresponding to the overlapped data segments exist in a preset feature cache region or not is inquired; calculating newly added features, and generating complete spectrum features of the current processing window by multiplexing the reusable features and combining the newly added features; and generating an audio feature vector, updating the audio buffer region and the feature buffer region, and completing real-time audio feature extraction of the current period. According to the method, the sliding step length is dynamically determined based on the audio content characteristics, so that the adaptive adjustment of the processing frequency is realized, and the real-time performance and the characteristic quality are effectively balanced; and through a cache index alignment and feature multiplexing mechanism, the calculation efficiency is remarkably improved.

The invention provides a cross-security domain computing power resource federation and privacy protection system. The system comprises a scheduling management node, a communication gateway, a plurality of computing nodes, a verification node and a key management node. The scheduling management node decomposes the calculation general task into a plurality of subtasks and plans different node execution paths; the communication gateway encapsulates the subtask data into an encrypted data packet containing a position identification segment and an encrypted data segment which can be independently decrypted; the computing node is integrated with the trusted execution environment to provide hardware-level security isolation; the summarizing node is used for receiving sub-task results which are processed by the computing nodes and comprise encrypted data segments; the verification node realizes calculation integrity verification; and the key management node confirms a task completion state by collecting the position identification segment, and coordinates and starts a data segment aggregation decryption process. According to the method, the data and code privacy of the computing task is ensured while the computing power island is broken, and a cross-domain computing power resource sharing mechanism with balanced safety and performance is established.

The invention provides a cloud side end collaborative cross-domain data transmission system and method, and relates to the field of data processing, and the method comprises the steps: based on the collaboration of a cloud center, an edge node and an equipment end, the cloud center carries out the fragmentation of to-be-issued transmission data to obtain a first data fragment, and carries out the recombination of an uploaded second data fragment; the edge node dynamically selects the optimal transmission path of the data fragments and schedules the transmission of the data fragments; and the device end fragments to-be-uploaded transmission data to obtain a second data fragment, and recombines the first data fragment. The system and the method provided by the invention are suitable for a heterogeneous network, and the data transmission efficiency is greatly improved by performing fragmentation processing on the transmission data and combining a priority transmission mechanism; by dynamically selecting the optimal transmission path of the data fragments, multi-link aggregation is realized, the throughput under multiple network scenes is greatly improved compared with a traditional scheme, and self-adaption of a heterogeneous network can be realized.

The invention relates to the field of press machine detection, and particularly discloses a press machine anomaly detection method, which comprises the following steps of: segmenting continuous force and displacement signal flow into independent stroke data segments, and extracting multi-dimensional feature vectors containing statistics, form and frequency domain information from the stroke data segments so as to comprehensively describe the state of each stamping stroke. Then, inputting the current stroke feature vector into an acute anomaly detector based on a heterogeneous auto-encoder, and sensitively capturing an emergent and isolated abnormal event in a high reconstruction error form by learning a reconstruction mode of normal data; besides, a chronic reference drift detector based on a Wasserstein distance is utilized, and systematic slow drift caused by equipment wear or working condition change is identified by analyzing distribution change of feature vector flow within a period of time. And finally, through an intelligent decision-making and arbitration mechanism, the acute abnormal score and the drift signal are fused, and a real equipment fault is identified.

A device for deep learning acceleration with mixed precision may include matrix-vector (MV) components that each include vector-vector (VV) components that are each configured to generate a respective VV output based on an input precision mode, an output precision mode, and an accumulation of products. The accumulation of products may be calculated by adding products based on the input precision mode. Each product may be calculated by multiplying, based on the input precision mode, a map data segment and a kernel data segment. Each MV component may include one or more components configured to concatenate VV outputs to generate a concatenated VV output. The device may include activation function components that are each configured to receive a corresponding concatenated VV output, generate an activation function output based on the corresponding concatenated VV output and the output precision mode, and output the activation function output.

The invention belongs to the technical field of exercise training and analysis, and particularly relates to an intelligent badminton action analysis system and method based on a motion sensor, and the system comprises a sensor data collection module, a real-time data processing module, an action recognition analysis engine, an intelligent data storage module and a visualization module. The method comprises the following steps: receiving and buffering sensor data in real time; periodically extracting data segments by using a sliding window to carry out multi-feature (an acceleration peak value, an angular velocity peak value, a peak value ratio and the like) calculation; based on a predefined multi-feature threshold rule set, identifying action types (such as kicking, flat ball drawing and the like) and evaluating the quality; an intelligent strategy is adopted to selectively store data so as to optimize the system load; and finally, a result is visually fed back in real time. The method realizes objective, quantitative and real-time analysis of badminton actions, has the advantages of high recognition accuracy, high system real-time performance, efficient data processing and the like, and is suitable for autonomous training of athletes and auxiliary guidance of coaches.