418 results about "Embedded processing" patented technology

The invention discloses an aircraft wheel cooling method and device, and the method comprises the steps: collecting the temperature of a brake disc and the air pressure of a hub in real time through a temperature and pressure sensor, calculating the temperature change rate and the pressure difference through an embedded processing module, and generating a thermodynamic initial parameter set; calculating a thermal stress distribution value of the brake disc based on the initial parameter set, and generating a cooling liquid flow and spraying angle instruction set in combination with a material threshold database; the opening degree of an annular pipeline piezoelectric valve is regulated according to an instruction set, an injection area is regulated according to flow velocity-surface area heat exchange efficiency, and waste heat is synchronously recovered to optimize power consumption; the thermal infrared imager obtains temperature field distribution, the initial parameter tolerance range is compared, the spraying angle and the flow instruction are corrected through a fuzzy PID algorithm, and temperature gradient errors are eliminated; and based on the temperature drop rate, the pressure stability and the waste heat recovery efficiency, an efficiency index is generated by using a neural network scoring model, and algorithm weights and piezoelectric valve response parameters are dynamically adjusted. The cooling efficiency is effectively improved, and the thermal stress risk is reduced.

The invention provides an unmanned aerial vehicle intelligent power line inspection system, and relates to the technical field of power line inspection. The unmanned aerial vehicle intelligent power line inspection system comprises an unmanned aerial vehicle platform which comprises an unmanned aerial vehicle body, a flight control system, a power system and a navigation system; the sensor module comprises a high-definition camera, an infrared camera, a laser radar and an ultrasonic sensor; the data processing module comprises an embedded processor, a data storage unit and a data transmission module and is used for processing and analyzing the data acquired by the sensor; the intelligent algorithm module is used for intelligently analyzing and judging the state of the power line; and the communication module comprises a wireless communication module and a satellite communication module and is used for realizing data transmission and remote control between the unmanned aerial vehicle and a ground control center. The unmanned aerial vehicle is used for replacing manual inspection, so that the work of personnel in dangerous environments such as high altitude and complex terrains is reduced, and the personnel risk is reduced.

The invention relates to a final-stage cache design method based on near memory calculation, and belongs to the technical field of near memory calculation. And compared with a near memory processor near a traditional main memory, data access and calculation can be carried out more quickly, and the calculation throughput rate can be improved. According to the near storage calculation, a near storage processor and a last-stage cache storage array are directly designed in an integrated mode, faster memory access is achieved through an internal bus interconnection mode, and the problem of a storage wall between the calculation speed and the memory access speed and the problem of a power consumption wall of data handling are relieved to a certain degree. The last-stage cache controller cooperates with the embedded processor, and can dynamically coordinate between a standard cache function and a calculation mode. Through a synchronization mechanism in a last-stage cache controller, a memory access request from a host CPU and a calculation task executed by an embedded processor in a last-stage cache are transparently balanced, so that efficient near-memory calculation under cache consistency is realized.

The invention provides a code quality adaptive evaluation and optimization method, system and device and a medium, and belongs to the technical field of computers. The method comprises the following steps: performing redundant information cleaning and standardization processing on an original code; selecting an adaptive parser to convert the processed code into an abstract syntax tree, and optimizing the abstract syntax tree; performing embedding processing on each node of the abstract syntax tree by using a pre-trained deep learning model to generate a node semantic vector, and performing aggregation, feature extraction and vector normalization operation to obtain a semantic vector representing code semantic features; extracting a dependency relationship from the abstract syntax tree to construct a program dependency graph, extracting node features by using a neural network model, and generating a dependency relationship analysis result; based on the semantic vector and the dependency analysis result, identifying a code potential problem by using an optimization algorithm and generating an optimization suggestion; codes are automatically modified according to optimization suggestions, the optimization effect is re-evaluated, and related models and algorithms are iteratively optimized according to feedback information.

The invention belongs to the technical field of radar human body tracking and recognition, and particularly relates to a rotary millimeter wave radar human body tracking and recognition device and method. The millimeter-wave radar module is mounted on the rotating holder module, and the main beam direction of an antenna in the millimeter-wave radar module is perpendicular to the rotating shaft; the millimeter wave radar module is connected with the control module, and the embedded processor is internally provided with a point cloud data processing algorithm and an identity recognition neural network which are used for overcoming the influence of environmental noise and multipath effect on data quality. The problems of fixed field angle and limited monitoring range of the traditional millimeter wave radar are effectively solved; and through an improved data processing algorithm and an identity recognition neural network, the influence of environmental noise and a multipath effect on data quality is overcome, and the reliability and adaptability of the system in a complex scene are remarkably improved.

A medical monitoring system comprises a wearable garment incorporating multiple physiological sensors for real-time patient monitoring. The garment integrates an embedded processing unit that receives sensor data and transmits it wirelessly to a base station or remote server. A multi-modal physiological signal processing unit analyses the data to generate composite physiological indices, including metabolic rate, work of breathing, and fall risk assessments. The garment includes optional monitoring features such as transcutaneous carbon dioxide measurement, electrocardiography, pulse oximetry, blood pressure monitoring, and accelerometer - gyroscope-based motion tracking. The system provides early-warning alerts for clinical deterioration, enabling proactive therapeutic interventions. A wireless communication module ensures real-time data transmission, supporting remote monitoring and integration with hospital networks. The garment's modular design allows for non-invasive physiological assessment while maintaining patient mobility, making it suitable for critical care, ambulatory monitoring, and telemedicine applications.

The invention relates to the technical field of artificial intelligence and thermal power production fault detection, in particular to a thermal power production fault detection data measuring point screening method and system based on a clustering algorithm, and the method comprises the steps: obtaining historical fault data of thermal power plant production equipment, and selecting related measuring points to form a multi-dimensional time series data set; and performing embedding processing on the time sequence data by using the time sequence large model to generate a feature vector with a set dimension. And then, carrying out clustering analysis on the feature vectors by adopting a clustering model to obtain a clustering result. And finally, screening out the measuring points related to the fault points of the production equipment of the thermal power plant according to a clustering result, obtaining a corresponding screening result, and providing support for equipment fault prediction and diagnosis.

The invention relates to the technical field of electrical equipment state monitoring and fault early warning, and particularly discloses an electrical cabinet condensation risk online monitoring method and system, and the method comprises the steps: synchronously collecting the physical parameters of the same microscopic space point in an electrical cabinet; calculating the classical dew point temperature by adopting a Magnus formula; performing quantification processing on the air pressure micro-fluctuation data in the time window, and calculating information entropy of micro-air pressure fluctuation; the wall surface temperature gradient and the air flowing state are fused, and the equivalent dew point temperature is calculated; real-time safety margin is calculated, and the instantaneous change rate of the safety margin is solved; constructing a condensation risk trend index; comparing with a preset threshold value, generating an early warning signal and a risk diagnosis label, and completing condensation risk monitoring; the system comprises a multi-physical field sensing unit, an embedded processing unit, a self-adaptive control unit, a man-machine interaction unit and a communication unit which are integrated in the same shell.

The invention relates to the technical field of robots, and discloses a humanoid robot capable of sweeping a floor and a control system thereof, and the humanoid robot comprises a cleaning module with a biped walking capability, a multi-degree-of-freedom joint humanoid body, an integrated dust collection device, a rolling brush, a mop and an automatic switching mechanism, a sensing system comprising a laser radar and the like, and a navigation and path planning system based on an SLAM algorithm. And human-computer interaction interfaces such as voice interaction are supported. The control system comprises an embedded processor, a task scheduling unit, a state monitoring unit, a fault processing module and an energy consumption management unit, and can coordinate the work of each module. The robot has the advantages that the problem that an existing sweeping robot is limited in cleaning capacity, and a humanoid robot lacks a cleaning function is solved, organic combination of flexible movement and efficient cleaning is achieved, and the intelligent cleaning requirement under the complex indoor environment is met.

The invention provides an energy coordination control system and scheduling method for multi-energy complementary new energy storage, and belongs to the technical field of new energy storage and energy technology management. The data acquisition and processing module is connected with a current and voltage sensor, a power transmitter, a meteorological station sensor and a battery management system sensor, and is used for acquiring voltage, current, power, state of charge and environmental data of photovoltaic power, wind power, energy storage and load through the sensors; and sliding window filtering and wavelet de-noising processing are carried out through the embedded processor. The method has the beneficial effects that the uncertainty of wind and light output is effectively processed by establishing a layered distributed optimization architecture and adopting a method of combining multi-objective optimization and robust optimization, and the consumption rate of renewable energy sources can be remarkably improved, the net load fluctuation can be reduced and the energy consumption can be reduced on the premise of meeting various system operation constraints. Therefore, dependence on conventional fossil energy is reduced, and the economic operation level and the environmental protection benefit of the whole energy system are improved.

The invention provides a collaborative data migration scheduling method based on topology awareness and deep reinforcement learning, and the method comprises the steps: initializing system configuration, constructing a weighted graph comprising a source server, a working machine, a target storage and network equipment, carrying out the embedded processing of a graph structure through a graph neural network, and generating a node and a path representation vector; in combination with source data, a working machine and a target storage state, constructing a system state vector, inputting the system state vector into a deep reinforcement learning agent for decision making, and outputting a composite action including data block selection, working machine selection, path indexing and migration rate suggestion; the network control module converts path information into flow rules, the flow rules are issued and executed through the software defined network controller, the task management module continuously monitors the migration state and processes failed data blocks, and the integrity and non-repeatability of migration tasks are ensured. According to the method, the defects of an existing data migration system in the aspects of task allocation, path selection, resource control and dynamic adaptability can be overcome.

The invention relates to a device and a method for assisting in detecting depression degree by utilizing emotional picture cognitive deviation, and relates to the technical field of biomedical engineering and medical instruments, and the device comprises an emotional picture library module which is used for storing a pre-selected emotional picture set; the display interaction module is used for displaying pictures to the user and recording classification selection and reaction time of the user; the data acquisition module is used for acquiring user facial micro-expressions and synchronously storing the facial micro-expressions and classification results; the data processing module is used for calculating positive picture error rate related features and normalizing behavior features and micro expression features into feature vectors; the depression index calculation module is used for outputting depression indexes; and the result output module is used for generating a visual report. The depression index is calculated based on behavior data, deviation is avoided, multi-modal features are fused to improve quantization precision, low-cost non-intrusive screening is achieved through a touch screen and a camera, and an embedded processor is integrated to improve the integration level and generalizability.

The invention provides a solid state disk data reading method and device, equipment and a storage medium. Relates to the technical field of storage. The method comprises the steps of obtaining a historical read operation data set of the solid state disk; the method comprises the following steps: dividing a historical read operation data set according to a set time window to obtain a plurality of samples, and performing embedding processing on the samples to obtain a corresponding embedded time sequence; performing stacking processing on the embedded time sequence through N stacking layers to obtain a stacking enhanced time sequence; linear prediction is carried out based on the stacking strengthening time sequences corresponding to all the samples, and a pre-reading operation sequence is obtained; according to the pre-reading operation information, reading corresponding data from a flash memory of the solid state disk and storing the data in a memory buffer area of a controller of the solid state disk; and receiving a data reading command, determining that the data reading command hits the data in the memory buffer area, and reading the hit data from the memory buffer area through the PCIe bus. Therefore, low-computing-power model prediction is realized and resource consumption and access delay are reduced.

The invention provides an embedded image enhancement method and system for a large-target-surface image sensor. According to the method, temperature gradient distribution and environment humidity data of a large-target-surface image sensor are collected, and an environment parameter matrix and a temperature noise mapping model are constructed in an associated mode; and correcting optical distortion of the incident light signal by using the liquid crystal modulator array, and outputting an original image. Then through a double-path module, a main path extracts noise features, and an auxiliary path separates color shift features caused by humidity; and verifying the physical consistency of the characteristics based on a thermodynamic equation and screening effective data. And finally, dynamically optimizing the image contrast, the dynamic range and the target details by the embedded processor, and outputting an enhanced image. Through environmental parameter modeling, dual-path feature separation and physical verification, temperature and humidity interference is effectively suppressed, and detail enhancement and color fidelity of the large-target-surface image sensor are improved.

The invention provides a multi-mode integrated auxiliary system and method for a special demand group, and the system comprises a basic function subsystem which is used for collecting basic information, transmitting the basic information to an embedded processing unit, and providing an auxiliary service for the special demand group according to the related content from the embedded processing unit; the embedded processing unit is used for driving each module in the basic function subsystem, transmitting the received basic information to the cloud server and downloading related contents from the cloud server to the basic function subsystem; the cloud server is provided with a two-way sign language translation subsystem, a visual impairment navigation subsystem and a text-voice two-way conversion subsystem, and is used for carrying out two-way sign language translation, visual impairment navigation and text-voice two-way conversion on the basis of the basic information and sending generated related contents to the embedded processing unit or the user side; and the user side is used for visually displaying the related content.

The invention discloses a millimeter wave barrier gate radar multipath interference suppression method and device, and belongs to the field of millimeter wave radar signal processing. The method comprises the following steps: S1, generating an RD image by an echo signal, and executing 2D-CFAR detection to obtain a candidate target unit; s2, eliminating ground reflection (GB) through an RD domain multipath sensitive area mask (MVZ); s3, converting the candidate target unit into a Cartesian coordinate system point cloud; s4, through point cloud domain mirror symmetry verification, removing a pole side mirror image (PSR); s5, clustering the reserved point cloud to obtain a target cluster; s6, through vehicle contour modeling, constructing a rear suppression area to reject vehicle-brake box multiple reflections (VBR); and S7, outputting effective target information after tracking processing. The device comprises a radar radio frequency front end, a signal processing unit, a point cloud post-processing unit, a tracking and control unit and an embedded processing platform. Through a three-layer cooperative inhibition mechanism, three types of main multipath interferences of a barrier gate scene are comprehensively covered, the detection accuracy is improved, static targets are prevented from being mistakenly eliminated, the engineering practicability is high, and safe and reliable barrier gate control can be ensured.

The invention discloses a voice and music collaborative generation method and system based on a dynamic mixed attention and expert architecture, terminal equipment and a medium, and relates to the technical field of audio generation. The method comprises the following steps: acquiring multi-modal input containing at least one of audio, text and vision, performing embedding processing on the multi-modal input, and mapping the multi-modal input to a unified space to obtain a fusion sequence; setting a plurality of attention heads, dynamically selecting the attention heads to activate experts through attention head routing gating, and weighting and aggregating calculation results to obtain mixed attention features; generating expert probability distribution through expert set routing gating, dynamically selecting an expert set to activate experts, and calculating task adaptation features; and performing audio generation processing based on the audio language modeling head to obtain a voice or music result. According to the method, collaborative generation of voice and music, dynamic allocation of computing resources, balance of domain-specific and cross-domain general knowledge learning are realized, the problems of task conflict and data imbalance are solved, and the audio generation quality and efficiency are improved.

The invention provides a time-space air quality prediction method based on Transform and a KAN network, and the method comprises the steps: in the time dimension, firstly, obtaining the high and low frequency data features of a main station through wavelet transform, then, constructing an Embedding module through the KAN network, enabling the model to better capture the dependency relationship of a sequence, and finally, carrying out the correlation calculation through ProAttention, and effectively extracting the important features of the time dimension. In the spatial dimension, firstly, a spatial region is divided into four sub-regions, station data in the four sub-regions are fused, then, embedding processing is performed on the data of each sub-region by using a KAN network, and finally, correlation calculation is performed by using ProAttention, and spatial features of each sub-region are obtained in a fine-grained manner. In terms of spatio-temporal feature processing, firstly, time features and spatial features are fused, then, correlation calculation is performed on the spatio-temporal features by using ProAttention, and finally, fitting is performed by using a KAN network, so that the prediction precision is improved. Experimental results show that the method provided by the invention has a better effect.

The invention relates to the technical field of three-dimensional reconstruction, in particular to an electric power facility surrounding environment three-dimensional reconstruction method and system. The method comprises the following steps: acquiring electric power facility data, performing electric power facility layout structure analysis, and constructing an electric power facility three-dimensional model; obtaining electric power facility environment data, carrying out dynamic environment evolution, and constructing an electric power facility environment dynamic evolution model; performing rendering and model alignment embedding processing on the electric power facility three-dimensional model based on the electric power facility environment dynamic evolution model to generate an environment-electric power facility fusion model, and performing electric power facility corona loss fluctuation detection and environment fluctuation factor analysis; corona environment disturbance area detection and error calculation are carried out on the environment-electric power facility fusion model, disturbance error data are generated, and corona environment disturbance reconstruction processing is carried out on the environment-electric power facility fusion model. According to the invention, the electric power facility environment combined corona loss dynamic three-dimensional model reconstruction can be realized.

The invention relates to an optimization method for improving the input processing and output efficiency of real-time image data, belongs to the technical field of embedded systems, and is particularly suitable for optimizing the input processing and output efficiency of the real-time image data in a resource-constrained environment. According to the method, parallel execution of image data processing tasks is realized by starting a multi-thread programming model; a hardware acceleration unit arranged in the embedded processor is used for carrying out batch processing on the image data, and the format conversion speed of the image data is increased; and direct access and transmission of the image data are realized by using a memory mapping technology and a DMA controller, so that the copying times of the image data in a memory are reduced, and the data transmission delay is reduced. Specific implementation details of how to realize a multi-thread synchronization mechanism, a hardware acceleration function and a zero copy technology are described, and efficient and real-time image data processing is ensured to be realized under the condition that resources are limited.

The invention relates to the technical field of chip delay error dynamic compensation scheme design, in particular to a chip delay error dynamic compensation method and system based on TDC. The method comprises the following steps: measuring the transmission delay of a critical path in real time by using a time-to-digital converter core integrated in a chip; dynamically predicting a future delay error through an embedded algorithm based on the delay measurement value; and according to the predicted delay error, dynamic compensation is carried out on the chip timing sequence, and the compensation comprises adjustment of a clock signal or modification of path delay. A corresponding system includes a TDC core for real-time measurement, an embedded processing unit for running a predictive algorithm, and a dynamic compensation unit for performing compensation. Closed-loop control is formed through real-time monitoring, prediction and compensation, the influence of process, voltage and temperature changes on the time sequence of the chip can be effectively resisted, the time sequence precision, working stability and reliability of the chip are remarkably improved, and the method is particularly suitable for high-frequency and precise measurement chips.

The invention relates to the technical field of physiological measurement, in particular to a remote physiological measurement method combining multi-scale periodic attention and MAE (self-supervised learning). The method comprises the following steps: S1, constructing an SRMFormer overall framework based on a space-time fusion module, an original video MAE module and a multi-scale periodic attention module; and S2, carrying out block embedding processing on the output of the space-time fusion module through Patch Embedding, and respectively inputting Token into an original video MAE module and a multi-scale periodic attention module combined with physiological priori for processing. According to the remote physiological measurement method combining the multi-scale periodic attention and the MAE, a self-attention mechanism is aligned with heartbeat intervals on multiple time scales through a multi-scale periodic attention module in combination with physiological priori bias, and the multi-scale periodic attention is obtained through an original video MAE structure. A shielded video Token can be reconstructed through periodic sensing loss under the condition of no preprocessing, so that a manually designed STMap (space-time diagram) process is removed.

A real-time temperature detection device for a liquid crystal display screen and a control method thereof, comprising the steps of: turning on a thermal imager, setting the distance between the liquid crystal display screen and the thermal imager to be l, and obtaining a thermal imaging image of the liquid crystal display screen; turning on a CCD camera to obtain a real-time video frame image of the liquid crystal display screen; a temperature calculation module of an embedded processing system obtains the operating temperature T of a corresponding pixel position of the liquid crystal display screen according to the thermal imaging image and the real-time video frame image; and outputting and displaying the temperature of the nth block of the liquid crystal display screen. The present invention dynamically adjusts the proportional coefficients of the temperature values ​​of the α and β thermal imagers and the CCD camera according to different weather, temperature, and material factors, thereby greatly enhancing the accuracy of temperature measurement.

The invention relates to a multi-parameter compensation type combustible gas detection method and system based on a thermal conductivity coefficient principle. The method comprises the following steps: measuring heat conductivity coefficient change of target gas by using a thermal conductivity detector to calculate gas concentration; the environment isolation technology is adopted to maintain constant temperature and humidity in the detection cavity; collecting data through an embedded processor, and dynamically correcting a gas concentration calculation result based on a pre-trained compensation algorithm model; and executing grading alarm control according to the compensated concentration value. The compensation algorithm model adopts a BP neural network model based on a particle swarm optimization algorithm. The system comprises a power supply circuit, a communication interface, a thermal conductivity sensor module, a constant temperature and humidity detection cavity, an embedded processor, a compensation algorithm module and a linkage control module, and a software multi-task hierarchical scheduling framework is also constructed. The device is suitable for hydrogen detection, and can effectively improve the accuracy and reliability of gas detection.

The invention relates to the technical field of thin-strip continuous casting production, in particular to a method for automatically controlling the height of a molten pool in the double-roller thin-strip continuous casting production process, which comprises the following steps: arranging a laser ranging module and a visual perception module; processing by an embedded processing module; establishing a molten pool height dynamic prediction model by adopting a deep learning algorithm, and outputting a molten pool height prediction value; self-adaptive sliding mode control is adopted as a main controller, and the displacement amount of the servo hydraulic stopper rod is output; and controlling the height of the molten pool based on the displacement of the servo hydraulic stopper rod. According to the method, laser is used for providing accurate distance reference, visual sensing is used for providing molten pool liquid level information, and the detection precision is improved and environmental interference is reduced through a machine vision algorithm; based on the process parameters and the real-time information of the sensor, constructing a molten pool height prediction model by using a neural network algorithm with physical constraints; and a reinforcement learning optimizer is added to adjust sliding-mode control parameters online, so that the sliding-mode control parameters adapt to working condition changes to carry out high-precision and fast-reaction molten pool height control.

The invention discloses a new energy automobile electric compressor preheating structure and a control method, and relates to the technical field of new energy automobile heat management. The method comprises the three steps of heat source collaborative recovery, heat demand dynamic distribution and self-starting collaborative execution, specifically, motor stator waste heat and compression refrigerant waste heat are recovered through a heat conduction fin array and a spiral flow channel, and a shared heat conduction medium is guided in; based on a fuzzy PID controller, a heat flow guide valve is dynamically adjusted to distribute heat flow according to the lubricating oil way temperature, the suction superheat degree and the shell temperature; the generator is triggered to be rapidly heated in a locked-rotor mode at the ultralow temperature, the preheating requirement is pre-judged in combination with a heat attenuation model, the structure comprises a temperature sensor set, a three-way heat flow guide valve and an embedded MCU processor, efficient waste heat utilization is achieved, external energy consumption is not needed, preheating energy consumption is effectively reduced, and the low-temperature starting performance and operation stability of the compressor are remarkably improved.

The invention provides a lithium ion battery electric energy prediction method and system, and belongs to the technical field of battery performance prediction, and the method comprises the steps: obtaining battery electric energy time sequence data, a prediction task demand and battery replacement application background information; determining statistical characteristics in the battery electric energy time sequence data, and generating a prediction instruction in a text format based on the statistical characteristics, the prediction task demand and the battery replacement application background information; performing embedding processing on the prediction instruction and the battery electric energy time sequence data to obtain a text embedding feature and a data embedding feature; aligning the text embedding features and the data embedding features to obtain aligned features; and inputting the alignment features into a pre-trained large language model to obtain an electric energy prediction result. According to the method, the set prediction instruction comprises statistical characteristics, prediction task requirements and battery replacement application background information, the reasoning ability of the pre-trained large language model is improved, and the accuracy of an electric energy prediction result is improved.

The invention provides a GIS bus electrical contact defect diagnosis method fusing magnetic field-current characteristics and a related device. The GIS bus electrical contact defect diagnosis method comprises the steps of obtaining GIS bus conductor current, shell ring current waveforms and shell space magnetic field distribution under different electrical contact defects; carrying out fault characteristic current extraction on the conductor current and the shell circulation waveform by adopting an inverse phase difference algorithm, and carrying out time embedding processing on the shell space magnetic field distribution through map coding to respectively obtain a characteristic current signal and magnetic field time sequence data; and taking the characteristic current signal and the magnetic field time sequence data as a comprehensive characteristic set, carrying out classification training on the comprehensive characteristic set by adopting a deep learning neural network, constructing a GIS bus electrical contact defect diagnosis model, and carrying out electrical contact defect diagnosis by utilizing the model. According to the invention, by obtaining the related current and magnetic field data of the GIS bus, multi-dimensional detection can be realized, technical support is provided for state maintenance, the intelligent operation and maintenance level is improved, and safe and stable operation of equipment is guaranteed.

The invention discloses a therapeutic footwear system comprising a footwear body (101) with insole (102), midsole (103), and outsole (105), integrated with a multi-modal sensor array (110) including a plantar pressure matrix (111), inertial measurement unit (112), photoplethysmography sensor (113) and temperature sensor (114). An embedded processor (120) houses an adaptive sampling module (121), an edge-AI inference engine (122), a federated learning client (123), and memory (124) for local analysis. A feedback interface (130) with zonal haptic actuators (131) under medial forefoot (132), lateral forefoot (133), medial heel, and lateral heel (135) delivers corrective prompts. A power subsystem (140) comprising a battery management circuit (141), energy harvesting elements (142), and battery (143) sustains operation. The integration reduces latency, conserves energy, preserves privacy, and enables closed-loop therapeutic intervention during ambulation.

The invention discloses an equipment health management method and device. The method comprises the following steps: acquiring multi-dimensional state data of target equipment in a target period, and preprocessing the multi-dimensional state data into a target state matrix; utilizing a multi-head self-attention module contained in the health state evaluation model to respectively carry out longitudinal and transverse analysis to obtain first state embedding reflecting a time change relation and second state embedding reflecting mutual influence between dimensions; integrating the two embedding and historical period embedding through an embedding processing module, and determining target state embedding; and finally, the health level of the target equipment is evaluated through a health analysis module, and when the level is too low, the fault risk source is embedded and identified based on the second state. The technical problems of poor accuracy and low interpretability caused by the fact that a traditional base station equipment health assessment scheme does not consider the full life cycle state of the equipment are solved.