5results about How to "Comprehensive recognition" patented technology

This invention relates to the field of robot control technology, specifically to an end-to-end closed-loop control method and system for robots based on multimodal perception fusion. It includes: a perception access layer for standardized access to data from multiple sensor types; a fusion processing layer for spatiotemporal alignment and multimodal fusion processing of the standardized access data; a decision control layer for generating execution instructions based on the fusion processing results; and an execution feedback layer for feeding back the execution status to the fusion processing layer to form closed-loop control. This technical solution can solve the problems of standardized access and real-time fusion of multi-source heterogeneous data, and construct a closed-loop control architecture with high fault tolerance and low latency.

The application relates to a transformer on-line fault diagnosis and early warning method based on transformer transformation ratio, which comprises the following steps: collecting primary side and secondary side voltages of the transformer in real time, and calculating corresponding real-time voltage transformation ratio; comparing the real-time voltage transformation ratio with the rated transformation ratio of the transformer to determine whether the transformer is faulty and the fault type; based on the real-time voltage transformation ratio, transformation ratio prediction is carried out to obtain a transformation ratio prediction value; the transformation ratio prediction value is compared with the rated transformation ratio of the transformer to determine whether early warning is needed; if early warning is needed, the transformation ratio prediction value is compared with the rated transformation ratio of the transformer to determine whether the transformer is faulty and the fault type. Compared with the prior art, the application directly uses voltage parameters to diagnose the transformer fault, can comprehensively, accurately and quickly identify the transformer fault, and simultaneously adopts a prediction algorithm, can realize timely fault early warning by transformation ratio prediction of the transformer.

The utility model discloses a kind of high-efficiency conveying detection systems. Including conveying unit, direction adjusting unit, sole pattern detection unit and control unit. Conveying unit adopts double-row chain type conveyor belt structure, two rows of conveyor belt are arranged in parallel, direction adjusting unit is arranged in the middle of conveying unit, including visual identification module and the rotary adjusting mechanism of being set on the top of conveyor belt being installed in the both sides of conveyor belt. Sole pattern detection unit is located in conveying unit middle rear section, including bottom visual detection component and defect identification system. Control unit adopts programmable logic controller. The utility model has the beneficial effect that: the utility model has automated direction adjustment, through visual identification module and rotary adjusting mechanism, realize the automatic detection and adjustment of slippers conveying direction, greatly improve production efficiency. The utility model detects quality accurately, can carry out comprehensive, accurate detection to sole pattern, ensure that product quality meets standard.

The application discloses a kind of neuron network hospital data correlation analysis systems, belong to disease prediction technical field.The application is by acquisition module, for obtaining the historical diagnosis and treatment data of the electronic medical record system record of hospital, health correlation state data and the illness data of target patient, the historical diagnosis and treatment data are divided according to disease category, obtain the disease diagnosis and treatment data related to each disease;Construction module is used to generate preset disease correlation model according to health correlation state data, the preset disease correlation model includes: each disease corresponding neuron node and the first connection between neuron node;Optimization module is used to adjust preset disease correlation model according to the disease diagnosis and treatment data, obtain target correlation model;Correlation module is used to determine the associated disease of target patient according to the target correlation model and the illness data, realizes the technical effect of effectively and comprehensively identifying illness.

This invention relates to a method for locating defects in forgings, aiming to improve the efficiency and accuracy of defect identification. The method first involves grinding the surface of the forging to ensure its smoothness and flatness. Next, an ultrasonic flaw detector with low-frequency and high-frequency emission capabilities is used to emit ultrasonic waves and acquire signals. Wavelet transform technology is used to process and analyze the acquired signals, extracting amplitude, phase, and frequency features to form a feature set. In the defect feature identification stage, feature vectors are constructed and input into a trained support vector machine (SVM) model, outputting the probability of various defects. The defect type is determined based on a set threshold; if the probability exceeds the threshold, the defect is considered to exist. Finally, samples are cut and subjected to tensile tests based on the detection results to confirm the defect characteristics. This invention combines wavelet transform and machine learning techniques, not only achieving precise location of forging defects but also improving detection efficiency, demonstrating promising application prospects and economic value.