7results about How to "Easy to refactor" patented technology

The application discloses a kind of robot gripper device and grabbing method for assisting spraying operation, device includes main gripper module, auxiliary gripper module, deflection angle adjusting module, rotation angle adjusting module, pitch angle adjusting module, auxiliary gripper movement module, drive module, motion transmission module, by drive module through motion transmission module, pitch angle adjusting module, auxiliary gripper movement module action is realized, combined with deflection angle adjusting module, rotation angle adjusting module itself action, realize the relative pose adjustment of main gripper module, auxiliary gripper module and the grabbing of part to be sprayed;Method is based on deep learning network to generate grabbing posture, finally through the cooperation of main gripper module and auxiliary gripper module, the grabbing of specific position such as part non-spraying surface is completed.The application design is novel, and flexibility is high, can realize more complex grabbing action, and auxiliary realization is to complex shape part multidirectional spraying operation.

The invention relates to a multi-section sewage treatment integrated system applied to the technical field of sewage treatment, which comprises a treatment section arranged along the flow direction, and a standard function module capable of dynamically gating is arranged in the treatment section; the intelligent fluid routing layer comprises a central multi-dimensional distributor and is used for dynamically reconstructing fluid connection between modules; and a space intelligent control hub. The space intelligent control center constructs a digital twinborn body and calculates a module connection topology with optimal efficiency based on real-time data through a process topology optimization model, and the model calculates a topology reconstruction comprehensive index by distributing dynamically configurable weights for a process efficiency memory index, a connection efficiency prediction index and a cross-module synergistic effect index. And selecting the topology with the highest comprehensive index as the optimal scheme. According to the invention, online intelligent dynamic reconstruction of a treatment process chain is realized, and the self-adaptive capability, the treatment efficiency and the operation energy efficiency of the system to water quality fluctuation are remarkably improved.

The application relates to a threshold dynamic quantum secret sharing method based on quantum walk and Lagrange interpolation, which comprises the following steps: setting a non-negative integer secret to establish a first classical binary sequence with a target length and taking the first classical binary sequence as a secret share; preparing a decoy state photon, mixing the decoy state photon into a coin particle sequence to form a mixed sequence, judging whether a participant passes verification or not through the mixed sequence, randomly selecting a target operator for determining a second classical binary sequence if the participant passes the verification, performing bitwise XOR operation on the first classical binary sequence and the second classical binary sequence to obtain a decryption value of a private secret share; inputting the decryption value into a Lagrange interpolation model to obtain an initial hidden secret; and when personnel change or secret updating occurs in the participants during the process of obtaining the initial hidden secret, quantum resource multiplexing characteristics of quantum walk are utilized to re-distribute quantum resources. The application can efficiently add or remove participants and update secrets, and can effectively resist various attacks.

This invention relates to noise reduction methods, and more particularly to a lightweight multi-axis Transformer single-channel noise reduction method. It can fully extract the latent time-frequency features of speech signals. A multi-head dynamic local self-attention module is employed to efficiently extract local features. The proposed method has fewer network parameters and lower computational cost, while remaining competitive with state-of-the-art methods in terms of speech signal quality and intelligibility. It can fully extract the latent time-frequency features of speech signals, effectively reducing computer resource consumption. Hint blocks are used to help the model better learn frequency feature information.

A small target detection method, system, device and medium based on a local attention feature correlation mechanism, the method comprising: data preprocessing, model construction, model training, small target detection; the system, device and medium are used to realize a small target detection method based on a local attention feature correlation mechanism; the application extracts spatial features and semantic features through a residual network, performs feature fusion through a feature pyramid mechanism, reconstructs features based on a local attention feature correlation mechanism, trains a model using a Hungarian matching algorithm, obtains an optimal matching model suitable for a data set, can ensure high accuracy with fewer iteration times, and has the characteristics of low network operation complexity, low training cost, easy model migration, and high model detection performance.

The invention belongs to the field of food processing, and particularly relates to a preparation method of a yeast-fermented fluffy fish cake easy to swallow. The table salt, the starch and the yeast are added into the minced fillet, the fish cake which is well formed is prepared through chopping and mixing, moisture adjustment, yeast fermentation and heating curing, and the fish cake is fluffy, soft, moist, easy to crush and swallow, good in taste, mild in color and luster and easy for people with dysphagia to eat. The preparation method is easy in condition control, simple and convenient to operate, easy in forming and suitable for large-scale production.

The invention belongs to the technical field of biomedicine, and discloses mucus excitation hydrogel with regeneration activity and a preparation method thereof, the mucus excitation hydrogel is composed of dihydrocaffeic acid modified boloxamide, hyaluronic acid, epigallocatechin-3-gallate and bioactive factors, and the bioactive factors are KPV tripeptide and epidermal growth factors. The PHE-EK is a flowable liquid at room temperature and is gelled within 10 seconds at the body temperature. And the strain of the PHE-EK hydrogel is 77.8%. The PHE-EK hydrogel specifically adheres to the inflammatory colon through electrostatic interaction. The PHE-EK solution provided by the invention has good temperature sensitivity. The PHE hydrogel has relatively high mechanical strength and relatively high fracture strain. The PHE hydrogel can specifically adhere to the inflammatory colon by electrostatic interaction. PHE-EK enema is a promising method for treating ulcerative colitis.