78results about How to "Strong memory" patented technology

The invention provides a multi-neural network control planning method for a robot path in an intelligent environment. The method comprises the steps that 1 a global map three-dimensional coordinate system is constructed for the carrying area of a carrier robot to acquire a walkable area coordinate in the global map three-dimensional coordinate system; 2 a training sample set is acquired; 3 the global static path planning model of the carrier robot is constructed; and 4 starting and ending coordinates in a transportation task are input into the global static path planning model based on a fuzzy neural network to acquire the corresponding optimal planning path for the carrier robot. According to the invention, the global static path planning model and a local dynamic obstacle avoidance planning model are separately established; the nonlinear fitting property of the neural network is used to find the global optimal solution quickly; and the problem of falling into a local optimum in common path planning is avoided.

The invention relates to a transformer fault diagnosis method based on a neural network. The method comprises the following steps: analyzing fault reasons, and extracting a sample set of dissolved gas in transformer oil and a corresponding a fault type inside a transformer; determining the number of Neurons in a neural network hidden layer according to experience and a successive optimization searching method, and building a neural network; adopting an improved Cuckoo optimization algorithm to train the neural network; and finally utilizing the trained neural network to diagnose a fault of the transformer. The entire design method has the characteristics of high search precision and high global search capability, and effectively improves the accuracy of fault diagnosis for a transformer.

The invention discloses a method for forecasting a ship wake based on an optimized support vector regression parameter, and the method comprises the following steps of: providing a training set firstly; selecting a type of a kernel function, and implementing a model training through a learning set to obtain a forecasting function; reading and judging whether a mean square error of the forecasting function reaches a precision requirement; if so, forecasting the ship wake through the function, if not, generating a parameter path searching map; searching an optimum parameter through an intelligent water drop algorithm until reaching the precision requirement. While searching an optimal parameter, the intelligent water drop has a positive feedback mechanism in an optimum searching process, therefore, a searching process is rapid and effective; a sand content of a path can change in accordance with the sand content of an optimum water drop after searching each generation of water drops, thereby avoiding a phenomenon of early maturing and improving a searching ability; a final forecasting function can be rapidly obtained; and rapid and effective forecasting of the ship wake is realized.

The invention discloses a preparation method of a high-strength corrosion-resistant copper pipe for a heat exchanger. The high-strength corrosion-resistant copper pipe comprises a copper matrix and alloy metals. The alloy metals comprise, by weight, 0.5-0.6% of magnesium, 0.08-0.09% of silver, 0.2-0.4% of tin. 0.012-0.014% of zirconium, 0.012-0.014% of niobium, 0.05-0.06% of nickel, 0.005-0.008% of titanium and 0.02-0.04% of mischmetal. The mischmetal comprises yttrium, berkelium and lanthanum according to a mass ratio of 1: (0.15-0.2): (0.4-0.5). The preparation method utilizes upward continuous casting, rolling and drawing processes to replace casting, sawing, heating, extrusion and pickling processes of the traditional preparation method so that processes are greatly simplified, continuous production is realized, a period is short, and efficiency is high. A copper pipe prepared by the preparation method has oxygen content satisfying the oxygen-free copper level standard, has fine grains, and has mechanical strength, high temperature resistance, corrosion resistance and wear resistance obviously superior to the same performances of the existing copper pipe.

The invention provides a classification model training method and device and a classification method and device. The classification model training method comprises steps that the preset number of training image frames in training videos with labels are acquired; the target neural network is utilized to carry out characteristic learning of the training image frames, and characteristic vectors are extracted for the training image frames; the training videos represented by the extracted characteristic vectors are classified based on memory characteristic vectors corresponding to different classifications stored in the memory network to acquire the classification result of the training videos; according to the classification result of the training videos and the comparison result between the labels of the training videos, the target neural network is trained, the memory network has strong memory ability, a problem of poor memory ability of the neural network can be made up, regardless of the input sequence of any length, characteristics can be memorized by the memory network well, therefore, regardless of the input sequence, identification accuracy of a classification model is not affected by the poor memory capacity of the neural network.

The invention relates to a coating method for preventing birds from building nests, in particular to a method which is capable of effectively preventing birds from building nests on power transmissionand transformation circuits and equipment. The coating method for preventing birds from building nests adopts coating of any two or three of a red colour system, a yellow colour system and a blue colour system to carry out interval coating. The invention applies the characteristics of biological spontaneous reaction and utilizes the sensitivity of the birds to colors to prevent the birds from building the nests and prevent bird damage. Meanwhile, the coating method has no environmental pollution or chemical pollution, and simple and convenient construction.

The invention discloses an individual cognitive ability prediction method and system based on dynamic function connection. The method comprises the following steps: acquiring functional magnetic resonance imaging data in a resting state; extracting an average time sequence signal of each brain region through a brain template after preprocessing of the functional magnetic resonance imaging data; calculating and constructing a dynamic function connection matrix by using a sliding time window; with the dynamic function connection matrix as input high-dimensional dynamic brain description originaldata, constructing a model combining convolution and a long-short-term memory network; and substituting a dynamic function matrix and corresponding cognitive behavior data into the model for trainingand testing so as to realize prediction of cognitive competence. According to the invention, dynamic information of brain function connection is fully utilized, the advantages of a deep learning method in big data processing are utilized, calculation is rapid, prediction effect is ideal, various behavior data of a population can be predicted, dependence on the absolute value of the amplitude of imaging signals is avoided, migrations among different MRI machines can be realized, and prediction performance is excellent.

The invention discloses a method for re-identifying pedestrians across scenes based on a multi-layer neural network, which comprises the following steps: (1) collecting pedestrian video in a camera under a current scene, and intercepting video frames; (2) feature extraction and dimension reduction of the pedestrian image in the video frame intercepted in the step (1) being carried out, and the training set Xt of the target domain being constituted by using sample pairs to obtain a test set Xo; (3) processing the identified data under the relevant scenarios, and using sample pairs to constitutethe source domain training set Xs; (4) establishing training set X, X= [Xs, Xt]; (5) using X to train the multi-layer neural network model; (6) identifying the sample to be identified in the test setXo according to the model obtained in step 5. The invention selects a multi-layer neural network with complex non-linear mapping as a learning model, and adds the identification data of the relevantscene to the model learning of the new scene by using the transfer learning idea, so that the learning of the new scene is more accurate and effective.

The invention relates to a reinforcement learning-based micro-grid optimal scheduling method, and the method comprises the following steps of obtaining the prediction information of a micro-grid wind-light element and the boundary information of a power supply, and constructing amicro-grid optimal scheduling reinforcement learning model based on reinforcement learning; carrying out self-learning on the micro-grid optimization scheduling reinforcement learning model based on reinforcement learning, and accumulating the scheduling knowledge learned in the self-learning process to obtain a micro-grid with prior scheduling knowledge; mining and utilizing the learned scheduling knowledge through transfer learning, and building a similarity calculation model for re-utilizing the scheduling knowledge; and performing fine adjustment learning in a new micro-grid optimal scheduling task by using the prior scheduling knowledge to obtain an optimal scheduling strategy of the new task. According to the method, the reinforcement learning and the transfer learning are introduced into the micro-grid optimization scheduling, the reinforcement learning has the strong self-learning and memory ability, and the experience knowledge learned in an optimization process can be stored in neural network parameters.

The invention relates to a PID neural network controller-based ultrahigh-speed permanent magnet synchronous motor rotating speed control method. The PID neural network controller-based ultrahigh-speedpermanent magnet synchronous motor rotating speed control method comprises structuring a double-closed-loop ultrahigh-speed permanent magnet synchronous motor control flow chart with a current innerloop and a rotating speed outer loop, wherein the rotating speed outer loop acquires actual rotating speed value through a rotating speed estimation module and generates control signals through a PIDneural network controller, and the current inner loop is provided with a PID regulator. The PID neural network controller structures a neural network model through static and dynamic functions and dynamically corrects neural network link weight through back-propagation algorithms, thereby achieving a learning and memory function and improving system self-adaptability. The PID neural network controller-based ultrahigh-speed permanent magnet synchronous motor rotating speed control method can effectively achieve reliable, rapid and stable rotating speed control of an ultrahigh-speed permanent magnet synchronous motor.

The invention discloses ginseng root and common jujube seed nerve calming and immunity enhancing health maintenance tea and a preparation method thereof. The health maintenance tea comprises ginseng roots, common jujube seeds, Siberian solomonseal rhizomes, palmleaf raspberry fruits, barbary wolfberry fruits, Chinese dates, ural licorice roots, lobed kudzuvine roots, cape jasmine fruits, Mongoliandandelion herbs, fleshfingered citron fruits and Indian bread. According to the tea, by the aid of synergistic effects of components such as the ginseng roots and the common jujube seeds, the functions of invigorating primordial energy, restoring pulse, relieving depletion, calming nerve and increasing intelligence of the ginseng roots are furthest played, the functions of nourishing heart, calming nerve, calming and hypnotizing of the common jujube seeds are furthest played, so that the health maintenance tea has good calming and hypnotizing functions and good effects on neurasthenia, insomnia and headache, sleep is effectively facilitated, sleeping time is prolonged, sleep quality is improved, the health maintenance tea has a physiological health care function of preventing and treatinginsomnia and does not have toxic and side effects on human bodies, dependence is avoided, medicinal and edible homologies are achieved, the health maintenance tea has the effects of nourishing spleen, tonifying qi, nourishing blood and calming nerve, palpitation and neurasthenia are relieved, human memory can be effectively enhanced, senility can be effectively resisted, and human immunity is enhanced.

The invention provides a peanut and walnut milk and a preparation method thereof. The peanut and walnut milk comprises the following raw materials: raw milk, peanut kernel, walnut kernel, chia seeds,momordica grosvenori concentrated liquid, momordica grosvenori powder, a stable emulsifier, a pH regulator, cream hickory nut essence and purified water. The sweetness of the momordica grosvenori concentrated liquid is 30 times as that of saccharose, and heat is not generated, and thus momordica grosvenori has effects of reducing blood glucose, depressing blood fat, resisting aging and beautifying. Dietary fiber in the momordica grosvenori concentrated liquid and the momordica grosvenori powder can promote intestinal tract movement and then facilitate digestion and absorption of walnut milk. The preparation method has high raw material utilization rate and production rate, and the prepared peanut and walnut milk does not have precipitate and oil slick in the process of long-term placement.The emulsion is fine and smooth, moderate in sweetness and long in quality guarantee.

The invention discloses a micro-electrolytic filler and a preparation method thereof. The micro-electrolytic filler includes iron powder, iron filings, purple sand, borax, bentonite, activated carbon powder, sodium hydroxide, zeolite, corn starch, sodium carboxymethyl cellulose, each The composition content is as follows (wt%): iron powder: 60‑70; iron filings: 15‑20; purple sand: 2‑3; borax: 2‑3; bentonite: 5‑8; activated carbon powder: 6‑8; : 0.2‑1; Zeolite: 0.1‑0.5; Corn starch: 0.2‑0.5; Sodium carboxymethylcellulose: 0.1‑0.8. The preparation method includes raw material screening; raw material mixing; pressing film forming; drying and sintering. Compared with the prior art, the invention has the characteristics of high micro-electrolysis efficiency, no hardening, low cost, long service life and good filtering and purification effect.

The invention relates to a ginkgo dew beverage prepared by the following components: 4kg of ginkgo powder/white nut, 0.007kg of ethyl maltol, 1.5kg of xylitol, 0.05kg of ginkgo essence, 0.25kg of stabling agent RB4, 0.02kg of pure milk essence, 0.04kg of sweetener TR50, a proper amount of baking soda, 2.5kg of white granulated sugar, and 91.00kg of water. The ginkgo dew beverage of the invention is a functional beverage used both as medicine and food. Containing the nutrient contents of ginkgo biloba extract, amino acid, protein, vitamin A, C and nicotinic acid, and abundant mineral calcium, phosphor, iron, potassium and a plurality of trace elements, the ginkgo dew beverage is the most promising functional beverage in the international beverage market.

The invention discloses an STO-BTCN thermal error prediction model modeling method. The method comprises the following steps: 1) initializing parameters of a swallow gull optimization algorithm (STO); 2) creating a BTCN neural network, and mapping the initial position of the swallow gull into the batch processing size and the number of filters of the BTCN neural network; 3) taking a mean absolute error (MAE) as a fitness function; 4) judging whether the mean absolute error (MAE) is smaller than a set threshold value or not; 5) judging whether the number of iterations reaches the maximum value or not, and if yes, terminating iteration; if not, adding 1 to the number of iterations, mapping the updated swallow gull position into the batch processing size and the number of filters of the BTCN neural network, and circularly executing the step 3); and 6) taking the batch processing size and the number of filters obtained by a swallow gull optimization algorithm (STO) as optimal hyper-parameters of the BTCN neural network, and constructing an STO-BTCN thermal error prediction model. The invention further discloses a transfer learning method of the STO-BTCN thermal error prediction model.

The invention discloses a micro-electrolytic filling material and a preparation method thereof. The micro-electrolytic filling material comprises the following components in parts by weight: 60 to 70 parts of iron powder, 15 to 20 parts of iron chips, 2 to 3 parts of purple sand, 2 to 3 parts of borax, 5 to 8 parts of bentonite, 6 to 8 parts of active carbon powder, 0.2 to 1 part of sodium hydroxide, 0.1 to 0.5 part of zeolite, 0.2 to 0.5 part of corn starch, and 0.1 to 0.8 part of carboxymethylcellulose sodium. The preparation method comprises steps of raw material screening, raw material mixing, film squeezing, moulding, drying, and sintering. Compared with the prior art, the micro-electrolytic filling material has the characteristics of high micro-electrolytic efficiency, no hardening, low cost, long service life, and good purifying and filtering effect.

The invention relates to a bottom-up resident energy consumption accounting method based on energy consumption equipment. The method is characterized by comprising the following steps: 1) pre-collected resident household energy consumption basic data are classified; 2) according to energy classes and an energy consumption activity classification result, the energy consumption amount of each energy class in each year in each class of energy consumption activity in each energy consumption household is calculated respectively; and 3) according to the obtained energy consumption amount of each energy class in each year in each class of energy consumption activity, the annual energy consumption amount of each energy consumption household or the annual energy consumption amount of each energy class or the annual energy consumption amount of each class of energy consumption activity are accounted. The collected basic data information is high in accuracy, commercial energies can be accounted, more importantly, non-commercial energies such as a biomass energy and granular coal can also be accounted, and the method can be widely applied to the field of resident energy consumption accounting.

The invention discloses an improved Elman neural network prediction method based on a noise reduction algorithm. The prediction method comprises the following steps: S1, dividing original data into a training set, a test set and a verification set in proportion; S2, performing noise reduction on the original data by adopting a noise reduction algorithm CPW, decomposing data of different dimensions in the original data by adopting CEEMDAN through the noise reduction algorithm CPW, performing noise reduction on the decomposed IMF by combining permutation entropy and using wavelet transform, and then reconstructing a sequence subjected to noise reduction processing into a time sequence subjected to noise reduction; S3, constructing an EAMC neural network combined with an attention mechanism; S4, putting the data of the training set after noise reduction into a neural network for training, storing the trained neural network after a loss value is smaller than a given threshold value, and ending the training; and S5, inputting the to-be-tested sample into the neural network obtained after training to obtain a prediction result, thereby improving the prediction precision.

The invention belongs to a corrugated continuously reinforced plastic insulation composite pipe. One or more layers of steel wires are knit on a transport layer 1 by a knitting machine to form a reinforced layer 2, a polyethylene protective layer 3 is clad outside the reinforced layer, the protective layer 3 is evenly coated with one layer of a thermoplastic elastomer solvent adhesive to form a bonding layer, and the bonding layer 4 is longitudinally and continuously coiled with a closed-cell foamed crosslinked polyethylene insulation layer 5 by a continuous insulation cladding machine; and a corrugated protective layer 6 is compounded by a corrugated pipe production line. The pipe designed by the invention is a continuous flexible pipe which has no weak point due to no joint; the insulation layer is made from the closed-cell foamed crosslinked polyethylene material, thus the pipe has good flexibility and strong memory, the pipe can spring back very well after compression deformation, and especially a corrugated structure is adopted for the protective layer, thus maximally improving the outer pressure resistance effect of the pipe, causing the pipe to have smaller bending radius and to more conveniently pass the bending; and a factory prefabrication manner is adopted for insulation, thus the construction is faster.

The invention provides a heterojunction structure with the in-plane magnetic anisotropy and the magnetoelectric coupling property and a preparing method and application of the heterojunction structure. The heterojunction structure sequentially comprises a (011) oriented single-crystal substrate, an amorphous SmCo film layer and a Cr coverage layer. An in-plane anisotropy stress field provided by the (011) oriented single-crystal substrate can induce an amorphous SmCo film to generate the anisotropy. When the substrate is PMN-PT, the heterojunction structure is the magnetoelectric coupling heterojunction structure specifically. The magnetoelectric coupling heterojunction structure is a brand new multiferroic compound heterojunction with an amorphous permanent magnet SmCo alloy as a ferromagnetic layer. Due to combination of the good plane anisotropy and nonvalatile anisotropy magnetic memory effect, the heterojunction has potential application value in the aspects of magnetic storage, magnetic recording, a high-sensitivity magnetic electricity weak signal detector, a magnetic energy product adjustable mini-type permanent magnet material, a sensor, a logic device, electric control magnetic recording and the like.