33 results about "Membrane computing" patented technology

Embodiments of the present invention are directed to an instruction set of an object memory fabric. This object memory fabric instruction set can be used to define arbitrary, parallel functionality such as: direct object address manipulation and generation without the overhead of complex address translation and software layers to manage differing address space; direct object authentication with no runtime overhead that can be set based on secure 3rd party authentication software; object related memory computing in which, as objects move between nodes, the computing can move with them; and parallelism that is dynamically and transparent based on scale and activity. These instructions are divided into three conceptual classes: memory reference including load, store, and special memory fabric instructions; control flow including fork, join, and branches; and execute including arithmetic and comparison instructions.

The invention relates to a PID controller parameter optimizing method, and especially discloses a PID controller optimizing design method based on a particle swarm membrane algorithm. The method comprises that an optimal particle is obtained from a swarm, different dimension values of the particle are successively assigned to to-be-optimized parameters Kp, Ki and Kd, a control system model is operated, performance indexes output by the system are obtained, and it is determined whether the evolution rule is used again to search for the optimal particle so that the optimal control effect can be obtained. The PID controller optimizing design method combines membrane computing with traditional PSO (Particle Swarm Optimization), fully utilizes the partition function and transport communication rules, can obtain a set of more proper PID control parameters, and finally enables a controlled system to achieve the optimal control effect. According the PID controller optimizing design method, the controlled system can obtain the set of more proper PID control parameters and achieves the optimal control effect under the condition that the controller structure is not changed.

The invention provides a memory computing-based customizable multimode big data processing system. The system comprises a data storage layer module, a memory-based data sharing and management layer module, a memory computing-based universal execution framework layer module and an access interface layer module, wherein a distributed memory abstraction mechanism, a position-sensing scheduling mechanism and a distributed hybrid column storage mechanism of data are adopted. According to the system, massive data storage is constructed; a cluster concurrent computing-oriented memory data management and sharing framework is provided, and an efficient and customized big data multimode universal processing framework is provided; batch processing and real-time data stream computing are supported; and the support is provided for flexible analysis and deep utilization of the data.

A method of implementation of a P-system in membrane computing comprising: placing three mutually immiscible liquids into a container, such that a first of said liquids provides a barrier between the second and third liquids; providing an initial substrate to at least one of said second and third liquids; and applying a transport rule for said first liquid to allow transport of substrate between said second and said third liquid.

The present invention relates to a local path planning algorithm based on membrane computing and particle swarm optimization, comprising the following steps: initializing a limited speed of a robot toestablish a speed-position coordinate system, sampling a position and a speed of a particle, initializing a membrane structure and allocating particles, calculating a fitness function of the particles and searching particle local optimum in an elementary membrane and global optimum in a surface membrane, for updating the position and the speed of the particle and local optimum and global optimum,performing iteration constantly, determining whether a fitness value reaches a required fitness threshold delta, if yes, stopping iteration, and outputting coordinates of a global optimum particle, otherwise determining whether the number of iteration times reaches a maximum value N, if no, continuing to perform iteration, otherwise stopping iteration and outputting coordinates of the global optimum particle, that is, a driving speed of the robot in a next moment is obtained. The entire algorithm obviously enhances reliability and a real-time characteristic of local path planning, so that a barrier can be safely avoided with a shortest driving distance and time.

The invention relates to the field of network search, and discloses a search method based on membrane computing. The method comprises the following steps of (A) obtaining the optimum weighing factor; (B) obtaining the similarity of document themes; (3) predicting and sequencing the optimum values. The method has the beneficial effect that the accuracy of predicting the similarity of the unvisited URL (uniform resource locator) themes is improved, and the recall ratio and accuracy of theme crawlers are further improved, so the theme crawlers are instructed to collect more webpage sets with higher quality from an internet, and the webpage sets which are more interested by a user can be more effectively collected.

The embodiment of the invention belongs to the technical field of artificial intelligence, is applied to the field of intelligent medical treatment, and relates to a text image super-resolution reconstruction method and related equipment thereof. The method comprises the steps of inputting a low-resolution picture into a scene text recognition model to acquire text position and text content information; generating a text mask based on the text position information and the text content information, and upsampling the text mask to obtain a target mask; inputting the low-resolution picture and the target mask into the adversarial network to obtain a discrimination result, and calculating discrimination accuracy based on the discrimination result; calculating a loss function based on the low-resolution picture and the target mask until the loss function converges and the judgment accuracy is lower than an accuracy threshold, and obtaining a trained adversarial network; and inputting a received low-resolution picture to be converted into the trained adversarial network to obtain a target super-resolution picture. The trained adversarial network can be stored in the block chain. Quality of super-resolution reconstruction of the text image is ensured.

The invention discloses a membrane computing-based mobile robot obstacle avoidance control method. The method involves the acquisition and processing of local environment information and the design ofan obstacle avoidance behavior enzyme numerical value membrane control system. The acquisition and processing of the local environment information specifically comprises the following specific stepsthat: the local environment information is acquired and is adopted as input, and n distance sensors on the periphery of a mobile robot are adopted to obtain the values xi of distances between an obstacle and the corresponding sensors; and linear transformation is performed on the xi through a formula that si=-xi+M, wherein si is the linearly-transformed values of the sensors, and M is the maximumdetection distance of the sensors. According to the membrane computing-based mobile robot obstacle avoidance control method of the invention, the distances from the obstacle to the robot and the navigation speed of the robot are calculated; the values of the sensors are multiplied by weights, and obtained products are accumulated; the speeds of a left wheel and a right wheel are calculated; and therefore, the control method is convenient to assist in enabling the navigation speed of the robot to adaptively change along with the distances between the robot and the obstacle and the angular velocity of the robot and enabling the robot to perform operation according to the distances between the robot and the obstacle and the angular velocity of the robot.

The invention discloses a method for realizing underground robot path optimization based on membrane calculation. The path optimization method comprises the following steps of constructing a coal mineunderground robot cell type membrane system, constructing a membrane controller structure, designing a membrane algorithm, simulating a mobile robot and analyzing an experimental result. According tothe robot path optimization method, by changing the evolutionary algebra, the moving speed error of the robot is smaller, the good convergence is achieved, and compared with a traditional algorithm,the speed error is smaller, and the accuracy of the membrane algorithm in speed calculation of the underground mobile robot is higher. The simulation and experiments show that the provided membrane algorithm has rapid convergence and effectiveness, the designed device can effectively solve the routing inspection problem under the complex environment of an underground coal mine, and the theoreticalsupport and practical application value are provided for the accurate mining of the coal mine in the future through the efficient calculation performance and the rapid optimization performance.

The invention provides a semantic segmentation method and device based on few samples, electronic equipment and a computer readable storage medium, and the method comprises the steps: enabling a target image combination to serve as the input of a trained feature extraction network, and obtaining the image features of each image in the target image combination; calculating an initial class prototype corresponding to the to-be-identified target class based on the image feature, the foreground mask and the background mask of each support image; determining a temporary class prototype of the queryimage based on the image features of a query image and the initial class prototype; fusing the initial class prototype and the temporary class prototype to obtain a final state class prototype; for the image features of the query image, calculating cosine similarity with the final state class prototype pixel by pixel to obtain a first cosine similarity graph; and determining a prediction foreground mask corresponding to the to-be-identified target category according to the first cosine similarity graph. According to the scheme, under the condition that the number of the sample images corresponding to the to-be-recognized target category is small, accurate semantic segmentation can be achieved with the help of a small number of supporting images.

The invention relates to a method for improving the time efficiency and implementing a minimum dominating set through membrane computing. The method comprises the following steps that a membrane computing (in a P system) minimum dominating set (MDS) model is constructed; membrane computing units used for implementing the MDS are initialized; the number of the membrane computing units is defined; alinear expression is input the membrane computing units; combined with an algorithm, iterative execution is conducted in a P-Lingua environment, and finally the MDS based on the membrane computing isimplemented in linear time. By means of the method, the minimum dominating set is implemented through the membrane computing, and through the results of the embodiment, it is shown that the method improves the time efficiency and solves that problem of the effective algorithm for solving the minimum dominating set in the linear time.

The invention discloses a false face video tampering detection method and system based on a time domain self-attention mechanism, and the method comprises a network training step and a sample testing step, and the network training step mainly comprises the steps of video preprocessing, segmentation network building and training. The sample test mainly comprises the steps of preprocessing a video, building a segmentation network, predicting a tampered area through three modules in sequence, and calculating a facial intersection ratio of a prediction mask to obtain a detection result. According to the invention, color features, noise features and time domain features of a spatial domain are input into a neural network comprising a double-flow feature extraction module, a time domain self-attention module and an up-sampling module to predict a tampered area, a detection object comprises a single picture and a plurality of video frame pictures, and relatively ideal accuracy is obtained in different databases. Compared with other existing algorithms, the cross-database test performance is obviously improved, and the invention has great potential application value.

The invention relates to a mobile robot path planning method integrating membrane calculation and RRT. The method comprises the steps of introducing membrane calculation on the basis of an original RRT algorithm to optimize the original RRT algorithm, setting a membrane calculation structure, expanding a random tree independently in each basic membrane, and increasing the number of sampling pointsduring random tree expansion in each basic membrane; and setting an inter-membrane communication rule to update a corresponding relationship between sampling points output by each basic membrane andfather nodes of the basic membrane, and returning an updated result to the corresponding basic membrane to continue to expand the random tree; and continuously iteratively expanding the random tree inthe whole membrane system until the random tree in one basic membrane reaches an end point, and stopping expanding. According to the method, the number of sampling points is increased and the inter-membrane communication rule is set during random tree extension in each basic membrane, so that the random tree extension in the membrane has higher directivity, and the real-time performance and reasonability of the path planning of the mobile robot are further improved.

The invention provides a semantic segmentation method and device based on few samples, electronic equipment and a computer readable storage medium, and the method comprises the steps of enabling a target image combination to serve as the input of a trained feature extraction network, and obtaining the image features of each image in the target image combination, wherein the target image combination comprises a to-be-identified target category, a plurality of support images and a query image, and the support images carry foreground masks and background masks corresponding to the to-be-identified target category; calculating a foreground class prototype and a background class prototype corresponding to the to-be-identified target class based on the image feature, the foreground mask and thebackground mask of each support image; and based on the image features, the foreground class prototype and the background class prototype of the query image, determining a prediction foreground mask corresponding to the to-be-identified target class on the query image. For the to-be-identified target category with few samples, semantic segmentation can be realized by means of the feature extraction network trained by the sample images marked with other categories.

The invention discloses a bridge crane neural network modeling method with object age feature film calculation. The method comprises the following steps: 1) obtaining input and output sampling data inthe operation process of the bridge crane, and establishing an RBF neural network model of the bridge crane; 2) respectively taking the quadratic sum of the difference between the estimated output ofthe position and the swing angle of the bridge crane in the operation process and the actual output sampling data as an objective function; 3) under the inspiration of biological cell membranes and intracellular substance aging phenomena, abstracting a bridge crane neural network modeling method with object age characteristic membrane calculation; 4) setting algorithm operation parameters; And 5)estimating unknown parameters in the bridge crane RBF neural network model by operating an object age feature film calculation optimization method, obtaining unknown parameter estimation values in the model by minimizing an objective function, and substituting the estimation values into the bridge crane RBF neural network model to form a nonlinear model. The modeling method disclosed by the invention has the characteristics of early maturing resistance, high local search precision and fast convergence.

The invention discloses an autophagy membrane-computing optimization modeling method for a fuel cell, which includes the following steps: 1, obtaining current input and voltage output sampling data of the fuel cell through site operation or tests, and enabling sum of squared error of estimated output and actual output of the sampling data of a fuel cell model to serve as minimum target function; 2, abstracting the autophagy membrane-computing optimization modeling method for the fuel cell and corresponding bionic rules by being inspired by recombination, variation and autophagy of substances and energy in biological cells; 3, setting algorithm running parameters; and 4, operating autophagy membrane-computing optimization algorithm to estimate unknown parameter in the fuel cell model, obtaining estimated value of the unknown parameter in the model through the minimum target function, and substituting the estimated value in the fuel cell model to form a mathematical model. The modeling method has the advantages of being anti-precocity, high in optimization precision, rapid in convergence and suitable for modeling in other complex chemical reaction processes.