36results about How to "Guaranteed Global Convergence" patented technology

The invention relates to an identification method of random structural damage based on genetic algorithm and static force measurement data. The identification method of the random structural damage based on the genetic algorithm and the static force measurement data comprises the following steps: (1) preliminarily obtaining statistical properties of random structural damage indexes; (2) defining the probability of unit damage as the random rigidity before occurrence of the damage or the probability that an elastic modulus Kai is larger than Kdi; (3) introducing damage probability indexes, determining several units of the damage probability indexes as non-damage units, and conducting corresponding adjustment on the statistical properties of the damage indexes; (5) going back to the governing equation of the initial damage identification in the step (1), and obtaining the objective function of the structural damage indexes after rearranging; and (6) solving the minimum value of the objective function in the step (5) by means of the genetic algorithm, and obtaining the statistical properties of the damage indexes. The identification method of the random structural damage based on the genetic algorithm and the static force measurement data has the advantages that multiple damage identification of different damage degrees can be carried out due to the fact that the genetic optimized algorithm has little limit on the type, the quantity and the size of the parameter.

Disclosed is a population dynamics optimization method with vertical-structure nutrition chains. The population dynamics optimization method with the vertical-structure nutrition chains is a PDO-NCVS algorithm. A solution space of an optimization problem is regarded as an ecosystem which has three vertical-structure nutrition chain types of an opened nutrition chain, a closed nutrition chain and a branch nutrition chain, the ecosystem is divided into a plurality of different sub-systems, and each sub-system has a specific vertical-structure nutrition chain type. For each sub-system, a plurality of populations live in each sub-system. The populations can not be transmitted among the sub-systems, and information transfer exists in the kindred populations among the sub-systems which have the same vertical-structure nutrition chain types. The populations living in a sub-system are connected in a predator-prey circulating mode or in a resource-consumption circulating mode. Behaviors during a population acts in a sub-system are constructed into evolution operators which are used for constructing evolution strategies of the populations. The algorithm has the advantages of being strong in search capability and having global convergence, and a solution scheme is provided for the solution of a complex function optimization problem.

The invention provides a polarity searching method of a fixed-polarity Reed-Muller logic circuit. The best polarity of an FPRM logic circuit is searched by using the new binary differential evolution algorithm, compared with a polarity searching method of the FPRM logic circuit based on the genetic algorithm, the capacity of local optimum and the capacity of premature convergence are avoided, and the convergence rate and the polarity searching efficiency are improved. The method comprises the following steps that firstly, a Boolean logic circuit is read; secondly, an evolution parameter is input; thirdly, an initial population is generated randomly, wherein the polarity is encoded into a binary individual; fourthly, the improved binary stochastic mutation operation is performed; fifthly, the binomial crossover operation is performed; an FPRM expression of the target individual and the FPRM expression of a test individual of the target individual are obtained; seventhly, the fitness value of the target individual and the fitness value of the test individual of the target individual are calculated; eighthly, the greedy selection operation and an elitism selection strategy are performed; ninthly, if the current evolution algebra is smaller than the maximum evolution algebra, the fourth to eighth steps are executed in sequence, and otherwise, the best polarity is output.

The invention discloses a robust controller of a permanent magnet synchronous motor based on a fuzzy-neural network generalized inverse and a construction method thereof. The construction method of the invention comprises the following steps of: combining an internal model controller and a fuzzy-neural network generalized inverse to form a compound controlled object; serially connecting two linear transfer functions and one integrator with the fuzzy-neural network with determined parameters and weight coefficients to form the fuzzy-neural network generalized inverse, serially connecting the fuzzy-neural network generalized inverse and the compound controlled object to form a generalized pseudo-linear system, linearizing a PMSM (permanent magnet synchronous motor), and decoupling and equalizing the linearized PMSM into a second-order speed pseudo-linear subsystem and a first-order current pseudo-linear subsystem; and respectively introducing an internal-model control method in the two pseudo-linear subsystems to construct the internal model controller. The robust controller of the invention has the advantages of overcoming the dependence and local convergence of the optimal gradient method on initial values and solving the problems of randomness and probability caused by using the simple genetic algorithm, obtaining the high performance control, anti-disturbance performance andadaptability of the motor and simplifying the control difficulty, along with simple structure and high system robustness.

An information fusion and adaptive matching method in a pan-correlation sparse matrix space belongs to the technical field of artificial intelligence. Firstly, the present invention 1) constructs a pan-correlation sparse matrix information space, and realizes the fusion representation of complex heterogeneous information; 2) provides a construction method of pan-correlation sparse matrix, and designs a multi-module sparse coding; 3) in The pan-correlation sparse matrix space defines the algorithm of heterogeneous elements; 4) On this basis, a system of multi-dimensional attribute intelligent matching and service discovery is realized. Solve the problem that the existing similar methods are difficult to integrate global elements due to the lack of a unified system architecture and heterogeneous information fusion method; the application layer has too much time and resource overhead, and it is difficult to realize the mobile terminal environment application. Problem; it is difficult to guarantee global stability and convergence at the technical level.

The purpose of the present invention is to provide a full waveform inversion method and system based on non-monotonic search technology. The method includes: firstly divide the multi-scale inversion frequency into N frequency groups according to the set frequency group range; secondly, use non-monotonic search technology In combination with the initial step size, determine the velocity model of the next iteration corresponding to the jth frequency group according to the seismic observation data and the initial velocity model; then judge whether j is greater than N; if j is greater than N, then set the next iteration corresponding to the jth frequency group The velocity model of the second iteration is output as the optimal velocity model; finally, seismic imaging is performed based on the optimal velocity model. The invention combines the non-monotonic search technology with the initial step size to determine the step size. This method can not only obtain high-precision inversion results under the premise of ensuring the inversion calculation efficiency; To a certain extent, the global convergence of the inversion is guaranteed, and better inversion results are obtained.

The invention discloses an attitude control method of an underwater robot based on BP neural network S-plane control, establishes a dynamic model for the underwater robot, models and analyzes the disturbance forces of the umbilical cable and the manipulator, and establishes an overall The kinematics control loop and the disturbance-compensating item-based coordinated controller for disturbance-resistant control use the S-plane control based on BP neural network to ensure the global convergence of the coordinated controller for underwater robots; in the design of the dynamic control law, the underwater During the operation, the robot is subjected to the disturbance force of the manipulator and the umbilical cable, and the movement of the propeller is controlled to compensate, so as to realize the stability and precise control of the underwater robot during the operation.

A multi-dimensional system contour error estimation method based on simplified Newton method, related to the field of numerical control machining technology, aimed at the large amount of calculation, long solution time and singularity of trajectory peak in the existing contour error estimation method based on Newton extremum search algorithm etc. Compared with the classical Newton extremum search algorithm, the simplified Newton contour error estimation algorithm proposed by the applicant ensures higher accuracy and does not need to calculate derivatives. In the process of contour error estimation, it avoids singularity and reduces amount of calculation. According to the specific experimental results, the estimation accuracy and efficiency of the multi-dimensional system contour error are about 30%-50% higher than those of the traditional scheme. The traditional contour estimation method based on the Newton extremum search method is a local convergence method, and the selection of its initial value will determine the convergence and accuracy of the algorithm. The present invention guarantees the overall Convergence.

The invention relates to the technical field of underwater positioning, in particular to a single-beacon positioning method for an underwater vehicle. A single-beacon positioning method for underwater vehicles with global convergence. Underwater vehicles are equipped with hydrophones, Doppler velocimeters, depth gauges, attitude and heading reference systems, and GPS; underwater acoustic beacons broadcast periodically Underwater acoustic signal; the present invention converts the nonlinear single-beacon positioning model in a discrete state into a linear time-varying model through state augmentation; when the underwater acoustic signal is not received, the underwater navigation is obtained through the underwater vehicle's own equipment. dead reckoning based on the relative speed and attitude of the vehicle and the water; after receiving the underwater acoustic signal, the transmission time of the underwater acoustic signal is obtained through the known transmission time of the underwater acoustic signal, which is used as an observation variable, and the dead reckoning data and various The sensor observation data is used to predict and update single beacon positioning based on Kalman filtering. Under the premise that the localization model is observable, this method has global exponential convergence.

The invention belongs to the field of robot control and nonlinear systems, and specifically relates to a method for adaptive impedance control of robots based on biologically inspired neural networks, aiming to solve the problem that the existing technology cannot realize real-time precise control of robots in complex nonlinear systems . The present invention includes: acquiring the initial control torque, expected impedance, and motion trajectory of the system; constructing the dynamic equation and expected impedance model of an n-degree-of-freedom manipulator system containing impedance to obtain the real state and expected state of the system at time t of the robot; based on full state feedback Build an adaptive controller with the bio-inspired network and obtain the control torque at time t+1; cycle through state acquisition, adaptive impedance control, and motion control until the robot arm completes the motion trajectory. The invention combines the biological heuristic network structure and time-delay feedback, adopts the sea-bian algorithm for adjusting the reward value, and the structure combining network estimation and full state feedback, so the system is stable and the control precision is high.