55 results about "Manifold optimization" patented technology

The invention relates to the technical field of multi-modal sensing, and discloses a multi-modal large model detection and recognition robot recognition system for a complex scene, the system constructs a dynamic manifold modeling module, realizes cross-modal joint denoising through a stochastic differential equation and depth score matching, constructs a drift term and an anisotropic diffusion term by using an optical flow field, and realizes multi-modal detection and recognition of a multi-modal large model. Dynamic noise interference such as rain fog and motion blur is eliminated; on the basis, designing an information geometric alignment module, and based on Riemannian manifold optimization and orthogonal projection matrix calculation, realizing geometric equidistant mapping of vision-Li DAR features through multi-scale measurement tensor fusion; a dynamic external parameter calibration module is further provided, SE (3) manifold Kalman filtering is combined with a noise self-adaptive scaling technology, and external parameter offset is tracked and compensated in real time. Compared with a traditional method, the method has the advantages that the core problems of cross-modal data geometric mismatch, external parameter drift accumulation, low semantic fusion efficiency and the like are solved, and the sensing precision and robustness of the automatic driving system in a complex dynamic scene are remarkably improved.

The invention discloses a low-overlapping-rate point cloud registration method based on topology-metric decoupling, and belongs to the technical field of computer vision and three-dimensional reconstruction, and the method comprises the steps: scanning an object through a three-dimensional laser scanning collection terminal, collecting point clouds at two different angles as a source point cloud and a target point cloud, and achieving high-robustness registration in a difficult scene with an extremely low overlapping rate; according to the method, through the asymmetric association strategy, the zero solution problem under the low overlapping rate is solved through the one-way matching union set, and the recall rate is remarkably increased; meanwhile, through a DGTG geometric pruning and probability manifold optimization mechanism, a voxel quantization error is eliminated by using anisotropic covariance and Lie algebra iteration, a translation error and a rotation error are greatly reduced, and a precision bottleneck caused by voxelization is effectively broken through. According to the method, the existence of a solution can be ensured under the condition that the overlapping rate is extremely low, voxel quantization errors can be eliminated through manifold optimization based on covariance, and high-precision and high-efficiency registration of low-overlapping-rate point clouds is achieved.

Certain aspects of the disclosure provide a method for code generation, including: selecting a code generation use case type; identifying the target entity in a graph knowledge base; extracting a graph segment of the graph knowledge base; identifying a set of methods in the graph segment; identifying a first method of the set of methods that has a largest context size of the set of methods; determining an overlapping context size between the first method and each other method of the set of methods, and selecting a maximum number of other methods of the set of methods in order of decreasing overlapping context size with respect to the first method, such that a total context size of the maximum number of other methods of the set of methods and the first method does not exceed a maximum context window size associated with a machine learning model.

The invention discloses a manifold optimization transmitting directional diagram synthesis method for an analog digital hybrid array. According to the manifold optimization transmitting directional diagram synthesis method for the analog digital hybrid array, the HAD array transmitting directional diagram synthesis problem with the criterion of minimizing the matching error between an expected beam directional diagram and an actual beam directional diagram is researched. According to the method, the solution is carried out by utilizing block coordinate descent (BCD) and Riemannian manifold optimization. Firstly, a constant modulus auxiliary variable is introduced to equivalently represent that an objective function is a smooth biquadratic function; then, alternately optimizing a scaling factor, a digital weight, a simulation weight and a constant modulus auxiliary variable under a BCD framework, reexpressing the simulation weight and the auxiliary variable as a quadratic programming problem under constant modulus constraint, and solving through a Riemannian gradient descent method; the globally optimal solution of the digital weight is obtained through a Lagrange multiplier method. A numerical result shows that compared with other representative algorithms, the algorithm provided by the invention has better beam matching performance.

The invention relates to the technical field of wireless communication perception, in particular to a user perception method based on a multi-layer intelligent omnidirectional metasurface, and proposes to realize accurate perception of reflection and refraction user orientations by utilizing heterogeneous characteristics of propagation paths on two sides of a multi-layer intelligent omnidirectional metasurface structure. Based on the architecture, a user reflection and refraction type judgment algorithm based on uplink pilot frequency information is provided, and the user reflection and refraction types are determined based on the heterogeneous characteristics of user channels on the two sides of the multi-layer intelligent omni-directional metasurface and based on least square optimization problem solving and target function size comparison. The propagation path angle estimation algorithm specifically comprises the following steps: firstly, based on uplink pilot frequency information, designing a channel estimation algorithm comprising alternating optimization, constant-rank manifold optimization, convex relaxation and other methods to estimate and obtain a channel matrix by utilizing channel constant-rank and angular domain sparse characteristics; and finally, based on a user type determination result, obtaining user angle information based on each channel matrix by using an orthogonal matching pursuit algorithm.

The invention discloses an unmanned aerial vehicle low-altitude channel planning method based on two-stage manifold optimization, and the method comprises the steps: obtaining a low-altitude operation task of an unmanned aerial vehicle, and determining obstacles in a three-dimensional map of the task, a virtual obstacle region where a complex building group is located, and intersection points which do not exist around the virtual obstacle region; carrying out initial path planning on the unmanned aerial vehicle to avoid obstacles and a virtual obstacle area, and carrying out double-stage popular optimization on the initial path; the two-stage popular optimization comprises first-stage optimization for establishing a first cost function according to obstacle avoidance, energy consumption, smoothness and constraints of approaching an intersection point of a path of the unmanned aerial vehicle and second-stage optimization based on the path of the first-stage optimization; and adopting the optimal path trajectory or the suboptimal path trajectory according to the task requirements of the unmanned aerial vehicle. The method has the advantages that airspace density control strategies can be conveniently formulated, hotspot airspace congestion and conflicts can be avoided, task priority scheduling schemes can be deployed and the like, and key technical support can be provided for safe, smooth and efficient flight under complicated constraints by the aid of two-stage path planning schemes on the basis of manifold optimization.

The invention discloses an RIS auxiliary sensing integrated beamforming design based on manifold optimization, relates to the technical field of wireless communication, and aims to meet the requirement of high-precision sensing service in a 6G wireless communication system. By using the RIS technology, the invention provides an alternative optimization algorithm, which effectively improves the communication rate and reduces the interference between users. According to the method, the trade-off factor rho is introduced to balance communication and sensing performance, and dual optimization of the communication and sensing performance is realized. In the method, the communication sum rate of a'trade-off 'mode (rho = 0.2) reaches about 14 bit / s / Hz, the mean square error (MSE) of a beam pattern is controlled to be-4dB, and the system performance is remarkably improved.

The invention discloses a knowledge graph embedding and link prediction optimization method based on matrix representation and Grassmann manifold optimization, and the method comprises the steps: making a high-quality matrix representation template as input, generating an embedded matrix through a plurality of models, carrying out the coding through a Grassmann manifold optimization algorithm, and generating a multi-dimensional space vector code. According to the method, adaptive iterative detection of information correction, semantic precision and deviation control and a similarity algorithm based on rule template matching are further combined, iterative correction optimization is carried out on model output, and comprehensive judgment is carried out on the model output and a gold label, so that the explanation accuracy and quality are improved. According to the method, the problems of information normalization, semantic precision deviation and rule consistency of an existing model can be effectively solved, the overall performance of knowledge graph embedding and link prediction is improved, output of different models is kept in a balanced state, and the defects of a single model are overcome.

The invention discloses a geometric consistency guided deep manifold network parameter initialization method and system, and relates to the technical field of manifold learning in machine learning, and the method comprises the steps: dividing an input feature into a plurality of independent modals, and building a feature-level modal energy distribution model; deriving a geometric consistency constraint formula of learnable parameters based on the energy distribution consistency constraint; the parameters are mapped to a high-degree-of-freedom fixed rank manifold for optimization, and updating of the parameters on the manifold and geometric structure maintaining are achieved through a designed Riemannian projection operator and a pull-back operator; and combining a geometric consistency constraint formula and a manifold optimization strategy, and carrying out iterative training until the model is converged. According to the method, the geometric consistency in the network training process can be effectively maintained under different manifold structures and data distribution, and the convergence speed, the stability and the cross-scene generalization ability of the deep manifold network are remarkably improved.

The present application relates to the field of intelligent garbage sorting, and discloses a robot hand-eye calibration method for intelligent garbage sorting, comprising the following steps: Step 1, system initialization, configuration of robot, visual sensor and conveyor belt parameters; Step 2, garbage is transported to the sorting area via a conveyor belt, and the location and category of the garbage are identified by machine vision; Step 3, based on the identification result, the robot controls the metal suction cup to grab the garbage and perform the sorting operation; Step 4, during the sorting process, the target point data of the visual coordinate system and the robot coordinate system are obtained, and the hand-eye calibration matrix is ​​optimized by the H∞ robust control theory; Step 5, the Riemann manifold optimization method is used to perform gradient descent iteration in the manifold space of the rigid body transformation matrix, and by adopting a real-time fault detection technical solution based on sensor networks and data analysis, and by real-time monitoring and fault warning of key components, the technical effect of improving the stability of the garbage sorting system is achieved.

The invention relates to the technical field of visual positioning, and discloses a visual positioning method and system based on manifold optimization and a two-dimensional code, and the method comprises the steps: collecting multi-source positioning data, and carrying out the time-space alignment; performing anomaly detection and quality evaluation on the observation data sequence; constructing an incremental optimization framework of the manifold space; carrying out adaptive weight fusion based on environment perception; predicting a positioning quality trend by adopting a deep learning model; establishing a motion model for time delay compensation and trajectory prediction; and the digital twin model is driven to realize high-fidelity synchronization and intelligent visualization. According to the invention, the positioning precision and robustness of the mobile device cluster can be improved, and real-time high-fidelity digital twin synchronization and intelligent operation and maintenance supervision are realized.

This application provides a cooperative control method and system for two humanoid robots, relating to the field of robot control technology. The method includes acquiring interaction data from two demonstrators; performing interactive perception motion redirection on the two humanoid robots based on the interaction data to generate a reference trajectory and topological prior that conform to robot kinematic constraints. Specifically, kinematic conflicts arising from heterogeneous redirection are resolved through dual-scale manifold optimization, accurately preserving the interaction intent; using the reference trajectory as the tracking target and guided by the topological prior, reinforcement learning is performed to obtain a cooperative control strategy for the two humanoid robots; the trained cooperative control strategy is deployed to the two humanoid robots, which exchange state information through real-time communication to perform physical cooperative tasks; this significantly improves the robustness of control and training efficiency; and ultimately achieves zero-sample transfer of the cooperative strategy from simulation to real-world operation.

The invention relates to the technical field of wireless positioning, in particular to a reconfigurable intelligent surface codebook configuration method based on manifold optimization, and the method comprises the steps: building a positioning error lower bound model of a wireless positioning system assisted by multiple reconfigurable intelligent surfaces; constructing a reconfigurable intelligent surface codebook optimization model by taking the lower bound of the positioning error as a target function; converting the original multi-scalar optimization model into a single-matrix manifold optimization model; and a codebook optimization algorithm based on a Riemannian manifold is used for solving, and optimized reconfigurable intelligent surface codebook configuration is obtained. According to the method, the reconfigurable intelligent surface codebook optimization model is converted into the manifold optimization model, the codebook configuration is obtained by using the Riemannian manifold-based codebook optimization algorithm, an efficient codebook can be found in a complex codebook space, and the precision upper limit of the multi-reconfigurable intelligent surface assisted wireless positioning system is improved.

The present invention provides an intelligent air compressor with adaptive variable capacitance and its control system, which relates to the technical field of air compressors. The control system includes a multi-source heterogeneous perception fusion module, a dynamic manifold optimization control module, a quantum probability fault prediction module, a cross-domain energy collaborative scheduling module, a distributed self-organization communication module, and an adaptive meta-evolution module. With the help of the cross-domain energy collaborative scheduling module, this system combines the grey Markov chain model with wavelet decomposition - support vector machine to accurately predict the real-time electricity price and the factory gas consumption load. Based on these predictions, through dynamic game equilibrium and multi-objective differential evolution algorithm, the operation mode of the air compressor is intelligently adjusted. During the low electricity price period, the operation load is automatically increased by 30% - 40%, and at the same time, the gas storage capacity is increased by 25% - 35%; during the peak electricity price period, the gas storage is reasonably released to reduce the operation load by 20% - 30%. It effectively avoids energy waste caused by unreasonable operation.

The invention discloses a coal rock scanning electron microscope image pore type classification method and classification model. The method comprises the following steps: training a generative adversarial network by using a first sample set and adopting a dual manifold optimization mechanism to obtain an image generation model; image expansion is carried out on the second sample set by using an image generation model, a third sample set is established, and the first sample set and the second sample set both comprise scanning electron microscope images of coal rock primary pores, metamorphic pores, mineral pores and strain pores; based on the third sample set, adopting a dynamic boundary smoothing mechanism to train an extreme learning machine to obtain a classification model; and determining the pore type of the coal rock scanning electron microscope image based on the classification model. According to the method, training data of an adversarial network expansion classification model is introduced, the problems of insufficient training data and poor model generalization ability are solved, and a dual manifold optimization mechanism maintains a topological structure of image data; and the extreme learning machine of the dynamic boundary smoothing mechanism is adopted to perform pore type identification, so that the classification accuracy is further improved.

The application discloses a kind of unmanned vehicle intelligent metasurface assisted satellite-ground physical layer security transmission method and system, construct signal transmission model based on communication system, obtain the instantaneous signal-to-interference-and-noise ratio of communication terminal and system instantaneous and security rate, form joint optimization satellite base station transmission precoding matrix, satellite-borne passive intelligent metasurface phase shift matrix, airborne active intelligent metasurface diagonal reflection matrix and unmanned vehicle flight path, transmission optimization problem model with the goal of maximizing system average and security rate;Through the analysis of problem structure, propose the solution framework of hierarchical processing, in inner layer, successive convex approximation, manifold optimization based on punishment and semidefinite programming algorithm are used to process transmission signal and intelligent metasurface reflection respectively;In outer layer, the best flight path design is obtained by using unmanned vehicle flight path algorithm based on deep deterministic policy gradient, and the security performance of satellite-ground communication system is improved by solving the optimization problem of inner and outer layers.

The invention belongs to the technical field of mobile communication, and relates to a method for reducing the PAPR (peak-to-average power ratio) of an OFDM (orthogonal frequency division multiplexing) system based on Riemannian manifold optimization and a gradient descent method, which comprises the following steps of: according to requirements, establishing a channel which not only can suppress the PAPR (peak-to-average power ratio) of a transmitted signal, but also can ensure that a legal user channel can receive information as correctly as possible, and can reduce the PAPR (peak-to-average power ratio); the method comprises the following steps of: firstly, solving a mathematical model in which an eavesdropper user does not receive information as much as possible by using a Riemannian manifold optimization and gradient descent method to obtain a constant envelope pre-coding sending vector of which the modulus length of all elements is 1, then, solving the mathematical model by using a Riemannian manifold optimization method and a gradient descent method to obtain a constant envelope pre-coding sending vector; the method not only suppresses the PAPR of the OFDM signal through the constant envelope precoding technology, but also ensures good BER performance.

The invention discloses a miniaturized high-performance satellite navigation intelligent anti-interference method, which relates to the technical field of satellite navigation and artificial intelligence, realizes anti-interference by combining compressed sensing MUSIC super-resolution DOA estimation and manifold optimization, and realizes super-resolution DOA estimation under an unknown steering vector through compressed sensing MUSIC. And an optimal beam forming weight is solved on a complex unit spherical surface in combination with manifold optimization, and finally efficient interference resistance is realized in the Beidou B1 and GPS L1 frequency bands. Balance of miniaturization, low power consumption and high performance is achieved through algorithm innovation, array processing Rayleigh limit constraint is broken through from the perspective of an anti-interference signal processing algorithm, and low-power-consumption and low-resource lightweight deployment can be achieved while the anti-interference performance is guaranteed.

The invention relates to a millimeter wave ISAC system perception performance optimization method based on sub-connection hybrid precoding. The method comprises the following steps: establishing an echo signal model containing an interference source; constructing an optimization problem with the aim of minimizing the Cramer-Rao bound; solving a digital and analog precoder by adopting an alternating optimization strategy; wherein the digital precoding is converted into second-order cone programming through successive convex approximation, and the analog precoding realizes efficient solution under constant modulus constraint through Riemannian manifold optimization. According to the method, collaborative optimization of the communication performance and the sensing performance of the millimeter wave ISAC system is realized through a hybrid precoding algorithm of alternate iteration optimization, a digital precoder design based on a successive convex approximation method and an analog precoder design based on Riemannian manifold optimization.

This invention discloses an angle-prior-assisted intelligent metasurface phase coding design method for two-dimensional DOA estimation. The method includes: constructing a system model for two-dimensional DOA estimation with a single receiving channel on the intelligent metasurface; deriving the explicit relationship between the CRB of the two-dimensional DOA estimation parameters and the phase coding matrix; utilizing the prior range information of the target two-dimensional angle, constructing a phase coding matrix optimization problem with a unit modulus constraint, aiming to minimize the average CRB; and efficiently solving this non-convex problem using the Riemannian manifold optimization algorithm to obtain the optimal phase coding matrix. This invention directly links the coding design to the theoretical performance lower bound. By optimizing the coding to focus beam energy in the target prior angle region, it significantly improves the accuracy of two-dimensional DOA estimation for azimuth and elevation angles within the angle prior range, reduces the required measurement overhead, and exhibits good robustness under one-bit phase coding.

The application relates to the field of bioinformatics and clinical auxiliary decision-making technology, in particular to a high-dimensional weighted membranous nephropathy gene co-expression network construction method, which integrates biomolecular interaction, disease pathways and health baseline data to construct an initial topological matrix, uses a Riemann manifold optimization algorithm to inject high-dimensional transcription sequencing data into the matrix and update edge weights, generates a high-dimensional weighted network of a target biological sample, accurately locks an abnormal gene module by calculating the topological residual error of the network and the health baseline and the dynamic offset of the eigenvector centrality, compares module features with a target point database to screen candidate drug targets, and stores network weight eigenvalues into a vector database, so that gene co-expression network construction and drug target screening based on biological priori and data driving are realized; and the application accurately retains weak but key local nonlinear cascade signals.

The invention provides a spherical manifold-based power distribution network optimization scheduling method and device, and belongs to the technical field of power system optimization scheduling. The method comprises the following steps: establishing a power distribution network optimization scheduling model composed of an objective function and constraint conditions; performing spherical manifold conversion on the power distribution network optimization scheduling model to obtain a corresponding power distribution network equivalent optimization problem based on spherical manifold optimization; and solving the equivalent optimization problem of the power distribution network based on spherical manifold optimization to obtain an optimal dispatching scheme of the power distribution network. The method is suitable for power system optimization scheduling in a distribution-microgrid cooperative operation scene, flexible resources such as energy storage, distributed renewable energy sources and switchable capacitors can be fully considered, new energy consumption maximization and system operation cost minimization are achieved, and the method has the advantages of being high in flexibility, high in calculation efficiency and comprehensive in optimization target.

The invention belongs to the technical field of information recommendation, and particularly relates to a recommendation method and system based on big language model enhancement and structured alignment, and the method comprises the steps: processing original texts of a user and an article through a big language model, and generating initial semantic embedding; constructing a graph neural network learning collaborative filtering signal, and learning a group of enhanced embedding specially used for recommendation through a manifold optimization method including constraints such as fidelity and smoothness in combination with the semantic embedding; an enhanced embedding space and a semantic embedding space are regarded as probability distribution, and a stable subspace relation expression coefficient matrix is found by solving a subspace self-expression problem; and an extended positive sample set is constructed, and enhanced embedding is guided and optimized by comparing a learning loss function, so that the enhanced embedding learns common characteristics of similar nodes. According to the method, the semantic comprehension capability of the large language model and the collaborative filtering capability of the graph model are deeply fused, so that the accuracy and robustness of a recommendation system are remarkably improved.

The invention relates to the technical field of antenna array hybrid precoding, and particularly discloses an MIMO system hybrid precoding method and device based on Riemannian manifold optimization, and the method comprises the steps: obtaining the channel state information of an MIMO system; determining an all-digital precoding matrix of the MIMO system according to the channel state information; determining a precoding optimization target of a hybrid precoder according to the full-digital precoding matrix of the MIMO system; performing optimization processing on the precoding optimization target according to a Riemannian adaptive quantum particle swarm algorithm until the Riemannian adaptive quantum particle swarm algorithm meets a termination condition, and obtaining a globally optimal analog precoding matrix; and according to the full-digital precoding matrix of the MIMO system and the globally optimal analog precoding matrix, constructing the hybrid precoder of the MIMO system. According to the MIMO system hybrid precoding method based on Riemannian manifold optimization provided by the invention, the high-dimensional, complex and popular calculation efficiency during hybrid precoding is improved.

This invention discloses an intelligent cooperative control method and system based on UAV swarms, belonging to the field of intelligent control technology for UAV swarms. The method includes collecting multi-source data to construct a sparse tensor, defining UAVs as nodes, and constructing a dynamic category graph. The nodes of the dynamic category graph are mapped to manifold points of a Riemannian manifold, UAV coordinates are converted to grid indices, an objective function is defined, the Riemann gradient of the objective function is calculated, and the manifold points are updated using exponential mapping to obtain updated spatial coordinates and task vectors. A weighted linear combination is used to generate control command vectors, which are then converted into control commands and executed. By constructing a sparse tensor and a dynamic category graph, and combining Riemannian manifold optimization to balance task priority and cooperative constraints, task coverage and cooperative robustness are improved. The combination of Riemannian gradient optimization and exponential mapping updates enhances autonomous cooperative capabilities.

The invention discloses an LPI radar waveform design and resource allocation method based on manifold optimization, and aims to improve the target tracking precision of a radar. According to the method, the multi-target tracking performance is quantified through the target function based on the posterior Cramer-Rao lower bound. In order to improve accuracy and calculation efficiency, a manifold optimization framework is constructed to carry out optimization post-processing on a constructed objective function: firstly, an accurate penalty function is introduced, an LPI constraint condition is smoothly fused into the objective function, and conversion from a constraint problem to a penalty form is realized; constructing a volume manifold space based on resource constraint characteristics, and reconstructing an original optimization task into an unconstrained optimization problem on a manifold; meanwhile, a parallel conjugate gradient descent algorithm with a self-adaptive step length mechanism is adopted to execute optimization solution. According to the algorithm, parallel computing can be fully utilized to realize rapid convergence, and efficient exploration can be realized in a product manifold space through a self-adaptive step length strategy, so that a better solution and more stable optimization performance are obtained.

The invention discloses a data dimension reduction feature selection method based on a swarm intelligence algorithm, and relates to the technical field of big data analysis, and the method comprises the steps: mapping an initialized binary code to a Grassmann manifold, generating an initial point position, obtaining a direction vector through a vector transfer operator, updating a particle position in combination with a mixed quantum revolving door, and generating an actual value scoring vector. Generating an update coding vector through a Sigmoid function, performing personal and global optimal update in combination with the fitness, and outputting an optimal feature subset after the maximum number of iterations is reached; a Shannon entropy and Laplacian score joint evaluation mechanism is introduced, the precision and robustness of feature contribution degree evaluation are improved, a topology retention index based on persistent coherence is constructed, the structural stability and interpretation after dimension reduction are improved, Grassmann manifold optimization and a hybrid quantum revolving door search strategy are fused, and the swarm intelligence algorithm is prevented from falling into local optimum.

This invention discloses a method and model for classifying pore types in coal and rock scanning electron microscope (SEM) images. The method includes: training a generative adversarial network (GAN) using a dual manifold optimization mechanism on a first sample set to obtain an image generation model; using the image generation model to augment a second sample set to establish a third sample set, where both the first and second sample sets contain SEM images of primary, metamorphic, mineral, and strain-induced pores in coal and rock; training an extreme learning machine (ELM) using a dynamic boundary smoothing mechanism on the third sample set to obtain a classification model; and determining the pore type of the coal and rock SEM images based on the classification model. This method introduces adversarial network training data to expand the classification model, solving the problems of insufficient training data and poor model generalization ability, while the dual manifold optimization mechanism maintains the topological structure of the image data; the use of an ELM with a dynamic boundary smoothing mechanism for pore type identification further improves classification accuracy.

The invention relates to the technical field of power system frequency regulation, and particularly provides a manifold optimization model-based dynamic frequency support optimization method for an energy storage power station, which comprises the following steps of: acquiring a power grid frequency fluctuation signal of the energy storage power station in real time, and dynamically acquiring time sequence data of the power grid frequency fluctuation signal in a target time period; performing variational mode decomposition on the time sequence data to obtain a fast frequency change component and a slow frequency change component; calculating a frequency deviation corresponding to each rapid frequency change component, and when the frequency deviation exceeds a dead zone range, correcting the frequency deviation to be within the dead zone range through the output power of the energy storage equipment; and calculating scheduling parameters of the current time period according to the low-speed frequency change component, the objective function and the manifold optimization model. According to the dynamic frequency support optimization method of the energy storage power station based on the manifold optimization model provided by the invention, an effective multi-band collaborative optimization mechanism is provided, and the resource utilization efficiency and the real-time performance can be improved.

The invention discloses an unmanned aerial vehicle positioning method and system based on manifold optimization, and the method comprises the steps: building a pose calculation model based on manifold optimization based on a principle that the projection center of a camera carried by an unmanned aerial vehicle, the coordinates of an ith control point and the coordinates of a corresponding image point are collinear, and enabling the camera carried by the unmanned aerial vehicle to shoot a picture, an industrial personal computer carried by the unmanned aerial vehicle automatically selects a plurality of control points in the picture, coordinate information of the control points is substituted into the resolving model, the position and the posture of the unmanned aerial vehicle in the picture shooting process are obtained through resolving of the resolving model, and the resolving model comprises two contents including a rotation matrix confirmation method and a spatial position confirmation method. And determining two parameters of the rotation matrix and the spatial position by adopting an alternative minimization mode until a set termination criterion is met, and obtaining the rotation matrix and the spatial position which are the attitude and the position of the unmanned aerial vehicle respectively. According to the method, the positioning precision and robustness are improved, and the method has relatively high engineering applicability.