196 results about "Improved algorithm" patented technology

This invention provides an intelligent optimization and collision avoidance method for powder coating lifting fixture paths, relating to the field of powder coating production automation technology, including the following steps: S1: Constructing an environmental map of the powder coating production line, generating an initial lifting fixture path based on an improved A* algorithm, the improved A* algorithm optimizing the heuristic function by introducing a direction cost factor; S2: Iteratively optimizing the initial path generated in S1 using an adaptive inertia weight PSO algorithm to obtain a multi-objective optimal path; This invention introduces a direction cost factor into the heuristic function of the A* algorithm, fully considering the matching degree between the lifting fixture's movement direction and the target direction, resulting in an initial path that better conforms to the lifting fixture's movement characteristics, reducing redundant segments and unreasonable turns; Combined with the adaptive inertia weight PSO algorithm, the inertia weight is dynamically adjusted according to the iteration process, which can accurately adapt to the path optimization needs of different production line scenarios, achieving multi-objective optimization of lifting fixture running time, path length, and workpiece sway amplitude.

This invention proposes an integrated beamforming method and system for communication and sensing based on reconfigurable smart surfaces, belonging to the field of wireless communication technology. The aim is to design suitable beamforming to maximize communication and data rate while ensuring adequate radar sensing performance. The method includes: first, constructing an ISAC system containing J RIS elements, a base station (BS), and M single-antenna mobile users (MUs), and completing the modeling of baseband signals and user received signals; second, constructing user communication and data rate models and radar target detection signal-to-noise ratio models, establishing unit mode constraints on RIS reflection coefficients and transmit power constraints; third, employing an improved algorithm combining alternating optimization and fractional programming to decompose the non-convex optimization problem into sub-problems for iterative solution, ultimately maximizing communication and data rate and radar sensing performance, while also meeting the dual functional requirements of communication and sensing. This invention is applicable to providing target detection and communication functions for vehicles.

The application discloses a power grid power transmission section identification method based on an improved PageRank algorithm and network conversion, first, an improved PageRank algorithm is used to calculate a node criticality index, and a network constraint coefficient is calculated according to a network topological structure and is used to correct a restart vector; second, a key power transmission section search topological graph is constructed through network conversion and node criticality index calculation, and identification of the key power transmission section is converted into search of a shortest path; finally, according to energy transaction between a generator node and a load node, a search range of the key power transmission section is set, a related cut set with the highest criticality is found, and through checking and correction, an identification result of the key power transmission section is obtained. The application can well improve the deficiency of the original PageRank algorithm when applied to the power grid, can accurately identify the power transmission section of the power grid, provides a key monitoring guide for power grid monitoring and dispatch personnel, and better guarantees the safe and stable operation of the power grid.

The application discloses a multi-four-rotor unmanned aerial vehicle gathering type task path planning method based on reinforcement learning, first constructs a multi-four-rotor unmanned aerial vehicle Gym environment based on PyBullet, abstracts a state space and an action space of the four-rotor unmanned aerial vehicle, sets a reward function mechanism, then uses an improved deep reinforcement learning algorithm to make path planning decisions, finally trains the improved deep reinforcement learning network, controls the four-rotor unmanned aerial vehicle through output action information, and makes each four-rotor unmanned aerial vehicle successfully reach a specified target task in the shortest time. The method uses N-step returns in the TD3 algorithm, obtains more accurate return estimates, faster learning speed and better generalization performance, uses priority experience replay to reduce sampling bias, reduces model bias caused by uneven sampling, improves algorithm stability, makes the TD3 algorithm more suitable for continuous multi-dimensional decision problems, and can plan an optimal route in the shortest time under the condition that real-time performance and accuracy are reached.

This invention proposes a multifunctional all-terrain intelligent welding robot for ships, comprising a quadruped chassis module, a six-degree-of-freedom robotic arm module, a vision recognition system, and a hierarchical path planning system. The quadruped chassis module employs a Mecanum wheel quadruped chassis, a liftable electromagnetic adsorption structure, and a guide rail electromagnetic foot collaborative structure to form an omnidirectional mobile platform supporting all-terrain and all-attitude operations. The six-degree-of-freedom robotic arm module is fixed above the quadruped chassis module and uses a floating coupling to isolate chassis vibration, achieving high-precision positioning of the robotic arm. The vision recognition system processes image data using a YOLOv8 model improved through Mosaic data augmentation and AFPN feature fusion, and combines OpenCV edge-morphology algorithms to optimize target recognition, developing a dedicated integrated detection algorithm for ship section assembly. The hierarchical path planning system performs three-dimensional path planning using an improved RRT*Connect algorithm, and combines force feedback data to correct the trajectory in real time, completing circumferential, blind-angle-free repair.

The application discloses a kind of reinforcement learning batch process control methods based on improved AC algorithm, it is related to the field of deep reinforcement learning and batch process control field.The method will be based on reinforcement learning method The batch process control is modeled as an optimal control problem on the basis of Markov decision process;Control action constraint is introduced in the reward function of reinforcement learning controller, the number of effective reward samples is increased to improve the learning rate of reinforcement learning controller, and the control cycle is shortened.Priority sampling method is introduced in the Actor-Critic algorithm of deep reinforcement learning, and a soft actor-critic algorithm with priority sampling is proposed to improve the sampling efficiency in the experience replay pool.The present application does not depend on prior knowledge and process model, and can realize model-free control of batch process.

The present invention relates to the technical field of target detection. Provided are a small target detection method and system based on an improved YOLOv7 algorithm, and a storage medium and an electronic device. In the technique, the performance and effect of a YOLOv7 model in an urban traffic road-end detection scenario is improved from two perspectives: introducing an adaptive bounding box loss weight without changing the network structure of the model, thereby increasing the detection accuracy of the model with regard to a small target; and adding a detection head for a small target with only a slight increase in the inference time, which also increases the detection accuracy of the model with regard to the small target. In the present invention, the detection accuracy of a YOLOv7 target detection algorithm with regard to a small target is improved with almost no increase in the detection cost, thereby ensuring both the speed and accuracy of small target detection.

This invention aims to optimize the input sequence of actuators with significant time delays in marine vehicles (such as variable buoyancy hydraulic pumps and valves, and slow-response actuators like rudders). Under complex sea conditions and model uncertainties, this method achieves rapid and stable attainment of the target operational state with minimal action cost, while simultaneously satisfying safety and lifespan constraints. The method disclosed in this invention uses a parameter identification model as the prediction kernel, incorporating terminal task deviation, the number of execution actions and non-zero duration, and the safety volume boundary into a multi-objective cost function. It employs an improved LM algorithm with adaptive damping for sequence-level optimization and rolling updates, thereby overcoming the shortcomings of conventional closed-loop compensation and pure gradient methods in terms of slow convergence, oscillation, and difficulty in balancing safety and wear on time-delayed objects. This significantly improves convergence speed, control accuracy, and reliability, while reducing energy consumption and mechanical wear.

The application belongs to the technical field of industrial system information, and provides an industrial target recognition and accurate positioning method based on an improved Hough algorithm, which comprises the following steps: adjusting parameters of an industrial camera to obtain clear image data of a workpiece, and pre-processing the image data; performing edge detection on the pre-processed image based on a morphological processing algorithm and a Canny edge detection algorithm, obtaining edge information, and judging whether there is edge loss; for the edge loss condition, performing target recognition and positioning based on an improved Hough algorithm or a template matching algorithm, and obtaining the center point coordinates of the target object; the application can more accurately and effectively detect the features of the target boundary, accurately deduce the boundary loss condition, and then realize accurate positioning of the detection target.

This invention discloses an intelligent optimization method for communication chip design parameters based on deep learning, comprising: S1, constructing chip topology graph data and encoding a graph structure input using Kirchhoff's laws; S2, collecting historical data to construct a standardized training sample library; S3, aggregating node information to learn parameter coupling relationships and outputting a performance prediction model; S4, introducing an improved QPSO algorithm to construct a physically constrained pseudo-gradient guiding term, using the first derivative to force particle movement in flat regions, and outputting a potential global optimal solution region; S5, using the MOEA / D algorithm to decompose a multi-objective problem and iteratively selecting non-dominated solution sets as candidate sets in parallel; S6, backfeeding a high-precision simulator for verification and incremental updates, outputting the optimal parameter combination. This invention achieves physical constraint modeling and gradient-guided optimization of design parameters, effectively improving optimization efficiency, prediction accuracy, and design feasibility.

The application discloses a kind of fusion ToF and monocular vision's robot joint layer perception system, belong to robot perception field, system uses Eye-in-Hand architecture, integrates ToF camera and RGB camera;Through joint bilateral filtering and Kalman filtering, improve depth map quality;Based on SURF feature matching and the ICP registration of KD-Tree acceleration is realized target six degree of freedom pose estimation;Combined with improved RRT algorithm and hierarchical collision detection planning capture path.The robot joint layer perception system of fusion ToF and monocular vision provided in the application, through the space alignment and fusion of two kinds of visual information, make up for respective limitations, improve the robustness and precision of overall perception system;In single-arm and double-arm capture experiment, system positioning accuracy reaches ±3mm, dynamic scene capture success rate is high.It is suitable for industrial automation, logistics sorting and the like.

This invention provides a simulation and pre-drilling method and system based on smart water conservancy. Through multi-source data integration and preprocessing, it fuses hydrological, geographical, and population data to output a standardized fusion dataset. Digital twin scenario construction, combined with the MARL algorithm, achieves bidirectional coupling between flood evolution and resident behavior. Dynamic simulation and pre-drilling drive bidirectional coupled simulation to predict flood inundation and disaster chain transmission. Strategy generation and optimization employ an improved NSGA-III algorithm to achieve multi-objective evacuation path optimization and generate resettlement plans. Specifically, it breaks through the limitations of traditional unidirectional simulation, achieving multi-objective dynamic optimization and improving the intelligence, precision, and fairness of flood disaster response. It can be widely applied to flood disaster prevention and control in various river basins and cities.

The application relates to an unmanned underwater vehicle path planning method based on an improved SAC algorithm, and comprises the following steps: obtaining state data of an unmanned underwater vehicle in an underwater environment; a simulation experiment environment is built based on a Gym framework, physical parameters of the unmanned underwater vehicle, geometric shapes of seabed obstacles and current characteristics are simulated, and an interactive feedback mechanism of an environment state and unmanned underwater vehicle action is constructed; a path planning algorithm based on an improved SAC is constructed, a heuristic exploration mechanism is introduced on the basis of the SAC algorithm, heuristic actions are generated based on Euclidean distance or actions are generated through a strategy network, and a multi-target reward function is designed; state data is input into the path planning algorithm, actions of the unmanned underwater vehicle are generated through the strategy network, the value of the actions is evaluated in combination with a value network, network parameters are updated by using the interactive feedback mechanism, and a path planning strategy is optimized. Compared with the prior art, the application has the advantages of fast convergence in a complex underwater environment, good adaptability and the like.

This invention discloses a mountain path planning method based on an improved A* algorithm, comprising: acquiring lidar or image data of the target area using a UAV and generating point cloud data to construct a digital elevation model containing elevation and slope information; subsequently, performing coordinate unification, noise reduction, and downsampling processing on the point cloud data, and combining the start and end points for region clipping and rasterization to construct a local raster map; employing the improved A* algorithm during path search, introducing a slope correction mechanism and an equivalent distance cost model, comprehensively incorporating distance, slope, and elevation factors into the path cost function, and dynamically adjusting the path cost based on the slope influence coefficient, while simultaneously constraining high-slope and impassable areas. This method can effectively reduce the search space and computational complexity, improve the efficiency and accuracy of path planning in complex mountainous environments, and is applicable to scenarios such as power transmission line planning and mountain engineering path design.

The application discloses a kind of based on spatial intelligence's active fault three-dimensional model automatic construction method, belong to geological exploration and earthquake engineering technical field;Method includes: collection data, analysis and extract minimum complete subdirectory;Through adaptive threshold hierarchical clustering combined with improved RANSAC algorithm, the exclusive small earthquake cluster of each fault is automatically identified;Based on local weighted regression, three-dimensional automatic slice is made to small earthquake cluster, and each profile fault interpretation line is fitted by moving least square method;Finally, active fault three-dimensional fine model is constructed;Test model rationality and output model file;The application realizes the automation, quantification of fault modeling whole process, can be fused multi-source data and realize multi-element constraint modeling, adapt complex fault zone modeling, and the fine three-dimensional model constructed can provide key data support for fault present-day deformation inversion, three-dimensional potential source research, earthquake geological disaster assessment, with higher engineering application value.

The present application belongs to the technical field of feed formula research, and more particularly relates to a sow nutrition feed formula screening method based on an optimization algorithm. By analyzing population initialization distribution, a Bernoulli mapping sequence is introduced to initialize the population, so that the population distribution is more uniform, and the search ability of the algorithm in the early stage is improved. An improved adaptive nonlinear convergence factor is introduced to balance and enhance the global search ability and local development ability of the algorithm. A neighborhood disturbance optimization mechanism is added to generate new neighborhood solutions in the iteration process to prevent the algorithm from easily falling into a local optimal solution in the iteration, especially in the later stage. Finally, by comparing the optimization of sow feed formula before and after the improvement of ChOA, the effectiveness of the BANChOA algorithm is further verified. Through the analysis of the improvement strategy of the present application, it can be seen that the improved algorithm requires a large number of parameters and the optimization performance of the algorithm is greatly affected by the neighborhood disturbance optimization. There is still room for improvement in population initialization and convergence factor.

This invention discloses an experimental droplet routing method for digital microfluidic chips based on an improved A* algorithm. The improvement to the A* algorithm includes: incorporating the number of electrode wear cycles into the heuristic function of the A* algorithm; adding a coefficient related to the number of wear cycles, which can be modified according to the chip wear level; reducing the coefficient when chip wear is minor, prioritizing routing distance, and calculating the shortest distance as the target for route calculation; increasing the coefficient when chip wear is severe, aiming to calculate the route through areas of less worn chip and avoiding areas of severe wear; and adjusting the path after obtaining the droplet routing path, moving the repetitive wear areas at the start and end points to the middle of the path to reduce repeated wear on electrodes in the start and end areas. This invention, by improving the A* algorithm, allows droplet routing to consider electrode wear factors simultaneously, further improving the chip wear uniformity and extending the chip's lifespan.

This invention discloses an artificial intelligence-based hot-dip galvanized layer uniformity detection system, comprising: a hardware acquisition layer (101), a data processing layer (102), an intelligent analysis layer (103), and an evaluation output layer (104) arranged from top to bottom, interconnected via an industrial Ethernet network; the hardware acquisition layer consists of a vision acquisition module and a mechanical transmission unit, used to simultaneously acquire images of the galvanized layer surface and data on the thickness distribution of the galvanized layer surface; the data processing layer consists of a data preprocessing module, used to process the images of the galvanized layer surface and the data on the thickness distribution of the galvanized layer surface; the intelligent analysis layer consists of an AI algorithm processing module, used for galvanized layer region segmentation, multi-scale feature extraction, and uniformity difference identification, employing an improved U-Net++ algorithm and attention mechanism; the evaluation output layer consists of a quantitative evaluation module, a result output module, and a calibration module, used for uniformity quantitative evaluation, result display, and system accuracy calibration, respectively. A corresponding detection method is also disclosed.

The application discloses a pump station group cooperative scheduling method and system based on digital twinning, and belongs to the technical field of optimization control based on digital twinning. The method comprises the following steps: collecting multi-source heterogeneous data of the pump station group, constructing a basic framework of a digital twinning model of each pump station group, and synchronously updating; deeply fusing health degree, energy efficiency ratio, load capacity and response speed data, and calculating a pump station scheduling response willingness factor; using an improved algorithm, embedding a dynamic scheduling capability weight vector into a balance fitness function and a response weighted search mechanism for solving; using the digital twinning model of the pump station group, pre-animating and verifying a cooperative scheduling optimization scheme of each pump station, confirming hydraulic balance and operation safety, generating target rotating speed and start-stop instructions, and delivering the instructions to each pump station for execution. The application quantifies the pump station scheduling response willingness by constructing a digital twinning model, thereby guiding cooperative optimization, and realizing willingness distribution driven scheduling.

The application relates to the technical field of target detection, in particular to a target detection method based on an improved YOLOX algorithm, which aims to solve the technical problem of insufficient detection precision of small target images. The target detection method comprises the following steps: obtaining a trademark data set, and converting the picture format in the trademark data set into a format that can be directly input into a YOLO network for training, thereby obtaining a usable trademark data set; pre-processing the data of the usable trademark data set, and dividing all trademark pictures into a training set, a verification set and a test set; analyzing the training set, the verification set and the test set to determine all trademark categories; improving the network structure based on YOLOX to obtain an improved YOLOX target detection algorithm; training the improved YOLOX target detection algorithm with a training data set, evaluating to obtain a best training weight file, and obtaining a final YOLOX target detection algorithm. The detection precision of small target images is improved.

The present application relates to a transformer fault diagnosis method based on improved sparrow search algorithm optimization SVM, through the pretreatment and feature engineering of transformer oil chromatographic data; according to the data of pretreatment and feature engineering, the model based on oil chromatographic data is constructed, and the model parameters are optimized and trained; the input transformer oil chromatographic data is carried out transformer fault diagnosis based on the optimized and trained model.The present application improves the sparrow search algorithm by introducing the best point set strategy, the golden regular update rule, the differential mutation disturbance and the reverse learning mechanism, so as to comprehensively improve the parameter optimization process of support vector machine, the global exploration ability of the algorithm is enhanced, the risk of falling into local optimum is avoided, and the problem of premature convergence is effectively avoided.In addition, the improved algorithm pays attention to multi-link optimization, and the parameter setting is simple, the algorithm is low in use difficulty, so that the optimized support vector machine (SVM) can effectively improve the transformer fault prediction precision.

The application provides an aircraft wing skin corrosion depth prediction method based on a GAN and an improved LightGBM algorithm, which comprises the following steps: S1, data enhancement: four fully connected layers are used to extract feature information between columns; S2, an improved condition generator and discriminator are used, and parameters are updated; S3, the LightGBM algorithm is improved, and the aircraft wing skin corrosion depth is predicted. The aircraft wing skin corrosion depth prediction method using the generative adversarial network for data enhancement and the gradient regression model for sample regression can accurately predict small sample data. Meanwhile, the gradient regression model LightGBM algorithm is improved, which is more suitable for the prediction of the aircraft wing skin corrosion depth, and can greatly improve the prediction accuracy of the aircraft wing skin corrosion depth prediction.

The application discloses an indoor distribution cable path intelligent generation method combined with building environment characteristics, and comprises the following steps: constructing a semantic grid map capable of distinguishing different building function areas by performing layer identification, semantic annotation and space discretization on an input building plan design map; based on the semantic grid map, different basic passing generation values are given to different types of areas, and a dynamic engineering constraint function comprising a corner turning penalty and a wall sticking reward rule is defined; path search of an improved A* algorithm is performed, in an engineering constraint cost field fused with the basic passing generation value and the dynamic engineering constraint function, a heuristic search algorithm with a dynamic penalty term is used to calculate and generate a candidate cable path meeting engineering specifications; virtual convergence points are determined through node clustering, the path search algorithm is applied hierarchically, a backbone path limited in a public area is generated, branch paths entering rooms are generated, and a complete indoor distribution system cable path topology is generated by merging.

This invention relates to the field of digital urban planning and intelligent path optimization technology, and discloses a method for optimizing urban path network replacement points based on an improved Dijkstra algorithm. First, a county-level facility-road network coupled graph model is constructed, and facility load levels and population weights are labeled. The standard Dijkstra algorithm is used to calculate the initial coverage rate and load rate. Based on the load level and coverage blind spots, a load-blind spot coupled gravitational field model is constructed, and AI adaptive weights are used to generate candidate replacement point heatmaps. Coverage gain, blind spot attraction, and load diversion factors are introduced to construct a multi-dimensional cost function for the improved Dijkstra algorithm, and the coefficients are dynamically adjusted using AI adaptive weights. The comprehensive cost is calculated by traversing the candidate points, and the point with the minimum comprehensive cost is selected as the optimal replacement point. The process is iterated until a termination condition is met, and then the solution is output. This solves the problem of not being able to automatically locate the optimal replacement point position that balances coverage improvement and load diversion.

The application discloses a kind of methods for unmanned aerial vehicle path planning based on improved differential evolution algorithm, comprising the following steps: 1) model is built;1-1) the representation method of unmanned aerial vehicle flight path is built;1-2) three-dimensional environment model is built, and the three-dimensional environment model includes terrain model and threat area model;1-3) constraint condition model is established according to unmanned aerial vehicle flight performance and terrain constraint condition;1-4) cost function is constructed according to distance, height, height variation and corner index;2) improved algorithm, the multi-variation strategy of self-adaptive weight adjustment is introduced in standard differential evolution algorithm, and scaling factor F is dynamically adjusted according to iteration number, and the standard differential evolution algorithm is improved;3) the flight path of unmanned aerial vehicle is obtained;3-1) the model parameters built in step 1) and the task parameters of unmanned aerial vehicle are loaded;3-2) improved standard differential evolution algorithm is used to obtain the flight path of unmanned aerial vehicle;3-3) the flight path of unmanned aerial vehicle is smoothed, and the final flight path of unmanned aerial vehicle is obtained.

The application belongs to the technical field of charging facilities, and discloses a charging station site selection and capacity configuration optimization method based on multi-source space-time data, which collects six-dimensional full multi-source space-time data of vehicles, power grids and the like, divides basic and vehicle-end characteristic data categories, performs full-process standardized processing, completes exception correction, missing completion and unification of CGCS2000 space coordinate system and dynamic time granularity, also realizes vehicle-network space-time calibration and builds an integrated index database of GeoHash coding; improves planning basic data precision, eliminates space-time matching deviation of different data sources, realizes rapid retrieval and linkage query of vehicle-network charging data; through construction of a three-level space-time constraint system, high feasibility alternative stations are screened, multi-dimensional decision variables are determined and a site selection-capacity-vehicle-end adaptation collaborative correlation matrix is built, five core multi-objective optimization directions are established, and an improved NSGA-III algorithm is used for solving and sorted by an entropy weight-TOPSIS method.