15results about How to "Achieve global optimization" patented technology

This invention provides a three-trailer collaborative aircraft transfer method and system based on multi-sensor fusion. This invention relates to the field of aircraft ground support technology, specifically to a three-trailer collaborative aircraft transfer method based on multi-sensor fusion, applicable to short-haul aircraft ground transport to hangars. Through multi-sensor data fusion and multi-trailer collaborative control, it improves transfer accuracy, safety, and efficiency. The method specifically includes the following steps: system setup and sensor calibration; trailer attitude initialization and collaborative parameter configuration; real-time acquisition and preprocessing of multi-sensor data; multi-sensor data fusion and state estimation; dynamic path planning based on fused data; attitude error quantification and calibration based on complex statistics; multi-trailer collaborative driving control and traction distribution; real-time environmental monitoring and risk warning; precise docking at the hangar entrance and attitude fine-tuning; and parking and system review and optimization.

The application discloses an active defense method, device and equipment for diffusion speech conversion and a medium, and relates to the technical field of speech security. The active defense method comprises the following steps: obtaining source speech and reference speech, introducing a constrained protection disturbance into the reference speech, constructing protected speech, and ensuring that the protection disturbance satisfies a disturbance imperceptibility constraint. The source speech and the protected speech are input into a speech conversion model with the reference speech as a speaker condition, content information analysis, speaker condition constraint acoustic generation and waveform synthesis are completed by the speech conversion model, and a protection generated speech is output. A joint loss function is constructed based on speaker embedding representations of the protection generated speech and the reference speech, gradients of a current disturbance position and a predicted midpoint position are calculated according to the joint loss function and are fused to obtain an update direction, the protection disturbance is iteratively optimized, and projection or clipping is performed to satisfy the disturbance imperceptibility constraint. Finally, the protected speech is output.

This invention relates to the field of intelligent battery swapping recommendation technology, and discloses an intelligent battery swapping recommendation method and system based on historical battery data. The method includes: a vehicle central control system identifying the vehicle from a user-triggered battery swapping request signal to obtain parameter information of the current vehicle; querying historical battery matching data corresponding to the parameter information from a cloud database, and calculating the compatibility score between each battery model and the current vehicle; a battery swapping cabinet control system performing multi-objective collaborative recommendation decision-making with the compatibility score and the inventory batteries to obtain the optimal recommended battery identifier; and the battery swapping cabinet opening the corresponding compartment door according to the optimal recommended battery identifier to execute the battery swapping process and generate a battery swapping operation record. This invention improves the accuracy of battery compatibility assessment, can discover performance correlations between different battery models, and thus improves the automation and intelligence level of the entire battery swapping process.

The present application provides a ladle scheduling method and device based on a heat plan, equipment and storage medium, the method comprises the following steps: obtaining the heat plan and the ladle information; formulating the corresponding ladle plan according to the heat plan; determining the planned start time and the planned end time of the ladle plan based on the time information corresponding to the multiple process stages of the heat plan; establishing a ladle matching model according to the ladle plan and the ladle information; taking the highest ladle utilization rate and the least number of matched new ladles as the optimization target; solving the ladle matching model to obtain the matching result of the ladle and the ladle plan; and matching the heat plan and the ladle based on the matching result and the corresponding relationship between the ladle plan and the heat plan, so as to schedule the ladle based on the heat plan. The method solves the technical problems of low efficiency and easy errors in manual ladle scheduling.

The deep learning-based video multi-level adaptive compression cross-layer transmission optimization method and system of the application relates to the technical field of video transmission. By designing a compression level selection model based on deep learning, the optimal compression level is output based on the original video data and network environment parameters, and the corresponding compression strategy of the optimal compression level is obtained. A transmission strategy selection model is designed to generate a transmission strategy based on video quality parameters and network environment parameters. A global utility function is defined to calculate the global utility value when the compression strategy and the transmission strategy are assumed to be executed. An optimization objective function is designed, which inputs the compression strategy and the transmission strategy and outputs the optimized compression strategy and the transmission strategy. A condition judgment logic is set to select the final strategy based on the compression strategy and the transmission strategy before and after optimization, thereby achieving global optimization of video compression and transmission.

This invention discloses a method for adjusting the end effector stiffness of a hybrid robot used in the machining of thin-walled structural parts. The method first collects the proportional gain of each active branch servo drive system as the axial servo stiffness; secondly, it establishes a mapping model based on the robot configuration to solve for the equivalent axial servo stiffness of the end effector in the reference coordinate system; subsequently, it introduces stiffness weighting coefficients and combines them with the mechanical stiffness and electromechanical coupling stiffness performance evaluation index; finally, it uses a one-dimensional discrete search algorithm to traverse and filter redundant rotation angles that satisfy displacement and rotation constraints while keeping the end effector pose unchanged, aiming to maximize the electromechanical coupling stiffness index and generate an optimal sequence of redundant rotation angles. This invention establishes an electromechanical coupling model integrating mechanical and servo systems and adaptively adjusts the end effector stiffness using redundant degrees of freedom, which can effectively suppress cutting deformation and chatter during the machining of thin-walled parts while reducing computational complexity, significantly improving machining quality.

The invention discloses an adaptive path training system and method based on capability evolution prediction, and the system comprises a data collection module which is used for collecting the real-time behavior data of a target training object in a training process; the capability state construction module is used for constructing capability state data according to the real-time behavior data; the capability trajectory prediction module is used for performing capability evolution prediction on the capability state data according to a preset time sequence prediction model to obtain a behavior evolution trajectory sequence; the training path generation module is used for constructing a target function on the basis of a preset training task set according to the behavior evolution trajectory sequence and in combination with training time and training cost, and solving the target function on the basis of a preset strengthening algorithm to obtain an optimal training path; and the training execution module is used for sequentially executing the corresponding training tasks based on the optimal training path. The technical problem that in the prior art, a training path is fixed, and a training object cannot be dynamically adjusted is solved.

This invention relates to the field of intelligent decision-making technology and discloses an intelligent decision-making system for agricultural irrigation based on artificial intelligence. The system constructs a collaborative architecture comprising a field perception layer, an edge computing layer, a cloud-based decision-making layer, and an irrigation execution layer, forming an irrigation task context through multi-source perception data. Based on this, a meta-reinforcement learning-driven multi-round agent-based search decision-making mechanism is introduced, combined with explicit self-reflection to achieve iterative optimization of the decision-making process. Simultaneously, a preference-aware reward evaluation model is constructed, achieving multi-objective trade-off modeling through latent variable mirror constraints and inverse autoregressive flow transformation, and globally optimizing the strategy using a cross-round benefit attribution mechanism to generate an irrigation scheduling scheme that meets water-saving, crop water requirement, and energy consumption constraints. This invention can improve the accuracy, stability, and adaptability of irrigation decisions.

The invention relates to a defect detection system based on a large projection model, and relates to the technical field of machine vision defect detection. The technical problems that an existing defect detection technology needs to depend on specific scene customization design, the cost is high, and the detection effect is limited are solved. The system comprises a hardware module, a data acquisition module, a model training module and a detection application module, the hardware module forms an imaging facility with a fixed pose by a DLP (Digital Light Processing) and an area-array camera, the data acquisition module obtains a 70-channel initial imaging effect picture and a structured cue word of multi-period fringe projection, and the model training module realizes cooperative training of a projection large model and a detection large model through reinforcement learning. And the detection application module completes full-process automatic defect detection. Manual customization design is not needed, the technical threshold is reduced, time and labor cost are saved, the detection precision and universality are remarkably improved, hardware equipment is small, exquisite and portable, and the method is suitable for detection scenes of various materials, surface types and defect types.

The application discloses a kind of network type converter and electric heat cold energy storage system collaborative optimization control method, contain steps as follows: construct "network type converter-electric heat cold energy storage-flexible load" multi-objective optimization model;Establish the collaborative constraint of network type converter and energy storage system;Establish demand response and network state linkage mechanism;Solve the multi-objective optimization problem by NNC algorithm and combine TOPSIS method to filter optimal operation scheme;According to optimal operation scheme, dynamically adjust the virtual inertia and droop coefficient of network type converter, synchronously dispatch electric energy storage charge-discharge power and cold energy storage and / or heat energy storage electric drive equipment operating state.The method can quickly and accurately suppress new energy output fluctuation and load mutation, strictly constrain system voltage deviation and frequency deviation within a more optimal operating range, fundamentally enhancing the transient stability and dynamic power quality of the grid with high proportion of new energy access.

The application discloses a method and device for identifying orthologous gene groups in a pan-genetic family, and belongs to the field of bioinformatics. The method first acquires a phylogenetic tree, a sequence set and a target species number of genes to be grouped, and calculates a similarity matrix of gene pairs; then, a data-driven 5th percentile (P5) is used to automatically infer a split threshold, and an initial gene grouping is performed in combination with the topological structure of the phylogenetic tree; subsequently, a post-processing pipeline including isolated leaf node redistribution, micro-group topological absorption, adjacent group merging and super-large gene group targeted re-splitting is executed; finally, based on a multi-dimensional quality score function and Latin hypercube sampling, parameter optimization iteration is performed, and the best orthologous gene group (OGG) is output. The application breaks the limitation of traditional clustering which only depends on sequence similarity, effectively eliminates over-merging, fragmentation and multi-lineage genome errors, significantly improves the accuracy and evolutionary rationality of gene grouping of complex polyploid species, and has important significance for dividing core and non-core genes in pan-genome analysis. The specific flowchart is shown in FIG. 1.

This invention discloses a weather-interference-adaptive point cloud 3D target detection method and apparatus, relating to the field of target detection technology. The method includes: constructing a weather-interference-adaptive point cloud 3D target detection model, comprising an interference sensing network and a hybrid expert target detection model; converting raw point cloud data into distance images and inputting them into the interference sensing network for interference level classification to obtain interference levels; matching the corresponding expert model in the hybrid expert target detection model using a sparse activation mechanism based on the interference levels; inputting the raw point cloud data into the matched expert model and outputting the corresponding detection results; based on the detection results and interference levels corresponding to the raw point cloud data, employing a staged training strategy, training the detection model through a constructed total loss function to obtain a trained model; inputting the point cloud data to be detected into the trained model and outputting the final 3D target detection result. Using this invention can improve detection accuracy.

The application discloses a low-resource environment large model deployment reasoning method based on deep learning, comprising the following steps: S1, acquiring hardware physical characteristic parameters and constructing a hardware perception cost model; S2, generating an initial lightweight model by using a neural architecture search algorithm; S3, monitoring the device environment state in real time, calculating the maximum theoretical calculation amount that can be borne, and generating a resource availability vector; S4, calculating semantic entropy values, judging sample complexity and allocating initial calculation resource quotas; S5, introducing an improved CMA-ES algorithm of semantic entropy guided population center initialization, adaptively searching and outputting an execution strategy; S6, performing high or low precision fixed-point operation on key layers and non-key layers respectively; S7, performing decoupling recovery and inverse quantization processing on feature maps, and outputting final results. The application realizes efficient deployment of a large model in a low-resource environment, significantly reduces delay and energy consumption while ensuring reasoning accuracy.

The application relates to the technical field of municipal pipeline wiring, and discloses an intelligent municipal pipeline wiring method and system based on path optimization, which comprises the following steps: preprocessing multi-source municipal pipeline related data to obtain a standardized data set, constructing a BIM and GIS model and fusing the two models to form a three-dimensional digital base; dividing three detection areas by hierarchical spatial indexing and establishing a pipeline detection unit, completing conflict checking and semantic topology analysis, and demarcating a constraint boundary; constructing a multi-constraint optimization model, adopting a multi-objective genetic algorithm and deep reinforcement learning hybrid optimization, checking conflicts in real time and adjusting parameters, and generating an optimal pipeline path scheme; generating construction instructions according to the scheme, completing field lofting through RTK and laser radar, collecting laying parameters in real time, comparing the laying parameters, and generating a correction scheme when the deviation exceeds the limit. The application can realize global optimization of a pipeline path under multi-constraint conditions and dynamic adaptation in a construction stage.