7results about How to "Improve task adaptability" patented technology

This invention discloses a method, device, and electronic device for evaluating the rationality of obstacle avoidance paths, relating to the field of UAV path planning and intelligent optimization technology. The method includes: acquiring information on the location and influence range of obstacles in the environment; constructing an obstacle avoidance path objective function, dividing the influence range into three levels to model a cost function, and dynamically adjusting weight factors based on task completion; repeatedly solving for multiple low-cost obstacle avoidance paths based on an improved whale swarm optimization algorithm to form a rational path sample set, balancing global exploration and local development, and refining the optimal solution in the later stages through pattern search; acquiring the path point sequence of the actual UAV flight path; using a Gaussian mixture model to fit the probability distribution of the offline-generated low-cost obstacle avoidance path feature vectors, extracting the common features of the actual flight paths and calculating their negative log-likelihood, and outputting the obstacle avoidance rationality level according to a preset threshold. This solves the problem of the lack of rationality evaluation for UAV obstacle avoidance paths in complex environments.

This invention discloses a method for constructing large-scale model applications based on configurable workflows and domain knowledge bases, comprising the following steps: S1, constructing a domain knowledge base, establishing a semantic graph, and generating a knowledge embedding dataset; S2, defining a configurable workflow, setting jump rules based on a semantic flow language, and binding task semantic tags; S3, configuring a Prompt adaptive generation mechanism, combining semantic tags and the knowledge embedding dataset to generate Prompt input text; S4, calling the knowledge embedding dataset, retrieving semantic fragments, and embedding them into the Prompt to form an enhanced Prompt input; S5, inputting the large language model to obtain return results and metrics, and performing jump judgment; S6, scheduling large language model service instances, dynamically selecting service interfaces based on metrics and task status; S7, processing process termination nodes, organizing output results, and recording logs and call data. This invention achieves intelligent scheduling and application construction of large-scale models based on workflows and knowledge bases.

The application discloses a kind of unmanned plane flight control method, device, unmanned plane and storage medium, the method includes: obtaining IMU data, laser radar data, determine unmanned plane current position and attitude information;Obtain the depth map data of camera, convert it into point cloud data, and construct unified environment model;Based on the dynamics constraint of environment model and unmanned plane, the feasible path from current position to target point is planned out using path planning algorithm;According to the path point sequence on feasible path, generate corresponding control instruction, drive unmanned plane to fly along feasible path autonomously;Use visual image to identify target, dynamically adjust the speed and attitude of unmanned plane according to the position deviation between unmanned plane and target point, realize accurate following to target point.The above scheme significantly improves the autonomy and robustness of unmanned plane in complex dynamic environment, so that it can complete positioning, obstacle avoidance, navigation and tracking tasks stably in variable scene.

The application discloses a multi-step instruction fine-tuning data quality improvement method based on a large model application technology, which comprises: in the data evaluation stage, scoring instruction fine-tuning data from four aspects of instruction compliance, logical coherence, context fidelity and fact accuracy; in the data correction stage, dynamically identifying the task type of the data according to the semantic characteristics of the instruction, and accordingly assigning appropriate repair priorities to different quality dimensions; performing targeted correction on the identified problems; in the illusion repair stage, detecting the illusion of the instruction fine-tuning data from the above quality dimensions, and according to the preset repair priority, performing multi-round iteration correction on the identified illusion problems; in the data labeling and screening stage, quantitatively labeling the instruction fine-tuning data from the quality score, the complexity score and the semantic intention label, and combining the threshold screening to obtain a high-quality target instruction fine-tuning data set and use it for model instruction fine-tuning training.

The application discloses a self-adaptive gravity compensation trajectory correction method based on a lead angle and belongs to the technical field of aircraft guidance control. The application constructs a dynamic mapping relation between a trajectory deviation and a gravity compensation amount by using the real-time measured lead angle information of the aircraft, adaptively adjusts a gravity compensation coefficient to correct the trajectory, reduces the compensation amount to inhibit the trajectory lifting when attacking a short-distance target, enhances the compensation to offset the falling effect when attacking a long-distance target, improves the problem that the gravity compensation is fixed and does not affect the guidance effect in the traditional method when attacking targets with different distances, reduces the dependence on the difficult-to-measure parameters such as the remaining time in other methods, and improves the striking precision and the shooting range of the aircraft on different targets.

This application discloses a hierarchical adaptive intelligent association and alignment method for multimodal flight data, comprising: acquiring multi-source airborne flight data and preprocessing it; mapping data of different modes in the preprocessed multi-source airborne flight data to a common feature space to obtain cross-modal alignment features, wherein an aircraft physical dynamic model is introduced as a constraint during the mapping process; performing local precise association and global semantic association in parallel based on the cross-modal alignment features; wherein the local precise association is used to obtain pixel-level or point-level matching results, and the global semantic association is used to obtain globally consistent semantic association results; and fusing the results of the local precise association and the global semantic association to obtain association information of corresponding granularity. This invention effectively solves the "data silo" problem caused by differences in format, spatiotemporal reference, and modality of multi-source flight data, significantly improving association accuracy, robustness, and system adaptability.

The application provides a kind of foot type robot obstacle avoidance navigation control method and system, it is related to robot control technical field, the method includes: obtaining building information model, construction stage and work task information;Component semantic analysis and traffic space extraction are carried out on the model, and multi-layer navigation prior map is generated;Initial pose estimation and real scene perception are completed using sensors to obtain online results, which are spatiotemporally registered with the prior map to identify environmental differences;Risk weight is updated according to environmental differences and construction stage information, and the prior map is refreshed;Based on the updated semantic topological constraints and task information, an initial global path is generated;Based on the path, online results, foot end state and motion posture, a comprehensive evaluation result is obtained;According to the result, local obstacle avoidance and trajectory correction are performed to obtain the final path and gait parameters and issue to the controller for execution.The application improves the environmental adaptability, path executability and long-term deployment self-maintenance capability of the robot in dynamic construction site.