81 results about "Autonomous Navigation System" patented technology

The invention discloses an unmanned aerial vehicle autonomous navigation system based on a hierarchical reinforcement learning strategy. The unmanned aerial vehicle autonomous navigation system is suitable for a three-dimensional flight task in an unknown environment. The system comprises a state sensing module, a hierarchical strategy network module, a control execution module, a data classification module and a data playback module. The state sensing module extracts obstacle position information based on a deep neural network, and fuses the target, the obstacle position and the flight state to generate a state vector and a time sequence. The hierarchical strategy network adopts a high-layer DQN to generate a navigation intention, and a low-layer LSTM and PPO are combined to output a continuous control action; the control execution module adjusts the attitude of the unmanned aerial vehicle according to the instruction and performs closed-loop correction. The system introduces a double dynamic memory mechanism (DDM), improves strategy training efficiency and stability through experience classification and proportional sampling, and adopts a multi-target award function guide strategy to optimize convergence among task completion, obstacle avoidance safety and flight rationality. The system has good environmental adaptability and generalization ability, and is suitable for autonomous navigation tasks in complex scenes.

The invention relates to a robot path planning method based on multi-sensor fusion and free space topology composition, and belongs to the technical field of robot autonomous navigation. The method comprises the following steps: firstly, fusing laser radar and millimeter wave radar data, constructing multi-modal point cloud data, extracting a three-dimensional obstacle boundary by combining depth and normal vector mutation features, and generating a three-dimensional obstacle map and a free region tree; according to the constructed space model, factors such as energy consumption, dynamic obstacles, path smoothness and the like are integrated, target selection weights are dynamically regulated and controlled, and self-adaptive screening of intermediate navigation targets is achieved. And based on the intermediate navigation target, generating a safe trajectory satisfying dynamic constraints by adopting geometric-dynamic dual-mode fusion modeling, and enhancing the feasibility and robustness of the trajectory through trajectory envelope optimization and pre-execution fault-tolerant control. The method can be widely applied to autonomous navigation systems such as mobile robots and unmanned vehicles, and has good environment adaptability and path execution stability.

The invention relates to the technical field of autonomous navigation, in particular to an AUV (Autonomous Underwater Vehicle) multi-sensor fusion autonomous navigation system based on ant colony-particle swarm collaboration.The AUV multi-sensor fusion autonomous navigation system is characterized in that a sensor weight decision module constructs a pheromone matrix based on an ant colony algorithm, retrieves a mapping relation between historical errors and pheromone concentration in real time and dynamically generates a sensor priority sequence; adaptive weight distribution in a data conflict scene is realized; the dynamic path correction module is combined with a particle swarm algorithm and pheromone concentration gradient constraint to balance global path optimality and real-time obstacle avoidance requirements in optical camera obstacle deviation correction; and the closed-loop coupling module is used for converting sonar and geomagnetic residual errors into pheromone concentration correction values and confidence coefficient regulation factors by fusing residual error compensation and path tracking error feedback, so that an error traceability-weight calibration-path optimization closed loop is formed, sensor drift is inhibited, and the long-endurance navigation precision is improved.

The invention discloses a real-time ocean simulation method and system based on sensing data, and relates to the technical field of ocean information engineering and intelligent sailing. A graph neural network is combined with a multi-head attention mechanism to carry out data fusion, importance of different sensors and data time steps is dynamically weighted, and real-time ocean simulation is realized. According to the method, the quality difference and missing conditions of heterogeneous data are adaptively processed, the deviation and noise influence in the fusion process are reduced, meanwhile, physical consistency constraints such as mass conservation and boundary non-penetration are introduced, the physical rationality and smoothness of an output ocean current field are guaranteed, non-physical artifacts are avoided, and the reliability and practicability of a result are enhanced; a streaming updating mechanism allows a graph structure and model parameters to be rapidly and incrementally updated when new data arrives, data processing and assimilation delay are greatly reduced, and nearly real-time ocean current forecasting is realized, which is crucial to a ship autonomous navigation system, and accurate ocean current information can be timely provided to support dynamic path planning and risk avoiding decision making.

The invention discloses a composite tracking method for an aviation dual-frequency satellite communication phased-array antenna. According to the method, the dual-frequency satellite communication phased-array antenna can be always aligned with high-frequency and low-frequency satellites in the flight process by utilizing navigation data calculated by an autonomous navigation system. The method does not depend on external data, is an independent system, and is easy to maintain and convenient to use. In the implementation process, the high-frequency antenna master control module and the low-frequency antenna drive and control integrated module share navigation data, the working states of opposite ends can be sensed mutually, pointing adjustment can be made in time, high-frequency links and low-frequency links can be backed up mutually, and the reliability of a communication system is improved.

The invention discloses a multi-source autonomous navigation system and method with embedded carrier dynamic characteristics, and the method comprises the steps: building a carrier dynamic model in a dynamic modeling stage; in a filter prediction stage, a Kalman filter driven by dynamics is established, a carrier execution mechanism control signal is periodically read as filter input, and a state vector and a covariance matrix are predicted in each sampling period; a filter updating stage: calculating information of each sensor and a carrier maneuvering strength index, constructing a dynamics consistency judgment quantity, adaptively adjusting an observation variance matrix, updating a filtering state according to observation information of the sensor, and realizing multi-source information fusion; and a fault detection and isolation stage: constructing a multi-source consistency index, and judging the health state of the navigation system. According to the invention, the problem that the performance of a traditional navigation system is reduced when a carrier platform violently moves or is disturbed by external force of the environment is solved, and robust high-precision navigation under GNSS denial, sensor failure or complex dynamic conditions is realized.

The invention discloses an unmanned aerial vehicle autonomous navigation system based on a rasterized world model, and relates to the technical field of unmanned aerial vehicle control. According to the system, observation information of an unmanned aerial vehicle is input into a trained rasterized world model through an observation acquisition module; through cooperative work of a sequence model, a multi-modal self-encoder, a hidden space dynamics predictor, a multi-modal information prediction head and a grid predictor in a rasterized world model, a cyclic variable containing historical information, a prediction hidden state vector and a prediction local grid map are provided for an agent model. And action decision making is carried out based on the data through an intelligent agent model so as to accurately output the actions of the unmanned aerial vehicle. According to the system, by introducing the rasterized world model, multi-modal observation information can be effectively fused, and a three-dimensional structure of a local environment is predicted in real time, so that the spatial perception capability of an intelligent agent model to a complex environment is enhanced, and the decision-making performance of the intelligent agent model is improved.

The invention relates to a real-time non-graphical robot autonomous navigation system, and belongs to the field of robots. The system comprises a main control module, an environment sensing module and a hybrid navigation module, the main control module is responsible for man-machine interaction, task flow control and cooperation of other modules, and storing and managing a visual semantic database; the environment sensing module is responsible for controlling the robot to collect image data and depth data under the condition of no prior map, identifying objects in a scene, extracting semantic information of the objects, and integrating identification results into a visual semantic database; and the hybrid navigation module analyzes the instruction sent by the main control module by using a large language model, matches visual memory, determines a navigation target, and completes autonomous navigation of the robot through a hybrid navigation strategy. According to the method, the problem of low navigation task accuracy caused by the sparsity of global information in map-free navigation is solved.

The invention belongs to the technical field of intelligent navigation of agricultural machinery, and particularly relates to an autonomous navigation system of field unmanned transportation equipment after potatoes (such as potatoes and sweet potatoes) are harvested, in particular to a potato operation line identification method based on binocular camera multi-modal information dynamic weighting. Aiming at a complex field environment (soil is loosened and turned over, scattered potato blocks / weeds are mixed, and an original ridge-shaped structure is locally damaged) after the operation of a harvester, a ridge line track required by the driving of a transport vehicle is difficult to stably identify in a scene of strong light overexposure, weak light noisy points and shadow alternation by a traditional visual method. According to the method, an IntelRealSenseD456 active stereoscopic vision binocular camera is carried, RGB images and depth information are fused in real time, and multi-modal data are dynamically weighted based on illumination intensity, so that the robust perception capability of the unmanned transport vehicle on the geometric boundary of the potato ridge is improved, the vehicle is ensured to accurately run along a harvested field ridge operation line, and the situation that the potato blocks are rolled and scattered or deviated from a path is avoided.

The invention discloses an autonomous navigation system and method for an electric power inspection unmanned aerial vehicle, and belongs to the technical field of artificial intelligence and unmanned aerial vehicle control. The system comprises a multi-mode sensing module, a VLM navigation module and an electric power scene adaptation and flight control instruction conversion module, an instruction semantic adaptation unit disassembles a natural language inspection instruction into triple features, and an electric power equipment knowledge base is called to map an equipment type into target equipment features which can be identified by an SPF; the multi-modal feature fusion unit carries out feature fusion to obtain fusion features; the decision output unit combines the fusion features and the triple features and then inputs the combined features and the triple features into a VLM navigation decision model based on SPF frame 2D space grounding logic, and an initial 2D waypoint is output; and the flight control adaptation unit performs flight control instruction conversion based on the 3D displacement vector corresponding to the 2D waypoint. The natural language inspection instruction is mapped into target equipment characteristics which can be identified by the SPF, so that accurate conversion from the natural language inspection instruction to the flight control instruction can be realized.

Autonomous Navigation System for Indoor Delivery Robots

The invention discloses a laser radar positioning and mapping method based on visual review assistance, and particularly relates to the technical field of deep space exploration, and the method comprises the steps: firstly, carrying out positioning and mapping by using a tightly coupled laser-inertial odometer framework FAST-L IO2, then accurately calculating self attitude information by using a visual vector, and finally, carrying out deep space exploration. And the information is used as a reliable attitude reference to be introduced into the correction step of the laser radar SLAM. The attitude calculated by the visual vector is used as an observation constraint, and the pose output by the laser radar SLAM is corrected in real time, so that the attitude drift caused by vibration or point cloud mismatch is reduced, and the overall precision of pose estimation is improved. According to the method, the advantage of high precision of the visual vector in attitude determination is utilized, the problem of insufficient robustness of a positioning and mapping technology based on the laser radar is solved, and an innovative solution is provided for a satellite-borne autonomous navigation system.

The application relates to the technical field of man-machine cooperation, in particular to a man-machine cooperative autonomous navigation system and method based on multi-constraint optimization, which comprises a navigation action generation module, an artificial interaction module, a confidence evaluation module and a man-machine cooperative optimization module, wherein the man-machine cooperative optimization module is used for constructing an action fusion optimization target according to the machine navigation action, the artificial navigation action, the safety confidence mean value and the action executed at the previous moment, and solving a fusion execution action in a feasible region containing linear velocity constraints, angular velocity constraints, curvature constraints and safety velocity constraints; the application further comprises a Tube MPC correction module and a weighted experience priority sampling module. In the optimization solver, velocity constraints, curvature constraints and dynamic safety velocity constraints based on collision time are explicitly introduced, the physical limits of the machine and the environment risks are simultaneously brought into the optimization feasible region, and the solved action has sufficient safety under the premise of physical realizability.

The application relates to a ship autonomous navigation intention interpretation method and device, which comprises the following steps: obtaining a composite data set based on dynamic environment data of the ship and an autonomous future trajectory output by an autonomous navigation system; performing state decoding and motion mode identification on a state vector in the composite data set to obtain a segmentation point set; performing navigation scene type judgment on the state vector of the ship and state vectors of other ships in all dynamic environment data to obtain a segmentation point set; segmenting the state vector sequence based on the union set of the segmentation point set and the segmentation point set to obtain a minimum intention unit sequence; performing unit merging and intention recognition on each minimum intention unit to obtain an autonomous navigation intention interpretation result; the application combines the dynamic environment data with the autonomous future trajectory, and performs multi-level understanding on the autonomous future trajectory by determining the segmentation point at which the motion mode changes and the segmentation point at which the navigation scene changes, thereby improving the effective trust of the crew in the system.

The invention provides a multi-IMU end-to-end navigation method based on a mask fusion mechanism, and the method comprises the steps: constructing a deep neural network architecture named MultiImuNet, comprising a plurality of independent bidirectional LSTM feature extraction branches, a feature fusion module based on learnable masks and a position regression module, and designing a special feature encoder for each IMU, according to the method, multiple IMU time sequence data are input in a sliding window mode, relative displacement estimation is output from end to end, external auxiliary signals are not needed, high precision, high robustness and good expandability are achieved, and the method has the advantages of being high in robustness and good in expandability. The method is suitable for an autonomous navigation system in a GNSS-free environment.

An introduced autonomous aerial vehicle can include multiple cameras for capturing images of a surrounding physical environment that are utilized for motion planning by an autonomous navigation system. In some embodiments, the cameras can be integrated into one or more rotor assemblies that house powered rotors to free up space within the body of the aerial vehicle. In an example embodiment, an aerial vehicle includes multiple upward-facing cameras and multiple downward-facing cameras with overlapping fields of view to enable stereoscopic computer vision in a plurality of directions around the aerial vehicle. Similar camera arrangements can also be implemented in fixed-wing aerial vehicles.

The invention provides a state estimation and error identification method and device for an autonomous navigation system, and the method comprises the following steps: obtaining an observability evaluation index of the system by using an observability criterion; according to the observability evaluation index of the system, obtaining the accurate compensation degree of the measurement error of the optical sensor, and determining the achievable state estimation accuracy of the autonomous navigation system; constructing an observability optimization function; analyzing a coupling relationship between a measurement error and a system state based on an observability optimization function, and establishing an error response model; according to the error response model, constructing a system state parameter joint optimization model through the observability evaluation index of the navigation system; and optimizing the system state parameter joint optimization model and realizing on-orbit identification of errors and system states. The accuracy of observability judgment of the navigation system is improved, observation information and calculation resource allocation can be autonomously optimized, and the problem of state estimation under the weak observation condition is solved.

The application discloses a kind of based on point cloud fusion's tuyere scene autonomous navigation system and method, belong to blast furnace production process control technical field.It includes: robot trolley, robot trolley is sequentially provided with 16 line laser radar, IMU inertial measuring instrument, depth camera, holder, temperature and humidity sensor, gas sensor, sound intensity sensor and database server.The present application solves the problem that existing artificial handheld temperature measuring gun is used to measure temperature at tuyere, which causes greater damage to the human body by the environment, the point cloud fusion technology is used to better obtain the tuyere and obstacle information, and the odometer fusion algorithm is used to give the robot trolley more accurate coarse positioning;Through IMU inertial measuring instrument, the rotation angle, horizontal angle information of robot trolley is collected, high-precision positioning work is completed, and accurate temperature measurement of temperature measurement point is ensured;Using positioning autonomous navigation scheme ensures that the robot trolley accurately reaches the temperature measurement point in high temperature and high pressure environment, and ensures high frequency work every day.

The invention relates to an underwater multi-source autonomous navigation digital twinning and semi-physical simulation test system and method, which are suitable for development and algorithm verification of a multi-source autonomous navigation system of underwater unmanned platforms such as an unmanned underwater vehicle. According to the system, a semi-physical simulation architecture is adopted, and a physical inertial measurement unit (IMU), a three-axis turntable and a computer simulation module are combined; signal simulation, motion scene simulation, sensor error and environmental noise injection, multi-source navigation information fusion and precision evaluation of sensors such as an inertial navigation system, a Doppler log, an ultra-short baseline positioning system and a satellite navigation system are realized. The platform has the simulation capability of two typical underwater scenes of mooring and motion, supports verification of multiple combined navigation modes, can simulate complex underwater acoustic environments such as Gaussian and non-Gaussian noise and abnormal outliers, and provides a highly reliable and high-precision semi-physical simulation verification environment for development and completeness evaluation of an underwater multi-source autonomous navigation algorithm.

The invention provides an observability analysis method and device for deep space exploration optical autonomous navigation, and the method comprises the following steps: constructing an orbit dynamics model and an optical observation model according to the measurement information obtained by a deep space probe; constructing an optical autonomous navigation system model according to the orbit dynamics model and the optical observation model; observability parameters of the deep space exploration optical autonomous navigation system are given according to the optical autonomous navigation system model; constructing an observability characterization model of the autonomous navigation system; carrying out dimensionality reduction characterization on observability of the deep space exploration optical autonomous navigation system; therefore, an observability criterion is obtained; and realizing the observability autonomous determination of the deep space exploration optical autonomous navigation system according to the observability criterion. By analyzing key factors of complete representation of the observability of the system, judgment conditions of the observability of the system are deduced, and the observability quantitative expression method which is universal, simple and easy to autonomously realize on a satellite is provided.

The present invention relates to the field of navigation and guidance, and relates to an intelligent multi-source autonomous navigation system. The system comprises: a multi-source perception and measurement module, which integrates a plurality of navigation sensors and outputs observations for navigation positioning; and an information processing module, which receives the observations from the plurality of navigation sensors, and uses corresponding navigation solution algorithms to obtain navigation positioning information of a carrier, wherein a combination of a single navigation sensor and a navigation solution algorithm corresponding thereto is denoted as a single navigation information source; and according to navigation task requirements, the information processing module evaluates the navigation positioning information output by each single navigation information source to obtain an optimal navigation information source configuration scheme under a current task and a current scenario, screens out an optimal information fusion algorithm applicable to the optimal navigation information source configuration from a stored navigation algorithm library, and uses the optimal information fusion algorithm to perform multi-source fusion on outputs of the optimal navigation information source, so as to obtain an optimized navigation positioning result.

The invention discloses a ship energy efficiency real-time optimization and autonomous navigation system based on deep reinforcement learning, which relates to the technical field of intelligent ships and maritime engineering, and comprises a tide sensing module, an energy efficiency identification module, a coupling threshold control module, a reference weight balance module and a phase traction module, the method comprises the following steps: collecting continuous flow velocity data to construct a time synchronization perception chain, analyzing sudden change of a tidal current direction and shear gradient change, generating a high-risk fingerprint zone and determining a tail anchor point; dynamic linkage control over the ship energy efficiency and the hydrodynamic state is achieved through deep reinforcement learning, real-time response to sudden change of the tidal current is achieved through a tidal current sensing and energy efficiency self-adaption mechanism, energy efficiency and structural safety are coordinated through a coupling threshold control and phase traction mechanism, vibration accumulation and energy efficiency locking are prevented, and the ship energy efficiency and hydrodynamic state dynamic linkage control method is suitable for ship energy efficiency and hydrodynamic state monitoring. Track balance and energy efficiency optimization are realized, and the navigation safety and the energy utilization rate are improved.

This invention relates to a formation trajectory tracking control algorithm for autonomous underwater robots based on a preset time control theory. The algorithm includes: modeling the autonomous underwater robot; designing an environmental disturbance observer based on the preset time control theory; designing a formation control scheme; and optimizing the formation communication scheme to reduce communication frequency. This invention introduces a preset time control theory, avoiding the drawback of existing predefined time control schemes where the system stability time cannot be unified. It utilizes an adaptive algorithm to estimate the environmental disturbance boundary, increasing the applicability of the designed scheme. An event-triggered mechanism is introduced to optimize the formation communication structure, reducing the number of formation communications, alleviating the burden of formation communication, and improving the formation control performance of the autonomous underwater robot. This achieves fast and stable formation trajectory tracking control, and can be applied to the autonomous navigation system of autonomous underwater robots.

The invention discloses a vision-driven inland waterway navigation environment reconstruction understanding method and a vision-driven inland waterway navigation environment reconstruction understanding system. The method comprises the following steps: acquiring an inland waterway image; performing feature extraction on the inland waterway image by using a neural network constructed based on ResNet and a feature pyramid to obtain a waterway multi-scale feature map; obtaining a starting point thermodynamic diagram according to the channel multi-scale feature map; according to the starting point thermodynamic diagram and the channel multi-scale feature map, a shoreline instance is generated based on a conditional convolution dynamic generation mechanism; and according to the shoreline instance and the channel multi-scale feature map, predicting the shape of the shoreline based on line grid expected value solution, and obtaining an inland waterway boundary topological graph. According to the technical scheme provided by the invention, the number of shorelines does not need to be preset through the starting point thermodynamic diagram, the line grid expected position calculation and the like, the detection error is reduced, the shoreline positioning precision and the curve curvature reconstruction error are improved, the processing speed is improved, the real-time navigation decision-making requirement is met, and an environment reconstruction result is provided for an autonomous navigation system.

The invention provides an autonomous navigation path planning capability evaluation method based on an analytic hierarchy process, belongs to the technical field of autonomous navigation, and aims to solve the problems that key factors which are determined by an existing autonomous navigation system and influence the path planning capability are not scientific, quantitative calculation of evaluation indexes reflecting the path planning capability is inaccurate, and the evaluation efficiency is low. The method comprises the following steps: acquiring ideal data of a target path; executing a target path task through the autonomous navigation system, and recording actual navigation data; performing structured processing on original path data in the actual navigation data; performing comparative analysis on actual navigation data and ideal data of the target path, and constructing a multi-dimensional evaluation index in combination with a path gradient matrix and a path roughness matrix; and constructing a hierarchical structure model, and performing combined weighting on the constructed multi-dimensional evaluation indexes by adopting an analytic hierarchy process to obtain a path planning capability evaluation result of the autonomous navigation system.

[Problem] To provide a channel marker identification device that can increase the identification accuracy for marker contents. [Solution] A channel marker identification device provided with: a first acquisition unit for acquiring a first image that includes a buoy from a camera installed in a ship; a first identification unit for identifying the position of the buoy in the first image; a second acquisition unit for acquiring a second image having a higher resolution than that of the first image and corresponding to a partial area of the first image which includes the position of the buoy; and a second identification unit for identifying marker contents of the buoy from the second image.

This invention discloses a highly adaptable autonomous navigation method for fruit and tea garden equipment based on low-precision maps, belonging to the field of intelligent navigation technology for agricultural machinery. It deploys an autonomous navigation system, acquiring global images of the fruit and tea garden through drone aerial photography, and manually setting key navigation points. Semantic segmentation algorithms are used to detect roads, fitting straight lines to obtain corner points, and automatically generating secondary navigation points connecting the key points. Latitude and longitude coordinates are obtained through automatic conversion between pixel coordinates, projected coordinates, and global geographic coordinates. This invention acquires global low-precision images through drone aerial photography and requires only a few manually set key navigation points. By combining semantic segmentation, automatically generated secondary navigation points, and global geographic coordinates, a usable navigation network can be constructed. This allows for rapid application to different fruit and tea garden environments, avoiding the need for extensive repetitive mapping work for a single garden, and significantly improving the versatility and deployment efficiency of the navigation system across different fruit and tea gardens.

The invention relates to the technical field of inspection robots, and particularly discloses an inspection robot autonomous navigation system based on deep learning, which comprises an environment sensing module, a dynamic map construction module, a path planning module, an obstacle avoidance control module and an execution feedback module. According to the scheme, a convolutional neural network and an attention mechanism are introduced, deep feature extraction and dynamic target processing are performed on multi-modal data of an inspection environment, and high-precision global dynamic map construction is realized; equipment importance and abnormal position detection are introduced to form a path priority, an important target is preferentially accessed, and the dynamic adaptability of an inspection path is improved according to a global dynamic map and a path risk value updated in real time, so that the inspection robot can efficiently complete a task in a complex dynamic environment; real-time obstacle avoidance and local path re-planning in a dynamic environment are realized, the future collision time is predicted by combining relative motion modeling with a kinematic model, the method has strong local path adaptability, and the stability of robot inspection control is improved.

The invention provides a material carrying vehicle based on multi-mode perception and an autonomous navigation system thereof. Belongs to the technical field of computer vision and intelligent factory logistics. The system comprises a data set construction pre-division module which is used for constructing a factory material carrying detection data set containing trays, obstacles, personnel and ground road sign targets, and dividing the data set into a training set, a verification set and a test set; the monitoring model construction module adopts YOLO11 as a reference model, integrates CBAM and GAM attention mechanism modules in a Backbone module, integrates CBAM, GAM, CA and ECA attention mechanism modules in a Head module, and constructs an intelligent material handling detection model; the four modules are connected in series to form an optimized link from position coding to feature fine adjustment to channel enhancement to global association, the detection capacity of tiny marks and obstacles is remarkably improved, core technical support is provided for high-precision navigation of the material carrying vehicle, and the real-time requirement of edge deployment is met through lightweight design while the fineness is improved.

An introduced autonomous aerial vehicle can include multiple cameras for capturing images of a surrounding physical environment that are utilized for motion planning by an autonomous navigation system. In some embodiments, the cameras can be integrated into one or more rotor assemblies that house powered rotors to free up space within the body of the aerial vehicle. In an example embodiment, an aerial vehicle includes multiple upward-facing cameras and multiple downward-facing cameras with overlapping fields of view to enable stereoscopic computer vision in a plurality of directions around the aerial vehicle. Similar camera arrangements can also be implemented in fixed-wing aerial vehicles.