126 results about "Navigation model" patented technology

The invention provides a visual language navigation and model training method, device, equipment and medium, and the method comprises the steps: obtaining a training sample, processing a to-be-executed instruction and environment information in the training sample through a core model, obtaining a navigation feature corresponding to the to-be-executed instruction, and inputting the navigation feature into an action prediction head. Obtaining predicted action information corresponding to the navigation features, and inputting the navigation features into a thinking chain generation head to obtain predicted thinking chain information corresponding to the navigation features; and then, based on the predicted action information, the reference action information in the training sample, the predicted thinking chain information and the reference thinking chain information in the training sample, training the core model, the action prediction head and the thinking chain generation head so as to take the trained core model and the trained action prediction head as a visual language navigation model. According to the method, the reasoning efficiency and the reasoning accuracy can be improved.

The invention relates to the field of data processing, and discloses a neurosurgery operation data processing method for precision medical treatment, which is used for solving the problem that mark points cannot be captured due to the fact that a navigation camera is shielded when intraoperative data is collected. The method comprises the following steps: acquiring preoperative image data, intra-operative navigation image data and spatial information of an intra-operative reference mark point, constructing a preoperative three-dimensional navigation model through the preoperative image data, obtaining a navigation coordinate system, obtaining an intra-operative image frame sequence through the intra-operative navigation image data, carrying out spatial registration, extracting intra-operative navigation information, and obtaining an intra-operative image frame sequence. The method comprises the following steps: acquiring intraoperative navigation information, judging whether occlusion occurs in an operation or not according to the intraoperative navigation information, if yes, generating missing image frame data during occlusion through a dynamic trajectory prediction model, and effectively judging the missing image frame data according to a newest instrument space pose, thereby effectively improving data integrity and coherence, improving operation data accuracy and improving operation accuracy. The operation risk is reduced.

A system for mapping road segment free spaces for use in autonomous vehicle navigation. The system includes at least one processor programmed to: receive from a first vehicle one or more location identifiers associated with a lateral region of free space adjacent to a road segment; update an autonomous vehicle road navigation model for the road segment to include a mapped representation of the lateral region of free space based on the received one or more location identifiers; and distribute the updated autonomous vehicle road navigation model to a plurality of autonomous vehicles.

The invention relates to the technical field of endoscopes, and provides a spectral image processing method for a three-dimensional endoscope. The method comprises the following steps: synchronously acquiring single-band spectral images returned by multiple sensors of the endoscope; after the single-band spectral image is corrected, a two-dimensional spectral image is generated by combining multi-level fusion of the spectral weight; constructing a three-dimensional point cloud coordinate set through parallax calculation; mapping the two-dimensional spectrum fusion image to a three-dimensional point cloud based on a back projection sub-pixel mapping algorithm to obtain a three-dimensional spectrum point cloud; and performing Poisson fusion driven inter-block splicing fusion on the three-dimensional spectrum point cloud, and outputting an interactive three-dimensional navigation model. The technical problem that image details are lost and three-dimensional reconstruction is inaccurate due to the fact that spectral image fusion precision is insufficient in an existing three-dimensional endoscope image processing method is solved, registration and reconstruction of high-precision three-dimensional spectral point clouds are achieved through multi-modal image collaborative correction and optimization fusion, and the image fusion precision is improved. And the accuracy and the real-time performance of navigation in the endoscope are improved.

The invention discloses a mobile robot autonomous search navigation method based on a visual language large model. The method comprises the steps of receiving a natural language instruction of a user, and synchronously collecting an RGB image and a laser point cloud; two-stage visual language reasoning is adopted, and target semantic analysis and existence judgment are completed in sequence; when the target does not exist, autonomous exploration is executed based on the point cloud data and is parallel to target judgment, and an exploration-judgment closed loop is formed; after determining that the target exists, switching to a visual language navigation model, generating an incremental motion instruction, and guiding the robot to approach the target; target state monitoring is continuously carried out during navigation, if the target is lost, a dynamic backtracking mechanism is triggered, the robot is driven to return to the historical state of the recently confirmed target, and navigation is restarted. The method does not need to depend on a prior map, can execute efficient and robust autonomous search and navigation tasks on edge equipment through task collaboration, model quantification and state backtracking, and is suitable for complex scenes such as industrial inspection and emergency rescue.

The invention discloses a ship navigation model deployment method and device, electronic equipment and a storage medium, and the method comprises the steps: determining a plurality of feature ship sets, and obtaining the ship data of the feature ship sets; wherein each characteristic ship set comprises a plurality of ships with the same characteristics; training an initial navigation model based on the ship data of the plurality of feature ship sets to obtain a plurality of basic navigation models; in response to a received model request of a target ship, determining a basic navigation model matched with the target ship from the plurality of basic navigation models as a target basic navigation model according to the ship data of the target ship; and deploying the target basic navigation model to the target ship, so that the target ship trains the target basic navigation model based on the ship end data to obtain a target collaborative navigation model. The method can achieve the end-cloud cooperation of the autonomous navigation model of the ship, improves the performance and adaptability of the model, and provides powerful support for the safe and efficient navigation of the ship.

The invention discloses a low-altitude unmanned aerial vehicle autonomous cruise method and system based on cloud edge basic model collaboration, and the method comprises the following steps: S1, collecting an environment RGB image through an airborne monocular camera of an unmanned aerial vehicle, and carrying out the preprocessing of the image, and obtaining a preprocessed gray image; s2, a neural scheduler based on deep reinforcement learning generates a scheduling instruction according to the environment data and the network state reasoned by the navigation model at the previous moment; s3, generating a preliminary flight instruction; s4, generating an optimized flight instruction; s5, generating a structured flight instruction; and S6, the unmanned aerial vehicle executes the preliminary flight instruction or the optimized flight instruction or the structured flight instruction. According to the invention, through dynamic on-demand cooperation and intelligent scheduling of the end-edge-cloud three-level model, resource consumption and delay are substantially reduced, and high-robustness and high-safety autonomous cruise of the unmanned aerial vehicle in a complex open environment is realized.

Systems and methods are provided for vehicle navigation. In one implementation, a navigation system for a host vehicle includes at least one processor programmed to: receive, from a camera of the host vehicle, one or more images captured from an environment of the host vehicle; analyze the one or more images to detect an indicator of an intersection; determine, based on output received from at least one sensor of the host vehicle, a stopping location of the host vehicle relative to the detected intersection; analyze the one or more images to determine an indicator of whether one or more other vehicles are in front of the host vehicle; and send the stopping location of the host vehicle and the indicator of whether one or more other vehicles are in front of the host vehicle to a server for use in updating a road navigation model.

The invention provides a power transmission line adaptive inspection method and system, and belongs to the technical field of power transmission inspection. According to the real-time image, a light-weight visual navigation model is adopted to determine and recognize a tower, and the unmanned aerial vehicle flies to the tower top of the tower in an imitated mode according to the tower shape features of the recognized tower; wherein in the simulated flight, a GPS and a visual SLAM are adopted to independently solve a pose, and an optimally estimated fusion height is output through extended Kalman filtering fusion; adopting an identification algorithm for the tower to obtain tower parameters; and determining a routing inspection waypoint plan according to the tower parameter matching database, and shooting a routing inspection image of the tower key equipment according to the routing inspection waypoint plan. According to the invention, algorithms of target identification, visual tracking, automatic obstacle avoidance, path correction, adaptive dimming, automatic focusing, multi-angle shooting and the like are integrated, and power transmission line inspection with low difficulty, high efficiency and accurate identification is realized.

The invention discloses an auxiliary operation positioning system based on AI intelligent navigation, and the system comprises an image data collection module which is used for obtaining the original image data and operation planning information of an operation part of a patient; the navigation model construction module is used for constructing a three-dimensional navigation model based on the original image data and generating a target key point and a target path through a preset AI model based on the operation planning information; the positioning and tracking module is used for acquiring the position information of the surgical instrument in real time in the surgical process and mapping the position information into the three-dimensional navigation model; and the real-time positioning correction module is used for acquiring continuous image frames acquired by the surgical instrument in real time, acquiring intraoperative motion deviation through optical flow estimation based on real-time image data, and correcting the target key point and the target path based on the intraoperative motion deviation. According to the method, the intraoperative motion deviation is obtained in real time through optical flow estimation, target positioning is corrected, and the accuracy and reliability of operation positioning can be improved through the real-time dynamic correction mechanism.

The invention discloses a robust adaptive fusion filtering astronomical attitude determination method and system, and belongs to the technical field of radio navigation. The method comprises the following steps: constructing an integrated navigation model of a starlight and inertial integrated navigation system based on Lie group description; obtaining an observation residual error and an observation residual error covariance of a corresponding mode; robust and adaptive mode parallel sub-filtering is carried out, and posterior states and covariances in the two modes are updated; updating the mode probability; and fusing the updated posterior state with the covariance to obtain fusion output of parallel sub-filtering, taking the fusion output of the parallel sub-filtering as final estimation of the starlight and inertial integrated navigation system based on Lie group description, and outputting an astronomical attitude determination result containing an azimuth angle, a pitch angle and a roll angle. The high-precision astronomical attitude determination method has strong adaptability, strong robustness and accuracy when coping with complex noise and measuring outliers, and can realize high-precision astronomical attitude determination of starlight and inertial integrated navigation in a complex dynamic environment.

Systems and methods are provided for collecting anonymized drive information. A processing device may be configured to receive outputs from one or more sensors; determine at least one motion representation for the host vehicle based on the outputs; receive at least one image representative of an environment of the host vehicle; analyze the at least one image to determine at least one road characteristic associated with a road section; assemble first road segment information relative to a first portion of the road section, wherein the first portion of the road section is separated from a starting point associated with a route traveled by the host vehicle; assemble second road segment information relative to a second portion of the road section; and cause transmission of the first road segment information and the second road segment information to a server for assembly of an autonomous vehicle road navigation model.

The invention belongs to the technical field of navigation methods, and discloses a self-error-correction navigation method, which comprises the following steps: on the basis of a basic navigation model, acquiring a model prediction track and a corresponding real track, carrying out time step alignment on the model prediction track and the corresponding real track, defining the moment when the start deviation of the track is greater than a preset threshold value as a track deviation point, and carrying out error correction on the track deviation point; the method comprises the following steps: acquiring navigation data in a preset time before and after a trajectory offset point, inputting a large language model, acquiring error analysis, converting the error analysis into structured training data, training a basic navigation model based on the structured training data, acquiring an optimized navigation model, and finally realizing navigation based on the optimized navigation model. According to the method, the deviation between the navigation action of the basic navigation model and the real track is detected, efficient error analysis is realized through the large language model, and the error analysis is applied to further training optimization of the basic navigation model, so that the navigation model has an autonomous error correction capability, and the reliability of the navigation capability is improved.

The invention provides an unmanned aerial vehicle self-adaptive route planning and dynamic obstacle avoidance method and system for distribution line inspection. According to the method, firstly, a special visual data set is constructed, a target detection network is optimized, and the detection precision is improved by fusing a double attention mechanism and a trainable weight feature fusion module; deploying a multi-sensor fusion system to realize three-dimensional positioning of a target, and inhibiting detection jitter in combination with video stream filtering; designing a deep reinforcement learning obstacle avoidance navigation model based on a TD3 framework, fusing GRU and an attention mechanism to process a sequence state, and realizing safe and efficient navigation through a multi-target reward function; a priority decision-making mechanism is established to synchronously execute feature recognition, attitude adjustment and dynamic obstacle avoidance, and model prediction control is adopted to realize closed-loop path planning. The system shows stable self-adaptive routing inspection and obstacle avoidance capabilities in a complex line environment.

A tasked robot for a deep reinforcement learning system using asymmetric self-play for robust multi-robot flocking is provided. The tasked robot includes a reinforcement learning control module and an auxiliary training module. The control module dynamically adjusts the robot's behavior to optimize decisions, featuring a target navigation model, a cluster behavior maintenance model, and a collision avoidance model. The auxiliary training module enhances environmental perception, enabling the robot to predict dynamic changes. The auxiliary training module includes a local environment grid estimation model to generate small-scale maps and a motion prediction model to forecast robot and obstacle trajectories.

A device may determine that transaction account information, associated with a transaction account associated with a user, is to be updated. The device may identify a merchant account that is configured to use the transaction account for a transaction associated with the user and a merchant, and navigate, using a navigation model, a merchant portal associated with the merchant, to access the merchant account. The device may perform, using the navigation model and login credentials for the merchant account, a login operation to access the merchant account via the merchant portal, wherein the login credentials are stored in a credential mapping that indicates an authorization to update the transaction account information to permit the transaction account information to be used in association with the merchant account. The device may update, using the navigation model, a transaction setting of the merchant account to include updated transaction account information for the transaction account.

The invention discloses a positioning and delimiting method and system for a cortex pathological focus based on an intracranial electroencephalogram signal. The method comprises the following steps: constructing an individualized three-dimensional navigation model; obtaining an original intracranial electroencephalogram signal; generating a plurality of representative segments according to the original intracranial electroencephalogram signals; analyzing the original intracranial electroencephalogram signal and the plurality of representative segments so as to obtain a network analysis core index; generating a fusion type three-dimensional brain model according to the network analysis core index and the individualized three-dimensional navigation model; generating a space-time three-dimensional electroencephalogram model and an electrophysiology functional area distribution diagram according to the fused three-dimensional brain model and the network analysis core indexes; and according to the space-time three-dimensional electroencephalogram model and the electrophysiology functional area distribution map, generating accurate electrophysiology positioning and delimiting information of the cortical pathological focus. According to the invention, accurate electrophysiological delimitation of intra-operative cortical pathological lesions (glioma, hemangioma, focal cortical dysplasia and the like) is realized.

The invention discloses an unmanned aerial vehicle low-delay navigation method based on multi-modal large model reasoning and reinforcement learning sequence optimization, and relates to the field of multi-modal large models and reinforcement learning, and the method comprises the following steps: constructing a target function fusing a supervision fine tuning loss item and a reinforcement learning loss item; performing gradient descent optimization on parameters of the multi-modal large model based on a loss function to generate a trained navigation model; inputting an environment image and a task instruction acquired by the unmanned aerial vehicle into a navigation model for reasoning, and outputting an action sequence consisting of k actions and corresponding time sequence identifiers; and storing the action sequence into an action buffer pool, sequentially taking out the head-of-queue action for execution, synchronously receiving a new action sequence, replacing the action at the corresponding position in the buffer pool according to the time sequence identifier, and maintaining the number of the actions in the buffer pool to be k. And a future action sequence is generated in real time through a multi-mode large model, so that the unmanned aerial vehicle realizes continuous, low-delay, stable and reliable sequence flight.

The invention discloses a decomposition type value evaluation multi-agent collaborative navigation method based on an attention mechanism. The method comprises the following steps: modeling a collaborative navigation task of multiple agents into a decentralized partially observable Markov decision process; designing local attention modules of a precoding module and an attention coding stage to extract and code environment feature information of multiple agents, and sharing the local attention modules by a value network and a strategy network; designing a multi-agent strategy-value method under decomposition type value evaluation; constructing a static target navigation environment, and guiding the mobile intelligent agent to actively search for an optimal position; constructing a cooperative hunting environment; a multi-agent cooperative navigation model is constructed, an optimal model is obtained through training and testing, and non-explicit and complete perception navigation of multi-agent cooperation with the reliability distribution problem is achieved. The collaborative navigation method provided by the invention has the advantages of faster convergence, higher success rate, shorter time and less collision, and shows good generalization ability and robustness.

The application discloses a kind of highway survey radar measurement systems based on electromagnetic wave, through quantum magnetic force gradient tensor analysis and space coordinate calibration, synthetic aperture radar echo feature extraction module cooperation, using quantum magnetic force navigation model and synthetic aperture radar algorithm, realize multi-source data space-time registration and weighted fusion, construct highway geology layered model and inversion geological parameter, after grid processing of fusion data, generate three-dimensional visual model.The system overcomes the defects of traditional technology multi-source data fusion deficiency and low precision of geological parameter inversion, improves the precision, efficiency and reliability of highway survey, provides accurate geological information and visual survey results for highway engineering design and construction, has important application value in the field of highway survey.

The invention provides a task robot for a deep reinforcement learning system. The deep reinforcement learning system can utilize an asymmetric self-game algorithm to realize robust multi-robot clustering. The task robot comprises a reinforcement learning control module and an auxiliary training module. The control module can dynamically adjust behaviors of the robot to optimize decisions, and is characterized by comprising a target navigation model, a cluster behavior maintenance model and a collision avoidance model. The auxiliary training module enhances the environment perception capability, so that the robot can predict dynamic changes. The auxiliary training module comprises a local environment grid estimation model used for generating a small-scale map and a motion prediction model used for predicting tracks of a robot and an obstacle.

The invention discloses a distillation evolution optical core navigation intelligent cutting, sewing and ironing robot system and equipment. The system comprises a cutting module, a sewing module, an ironing module, a model module, a navigation module and a driving module. Wherein the model module is mainly responsible for model generation, memory, storage and modification of cutting, sewing, ironing and driving of various materials; the navigation module is mainly used for generating a navigation model and instruction information including positioning, planning, control and decision making as well as reasoning, establishment and logic generation and control of a workflow; the driving module is mainly responsible for generating a driving model and providing driving force for each module of the system; each module can independently form a subsystem, and the modules can be combined or fused to form various new robot systems and devices. The method has the advantages that the automation, intelligence and precision levels of cutting, sewing and ironing are improved, safe guidance is achieved, energy consumption is reduced, development and innovation of an unmanned system are assisted, the bottleneck of space calculation is broken through, and the efficiency of space communication is improved.

The application discloses a nano unmanned aerial vehicle image navigation method and system based on an ultra-low-power edge chip, relates to the cross technical field of unmanned aerial vehicle autonomous navigation and edge computing, and comprises the following steps: constructing a lightweight feature extraction backbone network based on a deep separable convolution, generating a basic navigation model suitable for an edge chip based on early feature fusion and end-to-end decision calibration; obtaining a quantization model suitable for a chip instruction set through full-link fixed-pointing, operator smelting, and integer and inverse quantization collaborative optimization; combining a multi-level storage structure of the chip, demarcating a target memory space, establishing a memory alignment and hardware protection mechanism, and building a stable bottom-layer running environment; and achieving full-airborne high-frame-rate real-time navigation operation of the nano unmanned aerial vehicle through tensor block constraint planning and asynchronous inference pipeline design. Through optimization of a network architecture, power scheduling and a memory mechanism, the application realizes airborne independent low-power navigation, and solves the problems of high navigation power consumption of the unmanned aerial vehicle and difficulty in adaptation.

The application relates to the technical field of unmanned aerial vehicle visual language navigation, in particular to a visual language navigation model training data generation method, an electronic device and a storage medium. A processor can automatically generate a plurality of target trajectories and tracking trajectories corresponding to the target trajectories based on three-dimensional scene map data; based on the three-dimensional scene map data, the target trajectories and the tracking trajectories corresponding to the target trajectories, a simulation scene in which an unmanned aerial vehicle tracks a tracking target in a three-dimensional scene is constructed, and original perception data of the unmanned aerial vehicle is collected based on the simulation scene; and a visual language navigation model training data set carrying labels is generated based on the original perception data of the unmanned aerial vehicle. Therefore, the embodiment of the application can automatically generate target trajectories, tracking trajectories and a visual language navigation model training data set carrying labels, which is beneficial to batch generation of large-scale and high-quality visual language navigation training data sets and reduces the cost of manual labeling.

The invention discloses a ship energy optimization method considering energy storage life and flexible navigation. The ship energy optimization method comprises the following steps: step 1, constructing a ship flexible navigation model based on marine environment data and an electronic chart system; establishing an energy storage attenuation model of the battery energy storage system, wherein the model comprises quantitative calculation and linear processing of calendar life attenuation and cycle life attenuation; constructing a photovoltaic output model dynamically associated with the ship position, wherein the model comprises solar irradiation calculation, photovoltaic panel efficiency correction and environmental factor influence evaluation; 2, a flexible navigation model, an energy storage attenuation model and a photovoltaic output model are integrated, an SIES multi-objective optimization model is established, and the model aims at minimizing the total cost of the system; and 3, solving the SIES multi-objective optimization model by adopting a mixed integer linear programming algorithm to obtain an optimal route planning, energy storage configuration and energy scheduling scheme.

The application provides an electric vehicle fast charging navigation method and device, electronic equipment and storage medium, constructs a coupling network model of a traffic network-power grid in a region where a target electric vehicle user is located, a coupling edge is composed of a road node containing a charging station and a power grid node of the charging station, and the weight of the road node is defined as an average travel time, the weight of the power grid node is defined as a voltage variation, and both of them jointly determine a charging cost of the charging station; a charging navigation model is constructed with a target function of a comprehensive cost of the target electric vehicle user and constraint conditions of minimum residual allowable capacity of a battery and queue tolerance time; and through processing of the charging navigation model, a charging station with the lowest comprehensive cost is taken as a target charging station, and an optimal path to the target charging station is output. The application can realize best charging station recommendation and optimal path planning of the electric vehicle user by using multi-source information such as the traffic network-power grid, and promote benign collaborative operation of the traffic network-power grid coupling network.

The invention relates to the technical field of ship navigation model tests, and discloses a giant estuary ship navigation capability prediction model test method fused with machine learning, which comprises the following steps: initializing and defining a scene vector, training an agent model, generating a population, and setting archives and parameters; executing dynamic fitness evaluation to obtain a fitness score; generating filial generations based on score selection and genetic operators and ensuring physical effectiveness; selecting a scene with the highest score to obtain a predicted risk; when the error is greater than a threshold value, storing the candidate scene; if yes, outputting the file; and if not, replacing the population and returning evaluation. According to the method, a feedback closed loop is constructed through failure judgment and updating of a failure mode file, after a failure scene verified by a high-fidelity simulator is stored in the file, the areas are avoided when the novel fitness is calculated through dynamic fitness evaluation, the search direction is automatically adjusted, and automatic and self-adaptive exploration of a high-risk scene is achieved.

The invention discloses an intelligent control method and system for an ophthalmic surgical robot, a prediction model and a cooperative control method, and belongs to the technical field of medical robots. The system comprises a 3D display system, an OCT information system, a micro-touch system, an information comprehensive processing system, a control system, a mechanical arm and special inner boundary membrane tweezers. The core of the method is that an augmented reality surgical navigation model is constructed through preoperative and intraoperative OCT data fusion; the micro-operation force is monitored in real time through the novel inner boundary membrane tweezers integrated with the FBG sensor; the visual sense information and the force sense information are fused through an information comprehensive processing system, and a personalized control strategy is generated based on a pre-trained deep learning decision model; and finally, the control system adopts a self-adaptive impedance control algorithm to drive the mechanical arm to execute accurate and smooth force-position closed-loop operation. The problems that in the ophthalmologic operation, the visual depth is lost, the tactile feedback is deficient, and the operation precision depends on the experience of doctors are solved, and the safety, the precision and the standardization level of the retina inbound membrane stripping operation can be remarkably improved.

The invention discloses a navigation method fusing conditional value-at-risk and model prediction path integration, and relates to the technical field of navigation, the method comprises the following steps: modeling an underwater robot navigation process to obtain a navigation model; sensing information is dynamically collected based on the navigation model; wherein the sensing information comprises environment parameters and state parameters of the underwater robot; generating an initial control instruction according to the perception information by using the conditional value-at-risk; controlling the underwater robot according to the initial control instruction; when the distance between the underwater robot and the obstacle is smaller than a preset threshold value, performing safety filtering by using model prediction path integration, and further generating a safety correction instruction; and controlling the underwater robot according to the safety correction instruction. According to the method, the conditional value-at-risk and the model prediction path integral are fused, so that the robustness in the navigation process of the underwater robot can be improved, and the navigation precision is improved.

The application provides a kind of electromagnetic space and geographical space joint navigation optimization method for single mobile user, receives the current state and terminal point information of mobile user agent, current state includes the current position of agent, path obstacle, measured electromagnetic information and walkable path;Current state containing measured electromagnetic information is input to deep neural network, so that the output predicted moving direction tends to move in the direction of sufficient electromagnetic intensity to maintain communication.The application not only considers the navigation of geographical space, but also further considers the joint navigation of electromagnetic and geographical space, so that the mobile user in space can reach the destination faster while ensuring the communication quality.The application combines the characteristics of reinforcement learning during the training of the navigation model based on deep network, compares the real electromagnetic map after the movement of the agent with the predicted map for supervised learning, so that deep reinforcement learning and machine learning are organically combined together.