56 results about "Human–machine system" patented technology

The invention relates to the technical field of unmanned aerial vehicles, and discloses an intelligent logistics unmanned aerial vehicle and system, and the system comprises a space-time synchronization module which is used for obtaining an original data flow of a multi-source sensor on an unmanned aerial vehicle body, and carrying out the space-time synchronization processing of the data flow of the multi-source sensor; the quality evaluation module is used for carrying out quality evaluation on the sensor data subjected to time-space synchronization; the fusion positioning module adopts a multi-sensor fusion positioning method to obtain the high-precision three-dimensional position, attitude and speed state of the unmanned aerial vehicle; the semantic map module is used for constructing a three-dimensional semantic map comprising a static obstacle, a dynamic target and a landing area; the path planning module is used for performing intelligent path planning under a multi-constraint condition; the control execution module performs path execution and real-time adjustment; according to the invention, the technical problems of insufficient unmanned aerial vehicle positioning precision, limited environment perception capability and low path planning intelligence level in a complex urban environment are solved.

The invention provides a multi-unmanned aerial vehicle airborne cooperative control method, electronic equipment and a readable storage medium, the multi-unmanned aerial vehicle airborne cooperative control method comprises the following steps: S1, dispersing control logic to each unmanned aerial vehicle through a distributed control architecture, and establishing communication connection between the unmanned aerial vehicles; s2, adopting a real-time communication mechanism to realize information exchange and collaborative decision-making among the unmanned aerial vehicles on the communication connection among the unmanned aerial vehicles; s3, according to the task demand, the information exchange result and the collaborative decision result, generating a collaborative control scheme and executing the task; s4, adjusting the structure of the unmanned aerial vehicle system through a dynamic reconstruction mechanism and the distributed execution task; and S5, state monitoring is carried out on the adjusted unmanned aerial vehicle system structure, and abnormal conditions are processed. According to the invention, the distributed control architecture and real-time communication capability of the ROS2 are utilized to realize efficient, reliable and extensible multi-unmanned aerial vehicle cooperative control.

A system and method for strategic airspace deconfliction is disclosed, centered on a Cooperative Multi-Agent Platform (CMAP) embedded within an Automated Data Service Provider (ADSP) operating within a federated Unmanned Aircraft Systems (UAS) Traffic Management (UTM) network. The system's inventive feature resides in a specific technical architecture that synergistically integrates a real-time safety constraint into a strategic negotiation engine. A Regulatory Compliance Monitor (RCM) generates a real-time Conflict Risk Score that is incorporated directly into the Local Observation Vector of a Multi-Agent Reinforcement Learning (MARL) policy network. The MARL engine is thereby configured to select a strategic negotiation primitive (e.g., BID, YIELD, TRADE) that is dynamically determined based on this real-time Conflict Risk Score. This transforms abstract resource allocation into a safety-aware, risk-adaptive protocol, allowing an ADSP agent to maximize fleet efficiency while guaranteeing collective safe separation

The invention relates to the technical field of artificial intelligence and unmanned aerial vehicles, and discloses a multi-unmanned aerial vehicle cooperative navigation method based on deep reinforcement learning, the method is suitable for multi-agent task execution in a complex dynamic environment, the method is constructed by adopting a multi-agent dual deep Q network model, a centralized training and distributed execution architecture is combined, and in a training stage, the multi-agent dual deep Q network model is constructed; through a global state optimization cooperation strategy, each unmanned aerial vehicle autonomously decides according to local observation in an execution stage, and dynamically adjusts an information synchronization frequency based on a communication load sensing mechanism so as to reduce communication bandwidth occupation and improve system real-time performance, and the method also introduces an experience playback and priority sampling mechanism to improve model stability and convergence speed. The system can sense obstacle information, energy consumption level and other unmanned aerial vehicle states in real time, realizes path dynamic optimization and formation keeping, has good task adaptability and expandability, and is suitable for collaborative navigation application of a heterogeneous unmanned aerial vehicle system.

The invention discloses a dual-scale resource optimization method for a multi-unmanned aerial vehicle auxiliary edge computing system. The method comprises the following steps: establishing a dynamic mobile edge computing system model comprising multiple unmanned aerial vehicles and ground mobile users; establishing a joint optimization problem of task unloading, computing resource allocation and unmanned aerial vehicle trajectory planning by taking minimization of total energy consumption of an unmanned aerial vehicle system as an optimization target; based on a designed double-time-scale layered optimization framework, aiming at a joint optimization problem, on a small time scale, an improved clustering algorithm and a closed solution method are adopted, and a real-time optimal task unloading decision and a computing resource allocation strategy are efficiently obtained; and on a large time scale, a near-end strategy optimization algorithm in deep reinforcement learning is used to carry out autonomous learning optimization, and an optimized unmanned aerial vehicle flight path is obtained. The invention provides a resource optimization method which can give consideration to dynamic adaptability, multi-time scale collaboration, global energy efficiency optimality and low calculation complexity.

The invention discloses a dynamic platform unmanned aerial vehicle autonomous landing method and system based on deep reinforcement learning and digital twinning. The system comprises an unmanned aerial vehicle, a dynamic platform, a digital twin environment and a deep reinforcement learning agent. The deep reinforcement learning agent performs offline training through interaction with a simulation environment in a DT environment, utilizes an elaborately designed reward function to guide learning, and adopts a domain randomization technology to enhance the robustness of a strategy to model uncertainty and environment change. And after training is completed, a control strategy obtained through optimization is deployed to a real unmanned aerial vehicle system. In an actual landing task, the system fuses sensor data in real time for state estimation, the estimated state is input into a deployed DRL intelligent body, the intelligent body outputs a control action based on a learned strategy, the unmanned aerial vehicle propulsion system is driven after control distribution, and the unmanned aerial vehicle propulsion system is guided to safely and accurately land to a target area on a dynamic platform. According to the method, the autonomous landing precision, success rate and robustness of the unmanned aerial vehicle under the complex dynamic condition are remarkably improved, the research and development risk and cost are effectively reduced, and the strategy migration effect from simulation to reality is improved.

This invention discloses a spatial domain impedance learning control method for a five-bar parallel robot, comprising the following steps: Step 1, establishing an Eulerian-Lagrange model of the five-bar parallel robot; Step 2, establishing a human-machine system interaction force model based on the trajectory tracking error of the parallel robot; Step 3, estimating impedance dynamics based on the Eulerian-Lagrange model; Step 4, designing an iterative learning strategy for human-machine interaction force based on the human-machine system interaction force model and impedance dynamics; Step 5, conducting control tests on the five-bar parallel robot using the iterative learning strategy for human-machine interaction force. The learning control method proposed in this invention ensures the convergence of tracking errors, considers the uncertainties in robot modeling, and further guarantees the safety of human-machine interaction.

This disclosure provides a method and apparatus for predicting human-machine conflict, capable of proactively predicting human-machine conflict from multiple dimensions. By acquiring and analyzing various physiological and behavioral data of operators in historical human-machine interaction scenarios, significant characterization factors corresponding to fatigue states are screened, enabling a more accurate quantitative representation of operator status. Simultaneously, the significant characterization factors of operators are combined with environmental and machine characteristic dimensions (including the complexity of interface interaction) to establish a human-machine conflict prediction model suitable for the target scenario. This method can dynamically reflect changes in multiple factors during system operation, proactively predicting the probability of human-machine conflict occurrence, thereby improving the comprehensiveness, timeliness, and reliability of identification, avoiding lag and bias caused by single data or rules, and ultimately contributing to improving the safety and stability of complex human-machine systems.

The invention provides a method and a device for evaluating the workload of a pilot in a high-dynamic scene, and the method comprises the steps: inputting the audio and video information of a cockpit according to the real-time operation of a computer; the method comprises the following steps: decomposing uncertain actions made by a pilot when executing a task in a high-dynamic scene in a time window with a preset length into basic action elements with fine granularity and certainty; and evaluating the workload of the pilot according to the types of the basic action elements appearing in the time window with the preset length, the duration of each basic action element and the total resource consumption value of each basic action element. The method is real-time, accurate, quantitative and non-invasive, and a certain reference is provided for efficiency improvement and task arrangement optimization of new-generation aviation equipment. The method is suitable for flight training, task optimization and man-machine system design.

The invention discloses a three-dimensional modeling unmanned aerial vehicle system and method for dynamic scene change monitoring, and the method comprises the steps: collecting an RGB image, depth data, IMU inertial information and GNSS positioning data of a to-be-monitored scene, and carrying out the standardization processing, so as to output a standardized multi-source data set; calculating a pose corresponding to each frame of data of the unmanned aerial vehicle based on the RGB image after standardization processing, the IMU inertial information and the GNSS positioning data so as to output a continuous pose sequence; processing the RGB image, the depth data and the continuous pose sequence based on an optical flow method-semantic segmentation-geometric consistency verification mechanism to output static data; generating static three-dimensional models at different time points based on the static data and the continuous pose sequence; and generating a monitoring result based on the static three-dimensional model at different time points, wherein the monitoring result comprises a change thermodynamic diagram, a change area quantitative report and an alarm indication signal. The method gives consideration to precision and real-time performance, can be widely applied to the fields of emergency rescue, engineering construction, ecological environment and the like in industrial application, and has remarkable technical innovation and wide market prospects.

The application discloses a kind of based on environmental perception's exoskeleton robot gait autonomous decision-making method, belong to robot automatic control field.It is composed of data acquisition module, environmental perception module, gait switching module three parts, wherein data acquisition module obtains current scene information data by vision sensor, the type information and distance information of ground in front are measured by environmental perception module, gait switching module calculates the landing point of each step according to the track energy of man-machine system, and the end trajectory is planned using the method of dynamic motion primitive, and the step length is dynamically adjusted, so that man-machine system can smoothly reach the terrain in front;And the size information of terrain in front is obtained by environmental perception module, and the gait parameters of exoskeleton robot are dynamically adjusted, to ensure that exoskeleton robot can smoothly, safely and autonomously switch gait;Through motor execution position feedback information, whether the current task of exoskeleton robot is executed is judged in real time.

This invention relates to a pilot decision-making deduction method based on dynamic optimization of a multi-level fuzzy branch structure, comprising the following steps: 1) quantifying the set of decision states; 2) quantifying the set of fuzzy rules; 3) designing multiple key influencing characteristics such as hysteresis factors, pilot skill level stability factors, and search factors to establish a multi-level fuzzy branch structure relationship that reflects the pilot's decision-making logic, and using it as the pilot decision-making deduction mechanism in the human-machine system; 4) dynamically optimizing the multi-level decision logic and the decision-making deduction sequence results during the time-series recursion process. Compared with existing technologies, this invention introduces a logical thinking mode of human cognition mechanisms and decision-making characteristics by designing multiple decision levels and their rule changes, overcoming the shortcomings of traditional pilot models such as fixedness, static parameters, and poor transferability, while also possessing the advantages of adaptability to various aircraft types and wide application range.

This invention provides a functional architecture and method for an intelligent human-machine adaptive collaborative system in an aircraft cockpit. It features advantages such as hierarchy, modularity, and reusability, supporting detailed design of the physical architecture of the intelligent human-machine system in the aircraft cockpit and providing a technical means to achieve efficient human-machine matching and dynamic natural interaction in the aircraft cockpit. The intelligent human-machine adaptive collaborative method provided by this invention, by adjusting the human-machine interaction level, the autonomous dimension of the intelligent agent, the information processing dimension, and the task allocation dimension, and based on the decision steps of the information flow graph, can obtain the optimal intelligent human-machine adaptive collaborative solution for the aircraft cockpit. The system functional architecture and method provided by this invention can realize intelligent human-machine interaction that adapts to the pilot's physiological, psychological, behavioral state and task requirements. It can provide technical support for solving problems such as pilot cognitive load imbalance, loss of situational awareness, and frequent human errors, enhancing the pilot's situational judgment and decision-making capabilities, and optimizing the overall effectiveness of the cockpit human-machine system.

A human system interface (HSI) for contactlessly controlling a device for a medical procedure is configured to generate a control space, measure a position of a user in the control space, provide a first contactless haptic stimulus in the control space to the user based on the measured position, measure a contactless haptic input from the user in the control space, provide a second haptic stimulus to the user depending on the contactless haptic input, and, based on the contactless haptic input, change a parameter of the HSI and / or provide a parameter to the device connected to the HSI.

The invention provides an unmanned aerial vehicle bridge inspection comprehensive system and a bridge inspection method.The system comprises a hardware layer, a service layer and a user layer, and the hardware layer comprises an unmanned aerial vehicle system and a bridge sensor system and is used for monitoring key parts of a bridge based on the bridge sensor system to obtain multi-dimensional data of the bridge; generating a bridge abnormal area flight path corresponding to the unmanned aerial vehicle system under the condition of determining that abnormal mechanical parameters exist in the bridge multi-dimensional data; the service layer is used for carrying out bridge disease identification on a bridge abnormal area corresponding to the abnormal mechanical parameters according to the unmanned aerial vehicle image data to obtain a bridge disease identification result of the bridge abnormal area; and the user layer is used for generating a corresponding bridge inspection report according to the bridge disease identification result and the bridge multi-dimensional data. According to the invention, the accuracy and inspection efficiency of bridge inspection results are improved.

This invention discloses a multi-level human-computer interaction intent recognition method for complex task scenarios, belonging to the field of human-computer interaction intent recognition technology. It solves the technical problems of low accuracy and low interpretability of multi-level human-computer interaction intent recognition in complex task scenarios. The key technical solution proposes a multi-level intent recognition framework suitable for complex task scenarios, combining task analysis methods and machine learning methods. This framework possesses both sensitivity to human behavior and contextual information, as well as interpretability. Its input data consists of easily captured human behavior data and situational data that can be collected by machine sensors, offering the advantages of not interfering with the working process of the human-computer system and its operability. By simultaneously integrating human behavior, environmental, and task context features, it identifies multi-level human intents layer by layer from the bottom-level data, achieving a more comprehensive understanding of the true human intent. This effectively overcomes the limitations of existing methods that rely on a single level and only on human behavior characteristics for intent recognition, improving the accuracy of multi-level intent recognition while simultaneously achieving interpretable modeling of multi-level intents.

The application is directed to the problem of fault-tolerant formation control of fuzzy multi-leader UAV system. A fault-tolerant formation control method based on reduced-order observer is proposed, including: using T-S fuzzy model to model the multi-leader UAV system; transforming the original system into a reduced-order form, and then deriving an augmented form; designing a reduced-order fault observer based on the intermediate variable for the reduced-order augmented system to estimate the state of the follower UAV, process fault and system uncertainty; using the relative state information to design a tracking fault-tolerant formation controller, taking the estimation of process fault and system uncertainty as compensation term; designing adaptive parameters for the reduced-order fault observer and the tracking fault-tolerant formation controller; verifying the performance of the reduced-order fault observer and the tracking fault-tolerant formation controller; the application can more prominently deal with the nonlinear and uncertain control problem, while reducing the calculation amount and improving the efficiency of distributed decision-making.

This invention relates to an ankle joint motion recognition mechanism, comprising a first branch, a second branch, a third branch, a moving platform, and a fixed platform. All three branches are connected to the moving and fixed platforms, with the first and second branches symmetrical about the third branch. Each branch includes a motor and nine links, forming two parallelograms. The first branch has two encoders, and the second branch has one encoder. The mechanism has a fixed center and a moving center. The motors on the three branches control the pose of the moving platform, giving the moving center two degrees of freedom of movement around the fixed center and the moving platform three degrees of freedom of rotation around the moving center, thus achieving high-precision fitting of ankle joint motion. A human-machine motion mapping model is constructed based on the link angles measured by the encoders and the human-machine system mapping parameters. This model identifies the talus posture, effectively recognizing the talus movement patterns hidden within the ankle tissue, providing data for the design and optimization of the motion trajectory and posture transformation trajectory of subsequent ankle joint rehabilitation robots.

The invention provides an unmanned aerial vehicle multivariate heterogeneous virtual-real fusion simulation test method and system, and the method comprises the steps: 1, constructing an unmanned aerial vehicle virtual-real fusion simulation test system which comprises a simulation modeling computer, a simulation calculation node and a storage server; 2, constructing an unmanned aerial vehicle digital prototype model based on an unmanned aerial vehicle virtual-real fusion simulation test system; 3, performing unmanned aerial vehicle virtual prototype model integration on the unmanned aerial vehicle digital prototype model constructed in the step 2; 4, hardware resources are integrated, and virtual-real mapping of all systems and real objects in the unmanned aerial vehicle virtual prototype model is achieved; and 5, full virtual simulation and semi-physical simulation of the unmanned aerial vehicle system are realized through a virtual simulation engine and a real-time simulation engine. By applying the technical scheme of the invention, the technical problem that in the switching process of different fault states of ground simulation test flight in the prior art, the preparation of the physical state of the equipment is relatively time-consuming is solved.

An architecture including a human and a human AI agent designed to understand and interact with the human, a system and a system AI agent designed to understand and interact with the system. The human AI agent and the system AI agent are configured to communicate with each other so that the human AI agent learns about the system and the system AI agent learns about the human, optimizing the interaction between the human and the system. The human AI agent and the system AI agent are configured so that the human AI agent learns about the system and the system AI agent learns about the human during a configuration process before the architecture is operational and while the architecture is operating.

The invention relates to the technical field of photovoltaic panel cleaning, in particular to a full-automatic photovoltaic panel cleaning system based on artificial intelligence technology control, which comprises a vehicle-mounted platform, an intelligent unmanned aerial vehicle system, a multi-mode cleaning robot, an AI intelligent central control system and a wireless local area communication system. A to-be-cleaned area is scanned through the alpha unmanned aerial vehicle, three-dimensional information and pollution distribution are obtained, and an optimal cleaning scheme is generated in combination with real-time meteorological data; the beta unmanned aerial vehicle throws the cleaning robot and coordinates the gamma unmanned aerial vehicle to complete material supply; the cleaning robot executes multi-mode operation and feeds back the effect in real time; the system dynamically adjusts the strategy to cope with sudden weather conditions. The cleaning efficiency can be remarkably improved, resource utilization is optimized, intelligent and multi-scene adaptive photovoltaic panel cleaning is achieved, and therefore the overall performance and economic benefits of a photovoltaic power generation system are improved.

According to the man-machine permission dynamic allocation method based on the combination of the ESMD algorithm and the heart rate variability, through the technical combination of multi-dimensional data, the ESMD algorithm and a multi-layer model, internal pole symmetric interpolation and multi-scale decomposition of the ESMD method are utilized; a nonlinear trend is effectively discriminated, and random heart rate variability (HRV) data is decomposed into modal components and trend remainders of different time scales, so that a system can automatically sense the change of the state of a pilot, and authority distribution is dynamically adjusted according to a preset multi-attribute fuzzy rule. The method can solve the problems of single data processing dimension, insufficient physiological monitoring precision, insufficient authority distribution flexibility in a complex scene and the like in the prior art, has the remarkable advantages of comprehensive evaluation, intelligent decision making and high adaptability, and has important application value for improving the task execution efficiency and safety of a high-reliability man-machine system.

This invention relates to the field of complex human-machine systems technology, and particularly to an early warning method and device for detecting poor mental workload in operators. The method includes: determining the output state of a target mental workload recognition model to obtain a mental workload state result; determining whether the proportion of the same abnormal state in the mental workload state result exceeds a preset threshold; if not, no early warning response is triggered; otherwise, performance monitoring of the target operator is performed to obtain a performance state result; determining whether the performance state result deviates from the individual's historical performance range; if it deviates, an intervention response is triggered; otherwise, a prompt-level response is triggered. This solves the problems of existing early warning methods, such as difficulty in adaptively adjusting to changes in the operational context, leading to a high risk of false alarms or missed alarms, and difficulty in achieving targeted, sensitive, and effective support, thus affecting the intervention effect.

The application discloses a kind of intelligent lower limb prosthesis control mode leading switching method, it is related to rehabilitation robot control technical field, including the following steps: target terrain detection based on monocular vision, and the orientation description of the target terrain is carried out;Shift centroid in horizontal plane pose estimation;Combining the rule determination of motion mode switching with the orientation description and pose estimation.The present application can realize the leading prediction of prosthesis motion mode switching by constructing the effective perception channel of prosthesis to walking movement environment and the motion intention of prosthesis wearer, thereby reducing the delay required for human-machine system interaction communication, which helps to realize the natural switching of different motion modes of prosthesis wearer, and improves the reliability of prosthesis wearer walking.

A single-person driving dynamic function distribution method based on a Bayesian reinforcement Q learning algorithm comprises the steps that after modeling is conducted on an SPO man-machine system through a Bayesian network, triggering judgment is conducted after man-machine system performance prediction, and then closed-loop control is conducted through a dynamic function distribution strategy based on the Bayesian reinforcement Q learning algorithm. According to the invention, by adjusting the automation level of the cockpit automatic system, dynamic task allocation between a single pilot and the automatic system is realized, so that man-machine cooperation is more coordinated, and improvement of the overall operation efficiency is facilitated.

This embodiment discloses an unmanned aerial vehicle (UAV) system optimization method, electronic device, medium, and computer program product. The UAV system optimization method includes: constructing an optimization model for the UAV system based on its secure communication rate and energy consumption; wherein the value of the optimization model is positively correlated with the secure communication rate and negatively correlated with the energy consumption; constructing constraint conditions for the optimization model based on constraint parameters; wherein the constraint parameters include one or more of the following: communication parameters of the UAV system's intelligent reflector (IRS), beam parameters of the UAV base station, and movement parameters of the UAV; combining the constraint conditions, determining the optimal solution of the optimization model with the objective of maximizing its value; and optimizing the UAV system based on the optimal solution.

The invention discloses a scene awareness objective measurement method considering abnormal states, and relates to the field of human-intelligent interaction safety analysis. The method comprises four steps of experiment design and preparation, tested training and pre-experiment, formal data acquisition and monitoring, and data analysis and verification. Eye movement data before and after an abnormal state is collected through an eye movement tracking technology, a second-order transfer entropy index is innovatively introduced, parameters such as pupil variability and gaze entropy are combined, the relevance between eye movement parameters and scene awareness is verified through correlation analysis, and accurate quantitative evaluation of the scene awareness in the abnormal state is achieved. The method breaks through the limitation of a traditional measurement method, can truly reflect the real-time response of an operator to an abnormal state, provides support for the safety design and optimization of a man-machine system in the safety key fields of aviation, nuclear power and the like, and is remarkable in application value.

The invention discloses a task adaptive allocation method for an unmanned aerial vehicle group man-machine system. The task adaptive allocation method comprises the following steps: S1, constructing an adaptive man-machine system task allocation module framework; s2, establishing a human model for an operator based on a cognitive load theory and the like, and establishing an unmanned aerial vehicle model for an unmanned aerial vehicle group based on theories such as flight dynamics and the like; s3, forming a man-machine function preliminary distribution strategy of a tetragonal agent based on a top Agent perception and decision-making module of a self-adaptive man-machine system; the states of the unmanned aerial vehicle group and the operators are monitored in real time; and S4, based on an evaluation result and a control right switching rule, performing secondary task allocation on a man-machine function through an adaptive man-machine system task allocation module, and adjusting specific tasks of each Agent. According to the method, the dynamic problem of man-machine function distribution of an unmanned aerial vehicle group man-machine system under complex multiple constraints is solved, the reasonability and safety of man-machine function distribution under multiple influence factors are ensured, and the efficiency of executing complex tasks by an unmanned aerial vehicle group is improved.

The invention discloses an unmanned aerial vehicle inspection image real-time recognition device and method based on edge calculation, and relates to the technical field of unmanned aerial vehicle inspection and artificial intelligence, the device comprises an unmanned aerial vehicle system, and further comprises an edge calculation module carried by an unmanned aerial vehicle, a multi-source image acquisition module, a communication module, a storage module and a ground control terminal, according to the method, the problems of data transmission delay, dependence on cloud computing power and poor adaptability to complex terrains in traditional unmanned aerial vehicle inspection are solved, and the method has the advantages of being high in accuracy, high in accuracy and high in reliability. The method improves the inspection efficiency and recognition accuracy, and is especially suitable for highway inspection scenes of plateau high-altitude complex terrains.

The application provides a physical human-computer interaction control method based on impedance iterative learning, and comprises the following steps: establishing an Euler-Lagrange dynamics model of a five-link parallel robot; establishing a human-computer system interaction force model according to a trajectory tracking error of the parallel robot; designing a robot model reference adaptive control based on the Euler-Lagrange dynamics model; designing a human-computer interaction force iterative learning strategy based on the human-computer system interaction force model and the robot model reference adaptive control; and testing the control of the five-link parallel robot by using the human-computer interaction force iterative learning strategy. The physical human-computer interaction control method based on impedance iterative learning provided by the application can reduce the dependence on system information and ensure stable and safe human-computer interaction by realizing variable impedance adjustment.