371 results about "Kinematics equations" patented technology

The invention provides an equipment fault early warning method based on a neural network, a medium and equipment, and the method comprises the steps: collecting equipment state information at regular time, recognizing abnormal equipment, obtaining operation parameters and structure parameters of the abnormal equipment and a layout schematic diagram of a system where the abnormal equipment is located, constructing a dynamic coupling topology link, and extracting associated parameters; verifying and generating a first error weight in combination with a physical kinematics equation, generating a second error weight by matching historical fault data, and comprehensively generating fault state information; a detection mode is switched based on risk level evaluation, a system dynamic coupling topology link, associated parameters, abnormal equipment information and fault state information are input into a neural network model, fault link information is output, and finally first early warning information and second early warning information are generated according to a dynamic detection mode to realize two-stage early warning. According to the method, the fault conduction relation between equipment is modeled through the graph neural network, and fault path prediction and systematic risk early warning are realized.

The invention discloses a four-rotor unmanned aerial vehicle model prediction control method based on a deep Q-network (DQN). The method comprises the following steps: firstly, establishing an inertial coordinate system and a body coordinate system, and establishing a kinematic equation and a kinetic equation of the four-rotor unmanned aerial vehicle based on a Newton second law and an Euler-Lagrange equation; according to the method, an unmanned aerial vehicle robust controller based on model predictive control (MPC) is designed, optimal control input is generated through dynamic optimization, and the trajectory tracking performance of the unmanned aerial vehicle is improved. For various disturbances existing in the whole tracking process, a DQN reinforcement learning algorithm and an MPC method are combined to design a flight control system, control rate errors caused by the disturbances are compensated, the stability of attitude control is enhanced, and the tracking precision and the anti-interference capability are improved. And finally, verifying the robust performance of the flight control system through a simulation experiment.

The application discloses a modular robotic service vehicle (MRSV) including a chassis and body comprising one or more robotic arms and robotic legs with drive wheel providing stepping, walking and driving capability to transport passengers and / or cargo and a control system comprising kinematics equations providing real-time administration involving controlling one or more robotic legs and / or robotic arms to transition in a retracted position and / or in protracted position for achieving walking, driving, attaining and handling objects. Accordingly, an operator or a Network associating with providing real-time administration involving controlling motion and position of the MRSV according to an assignment relative for walking, driving, attaining and handling objects, and aiding or assisting a user.

The present disclosure discloses a method for designing a terminal guidance law based on deep reinforcement learning, and relates to the field of missile and rocket guidance. The method includes: establishing a relative kinematics equation between a missile and a target in a longitudinal plane of a target interception terminal guidance section of the missile; to adapt to a research paradigm of reinforcement learning, abstracting a research problem and modeling as a Markov decision process; building an algorithm network, and setting algorithm parameters, where a selected deep reinforcement learning algorithm is a deep Q-network (DQN); in a terminal guidance process of each round, obtaining a sufficient number of training samples through Q-learning, training a neural network and updating a target network at fixed frequencies respectively, and continuously repeating the above process before set learning rounds are reached.

The invention discloses a muscle strength prediction method and device based on fusion of a physical model and deep learning, and the method comprises the steps: taking the physical information of neuromusculoskeletal as the soft constraint of a data-driven model, fully considering the dynamic time relation of muscle strength distribution, converting the time series input data into time slice data, and carrying out the prediction of the muscle strength. Kinematic data and EMG signals are used as input, advanced features are automatically extracted through a deep learning algorithm, the relation between the input signals and muscle force is established, muscle force prediction meeting a kinematic equation is achieved, and the defect that an existing model for muscle force calculation generally does not contain definite physical modeling for the potential neuromechanical process, and the calculation accuracy is poor is overcome. And the physical mechanism in the joint movement process is not reflected.

The invention discloses a high-speed train multi-information fusion positioning method based on a Beidou satellite, and relates to the technical field of satellite navigation and positioning. The method comprises the following steps: designing a high-speed train multi-information fusion positioning scheme; based on a train multi-information fusion positioning scheme, establishing a nonlinear train state model based on a train kinematics equation; the position, speed, acceleration, course angle and pitch angle of the train are used as state vectors, and a train measurement model is constructed; based on a tracking target set by a train state model and a train measurement model, the train state is estimated by fusing IMM-UKF / PF models of unscented Kalman filter (UKF) and particle filter (PF) and introducing ACW and a short-term error feedback mechanism in state fusion. According to the method, the positioning precision and robustness can be improved, and the dynamic response capability in a complex scene can be enhanced.

The invention discloses an unmanned vehicle formation dynamic control method and system. The method comprises the following steps: collecting perception information of a current environment through a sensor and communication equipment; analyzing and determining the formation with the highest efficiency under the current condition according to the collected sensing information; establishing a master-vehicle and slave-vehicle formation model, and enabling each unmanned vehicle to move according to a kinematical equation of the master-vehicle and slave-vehicle formation model; an error dynamic model is established, and an error control method is designed, so that the tracking error between the slave vehicle and the master vehicle is converged to a minimum neighborhood; when a task vehicle needs to drive away from the formation, the task vehicle predicts the tracks of other vehicles and performs obstacle avoidance driving based on a dynamic prediction obstacle avoidance control method; and performing formation switching according to the current environment conditions and task requirements. According to the invention, the formation generation and maintenance capability of the unmanned vehicle formation can be improved in a complex environment, the formation obstacle avoidance strategy is optimized, and the balance between the formation and an independent task is realized.

The invention provides a mechanical arm control method based on a double evaluation network and a stability evaluation method thereof, and relates to the technical field of friction coefficient calculation. The method comprises the steps that a kinematics equation of the flexible double-connecting-rod mechanical arm is established; designing a reference neural network to obtain an internal enhanced signal of the reference neural network; designing an evaluation neural network based on the internal reinforcement signal of the reference neural network to obtain a long-term cost control function, the long-term cost control function being used for finding optimal control for minimizing the operation cost; according to the method, the initial torque of the flexible double-connecting-rod mechanical arm is obtained based on the kinematics equation of the flexible double-connecting-rod mechanical arm, the replaced control force is obtained based on the action neural network, trajectory tracking control is conducted on an FTLM system, meanwhile, elastic vibration of the system is restrained, and the requirement that historical data or pre-existing knowledge is assigned for strengthening signals is eliminated.

The invention relates to an automatic guided vehicle (AGV) adaptive path tracking method based on MPC. The method comprises the following steps: S1, establishing a kinematic discretization error model based on the kinematic characteristics of the two-wheel differential AGV, processing a continuous kinematic equation by adopting an Euler discretization method, expressing a dynamic change relationship between a transverse deviation distance and an angle deviation in a state-space equation form, and generating a kinematic discrete state-space model; the method has the advantages that the continuous equation is processed by establishing the kinematics discretization error model and adopting the Euler discretization method, the deviation dynamic relation is expressed through the state space, the calculation process is simplified, the precision is kept, sensor data are fused, the wheel type odometer, IMU and laser data are integrated through the extended Kalman filtering algorithm, and the precision is improved. The real-time position and angle deviation are calculated, the positioning accuracy is improved, a model prediction controller objective function is designed, a weight matrix and boundary constraint are introduced according to constraint conditions, and the effect of obstacle avoidance constraint is combined.

The invention discloses a compensation operation simulation method for a telescopic active wave compensation trestle, and the method comprises the steps: taking sea condition parameters as input, combining a wave frequency response factor of a ship, and simulating the six-degree-of-freedom wave frequency motion of the ship; establishing a series coordinate system connected with a compensation mechanism on the telescopic active wave compensation trestle, and constructing a forward motion equation and a reverse solution equation of compensation motion; according to the compensation target position of the tail end of the telescopic active wave compensation trestle, the six-degree-of-freedom motion amplitude of the ship is combined, and the motion stroke of a compensation mechanism is reversely solved; differentiating the motion stroke of the compensation mechanism, calculating the speed and acceleration of the compensation mechanism, and determining the statistical value of the motion time sequence of the compensation mechanism; and finally, judging whether the telescopic active wave compensation trestle can compensate the ship motion caused by the sea condition or not according to the statistical value. Whether the compensation mechanism design of the trestle can compensate the ship motion under the sea condition or not can be judged according to the simulation result.

The invention relates to the technical field of two-arm robot pipeline butt joint cooperative motion planning and control in a complex environment, in particular to a two-arm cooperative robot pipeline butt joint method based on relative Jacobi. According to the method, a double-arm kinematics equation relative to a Jacobian matrix is constructed, constraint conditions are constructed in combination with pipeline butt joint requirements, pose errors and integral information are introduced to design an anti-noise motion planning scheme, an end effector obstacle avoidance strategy is designed, the anti-noise scheme is converted into a quadratic optimization problem, and the quadratic optimization problem is solved through a numerical algorithm. According to the method, the problems that an existing scheme is insufficient in precision, inaccurate in constraint condition and poor in anti-noise and obstacle avoidance capacity are solved, the uniformity and accuracy of butt joint of the two-arm pipelines are improved, the influence of environmental noise is effectively restrained, butt joint safety is guaranteed, butt joint efficiency and reliability are improved, and the high-precision pipeline butt joint requirement under the complex environment is met.

The invention discloses a path planning method based on a special unmanned vehicle under an unstructured road, and the method comprises the steps: calculating the comprehensive cost of each node uniformly distributed in a space through dynamic planning logic according to a grid map containing dynamic and static obstacle information in a sensing range, and generating a preliminary path; and performing nonlinear optimization through a high-performance nonlinear programming solver in combination with the explored space corridor and a self-vehicle kinematics equation to generate a path for safely arriving at an expected target pose. The method has the advantages that the feasibility, comfort and safety of the path are ensured through dynamic planning logic and nonlinear optimization.

The invention discloses a multistage voxel adaptive laser radar pedestrian tracking method, which comprises the following steps: S1, acquiring laser radar point cloud data, and adding ground filtering and elevation filtering; s2, performing hierarchical voxel grid processing on the basis of the point cloud data after ground filtering and elevation filtering, inputting the point cloud data after hierarchical voxel grid processing into a DV-Det model to detect a point cloud pedestrian target of a current frame, and adding a Soft-NMS for post-processing optimization; s3, designing an unscented Kalman filter based on a pedestrian kinematics equation, predicting the motion state of a pedestrian target, and adding a joint probability data interconnection algorithm and an adaptive life cycle management strategy; and S4, the mass center of the pedestrian target is calculated and detected, the relative position and the movement speed of the pedestrian target are solved, and a display panel of pedestrian state information is added to a visual interface. In conclusion, the defects in the prior art are overcome, high-precision and low-delay pedestrian detection and tracking can be achieved in multiple indoor and outdoor scenes, and the method has high social use value.

The invention discloses an Internet of Things product intelligent management system and method based on Beidou positioning, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring Beidou positioning data, acquiring vehicle information and cargo information, and forming a Beidou positioning associated data set; analyzing whether the Beidou positioning data has cargo transportation route abnormity or not, and judging whether the cargo transportation state and the vehicle driving state are abnormal or not; when the cargo transportation route or the cargo transportation state is abnormal in the Beidou positioning data, calculating the matching degree of the loading condition of the vehicle and the cargo weight based on the vehicle information and the cargo information; predicting the current position of the cargo by using a linear regression algorithm; if the superposed vehicle driving state is abnormal, dynamically adjusting the linear regression model, and further adjusting cargo position prediction in combination with a vehicle dynamics model; and when only the running state of the vehicle is abnormal, predicting the current position of the cargo by combining the kinematics equation of the vehicle and the cargo dynamics model.

The invention relates to the technical field of magnetic suspension conveying lines, in particular to a variable curvature trajectory prediction control method and system for a magnetic suspension conveying line and a storage medium. The method comprises the steps that a variable-curvature trajectory prediction model of the magnetic suspension conveying line is established, and the variable-curvature trajectory prediction model describes the dynamic behavior of a conveying carrier on a variable-curvature path through a kinematics equation and geometric constraints and is used for predicting trajectory deviation in real time. According to the invention, by constructing a variable curvature trajectory prediction model and a multi-target rolling optimization control system, the problems of trajectory tracking and suspension stabilization of a magnetic suspension carrier under a complex path are systematically solved. The method has the advantages that a time domain motion equation is mapped to a space domain path coordinate system based on a curve coordinate system of arc length parameterization and a prediction model constructed by a French-Serret formula, the influence of curvature change on lateral displacement dynamics is accurately described, and dynamics mismatch caused by traditional fixed curvature approximation is eliminated from the model level.

The invention relates to the field of unmanned underwater vehicles, and discloses an unmanned underwater vehicle finite time trajectory tracking control method based on event triggering, which comprises the following steps: establishing two coordinate systems to describe the motion state of an unmanned underwater vehicle, and constructing a kinematical equation and a kinetic equation of a six-degree-of-freedom full-drive unmanned underwater vehicle; the method comprises the following steps: establishing an underwater unmanned vehicle kinematics and dynamics model, determining a control target, setting an event triggering finite time backstepping controller and event triggering conditions in combination with the underwater unmanned vehicle kinematics and dynamics model, and adopting a virtual controller of the event triggering finite time backstepping controller and the event triggering controller to control the event triggering conditions. Enabling the kinematics and dynamics model of the underwater unmanned vehicle to reach a stable state within the convergence time, and enabling the trajectory tracking error of the underwater unmanned vehicle to obtain a convergence time upper bound within the convergence time; according to the method, rapid convergence in finite time is realized, so that the accuracy and rapidity of trajectory tracking control of the underwater vehicle are realized.

The invention relates to the technical field of underwater robot control, in particular to an autonomous line patrol motion control method of an underwater cable detection robot, which comprises the following steps of: establishing a six-degree-of-freedom kinematics equation and a kinetic model of the robot, and collecting position information of the robot and a current heading angle of the robot; the target heading angle of the robot is calculated through an optimal deflection angle fusion algorithm, the current heading angle, collected by a sensor, of the robot is compared with the target heading angle, an error epsi is obtained, and the error epsi is added into a single-ring PID algorithm for adjustment; and controlling the thrust and direction of a propeller in the underwater cable detection robot according to a control signal tau (t) output by the single-loop PID algorithm and robot position information, judging whether a target heading angle of the underwater cable detection robot is consistent with a current heading angle, if so, completing autonomous line patrol, otherwise, continuing to correct, and if not, completing the detection of the underwater cable detection robot. The attitude control precision and the path navigation stability of the robot in the underwater environment are improved.

The invention discloses a mechanical arm repetitive motion inverse solution solving method and system considering iteration initial value selection. The problems that when a mechanical arm inverse solution is solved through a numerical iteration method based on a Jacobian matrix, dependence on iteration initial value selection is high, and motion cannot be repeated are solved. Comprising the steps that an accurate reachable working space of the mechanical arm is obtained, a certain number of reference inverse solutions optimized through an operability index are additionally stored in the accurate reachable working space, and an enhanced reachable working space is obtained; according to the expected pose information of the mechanical arm and the enhanced reachable working space, an iteration initial value for solving the inverse solution of the pose is selected and obtained; acquiring a tail end pose error kinematical equation of the mechanical arm at a joint position layer, and deducing to obtain an iterative updating equation of a joint position; and a repeatable motion index function is designed, a joint position iteration updating equation with the repeatable motion function is obtained, and according to the iteration updating equation and the selected iteration initial value, a mechanical arm repeated motion inverse solution solving scheme selected by considering the iteration initial value is obtained.

The invention provides an unmanned aerial vehicle six-degree-of-freedom hardware acceleration IP core implementation method and electronic equipment, and belongs to the technical field of calculation acceleration, and the method comprises the steps: packaging a six-degree-of-freedom kinematics equation of an unmanned aerial vehicle and parameters thereof to form an IP core; an IP core is loaded on the FPGA; the six-degree-of-freedom calculation of the unmanned aerial vehicle is carried out on the FPGA through the loaded IP core, and the calculation period of the six-degree-of-freedom of the unmanned aerial vehicle is controlled based on the clock period of the FPGA. According to the invention, the six-degree-of-freedom calculation period of the unmanned aerial vehicle is controlled through the clock period of the FPGA, so that the updating process from the input data of the IP core to the output data of the IP core is a fixed period and does not fluctuate along with the change of the input data or the calculation state; therefore, the technical problem of serious periodic jitter during six-degree-of-freedom calculation of the existing unmanned aerial vehicle is solved, and the credibility and the control stability of the simulation model are improved.

A method for improving unmanned aerial vehicle state estimation and mapping robustness comprises map construction and state updating, and the map construction comprises the following steps: S1, obtaining real-time data through a laser radar scanner and an IMU sensor carried on an unmanned aerial vehicle, performing forward propagation on real-time data through an IMU kinematics equation to predict a predicted pose of the unmanned aerial vehicle system; s2, correcting the laser point cloud by using the predicted pose to obtain corrected point cloud data; s3, performing voxel map construction by using the corrected point cloud data; through a pose prediction and correction mechanism, efficient voxel map construction and real-time state feedback and adjustment, the robustness of unmanned aerial vehicle state estimation and mapping is significantly improved. Even in a complex and dynamic environment, the unmanned aerial vehicle can keep stable pose estimation and accurate map construction, and a solid foundation is provided for subsequent tasks such as autonomous navigation and obstacle avoidance.

The invention discloses a satellite attitude system anti-interference finite time control method under complex dynamic disturbance, and the method comprises the steps: building a rigid satellite attitude maneuver control model according to satellite attitude dynamics and kinematics equations; designing a self-adaptive fine disturbance observer to estimate the total disturbance of a satellite attitude system by combining the influence of the external disturbance variable on the rigid satellite attitude maneuver process, and outputting an estimated value of the external disturbance variable; a three-stage segmented sliding mode control method is designed, dynamic continuity of a three-stage sliding mode is ensured, and segmented sliding mode control output is obtained; and compensating the segmented sliding mode control output by using an estimated value of an external disturbance variable output by a self-adaptive fine disturbance observer to realize satellite attitude stable control. According to the method, the problem that the stability of the satellite is reduced in the attitude maneuvering performance process when external disturbance exists is solved, better observation performance and shorter attitude stabilization time are achieved, and the control precision of the satellite is effectively improved.

The invention discloses a feeding motion calculation method for a double-output-shaft five-axis numerical control tool grinding machine, and the method specifically comprises the steps: building a unified coordinate system according to the mechanism characteristics and solving demands of the grinding machine, and the unified coordinate system comprises a machine tool coordinate system, a machining coordinate system, a workpiece coordinate system and a grinding wheel coordinate system; grinding machine structure parameters are set and comprise the offset of the original point of a machining coordinate system relative to the C axis, grinding machine structure offset examples of the offset of the original point of the machining coordinate system relative to the B axis, the grinding wheel installation distance and the grinding wheel installation direction; a universal grinding machine feeding kinematics equation is established by considering the inward and outward mounting conditions of main shafts at the left end and the right end of the grinding wheel; and solving based on grinding wheel installation transformation and grinding machine stroke constraint. According to the method for solving the partitioned selective solution based on the actual travel limit of the machine tool, the solution is more universal and practical, the actual travel limit of the machine tool is met, and the method has good adaptability.

The invention provides a space target trajectory parallel mapping method based on a physical information network, and belongs to the field of time sequence data processing in deep learning. Aiming at the problems that a large amount of repetitive work exists and time and labor are consumed in the process of space target motion physical system analysis modeling and accumulative error correction, the method comprises the steps of fitting a kinematics equation of a space target in a parallel digital space through a parameterized neural network, and introducing known physical equation information; and the constraint network searches an optimal hypothesis process in a hypothesis space, and a parallel space modeling mapping process is converted into solution of a parameter inversion problem in a target domain, so that rapid construction of a parallel digital space is realized. Meanwhile, along with the increase of input sample data, the ability of the cross-cycle space-time attention module designed by the method for capturing the motion characteristics of the space target is stronger and stronger, and dynamic correction of errors is realized in a continuous iteration process.

The invention relates to a track vehicle ground feature recognition method, belongs to the technical field of vehicle control, and solves the problem of inaccurate ground feature recognition in the prior art. Comprising the steps of performing time and space synchronization on ground coverage features and ground state features to obtain synchronization features; constructing a vehicle kinematic equation based on the vehicle kinematic model of the instantaneous steering center, and estimating vehicle sliding parameters at all moments by using a synchronous feature solving equation; performing principal component analysis on ground elevation information in the synchronization features to obtain ground unit normal vectors at all moments; estimating a vehicle pose sequence according to the ground unit normal vector at each moment; obtaining a ground texture type according to the deviation between the estimated vehicle pose sequence and the vehicle pose sequence in the synchronization feature; according to the vehicle speed, the vehicle sliding parameters, the ground surface material and the ground texture type, a ground adhesion coefficient interval is obtained through a clustering algorithm; and accurate identification of the ground texture type and the ground adhesion coefficient interval is realized.

The invention provides an intelligent parking dead reckoning method and device and a storage medium. The method comprises the following steps: initializing a vehicle state vector, a system error covariance matrix, a system error matrix and an observation noise matrix of a vehicle at the current moment; according to the vehicle state vector at the previous moment, a predicted value of the vehicle state vector at the current moment is calculated through a vehicle kinematic model, and a state transition matrix is calculated based on a kinematic equation of the predicted value; calculating an observation matrix based on the front wheel driving distance, the rear axle center speed and the vehicle angular speed, and calculating a Kalman gain; on the basis of the predicted value, the Kalman gain and the observed value of Kalman filtering, updating the vehicle state vector to an optimal estimated value, and updating a system error covariance matrix; and performing iterative operation so as to continuously optimize the optimal estimation value of the vehicle state vector, and continuously outputting a dead reckoning result. According to the invention, the cost is reduced, and the precision and accuracy of the parking dead-reckoning positioning result are improved.

The invention discloses a self-adaptive recognition method and system based on a four-bar linkage single station machine motion trail, and belongs to the technical field of motion trail recognition. The method comprises the following steps: acquiring a multi-source synchronous data stream, detecting a peak interference frequency, adjusting an adaptive notch filter in real time, obtaining a pure geomagnetic field signal, calculating the equivalent length of a connecting rod by using angle and force data, and establishing a kinematics equation to solve track coordinates; establishing a magnetic field space gradient tensor model through a magnetometer array, fitting a nonlinear mapping relation between an angle and a gradient tensor, and generating a metal magnetic interference compensation model; judging a motion mode and dynamically adjusting filter parameters; constructing a fast and slow dual-time scale filter to fuse multi-source information, and introducing a fading factor to prevent divergence; a future angle is predicted by using a long-short-term memory neural network to serve as a virtual observed quantity to enhance the fault-tolerant capability; and optimizing the reconstruction trajectory through multiple physical constraints. According to the method, the track identification precision, the anti-interference capability and the system robustness are remarkably improved.

The invention relates to the technical field of robot control, in particular to a trajectory tracking control method for a mobile robot. The method comprises the steps of determining a kinematics equation of a robot model under the control of a control system based on the robot model, setting a reference trajectory as a global reference trajectory, defining trajectory tracking errors based on the kinematics equation and the global reference trajectory, and establishing a position error model based on the trajectory tracking errors; discretizing the kinematic equation and the position error model to obtain a kinematic discrete model and an error discrete model of the mobile robot; and defining a cost function, setting a triggering condition and a triggering interval of an event triggering mechanism based on the cost function, and when the triggering condition is met, solving an optimal control problem of a limited prediction time domain to obtain an optimal control sequence and a corresponding state trajectory. According to the invention, the MPC online calculation amount is reduced, and the adaptability and robustness of a control algorithm are improved.

The invention discloses a simulation method and system for a spraying mechanical arm based on a Bullet engine. The simulation method comprises the following steps that SolidWorks software is combined with a URDF plug-in, three-dimensional modeling is conducted based on spraying mechanical arm model data, a basic axis and a coordinate system are established at the joint of the spraying mechanical arm, kinematics parameters are set, and a URDF file of the spraying mechanical arm is exported; on the basis of a Bullet engine, motion control over the spraying mechanical arm is achieved in a simulation environment in combination with a forward kinematics equation or a reverse kinematics equation; on the basis of a Bullet engine, multiple working scenes of the spraying mechanical arm are built and added into a model of the spraying mechanical arm; and scanning the steel structural part and forming point cloud data, generating a spraying track of the spraying mechanical arm by combining the spraying process parameters of the steel structural part, and importing the spraying track into the spraying mechanical arm model in the simulation environment for verification.

The invention relates to the technical field of unmanned system multi-agent control, in particular to a wheel type mobile mechanical arm coordination control method, system and device. Comprising the following steps that a wheeled mobile mechanical arm model is established, and a target reachability area is solved; the operability index of the wheeled mobile mechanical arm is calculated; calculating the maximum linear velocity and angular velocity of the wheel type mobile mechanical arm to obtain a velocity ellipsoid model of the wheel type mobile mechanical arm; designing a multi-agent combined state space, and designing a multi-agent combined action space; constructing a control algorithm flow; setting a task scene of a coordination control algorithm of the wheeled mobile manipulator, setting training hyper-parameters, and designing task evaluation indexes; and carrying out training setting, and recording the success rate of the task and the validity of the process when the training period is finished. According to the method, the problem that the model structure is complex and depends on kinematics equation solving or objective function optimization under the high-dimensional condition is solved. According to the invention, the technical effect of reducing complexity calculation is achieved.

The invention belongs to the technical field of space flight and aviation, and relates to a space deployable antenna system dynamics modeling method comprising the following steps: using finite element simulation software to establish a space deployable antenna finite element model; correcting a linear substructure and a nonlinear hinge; after correction is completed, rigidity and mass matrixes are exported; performing first coordinate transformation on the linear substructure to complete order reduction of the linear substructure; second coordinate transformation is carried out, and all the linear substructures are fused and connected to obtain an overall structure kinematics equation; connecting the corrected linear substructure and the nonlinear hinge, analyzing dynamic response, constructing an input-output mapping agent model through data of restoring force and relative displacement at a connecting interface, and fusing an overall structure kinematics equation with the agent model to obtain a fused dynamic model; the dynamic response can be quickly predicted, computing resources are saved, and the analysis efficiency is improved.