73 results about "Neural network controller" patented technology

An adaptive impedance-based multi-mobile-robot collaborative transportation control method. Each mobile robot estimates the actual pose and ideal pose of a reference point and the first and second derivatives of the ideal pose by means of two finite-time fully-distributed observers, respectively; and then, on the basis of the estimated poses of the reference point, the pose of an end-effector of a mechanical arm, and closed-chain constraints for collaborative transportation, an ideal trajectory of the end-effector of the mobile robot, and an estimated value of a pose deviation between the end-effector of the mobile robot and the reference point are obtained. An adaptive impedance system of each mobile robot is interconnected with a virtual energy tank, and the energy tank is used to guide the updating of impedance parameters, thereby ensuring the passivity of the entire collaborative adaptive impedance system. To process unknown system dynamics of mobile robots, an asymptotic tracking adaptive neural network controller is designed using a neural network, thereby asymptotically achieving an ideal adaptive impedance relationship. The operational accuracy of multi-robot collaborative transportation systems is improved while ensuring safe collaboration.

The invention provides a motion control method of a coupled output limited Mecanum wheel trolley based on a neural network and event triggering, a storage medium, equipment and a computer program product. The motion control method comprises the following steps: adopting a conversion matrix of Mecanum wheels; establishing a dynamic model of the Mecanum wheel trolley; decoupling the coupling output limitation, and converting the coupling output limitation into time-varying non-coupling limitation; improving the design of a sliding mode surface by adopting an error conversion function and a barrier function, and constructing a sliding mode controller of a trolley with limited coupling output; constructing a neural network controller; introducing an event triggering mechanism, and constructing a neural network sliding mode controller for the coupled output limited trolley based on the neural network and sliding mode control; and verifying the stability of the designed controller. According to the invention, the trajectory tracking problem of the Mecanum wheel trolley with limited coupling output based on the neural network and event triggering is solved.

The invention relates to the technical field of industrial automation control, in particular to a boiler combustion stability intelligent control method based on a self-adaptive neural network, and the method comprises the steps: collecting multi-source heterogeneous data of a boiler system, the multi-source heterogeneous data comprises a high-frequency vibration data flow and a low-frequency process data flow, carrying out space-time alignment preprocessing on the multi-source heterogeneous data; and calculating a self-adaptive penalty factor based on the current load characteristics and the vibration energy, and performing variational mode decomposition on the high-frequency vibration data stream by using the self-adaptive penalty factor to obtain a combustion characteristic component. According to the method, the gain of the neural network controller is dynamically adjusted through the comprehensive stability risk index, smooth switching of a control strategy between steady-state fine adjustment and fault strong control is achieved, and therefore the fast inhibition capacity for combustion instability is remarkably improved while the steady-state precision of a combustion system is guaranteed.

The invention provides a ship path planning and tracking control method based on an improved differential evolution algorithm, and the method comprises the steps: constructing a collision risk degree model, and improving the calculation precision of the collision risk degree through combining with parameters of a quaternary ship domain optimization model; determining a fitness function Fitness to evaluate the quality of the path points according to the constraints of the ship collision risk degree, the voyage, the steering angle, the international marine collision avoidance rule and the optimal collision avoidance distance; a self-adaptive cross factor dynamic adjustment strategy based on individual fitness and population statistical indexes is introduced, candidate path points are screened according to an improved differential evolution algorithm, and collision with obstacles and dynamic ships is avoided; and establishing a ship course motion mathematical model, taking the reference path as an input instruction of a ship motion control subsystem, and tracking a ship planning path by using an adaptive neural network controller. According to the invention, the ship collision avoidance path planning is more in line with the actual navigation, and the safety of the collision avoidance path and the accuracy of tracking control are improved.

The invention relates to the technical field of distributed rendering synchronization, and discloses a distributed rendering node display picture synchronization method. The method comprises the following steps: acquiring real-time display picture data and timestamp information of a plurality of nodes, extracting picture display delay characteristics of each node and determining a delay change period through time domain statistical analysis of timestamps, and further constructing a synchronous fluctuation degree of each node; calculating the current overall network interference degree by combining the relevance of the synchronization fluctuation degree in the historical data and the historical mean value, and analyzing the delay growth mode of the display timestamp in the current continuous time point sequence to obtain the synchronization lag trend degree; and combining the two to obtain a synchronization offset index, calculating a feedback synchronization adjustment amount by using the variation of the synchronization offset index, the current synchronization control intensity and a preset adjustment increment, and finally performing dynamic synchronization adjustment on a display image of the distributed rendering node based on the adjustment amount and an actual image synchronization measurement value by applying a neural network controller.

The invention discloses an electric vehicle hybrid energy storage system energy management method fused with a Hemma particle swarm algorithm LSTM, and the method comprises the steps: constructing a multi-layer LSTM neural network controller to learn the characteristics of a driving condition, and minimizing the stress of a battery; constructing a fractional order integral derivative (FOID) controller, outputting a power correction component, and decomposing the component into a battery power reference component and a supercapacitor power reference component; according to the method, an improved He-Marx particle swarm algorithm is utilized, LSTM weight and FOID controller parameters are optimized in a combined mode through a multi-objective optimization function (including battery root-mean-square current and voltage errors), optimization training is carried out on the weight and bias of a neural network controller, and power distribution between a battery and a supercapacitor in the hybrid energy storage system of the electric vehicle is controlled in real time; an LSTM power distribution result is adaptively corrected through a dynamic weight mechanism, and bus voltage fluctuation is controlled within a range of + / -3%; and the final power reference value is synthesized through dynamic weighting, so that the peak current of the battery is remarkably reduced.

The method is a neural network-based online control method for a single-switch DC-DC converter, comprising: constructing a neural network model comprising a neural network training structure, a neural network training mode and a neural network training algorithm; constructing a neural network controller with the single-switch DC-DC converter as a control object and with a converter output voltage as a control target; collecting running state data and control data of the single-switch DC-DC converter in real time, obtaining gradients online, and realizing online training of the neural network and online control of the converter. The method does not need to model the converter, nor does it need a large number of data samples to perform offline training of the neural network, but realizes online acquisition of a control law (weights and biases) in the running process of the converter, achieves the purpose of real-time control, and the neural network controller can cope with the influence of step changes in input voltage or output load, has good robustness and dynamic response performance.

This invention relates to the field of spacecraft orbit control technology, and particularly to a spacecraft approach control method based on the fusion of imitation learning and reinforcement learning. The method includes: S1: generating initial and target states, solving for the optimal fuel transfer trajectory offline using a direct method, and constructing an expert dataset; S2: based on the expert dataset, training a neural network controller through imitation learning to obtain a pre-trained model; S3: loading the weights of the pre-trained model, constructing a composite reward function, and training the policy network using a near-end policy optimization algorithm to obtain an optimized policy network model; S4: acquiring relative state data, optimizing the policy network model, and outputting the control quantity for the optimal fuel consumption trajectory transfer in real time. This invention rapidly masters an approximately optimal policy through imitation learning and surpasses the performance of expert data through reinforcement learning, completing control command solutions in milliseconds. It balances terminal control accuracy and fuel consumption optimization, offering advantages such as high solution efficiency, excellent control accuracy, and strong generalization ability.

The invention discloses a multi-valve equipment continuous variable mach number test control method and system, and belongs to the technical field of wind tunnel aerodynamic force tests. The invention aims to solve the problem of regulation and control strategies of test key parameters of multi-valve equipment. The method comprises the following steps: establishing an initial physical relationship model of valve opening and Mach number, and correcting based on a machine learning algorithm; a traditional PID controller, a fuzzy PID controller and a neural network PID controller are integrated, a simulation control system is constructed, and the output of the simulation control system is the valve opening degree; building a combined control system; defining multi-objective optimization parameters of the combined control system; optimizing the multi-objective optimization parameters by adopting a particle swarm algorithm or a genetic algorithm, and solving an optimal multi-objective optimization parameter set; and outputting a control result of the continuous variable Mach number test of the multi-valve equipment based on the multi-target optimization parameter set, and realizing adaptive control of the continuous variable Mach number test. According to the invention, adaptive regulation and control of key parameters are realized, the test efficiency, precision and safety are improved, and support is provided for verification of a combined power technology.

This invention provides an integrated method for pruning and controlling large-scale models of unmanned aerial vehicles (UAVs), belonging to the field of model lightweighting technology. The proposed method significantly reduces model size and computational load, greatly improves inference speed and control response performance, and enables the deployment of complex neural network controllers on resource-constrained airborne platforms. Simultaneously, the lightweight model maintains or approaches the control accuracy and stability of the original uncompressed model. This method has wide applicability and strong scalability, covering various scenarios from basic flight control to complex mission control, providing a feasible, efficient, and cost-effective solution for applying large models to airborne controllers of UAVs, and possesses high engineering practical value.

The utility model discloses a household medicine taking auxiliary device which comprises a neural network controller and a database connected with the neural network controller. The household medicine taking auxiliary device further comprises a base, a connecting rod, a shell, a touch display screen, a timing module, a voice broadcast module, a recognition camera and a medicine placing tray. Wherein the shell is arranged above the base, and the shell is connected with the base through a connecting rod; the neural network controller, the database, the timing module and the voice broadcast module are all arranged in the shell, the touch display screen is arranged on the side wall of the shell, and the recognition camera is arranged on the lower surface of the shell; the touch display screen, the timing module, the voice broadcast module and the recognition camera are electrically connected with the neural network controller; and the medicine placing tray is arranged on the base. The system is based on the neural network technology and is used for assisting the elderly at home to accurately take medicine on time.

The application provides a flexible joint robot arm tracking control method and device based on an adaptive neural network, and relates to the technical field of robot intelligent control. The method comprises the following steps: establishing a robot kinematics model, defining a trajectory tracking error, and designing a control input based on the robot kinematics model; adopting a neural network compensation model to compensate for model uncertainty, and adopting a disturbance observer to estimate unknown disturbance, so as to obtain an adaptive bounded neural network control based on state feedback; building a robot system platform to verify the feasibility and effectiveness of the method, and defining a Lyapunov function to prove the stability of the closed-loop system. The application constructs an adaptive bounded neural network controller, which can estimate the uncertainty in the model parameters and adjust the controller gain to adapt to the actuator limit, so as to ensure the effective trajectory tracking of the system. Subsequently, in order to further enhance the stability and robustness of the system, the application also designs a disturbance observer to estimate and compensate for unknown external disturbance.

To provide a decoder, an encoder, a neural network controller, and a method which allow efficient representation and transmission of neural network parameters for learning or updating processes.SOLUTION: A decoder for decoding parameters of a neural network is configured to: obtain (1010) a plurality of neural network parameters of the neural network on the basis of an encoded bitstream; obtain (1020), from an encoded bitstream, node information describing a node of a parameter update tree, the node information including a parent node identifier and parameter update information; and derive (1030) one or more neural network parameters using parameter information of a parent node identified by the parent node identifier and using the parameter update information.SELECTED DRAWING: Figure 10

The invention provides a method for generating a maneuvering instruction of a fixed-wing unmanned aerial vehicle, and the method comprises the steps: 1, carrying out the generation of a three-dimensional track, and forming a training data set; step 2, establishing a deep deterministic policy gradient algorithm (DDPG)-based actor network as a neural network of an off-line training network structure, and performing off-line training by using the training data set to obtain initial actor network parameters; step 3, establishing an unmanned aerial vehicle simulation model and a lower controller, establishing a corresponding network structure based on a depth deterministic strategy gradient algorithm DDPG, and designing different maneuvering trajectory lines for online training to obtain a neural network controller; and 4, deploying the neural network controller as an upper-layer controller in flight control, and performing combined control by combining with an overload controller and a rolling controller of a lower-layer controller. According to the method, the multi-task requirement of the unmanned aerial vehicle in a training scene is met, development time is shortened, and task adaptability is improved.

The invention provides a power system neural network controller design method and system, and the method comprises the steps: carrying out the dynamic modeling of a power system, and introducing and employing a Lyapunov stability theory; performing neural network control input and stability verification; and controller optimization and controller design are carried out. The technical problems that large-range nonlinear system behaviors cannot be processed due to the fact that the degree of dependence on system linearization is high, only stability of the system in a local area can be guaranteed due to the fact that the system attraction domain is small, and the adaptability to uncertainty and disturbance is weak are solved.

The invention belongs to the technical field of power system control, and particularly relates to a low-communication-cost neural network controller of a nonlinear power system and a design method of the low-communication-cost neural network controller. According to the method, the event-driven regulation and control algorithm is designed for the nonlinear power system by using a machine learning method, so that optimal triggering is realized while calm control is performed, and effective control under limited communication resources is realized. An event-driven mechanism is considered, that is, whether the control strategy is updated or not is determined according to the real-time state of the target system, so that the communication cost generated by updating the control signal is reduced; the optimal triggering problem in event-driven regulation is considered, that is, the event triggering frequency in specified time is minimized; according to the invention, two schemes of a direct method and an indirect method are designed to realize optimal trigger control under an event-driven mechanism; and a strict stability and optimality guarantee is provided for a control strategy obtained by neural network training by using an approximate projection method. Finally, the superiority of the technical scheme is verified by taking an industrial heat exchanger as an example.

The present application relates to a kind of self-adapting evaluation control method of event-triggered double-link mechanical arm, belong to double-link mechanical arm system tracking control field.The method utilizes the omnipotence approximation theorem of neural network and the design method of backstepping method, realizes the construction of adaptive neural network controller.At the same time, introduce event-triggered mechanism, reduce the communication overhead between controller and actuator, improve the utilization efficiency of network resources.Evaluation network and execution network are combined, utilize smooth utility function and optimal tracking controller, improve the control performance and fault tolerance of system.In addition, actuator fault compensation term is designed to suppress the influence of actuator fault on system performance.The present application can overcome the difficulty of accurate modeling in conventional method and the problem of limited network resources, realize the efficient control of double-link mechanical arm system and the improvement of fault tolerance.The present application has wide application potential, is applicable to industrial automation, robot and other fields.

This invention discloses a collaborative training method and system for autonomous driving planning and control, belonging to the field of autonomous driving technology. The method includes: acquiring the current state of the vehicle, the states of surrounding traffic participants, and reference road information; using a Transformer-based planning network to output a control sequence, which is then used to generate a planned trajectory via a vehicle kinematics model; inputting the planned trajectory into a model predictive control-based neural network controller for dynamic tracking optimization to obtain the executed control quantity and predicted vehicle dynamic state; constructing a control loss and backpropagating its gradient to the planning network; and, during training, layering constraints on the planning cost term, using safety and traffic rule-related terms as hard constraints and comfort-related terms as soft constraints, and employing a multi-stage cost scheduling strategy to adjust the constraint strength or weight. This invention achieves collaborative optimization of planning and control, improves the dynamic executability and safety compliance of the trajectory, and enhances training stability and closed-loop robustness.

The application discloses an unmanned bicycle motion control method and system with a balance flywheel and an unmanned bicycle, and solves the control problem of the unmanned bicycle under the condition that system parameters cannot be accurately measured. The method decouples the control of the unmanned bicycle into three parallel tasks of speed control, steering control and balance control, adopts a proportional-integral (PI) controller to realize the speed control, adopts a first-order filter to realize the steering control, and adopts an adaptive neural network controller to realize the balance control. The control method proposed by the application effectively solves the balance control problem of the unmanned bicycle under different speeds and different steering angles under the condition that system modeling is inaccurate, and can identify uncertain parameters and functions in the system model.

The invention provides a multi-mode-driven super-large-scale neural network controller and a danger early warning method thereof. The multi-mode-driven super-large-scale neural network controller comprises an audio and video acquisition module, a multi-sensor module, a data fusion module, an industrial data driving basic model, a controller decision module and an early warning and communication module. The intelligent controller and the method thereof have the beneficial effects that the intelligent controller and the method thereof integrate an audio and video system and multi-sensor information, perform cognitive calculation by using an audio and video neural network compatible with multiple data modalities, and complete environment intelligent evaluation and danger early warning; according to the invention, the audio and video neural network compatible with multiple data modals is applied to multi-modal data cognition and analysis of high-risk environments such as a coal mine for the first time; efficient fusion and cooperative processing of audio and video data and multi-sensor data are realized; an accurate and real-time danger early warning and accident response mechanism is provided, and the safety is remarkably improved; and an embedded platform is used to complete early warning of the intelligent controller.

The invention relates to the crossing field of artificial intelligence and industrial process control, and discloses a desulfurization pH value and slurry supply amount neural network coupling control method. The method comprises the following steps: synchronously collecting original pH values, flue gas parameters and slurry physical property data from multiple sources; constructing a dynamic compensation model based on the slurry temperature, the density and the historical drift trend, and generating a compensated pH estimation value; inputting the estimated value and the smoke parameters into a pre-trained neural network model, and outputting a target slurry supply flow; and the outlet sulfur dioxide concentration is used as a feedback signal, and a slurry supply instruction is finely adjusted on line to form closed-loop control. And three redundant pH electrodes, a self-adaptive neural network controller and a multiple fault-tolerant strategy are adopted, so that the control precision and robustness are remarkably improved.

The application discloses a neural network controller based on data mining, and belongs to the field of power electronics. First, a data subset under different characteristic states is obtained from a data set under different working points of multiple different control algorithms through a data mining method; then, a new data subset is formed by combining data subsets with optimal performance under different characteristic states through performance evaluation, and then the data subsets under different working points are recombined into an optimized complete data set; finally, the input-output relationship of the optimized complete data set is fitted in a supervised learning mode to obtain the neural network controller based on data mining. The neural network is trained by the optimized data in the supervised learning mode, and a neural network controller with better performance can be fitted, so that the performance of the neural network controller is further improved.

The invention provides an industrial equipment multi-modal cooperative control method and system based on a neural network, and the method comprises the steps: synchronously collecting the multi-modal sensor data of industrial equipment, carrying out the preprocessing, and generating a structured data frame; performing dynamic alignment processing based on a variable time window on the structured data frame to generate a standard modal frame, inputting the standard modal frame into a pre-constructed lightweight neural network confidence evaluation model, and outputting a confidence score; according to the weighted modal input, utilizing a pre-constructed neural network controller to generate a multi-device cooperative control instruction; and calculating modal contribution feedback according to the attention weight and a multi-device cooperative control instruction, adjusting sensing configuration through the modal contribution feedback, and optimizing a sampling strategy of multi-modal sensor data so as to optimize generation of the multi-device cooperative control instruction.

To provide a decoder, an encoder, a neural network controller and a method allowing for efficient representation and transmission of neural network parameters of a learning or update process.SOLUTION: A decoder for decoding parameters of a neural network obtains a plurality of neural network parameters of the neural network on the basis of an encoded bitstream (1010), and obtains, from the encoded bitstream, node information describing a node of a parameter update tree (1020). The node information includes a parent node identifier and parameter update information. The decoder also derives one or more neural network parameters using parameter information of a parent node identified by the parent node identifier and using the parameter update information (1030).SELECTED DRAWING: Figure 10

The application discloses a kind of upper limb rehabilitation robot discrete neural network control method and system, the method includes: each joint module of upper limb rehabilitation robot is regarded as a subsystem, adaptive neural network controller is constructed based on discrete barrier lyapunov function;Determine weight update rate, according to discrete determination learning theory, experience-based learning controller is constructed;On the basis of experience-based learning controller, the fusion of knowledge is realized using least square method based on space constraint, and experience-based learning controller based on knowledge fusion is constructed;Variable admittance control model is constructed, and the output of variable admittance control model is used as the reference trajectory of experience-based learning controller based on knowledge fusion, and the training of corresponding rehabilitation action is executed according to the rehabilitation demand of user.The application realizes the intelligentization and safety high-performance control of upper limb rehabilitation robot in different individuals and different recovery periods.

The invention discloses a semi-physical simulation method and system based on an FPGA neural network control method, and the method comprises the steps: operating an aircraft dynamics model at a simulation end, and transmitting state variables, such as an attitude angle, a speed and an angular rate, to an FPGA end in real time through Socket communication; a fixed-point neural network controller is deployed at the FPGA end, a control instruction is reasoned in a parallel pipeline mode and is transmitted back to the simulation end through a TCP / IP protocol, the system adopts a PS-PL cooperative bus to realize zero-copy data transmission, hardware-level reasoning acceleration is realized in combination with a Viis HLS and PYNQ framework, and the communication delay is less than 400 s. Through closed-loop verification, the error between the FPGA reasoning result and the floating point simulation is smaller than 4.5 * 10 <-5 >, and microsecond-level real-time response and high consistency of the flight control algorithm are achieved.

This invention discloses a precise deagglomeration and dispersal modification control system for non-metallic powders, belonging to the field of intelligent control technology based on deep learning. Specifically, it includes: a non-metallic powder identification module, a non-metallic powder deagglomeration module, an atomization coating module, and a coordination control module. The non-metallic powder identification module deploys hardware units to collect raw powder data; the non-metallic powder deagglomeration module constructs a calculation model to calculate the energy required for dispersing the non-metallic powder and implements precise deagglomeration; the atomization coating module uses reverse reasoning to charge and directionally adsorb droplets, completing the atomization of the modifier and coating of the newly formed surface in the same space where deagglomeration occurs; the coordination control module, through the construction of a time-window preemptive scheduling control strategy and the embedding of a main control board with a built-in feedforward fuzzy neural network controller, triggers the injection of the modifier during the interval of deagglomeration energy release, achieving precise synchronous control of "deagglomeration equals modification".

The application discloses a kind of based on adaptive model predictive control curve lane keeping method and system, first according to the curve working condition information of vehicle travel determines the safe vehicle speed of vehicle over curve;Then the optimal time domain parameters of MPC controller under different road information are matched using genetic algorithm;Again, neural network MPC controller is constructed, so that the safe vehicle speed of vehicle over curve and the time domain parameters of MPC controller are adaptively adjusted according to the change of curve working condition;Finally, according to the trained neural network, the vehicle longitudinal controller and vehicle lateral controller are integrated, to realize the control of curve lane keeping.The application effectively solves the problem of vehicle sideslip or rollover caused by high speed over curve, and the poor lane keeping accuracy caused by fixed MPC time domain parameters, improves the accuracy of unmanned vehicle curve lane keeping, ensures the stability of vehicle driving, and has practical significance for promoting the rapid development of unmanned technology.

The invention discloses a double-closed-loop model-free control method for a high-gain DC-DC converter based on data driving. The method comprises the following steps: acquiring input duty ratio, output voltage and current data of the DC-DC converter under an open-loop condition, and filtering; a data twinborn model of the converter is constructed by adopting NARX, and dynamic characteristics of the converter are accurately represented through open-loop training and closed-loop conversion; a double-closed-loop neural network controller composed of a voltage outer loop controller and a current inner loop controller is designed, the voltage outer loop controller and the current inner loop controller are both three-layer feed-forward networks, input comprises errors, error integral terms and error differential terms, and nonlinear mapping capacity is achieved; an ADP algorithm is adopted, a performance index function is minimized, and a batch gradient descent method is combined to automatically optimize the weight and bias parameters of the controller; and the trained controller parameters are imported into a data processor to realize real-time high-performance control of the DC-DC converter. The method does not need to depend on an accurate mathematical model, parameters are automatically optimized, the dynamic performance is excellent, the method is suitable for various DC-DC converter topologies, and engineering implementation is easy.

The invention relates to the technical field of temperature detection and automation, and discloses an intelligent temperature detection and control method and a big data Internet of Things system thereof, and the method comprises the steps: collecting a temperature value sequence of a detected object, and predicting the temperature; acquiring the actual temperature regulation control quantity to obtain an ideal temperature value; acquiring a temperature prediction value and a temperature disturbance quantity based on the actual control quantity and the temperature value sequence; the difference value between the temperature disturbance quantity and the temperature disturbance predicted value is fed back to the neural network for learning, the difference value between the temperature predicted value and the temperature predicted value in a period of time is fed back to the neural network for learning, and the temperature disturbance quantity and the temperature predicted value which are fed back in real time are obtained; the difference value between the expected temperature value and the predicted temperature value passes through the neural network controller to obtain a linear control quantity and a nonlinear control quantity, and the difference value between the accumulated value and the temperature disturbance quantity fed back in real time is processed to obtain a temperature regulation actual control quantity. Compared with the prior art, the dynamic nature, robustness and accuracy of temperature detection and intelligent control are improved.