31results about How to "Good tracking effect" patented technology

The invention discloses a robot reinforcement learning control method and system based on a gating circulation unit, and relates to the technical field of artificial intelligence. Based on a reinforcement learning state machine, a trained student network is deployed to a real quadruped robot in a zero mapping manner, and movement control of the quadruped robot is realized; the training process of the student network is as follows: building a simulation training terrain environment of the robot, and setting different levels of terrain difficulty for each terrain; based on the distillation learning thought, establishing a teacher network and a student network based on a gating circulation unit, constructing a total loss function, and based on the constructed total loss function, fitting a motor joint angle output by the teacher network through an omnibearing simulation training terrain environment, a large-scale parallel training method and the student network; according to the robot reinforcement learning control method and system, robust motion of the robot in an unstructured terrain is realized, and the robot reinforcement learning control method and system have good anti-interference capability and self-recovery capability for a dynamic environment in which traditional control cannot be competent.

The invention provides a multi-target tracking algorithm of a wide-view-angle feature fusion memory network, and belongs to the technical field of computer vision and artificial intelligence. The technical problem that the tracking performance is insufficient when a long-distance small-size target and a fast moving target are processed under the view angle of the unmanned aerial vehicle is solved. According to the technical scheme, the method comprises the following steps: S1, video frame preprocessing and target detection; s2, carrying out deep fusion on the picture feature information in a wide-view fusion memory network; s3, track confidence modeling is introduced in the filtering process of adaptive motion filtering AMF; and S4, in the first step of association, performing first association through a motion feature extraction model MFEM. The method has the advantages that the target position is adjusted and optimized step by step through dynamic updating of a tracking algorithm, and finally information of all tracks of the target is obtained.

The invention relates to the technical field of image data processing, and provides a target tracking method, device and equipment based on AI visual identification, and the method comprises the steps: tracking a target in real time through satellite positioning, obtaining the image feature information of the target and the map position information of the target during missing tracking when the target is not tracked, and generating the associated feature information; inputting the associated feature information into a preset motion prediction model, outputting motion prediction information of the target, and performing multi-size feature analysis on the associated feature information to obtain multi-scale feature information; and fusing the motion prediction information and the multi-scale feature information, and inputting a preset AI visual identification model to obtain real-time position information and motion track information of the target. The data are analyzed through an AI visual identification model, the real-time position information and the motion trail information of the target are output, and the problem that the target is lost and tracking fails when the target is in complex environments such as sheltered, illumination changing or rapid motion is solved.

The invention discloses a mechanical arm trajectory tracking sliding mode control method and system based on a predetermined sliding set. The method comprises the steps that a motion equation representing a mechanical arm dynamic model is established; according to the motion equation, mechanical arm tail end joint angular displacement and angular velocity are selected as state vectors, and a mechanical arm state space model is established; according to the expected tracking trajectory, a mechanical arm tracking error vector is obtained, and a mechanical arm tracking error kinetic equation is established in combination with the mechanical arm state space model; selecting a sliding variable according to the tracking error information of the mechanical arm; a self-adaptive sliding mode tracking control law is established in combination with the sliding variable, and the sliding variable of the tracking error of the tail end joint of the mechanical arm is restrained in combination with a Lyapunov function stability analysis method; according to a Lyapunov function direct method, stabilization control integration is carried out on the constrained sliding variables, it can be ensured that a mechanical arm tail end joint tracking system cannot lose a preset sliding mode, and therefore the anti-interference performance, the fault tolerance performance and other robust performance are improved.

The invention discloses an optimal trajectory planning method and system for an unmanned surface vehicle (USV) in a complex environment, and the method comprises the steps: carrying out the state initialization of a target, and obtaining the initial state information; constructing a system model based on the initial state information; constructing a knowledge graph based on threats, and reasoning to obtain threat degree information; combining the multi-target posterior Cramer-Rao lower bound and threat degree information to construct an evaluation index; and based on the evaluation indexes, completing optimal trajectory planning. According to the method, a knowledge graph based on marine threats is constructed in combination with a target and an ocean current movement situation. Threat semantic information in the atlas is reasoned through a reinforcement learning method, the defect that a traditional threat assessment method is lack of self-learning ability for threat features is effectively overcome, and the correlation between assessment indexes can be deeply mined. According to the method, the target tracking effect can be optimized while target threat and ocean current disturbance are effectively avoided.

The invention discloses a waveform design method and system for tracking a sea surface maneuvering target by an unmanned ship-mounted radar, relates to the technical field of radar communication, and aims to solve the technical problem of inaccurate maneuvering target tracking in a sea surface clutter environment. The method is technically characterized by comprising the following steps: predicting a target state by using a marginalized particle filter algorithm, and obtaining a predicted measurement value; generating a hypothesis tree according to the observation data; performing target state estimation on each assumed branch by using a marginalized particle filter algorithm; a multi-hypothesis tracking algorithm is used to calculate an association probability and update a hypothesis weight, hypothesis pruning and fusion are carried out, and final target state estimation is obtained; a plurality of waveform selection criteria are used as a model, and an interactive multi-criterion selection method is utilized to determine an optimal criterion with the highest effective probability and an optimal transmitted waveform parameter. According to the method, false alarms caused by sea clutters can be effectively handled, the performance of the unmanned ship-mounted radar for processing complex tracking scenes can be improved, and a more accurate and robust target tracking result can be obtained.

The invention discloses a vehicle and pedestrian tracking algorithm recommendation method based on Deep-sort improvement. Relates to the field of recommendation methods, an improved Deep-Sort serves as a tracker, due to the fact that a tracking object of a Deep-Sort algorithm is a person, a research object of the invention is a vehicle and an appearance model is not applicable, the improved Deep-Sort improves a feature extraction network of the Deep-Sort in three aspects: the feature extraction network is rebuilt based on ResNet-34, so that the feature extraction network is more suitable for extraction of vehicle features; a center loss function and a triple loss function are introduced and are combined with a Softmax loss function for use, so that intra-class and inter-class distances of the network are better optimized; and a cosine annealing learning rate attenuation method is used, so that the final convergence effect of the network training model is better. The improved feature extraction network is trained, and experimental results show that the vehicle appearance model obtained through training has good classification capability, and the recognition capability of vehicle features is improved compared with an original model. In a multi-moving vehicle tracking experiment, the D-MMVT algorithm is tested by using an OTB-100 data set and a daily monitoring road video, the accuracy rate MOTA, the precision rate MOTP and the ID switching frequency are used as evaluation standards, a comparison experiment is performed with other tracking algorithms, and an experiment result shows that the D-MMVT algorithm has a better tracking effect on a severely shielded target and a small target; and the tracking accuracy and precision are improved.

The invention belongs to the technical field of magnetic suspension, and discloses an axial magnetic suspension bearing system modeling and controller design method based on different-element fractional order, which comprises the following steps of: 1, deducing a fractional order axial magnetic suspension rotor mathematical model which considers an eddy current effect and has four denominators through a Maxwell differential equation set; 2, designing a fractional order expansion state observer according to the mathematical model so as to obtain accurate state estimation and system disturbance; and step 3, designing a fractional order sliding mode controller based on the mathematical model of known disturbance. According to the method, the axial fractional order magnetic suspension bearing with four denominators is established by adopting a method based on Maxwell's equations, the adaptive capacity of a system to parameter change and uncertainty is enhanced by combining a fractional order sliding mode control strategy (FSMC) of a fractional order extended state observer (FESO), and it is ensured that the system can still work normally even under extreme working conditions. And the system can maintain good dynamic response and stability.

The invention relates to a system for realizing a voltage control attenuator based on a radio frequency broadband PIN diode, which comprises a first PIN diode, a second PIN diode, a first inductor, a second inductor, a first blocking capacitor and a second blocking capacitor, the first PIN diode and the second PIN diode are sequentially connected in series, one end of the first inductor is connected with the positive electrode of the first PIN diode, and the other end of the first inductor is connected with the negative electrode of the second PIN diode. The other end of the first inductor is connected with bias voltage input, one end of the second inductor is connected with the negative electrode of the second PIN diode, the other end of the second inductor is grounded, one end of the first blocking capacitor is connected with the positive electrode of the first PIN diode, and one end of the second blocking capacitor is connected with the negative electrode of the second PIN diode. According to the system for realizing the voltage control attenuator based on the radio frequency broadband PIN diode, the electrical modulation characteristic of the PIN diode is utilized, a microwave capacitor inductor is selected as a blocking bias device, the attenuator is simple in structure, short in response time and good in tracking performance, and the requirements of accuracy and stability can be met.

Provided is an image processing device, an optical communication device, an optical communication system, a recording medium, and an image processing method. An image processing device (5) includes a processing section that acquires images (44) that are continuous in time, acquires the position of a visible light source in an on state among the continuous images (44) as the position of a mobile body (3), and acquires identification information of the mobile body based on the flicker pattern of invisible light of an invisible light source provided to the mobile body (3).

The invention belongs to the technical field of permanent magnet synchronous motor tracking control, and particularly discloses a permanent magnet synchronous motor discrete finite time self-adaptive backstepping control method and a permanent magnet synchronous motor discrete finite time self-adaptive backstepping control system. The method comprises the following steps: firstly, establishing a discrete time system model of the permanent magnet synchronous motor, and designing a discrete finite time adaptive backstepping controller of the permanent magnet synchronous motor based on a minimum operator according to a backstepping method principle and in combination with an uncertain nonlinear function of an adaptive neural network technology processing system; then proving the finite time stability from the last subsystem to the first subsystem by using a Lyapunov stability theory, and estimating the stability time; and finally, enabling the permanent magnet synchronous motor to quickly track an expected signal by utilizing a discrete finite time self-adaptive backstepping controller of the permanent magnet synchronous motor based on a minimum operator. The feasibility of the method is verified through simulation and experimental results, and the designed controller has good tracking effect and robustness.

The invention discloses a photovoltaic power generation system voltage stability control method and system, and relates to the field of photovoltaic power generation system stability control. The method comprises the following steps: acquiring a real-time storage energy state parameter by using a dynamic model according to a real-time operation parameter of a nonlinear boost converter system; determining real-time mismatching interference and matching interference by using an extended disturbance observer; determining a real-time sliding mode variable in a sliding mode controller by using a new sliding mode function according to the real-time mismatching interference; mismatching interference estimation and exponential function terms are introduced into the new sliding mode function; according to the real-time sliding mode variable and the matching interference, the duty ratio of PWM is adjusted through the control law, the conduction time of a switch device in the nonlinear boost converter system is changed, and the output voltage of a capacitor in the nonlinear boost converter system is made to track the reference voltage of the nonlinear boost converter system. According to the invention, the voltage transient response rate and stability of the nonlinear boost converter system can be improved.

The invention discloses a multi-extended target joint tracking and classification method based on star convex RHM and LMB filters, and belongs to the field of radar target tracking. According to the method, an LMB parameter set is initialized by using target prior information, and then a sensor measurement set is divided through a mean shift algorithm; then, state prediction is achieved by combining survival target parameter updating and new target parameter sampling, and then measurement updating is completed through LMB-to-GLMB, GLMB updating and GLMB-to-LMB; and then trimming and fusing the LMB parameter set, estimating the number of targets, extracting state information, and circularly executing until observation is finished. According to the method, a star convex RHM modeling expansion state is adopted to reduce dimensions, the low detection probability / high clutter scene performance is improved based on an LMB framework, the tracking classification effect and the real-time performance are both considered, and the engineering application value is high.

The invention belongs to the field of radar signal processing, and discloses a joint resource scheduling and transmission zeroing method and system based on reinforcement learning. According to the method, in a scene that a plurality of transmitting nodes and a plurality of receiving nodes cooperatively execute a target tracking task in a receiving and transmitting separated MIMO radar network, aiming at the problem that a traditional radar resource allocation method lacks dynamic real-time adjustment capability and lacks comprehensive consideration on long-term performance of radar system resources, firstly, the dynamic real-time adjustment capability and the comprehensive consideration on the long-term performance of the radar system resources are considered; according to the method, joint optimization is carried out by considering transmission zero setting and resource scheduling, emission waveforms, residence time and power distribution of nodes are taken as optimization variables, then, cumulative signal-to-noise ratio constraints and the low interception probability of a whole task are considered at the same time, an optimization model for minimizing the interception probability is provided, and finally, for the optimization model, the optimization model is optimized. An algorithm based on reinforcement learning is adopted for solving, specifically, a PPO algorithm is adopted, and finally the low interception probability and the resource management efficiency are improved.

The application provides a three-dimensional gesture tracking method based on an RGB camera, which comprises the following steps: taking the RGB camera as an input, performing standardization processing on a picture, and sending the picture into a three-dimensional joint detection module to extract image features, predict a preliminary three-dimensional joint and calculate a hand skeleton length; using a particle swarm optimization algorithm to iteratively fit the hand skeleton length and a MAMO hand model, finding an optimal hand shape and obtaining a new three-dimensional joint; sending the new three-dimensional joint into an inverse kinematics module to solve posture parameters required by the MAMO model; using a MANO model to combine the optimal hand shape and the posture parameters θ to obtain final joints and grid vertices; and performing reconstruction rendering on the obtained grid vertices and joints to output a final three-dimensional hand real-time tracking posture. The method provided by the application has high prediction accuracy and high three-dimensional gesture tracking effect precision, and the tracking process and result are easy to realize.

The invention discloses a data processing method for an AR intelligent camera device, and the method specifically comprises the steps: obtaining original image data shot by the AR intelligent camera device in real time, and carrying out the processing of the original image data through employing a self-adaptive illumination compensation algorithm, and obtaining first image data; performing denoising processing on the first image data to obtain second image data; whether any athlete in the second image data is shielded or not is judged, if yes, complementation is carried out according to the historical track of the athlete, and third image data is formed; marking illegal behaviors in the third image data to form fourth image data; and putting the fourth image data into an AR information display template for fusion processing to generate AR display image data. According to the invention, the ornamental value and interactivity of sports event live broadcast can be effectively improved, high-quality and personalized AR enhanced experience is provided for audiences, and the satisfaction degree and participation degree of users are greatly improved.

The application discloses a kind of under the FDI network attack of reinforcement learning dual-motor safety coordination control method, comprising the following steps: (1) the double-time scale mathematical model of flexible coupling dual permanent magnet synchronous motor system is established;(2) the FDI network attack model of double-time scale mathematical model is established;(3) design safety coordination controller based on LQR theory;(4) according to the FDI network attack model of step (2) established, design model-free reinforcement learning, the safety coordination controller of step (3) is iteratively solved;The method avoids the influence caused by FDI network attack to system, using reinforcement learning algorithm can not depend on model data, avoid the influence of motor parameter change to system;For improving the coordination control performance of networked dual permanent magnet synchronous motor coordination control system and guarantee network security has important significance.

The invention relates to the technical field of drowning identification, and discloses an anti-drowning risk identification method and system based on machine vision. The method comprises the following steps: firstly, acquiring an initial swimming pool monitoring image, performing noise reduction processing by using a Mean Shift algorithm, and then performing adaptive multi-scale image enhancement to obtain a processed swimming pool monitoring image set; secondly, establishing a codebook background model by using a background difference method, comparing whether pixel points in the processed swimming pool monitoring image set are matched with the codebook background model or not, obtaining foreground points, and completing the detection of the swimmer; establishing a tracking target model to obtain a tracking target, and tracking the swimmer by calculating the distance of the tracking target; and finally, performing drowning identification on the swimmer in three aspects of the movement speed of the swimmer, the image saturation component value and the continuous disappearance frame number of the swimmer. According to the method, the swimming pool monitoring image is processed, the swimmer is detected and tracked, anti-drowning risk identification is realized, and the method is objective and accurate.

The present invention relates to a knowledge distillation ultrasound sequence tracking method based on optical flow assistance and feature accumulation, and belongs to the field of ultrasound image tracking. It includes the following steps: obtaining ultrasound sequence data, and dividing the data into a training set and a test set; constructing a knowledge distillation ultrasound sequence tracking model based on optical flow assistance and feature accumulation, the model includes a teacher network and a student network; the teacher network includes a feature encoder, a visual converter, a feature accumulation module and a prediction layer; the student network is a lightweight network based on the visual converter, including a projection layer, a visual converter and a prediction layer; the ultrasound sequence in the training set is respectively input into the teacher network and the student network for image analysis; the model is trained by the total loss of the teacher network and the student network to obtain a trained model; the ultrasound sequence in the test set is input into the trained model to obtain the final tracking result of the target. The present invention can improve the accuracy of ultrasound image target tracking.

The utility model discloses a key assembling and testing machine, which comprises a case formed by an upper rack and a lower rack, a pressing plate module and a carrier transferring module which are arranged in the upper rack, and a key pulling module and a visual detection module which are arranged in the lower rack, a key pulling test station A and a key pulling test station B are arranged between the feeding station and the discharging station, and a carrier moves among the stations through a carrier transferring module. The key assembling detection device has the advantages that the quality of the key is judged by pulling the fixed stroke of the four corners of the key in cooperation with visual detection, the requirement for detecting poor key assembling can be met, production merchants are helped to improve the product quality, the production cost is reduced, and the introduction difficulty of automation in the later period is reduced.

The invention belongs to the technical field of battery parameter identification, and particularly relates to a battery model parameter identification method and device, electronic equipment and a storage medium. A multi-information theory is introduced into the memory-limited recursive least square algorithm, an identification model is determined, the multi-information theory is used for obtaining information length data to serve as input of the identification model, and the memory-limited recursive least square algorithm is used for limiting the memory length of the identification model; based on the identification model, the real-time voltage and the real-time current, parameter identification is carried out on the second-order RC battery model, the identification model has better tracking performance and stability, the identification result can quickly approach and stably converge near a reference value, and a parameter identification result with a better effect can be obtained.

The invention discloses a filtering method based on probability data association in a clutter environment with unknown observation noise, belongs to the field of adaptive filtering, and provides a filtering method based on a probability data association algorithm for solving the problem that the observation noise in the clutter environment is difficult to estimate. The method aims to solve the problems that observation noise is difficult to estimate and the accurate position of a target is difficult to determine in a clutter environment. The core technology of the method comprises the following steps: tracking a target in a clutter environment by using a probability data association algorithm to obtain a target position, and using obtained information for observing noise calculation and filtering updating. Simulation results show that the method has good estimation performance in a clutter environment in which observation noise is far greater than process noise.

The utility model belongs to the technical field of human body track detection, and particularly relates to a microwave device human body track detection induction analysis device which comprises a shell mechanism, a first adjusting mechanism is installed above the shell mechanism, a second adjusting mechanism is connected to the first adjusting mechanism, a microwave detection device is installed on the second adjusting mechanism, and the microwave detection device is connected to the shell mechanism. A control mechanism is arranged on the side edge of the top end of the shell mechanism. Wherein the shell mechanism comprises an outer shell and a mounting platform rotationally connected to the outer shell; the first adjusting mechanism comprises a base fixedly installed on the installation platform and a bearing rod rotationally arranged in the base. The second adjusting mechanism comprises a supporting shell fixedly connected to the bearing rod, and through the arrangement of the shell mechanism, the first adjusting mechanism and the second adjusting mechanism, the microwave detection device can be subjected to deflection adjustment in the horizontal direction, height adjustment and pitching angle adjustment, so that the induction detection range of the microwave detection device is wider.

The invention discloses a magnetic bearing rotor vibration compensation method of a variable step-size LMS adaptive filter, which compensates for the imbalance of a six-pole radial hybrid magnetic bearing rotor, and makes a difference between a displacement difference at a current time n and a compensation signal value at a corresponding time n-1, respectively, to obtain instantaneous values ​​of errors in the X and Y directions at time n, and adopts a quasi-spike function model to calculate a step-size factor, and adaptively adjusts a weight coefficient based on the step-size factor to calculate the compensation signal values ​​at the current time n in the X and Y directions, and when the instantaneous error value is greater than or equal to an error threshold, makes a difference between the displacement difference and the compensation signal value at the current time n, to obtain a rotor vibration compensation difference; the invention adopts a variable step-size LMS algorithm, provides a longer step-size in the initial convergence stage of the algorithm, ensures that the algorithm has a faster convergence speed, and provides a shorter step-size in the convergence completion stage of the algorithm, so that the algorithm maintains better stability in the steady-state stage, ensures control accuracy, and makes the algorithm converge quickly and has low steady-state imbalance.

This invention provides a multi-target tracking algorithm based on spatial attributes, belonging to the field of computer vision technology. It solves the problem of poor tracking performance in complex situations. The technical solution includes the following steps: S1: Strengthening the positional correlation between the same target by adding the positional features of the target from the previous frame to the current frame to improve anti-occlusion capability; S2: Enhancing detection capability with spatial information by proposing global and local positional feature extraction modules in the detector part to obtain detailed spatial positional information between different targets; S3: Proposing a new calculation standard in the target matching part of the tracker, utilizing the spatial relationship between targets to make the matching method between intersecting targets more accurate; S4: Matching the calculated target anchor boxes and combining them with a re-identification model to obtain the final tracking result. The beneficial effects of this invention are: This method strengthens spatial information and makes it more adaptable to targets with complex movements, resulting in more stable tracking.

The application belongs to the technical field of electric-gas servo control, and discloses a control method for enhancing robust anti-interference of a voice coil motor direct drive system. S1, a dynamic equation, a model error and a state equation of the voice coil motor direct drive system are established; S2, the lumped interference is expanded into an additional state variable to update the state equation, a condition that an extended state anti-interference observer and an estimation error thereof satisfy is set on the basis of the updated state equation, and the estimation of the lumped interference is calculated by using the extended state anti-interference observer; S3, a tracking error and an auxiliary error signal are set, a controller for enhancing robust anti-interference is constructed, and a difference between the estimation of the lumped interference and the lumped error is compensated, so as to realize the control of the voice coil motor direct drive system. By the application, the selection limit of a robust gain parameter is reduced, and the control precision and the robustness of the system are ensured in the case that the system suffers from strong interference.

This invention relates to an AI vision-based method for monitoring static electricity release behavior. It solves the technical problem of lacking a method for monitoring whether static electricity is being released and whether the release behavior is standardized in industrial safety production. The method includes the following steps: 1) Initializing parameters and starting the algorithm service; 2) Acquiring scene images; 3) Sending the scene images to the algorithm detector for detection; 4) Deleting personnel data information not in the static electricity release area; 5) Obtaining tracking numbers; 6) Binding hand detection boxes to corresponding human body detection boxes; 7) Performing minimum pixel filtering; 8) Determining whether the remaining personnel are entering or leaving the workshop; 9) Obtaining the determination result of whether the remaining personnel are releasing static electricity; 10) Obtaining the determination result of whether the remaining personnel have completed releasing static electricity; 11) Obtaining the determination result of whether the remaining personnel have left the static electricity release area; 12) Determining whether an alarm is triggered; 13) Repeating steps 2) to 12) until the monitoring period ends.

The present invention relates to a current control system and method for a permanent magnet synchronous linear motor with multiple movers and independent windings. The D-Q axis given current and a motor speed signal of a motor winding are obtained, and the D-Q axis given current is transformed to obtain a three-phase given current; the three-phase given current and a three-phase current signal output by an inverter are differentially input into a proportional resonant current controller, and the proportional resonant current controller performs parameter adjustment based on the motor speed signal and outputs a modulation wave; the modulation wave generates a digital switch quantity through sinusoidal pulse width modulation and is output to an inverter, and the inverter outputs a three-phase current signal to the motor, and current closed-loop control is performed on the motor; the single-phase current control method based on the proportional resonant current controller enables a controlled object to achieve error-free current tracking under high-speed conditions.

The invention discloses a radar distance fuzzy point trace classification method based on a support vector machine, comprising: receiving radar multi-pulse echo complex data formed by scattering of all objects in a radar detection area, processing the radar echo complex data to obtain radar detection point trace data, obtaining logarithmic amplitude data of N pulses in a distance unit where the radar detection point trace is located according to the radar detection point trace data and the radar multi-pulse echo complex data; constructing and training a binary classification support vector machine; and applying the binary classification support vector machine to a radar detection point trace preprocessing link in radar signal processing to judge whether the radar detection point trace belongs to a distance fuzzy point trace, thereby improving the radar's suppression performance on the distance fuzzy point trace.

The application discloses a dynamic liquid level detection system and method based on a stirred tank, and belongs to the field of liquid level detection. The system comprises a differential pressure sensor module, a rotating speed acquisition module and a signal processing unit. The signal processing unit comprises a vortex dynamic compensation module, an adaptive filtering module and a delay prediction compensation module connected in series. The application collects the differential pressure signal of the liquid column in the tank and converts it into an apparent liquid level value, and simultaneously collects the rotating speed signal of the stirring paddle. The liquid level deviation caused by the vortex at the tank wall is calculated according to the rotating speed signal, and a first-order recursive tracking output dynamic compensation is used to eliminate the vortex liquid level deviation. The low-pass filter cutoff frequency is adaptively determined according to the rotating speed signal to suppress the stirring fluctuation noise. The change rate of the filtered liquid level value and the group delay are used for forward prediction correction to reduce the filtering delay. The application realizes dynamic cooperation of each link by using the rotating speed signal as a link, and improves the accuracy and response speed of the liquid level measurement under the stirring condition.