CN116796110A - Tea leaf storage environment state estimation method, system and storage medium - Google Patents
Tea leaf storage environment state estimation method, system and storage medium Download PDFInfo
- Publication number
- CN116796110A CN116796110A CN202310817685.8A CN202310817685A CN116796110A CN 116796110 A CN116796110 A CN 116796110A CN 202310817685 A CN202310817685 A CN 202310817685A CN 116796110 A CN116796110 A CN 116796110A
- Authority
- CN
- China
- Prior art keywords
- state
- value
- input
- real
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 58
- 238000003860 storage Methods 0.000 title claims abstract description 46
- 241001122767 Theaceae Species 0.000 claims abstract description 40
- 238000005259 measurement Methods 0.000 claims abstract description 33
- 238000001914 filtration Methods 0.000 claims abstract description 30
- 239000013598 vector Substances 0.000 claims description 23
- 238000005070 sampling Methods 0.000 claims description 22
- 230000008569 process Effects 0.000 claims description 20
- 239000011159 matrix material Substances 0.000 claims description 14
- 238000012937 correction Methods 0.000 claims description 8
- 238000006467 substitution reaction Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 238000013016 damping Methods 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 230000008901 benefit Effects 0.000 abstract description 5
- 235000013616 tea Nutrition 0.000 description 27
- 238000012544 monitoring process Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 10
- 238000012545 processing Methods 0.000 description 10
- 238000001514 detection method Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- 235000019224 Camellia sinensis var Qingmao Nutrition 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 235000020339 pu-erh tea Nutrition 0.000 description 4
- 230000001133 acceleration Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 238000000354 decomposition reaction Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000000813 microbial effect Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000855 fermentation Methods 0.000 description 1
- 230000004151 fermentation Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Abstract
The application discloses a method for estimating the state of a tea storage environment, which comprises the following steps: s100, obtaining a predicted system dynamic state of a tea storage environment according to an extended Kalman filter UKF; s200, acquiring an input predicted value according to the dynamic state of a prediction system and the real-time parameters of the system; s300, correcting an input predicted value according to the real-time measurement state and the real-time parameters to obtain a final estimated value; s400, obtaining an output value according to the improved sliding average filtering. The application has the advantage of providing a method for acquiring the state estimation of the input parameters of the Kalman filter based on real-time input data, wherein the unknown input has two-stage estimation in each step. The first predicted dynamic state and system parameters provide an estimate of the input. And secondly, obtaining a final estimated value through measuring the state and correcting the parameters. And finally, the size of the optimal sliding window is obtained according to the combination of the estimated value and the measured value, so that the interference of fluctuation obvious data is effectively reduced.
Description
Technical Field
The present application relates to the field of tea leaf storage environment state estimation, and more particularly, to a tea leaf storage environment state estimation method, system and storage medium.
Background
The tea quality is formed without leaving proper storage environment and storage time. For example, pu 'er tea, the tea aroma of Pu' er tea is released optimally when the humidity value is 45-65%, and after the humidity exceeds 70%, the aroma released by the tea can be absorbed in a large amount by air humidity, so that the Pu 'er tea aroma release is accelerated, and the Pu' er tea aroma release is unfavorable. After the microbial growth rate exceeds 80%, the microbial growth rate is gradually accelerated to enable the puer tea to be subjected to fermentation of external decomposition, so that mildewing and deterioration are caused. At present, most of the monitoring of the tea storage environment by users adopts a mode of collecting data by a sensor and transmitting the data through a wireless network. However, in a wireless network, when the sensor collects data for transmission, the sensor is affected by signal fluctuation. It is therefore necessary to study the effect of interference data in the sensor acquisition data.
Currently, the prior art mostly adopts a kalman filter to filter the collected data. Most of the existing Kalman filtering, extended Kalman filtering and the like, UKF is extended Kalman filtering, is extended based on Kalman filtering, and is applied to a nonlinear system under the linear assumption by utilizing lossless transformation. In practical application, there is a Jacobian matrix which needs to calculate a nonlinear model, the calculated amount is large, errors are easy to occur, an optimal solution is difficult to obtain, a high-order term is ignored, estimation accuracy is affected, robustness of model uncertainty is poor, and when a system reaches a stable state, the tracking capability of an abrupt state is lost. If the error propagation function of the system does not work well with a linear function, it may lead to filter spread.
Disclosure of Invention
The summary of the application is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. The summary of the application is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.
Some embodiments of the present application provide a method, a system and a storage medium for estimating the state of a tea storage environment, so as to solve the technical problems mentioned in the background section above. A method for estimating the state of tea storage environment includes such steps as obtaining the state of input parameters of Kalman filter based on real-time input data, and two stages of unknown input. The first predicted dynamic state and system parameters provide an estimate of the input. And secondly, obtaining a final estimated value through measuring the state and correcting the parameters. Compared with the traditional assumed known system state, based on the method, all dynamic states in the monitoring system are estimated in real time, and the output sampling value is more in line with the actual value of the detection environment. And finally, the size of the optimal sliding window is obtained according to the combination of the estimated value and the measured value, so that the interference of fluctuation obvious data is effectively reduced. The collected monitoring data are higher in accuracy and more stable, and a user can control the tea storage environment more accurately based on the collected value, so that the tea with better quality is obtained.
As a first aspect of the present application, some embodiments of the present application provide a method for estimating a state of a tea storage environment, including: s100, obtaining a predicted system dynamic state of a tea storage environment according to an extended Kalman filter UKF; s200, acquiring an input predicted value of the system according to the dynamic state of the prediction system and the real-time parameters of the system; s300, correcting and inputting an estimated value according to the real-time measurement state and the real-time parameters of the system to obtain a final estimated value of the system; s400, obtaining the output value of the system according to the improved sliding average filtering. Wherein, according to the output value of the improved sliding average filtering acquisition system comprises: acquiring a sampling value and a final estimated value of a tea storage environment at the moment t; acquiring the magnitude of an improved sliding mean value according to the sampling value and the final estimated value; obtaining an output value according to the improved sliding average value; wherein, the expression for improving the sliding average is as follows:
wherein m is the acquisition times, O is the size of the sliding average window, O i To obtain the sliding window size according to the final estimated value, O c For the sliding window size obtained from the measurement.
wherein ,Vn Output value at time t, V i The output value of the last time of the t time is n times of sampling.
Further, the step of S100 obtaining the dynamic state of the prediction system according to the UKF comprises the following steps: s101, acquiring a dynamic state expected value and a covariance value of a system at a moment t; s102, acquiring dynamic state estimation and covariance estimation of a system according to a dynamic state expected value and a covariance value; s103 executes S102 until an optimal value of the dynamic state estimation is obtained, and the system dynamic state corresponding to the optimal value is set as the predicted system dynamic state.
Further, S200 includes: s201, acquiring a corrected dynamic state expected value and a corrected covariance value of the UKF of the system according to a real-time parameter at the time t; s202, estimating a corrected dynamic state and a corrected covariance according to a corrected dynamic state expected value and a corrected covariance value system; s203 executes S202 until an optimum value of the corrected dynamic state estimation is acquired, and the optimum value is set as the input predicted value.
Further, S300 includes: s301, acquiring equivalent substitute data of error data in real-time parameters according to the real-time measurement state; s302, correcting the input predicted value according to the equivalent substitution data replacement error data to obtain a final estimated value.
Further, the expression of UFK is as follows:
z(t)＝fz(t),u(t)+ν(t)
y(t)＝Hz(t),u(t)+η(t)
z k ＝F(z k-1 ,u k-1 )+ν k-1
y k ＝h(z k ,u k )+η k
wherein z (t) is a state vector, < >>θ is a parameter of the system, and y (t) is an observation vector; u (t) is an input vector, v (t) is process noise, and η (t) is measurement noise; f (x) is a state transition function, h (x) is a state observation function, f (x) and h (x) take into account the input vector u (t); q (t) is a covariance matrix of v (t), and R (t) is a covariance matrix of eta (t); k is kΔt is the time instant and Δt is the sampling period.
Further, the solving process of the dynamic state of the prediction system obtained by the S100 according to the UKF is as follows:
k=0 (time step)
z k ＝E[z 0 ]
P k ＝E[(z 0 -z k )(z 0 -z k ) T ]
Wherein E represents an expected value; p (P) k Representing the covariance matrix of the image and,
k＝k+1
Z p ＝F(z k-1 ,u k1 )
wherein p represents prediction
Y i ＝h[Z i,p ,u k ]
When (when)Is the measurement vector of time step k:
wherein ,Zk Is a state estimate; p (P) k Is a state estimate;
go to S102 until k=k max
Further, the solving process of the S200 for obtaining the input predicted value according to the dynamic state of the prediction system and the real-time parameters is as follows:
k=0 (time step)
z k ＝E[z 0 ]
P k ＝E[(z 0 -z k )(z 0 -z k ) T ]
k＝k+1
Wherein G (x) relates to the estimated parameters and each step is updated to replace all known parametersA known input line of zero/non-zero;
when (when)Is the measurement vector of time step k:
wherein ,Zk Is a state estimate; p (P) k Is a state estimate;estimating for the final input; go to S202 until k=k max 。
Further, the solving process of the final estimated value of the input estimated value obtaining system according to the real-time measurement state and the real-time parameter correction in S300 is as follows:
wherein in case of a wrong damping parameter c, stiffness parameter k or input u (t), the equation of motion is modified as:
when deltau (t) =0,
when ak (t) =0,
when deltac=0, the current value,
substituting c+Δc, k+Δk, and u (t) +Δu (t) for the true values, Δc, Δk, and Δu (t) represent erroneous portions;
this represents that even if the error portion of one unknown is minimized, there is a second error combination:
wherein ,
the linear equation of motion thus modified, assuming zero at the known input of the system, is modified as follows:
as a second aspect of the present application, some embodiments of the present application provide a tea leaf storage environment state estimation system, comprising: the prediction system dynamic state acquisition module is used for acquiring the prediction system dynamic state of the tea storage environment according to the extended Kalman filtering UKF; the input pre-estimation value acquisition module is used for acquiring an input pre-estimation value of the system according to the dynamic state of the prediction system and the real-time parameters of the system; the final estimation value acquisition module is used for correcting the input estimated value according to the real-time measurement state and the real-time parameter of the system to acquire the final estimation value of the system.
As a third aspect of the present application, some embodiments of the present application provide a computer storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.
The application has the beneficial effects that: a method for obtaining state estimation of input parameters of a Kalman filter based on real-time input data, wherein the unknown input has two-stage estimation in each step. The first predicted dynamic state and system parameters provide an estimate of the input. And secondly, obtaining a final estimated value through measuring the state and correcting the parameters. Compared with the traditional assumed known system state, the system based on the method has the advantages that all dynamic states are estimated in real time in a combined mode, the output sampling value better accords with the actual value of the detection environment, and the acquired monitoring data is higher in accuracy.
More specifically, some embodiments of the present application may have the following specific benefits:
based on the joint estimation of the dynamic state of the system and the real-time input parameters, the influence of the initial condition of the system is overcome, and the phenomenon of data divergence is avoided.
Furthermore, assumptions of system state are mostly empirical, so that the environment in which the assumed known system state parameters can be applied is limited. The application is based on the joint estimation of the real-time input parameters and the dynamic state of the system, and can be suitable for different monitoring environments.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application, are incorporated in and constitute a part of this specification. The drawings and their description are illustrative of the application and are not to be construed as unduly limiting the application.
In addition, the same or similar reference numerals denote the same or similar elements throughout the drawings. It should be understood that the figures are schematic and that elements and components are not necessarily drawn to scale.
In the drawings:
FIG. 1 is a schematic diagram of the main steps of a system for improving the quality of stored tea based on improved Kalman filtering according to one embodiment of the application;
FIG. 2 is a schematic diagram of obtaining the predictive system dynamic state from the UKF in accordance with one embodiment of the application;
FIG. 3 is a schematic diagram of the input predictive value being obtained from the predictive system dynamic state and the real-time parameters in accordance with one embodiment of the application;
FIG. 4 is a schematic diagram of the system's final estimate obtained by modifying the input pre-estimate based on the real-time measurement status and the real-time parameters, according to one embodiment of the application;
fig. 5 is a schematic diagram of a system utilizing equations of motion instead of acceleration in accordance with an embodiment of the present application.
Detailed Description
Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete. It should be understood that the drawings and embodiments of the present disclosure are for illustration purposes only and are not intended to limit the scope of the present disclosure.
It should be noted that, for convenience of description, only the portions related to the present application are shown in the drawings. Embodiments of the present disclosure and features of embodiments may be combined with each other without conflict.
It should be noted that the concepts of "first," "second," etc. mentioned in this disclosure are merely used to distinguish between different devices, modules, or units, and are not used to define an order or interdependence of functions performed by such devices, modules, or units.
It should be noted that references to "a" and "an" in this disclosure are intended to be illustrative rather than limiting, and those of ordinary skill in the art will appreciate that "one or more" is intended to be understood as "one or more" unless the context clearly indicates otherwise.
The names of messages or information interacted between the various devices in the embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.
The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.
The Kalman filter uses the input and output of the system as data to construct a state estimator for estimating the state of the random system to minimize the steady state error covariance matrix. The basic purpose of the kalman filter is to reconstruct the states that are actually needed but cannot be measured, while minimizing the effect of noise on state reconstruction. There are three types of kalman filtering in common use: the optimal Kalman filter is mainly applied to a linear time-varying system; sub-optimal Kalman filtering is mainly applied to a linear steady system; the extended Kalman filtering is mainly applied to a nonlinear system. In the nonlinear system filtering field, the expansibility Kalman filtering is an important tool in the modern signal processing field and is applied to engineering calculation problems of signal processing and signal prediction estimation. However, in practical applications, like extended kalman filtering, robustness with respect to model uncertainty is poor, and state estimation inaccuracy, even divergence, and the like easily occur due to sensitivity to initial conditions.
In the method, the improved extended Kalman filter is adopted to process the data with larger interference in the data collected by the sensor. The improved extended Kalman filtering mainly uses a method of jointly estimating a system dynamic state and real-time input parameters to replace the assumed known system state in the traditional extended Kalman filtering. This approach enables the output data to better conform to the current monitored system because all system dynamics and real-time input parameters are jointly estimated. Specifically, the unknown input has two phase estimates in each step. First, the predicted system dynamics and real-time parameters of the system provide an input estimate. Second, the final estimate is provided by measurement state and parameter correction.
Figure 1 is a schematic diagram of the main steps of a system for improving the quality of stored tea based on improved kalman filtering according to an embodiment of the present application, as shown in figure 1, the main steps of a system for improving the quality of stored tea based on improved kalman filtering are as follows: s100, obtaining a predicted system dynamic state of a tea storage environment according to an extended Kalman filter UKF; s200, acquiring an input predicted value of the system according to the dynamic state of the prediction system and the real-time parameters of the system; s300, correcting and inputting an estimated value according to the real-time measurement state and the real-time parameters of the system to obtain a final estimated value of the system; s400, obtaining the output value of the system according to the improved sliding average filtering.
Wherein, according to the output value of the improved sliding average filtering acquisition system comprises: acquiring a sampling value and a final estimated value of a tea storage environment at the moment t; acquiring the magnitude of an improved sliding mean value according to the sampling value and the final estimated value; obtaining an output value according to the improved sliding average value; wherein, the expression for improving the sliding average is as follows:
wherein m is the acquisition times, O is the size of the sliding average window, O i To obtain the sliding window size according to the final estimated value, O c For the sliding window size obtained from the measurement.
wherein ,Vn Output value at time t, V i The output value of the last time of the t time is n times of sampling.
For continuous sampling data, the sliding average value is defined as a certain temporary storage length of the sampling data, new measurement is carried out every time, newly acquired data are placed on the right side of a queue, first data on the leftmost side of an original queue are deleted, then the average value of all data in the queue is calculated, and the average value is used as an output value after filtering. Therefore, the method combines the sliding window of the estimated value and the measured value to find the optimal sliding window, obtains the final output value according to the optimal sliding window, has better filtering effect on the data with obvious disturbance in the acquired data, and has better average effect.
A method for obtaining state estimates of input parameters of a Kalman filter based on real-time input data, the unknown input having two phase estimates in each step. The first predicted dynamic state and system parameters provide an estimate of the input. And secondly, obtaining a final estimated value through measuring the state and correcting the parameters. Compared with the traditional system state which is assumed to be known, in the system based on the method, all dynamic states and real-time input parameters are estimated in a real-time combined mode, an output sampling value is more in line with an actual value of a detection environment, the acquired monitoring data is higher in precision, and a user can control a tea storage environment more accurately based on the acquired value, so that tea with better quality is obtained.
FIG. 2 is a schematic diagram of obtaining the dynamic state of the prediction system according to the UKF according to one embodiment of the present application, as shown in FIG. 2, S100 obtains the dynamic state of the prediction system of the tea storage environment according to the extended Kalman filter UKF;
consider first the nonlinear process equations in continuous time domain and state space formats:
z(t)＝fz(t),u(t)+ν(t) (3)
nonlinear observation equation:
y(t)＝Hz(t),u(t)+η(t) (4)
wherein ：
z (t) is the state vector and,θ is a parameter of the system, and y (t) is an observation vector; u (t) is an input vector, v (t) is process noise, and η (t) is measurement noise;
f (x) is a state transition function, h (x) is a state observation function, f (x) and h (x) take into account the input vector u (t); q (t) is a covariance matrix of v (t), and R (t) is a covariance matrix of eta (t); formulas (3) and (4) can be discretized into:
z k ＝F(z k-1 ,u k-1 )+ν k-1 (5)
y k ＝h(z k, u k )+η k (6)
k is kΔt is the time instant, Δt is the sampling period; the discretization process and the observed covariance matrix are as follows:
in the present method, the first predicted dynamic state and system parameters provide an estimate of the input. And secondly, obtaining a final estimated value through measuring the state and correcting the parameters. Compared with the traditional assumed known system state, the real-time input parameters are estimated in real time in a combined way based on all dynamic states in the system of the method, and the output sampling value is more in line with the actual value of the detection environment. Wherein λ is represented by λ=α 2 (l+k) -L, k=0 or 3-L, wherein L is z k And (3) the dimension of (c). Constant (constant)
α∈[10 -4 ，1]The weight V is obtained from the following formula:
where β is a constant containing a priori information z of z k Is a distribution of (a).
Unknown inputs can be estimated simultaneously using UKF, with the predicted state of time step k, using the continuous equation of motion for time instant kΔt to estimate the inputs:
where G (x) is a linear or nonlinear function, G (x) also contains the estimated parameters, so G (x) is updated iteratively at each step. In addition, the predicted state is determined by equation Z p ＝F(z k-1 ,u k-1 ) Estimated. Known input linesIs replaced by a known zero or non-zero input, which uses the measured value Y in the updating process i ＝H k Z i,p 。
However, this estimatedIs erroneous because the measured state has not been updated with the measured value and the system parameters have not been updated. Thus, eta k Both the model of the measured noise and the model of the input derived noise. The final step is to further correct the input estimate with updated measured dynamic states and parameters:
then the known input line is replaced with the final oneThis final +.>Is used continuously in the next predictive calculation of step k+1, and the whole process is iterated.
Assuming zero gaussian noise, the extended kalman filter (UKF) computing system dynamic state steps are as follows without loss of generality:
s101, acquiring a dynamic state expected value and a covariance value of a system at a moment t;
k=0 (time step)
z k ＝E[z 0 ] (13)
P k ＝E[(z 0 -z k )(z 0 -z k ) T ] (14)
Wherein E represents an expected value; p (P) k Representing a covariance matrix;
s102, acquiring dynamic state estimation and covariance estimation of a system according to a dynamic state expected value and a covariance value;
k＝k+1
wherein p represents a prediction;
when (when)Is the measurement vector of time step k:
wherein ,Zk Is a state estimate; p (P) k Is a state estimate;
s103 performs S102 until the optimal value of the dynamic state estimation is acquired.
Go to S102 until k=k max 。
FIG. 3 is a schematic diagram of the input pre-estimation values obtained according to the dynamic state of the prediction system and the real-time parameters according to one embodiment of the present application, as shown in FIG. 3, S200 obtains the input pre-estimation values of the system according to the dynamic state of the prediction system and the real-time parameters of the system; s201, acquiring a correction dynamic state expected value and a correction covariance value according to a real-time parameter at the time t;
k=0 (time step)
z k ＝E[z 0 ]
P k ＝E[(z 0 -z k )(z 0 -z k ) T ]
S202, estimating a corrected dynamic state and a corrected covariance according to a corrected dynamic state expected value and a corrected covariance value system;
k＝k+1
wherein G (x) relates to the estimated parameters and each step is updated to replace all known parametersA known input line of zero/non-zero;
when (when)Is the measurement vector of time step k:
wherein ,Zk Is a state estimate; p (P) k Is a state estimate;estimating for the final input; s203 performs S202 until the optimum value of the corrected dynamic state estimation is acquired. Go to S202 until k=k max
FIG. 4 is a schematic diagram of obtaining a final estimated value of the system according to the real-time measurement state and the real-time parameter correction input predicted value according to one embodiment of the present application, as shown in FIG. 4, S300 obtains a final estimated value of the system according to the real-time measurement state and the real-time parameter correction input predicted value of the system; s301, checking the identifiability of the real-time measurement state of the system; when the input is known, the parameter vector is checked by local recognizability; when the input is unknown, the parameter vector is checked by perturbation analysis. Checking the recognizability of the parameter vector by local recognizability without measurement noise for a linear system when the input is known; for nonlinear systems, the local recognizability and thus the recognizability of the parameter vector is checked with an observability class condition. When the input is unknown, the parameter vector is checked by perturbation analysis.
Specifically, S301 obtains equivalent substitute data of error data in the real-time parameters according to the real-time measurement state; in the case of incorrect damping parameters c, stiffness parameters k or inputs u (t), the sensitivity of the recognition program is investigated. The true values are replaced by c+Δc, k+Δk and u (t) +Δu (t), where Δc, Δk and Δu (t) represent the wrong parts. By doing so, the equation of motion is modified to:
when deltau (t) =0,
when ak (t) =0,
when deltac=0, the current value,
this means that even if the erroneous portion of one unknown quantity is minimized, there is a second erroneous combination which can stably satisfy the equality in the equation of motion. In addition:
wherein ,/>
by identifying the different erroneous parts, an equivalent "input" u (t) can always be found. As a sequence, a system with erroneous "inputs" u (t), whose quality and dynamic state are identical, can be identified.
Without known inputs, existing solutions, such as frequency domain decomposition, the inputs or dynamic estimates are not estimated, whereas linear systems are also only processed, especially if the frequency domain decomposition is off-line, making automation difficult.
However, it may be identified in a system with known zero or non-zero inputs, assuming c+Δc, k+Δk, and u (t) +Δu (t), for disturbance analysis of a linear two-dimensional system, where Δc, Δk, and Δu (t) represent erroneous parts.
S302, correcting the input predicted value according to the equivalent substitution data substitution error data to obtain a final estimated value;
the linear equation of motion thus modified, assuming zero at the known input of the system, is generally modified as follows:
from the above formula, it can be seen that the equivalent "input" u 1 (t) bringing about that whether the system knows that the input is zero or non-zero, it is possible to cause the true parameters, and thus the true input, to be correctly identified, as well as for non-linear systems.
FIG. 5 is a schematic diagram of a system using equations of motion instead of acceleration according to one embodiment of the present application, as shown in FIG. 5, below with an initial condition of x (0) = [ 00 00] T ,A pulse of 100N was applied for 0.01s at a time point of 5s, which was not known in advance. To create the composite measurement, the system response for 30s was calculated using the Runge Kutta method, 4 th order integration method. The sampling frequency of the dynamic state measurement is generally known as 100 hz, so the time discretization Δt used in the range Kutta numerical solution is 0.01s. Finally, to account for measurement noise, each response signal is contaminated with a gaussian white noise sequence, the ratio of noise to signal taking a root mean square of 5%. White noise is noise with a mean value of zero. In discrete time, the recursive form of the system is solved by the following formula:
/>
substitution of acceleration by equation of motion (30)The following formula (35) is obtained:
the error vector η is not included in the formula (6) k The formula of the error:
wherein for unknown inputs
The predicted state may be determined from the formulaObtained in the middle (.) m and (·)p Representing measurement and prediction, respectively, the predicted state can be derived from equation (36). The c and k matrices relate to estimated parameters, updated iteratively in each step. All that is used in the formula is the case of state measurement, process covariance Q k-1 And measuring covariance R k The matrix is identified in the processSetting to constant, respectively taking the value of 10 -9 ·I 12×12 and 10-3 ·I 9×9
Specifically, the equation of motion is used instead of accelerationThe method overcomes the defects that the traditional extended Kalman filtering formula has larger truncation error caused by linearization processing when processing a nonlinear system, and does not need to solve a complex Jacobian matrix, is convenient for processing nonlinear signals, saves the amount of calculation, and improves the calculation speed. The first predicted dynamic state and system parameters provide an estimate of the input. And secondly, obtaining a final estimated value through measuring the state and correcting the parameters. Compared with the traditional assumed known system state, the system based on the method has the advantages that all dynamic states are estimated in real time in a combined mode, the output sampling value better accords with the actual value of the detection environment, and the acquired monitoring data is higher in accuracy. The improved extended kalman filter is therefore used for tea storage environment monitoring data processing.
Taking environmental monitoring data of a certain tea storage system as an example:
and 48 humidity data acquired by the temperature and humidity sensor every half an hour in one day are compared with data output by an improved extended Kalman filter equation and data acquired by a precise instrument. The system processor adopts STM32, wherein STM32 is a controller kernel manufactured by ST (STMicroelectronics, semiconductor manufacturing method) company, and mainly comprises a serial port transceiver program, a 4G wireless communication program, a TFT display screen program, an SD card data storage program, an alarm program and an environment regulation program. STM32 is the core of the system, which processes, displays, stores data received from the data processing device, and also receives instructions set by the user via the display screen to set different storage environment monitor values. More specifically, a STM32F103 chip is adopted, a 32-bit ARM Cortex-M3 microprocessor is used, and the STM32F103 chip is integrated between a 2.4GHz transceiver and a 2.4GHz IEEE802.15.4 transceiver. Flash, RAM RAM and built-in interface based on gateway system. The largest chip is one that maintains low power consumption while improving processing power. Different versions of proxy protocols are reserved in the chip, and users can develop products by using network protocols conforming to the network. The improved extended kalman filter uses CC2530 manufactured by TI (Texas Instruments ). The CC2530 comprises a chip, a data receiving module, a data processing module, a radio frequency antenna and a power module, is a system-on-chip solution, can establish a node with strong performance in the ZigBee wireless sensor network by using very low material cost, and has low power consumption, and is stable and reliable to meet the requirements of system design. More specifically, the chip is a core unit of the whole data processing device, and is responsible for filtering received collected data according to an improved kalman formula, and transmitting the processed data to a system processor through a radio frequency antenna. The temperature and humidity sensor adopts DHT11, and adopts the data of a temperature and humidity recorder of a clear-of-precision company as a standard for comparison, and the humidity error of the precise instrument is +/-0.1%. The acquired data is subjected to an improved extended kalman filter process. As the results are shown in table one, it can be seen that the improved extended kalman filter formula provides a higher accuracy than the data collected by conventional sensors, with the first predicted dynamic state and system parameters providing an estimate of the input. And secondly, obtaining a final estimated value through measuring the state and correcting the parameters. Compared with the traditional assumed known system state, the system based on the method has the advantages that all dynamic states are estimated in real time in a combined mode, the output sampling value better accords with the actual value of the detection environment, and the acquired monitoring data is higher in accuracy. For continuous sampling data, the sliding average value is defined as a certain temporary storage length of the sampling data, new measurement is carried out every time, newly acquired data are placed on the right side of a queue, first data on the leftmost side of an original queue are deleted, then the average value of all data in the queue is calculated, and the average value is used as an output value after filtering. Therefore, the method combines the sliding window of the estimated value and the measured value to find the optimal sliding window, obtains the final output value according to the optimal sliding window, has better filtering effect on the data with obvious disturbance in the acquired data, and has better average effect.
Table one stores the ambient humidity monitoring value at each moment
/>
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises an element.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
While the application has been described above with reference to various embodiments, it should be understood that many changes and modifications can be made without departing from the scope of the application. It is therefore intended that the foregoing detailed description be regarded as illustrative rather than limiting, and that it be understood that it is the following claims, including all equivalents, that are intended to define the spirit and scope of this application. The above examples should be understood as illustrative only and not limiting the scope of the application. After reading the description of the application, the skilled person may make various changes or modifications to the application, which equivalent changes and modifications also fall within the scope of the application as defined in the claims.
Claims (10)
1. The method for estimating the state of the tea storage environment is characterized by comprising the following steps:
s100, obtaining a predicted system dynamic state of a tea storage environment according to an extended Kalman filter UKF;
s200, acquiring an input predicted value of the system according to the dynamic state of the prediction system and the real-time parameters of the system;
s300, correcting the input predicted value according to the real-time measurement state and the real-time parameters of the system to obtain a final estimated value of the system;
s400, obtaining an output value of the system according to improved sliding average filtering;
wherein obtaining the output value of the system according to the improved sliding average filtering comprises:
acquiring a sampling value and the final estimated value of the tea storage environment at the moment t;
acquiring the size of the improved sliding mean value according to the sampling value and the final estimated value;
acquiring the output value according to the improved sliding average value;
wherein the expression of the improved sliding average is as follows:
wherein m is the acquisition time number,for the size of the sliding mean window, +.>For the sliding window size obtained from the final estimate +.>To be based on the measured valueThe resulting sliding window size.
2. A method for estimating a tea leaf storage environment state as claimed in claim 1, wherein:
s100, obtaining the dynamic state of the prediction system according to the UKF comprises the following steps:
s101, acquiring a dynamic state expected value and a covariance value of the system at a moment t;
s102, acquiring dynamic state estimation and covariance estimation of the system according to the dynamic state expected value and the covariance value;
s103, executing S102 until the optimal value of the dynamic state estimation is obtained, and setting the system dynamic state corresponding to the optimal value as the prediction system dynamic state.
3. A method for estimating a tea leaf storage environment state as claimed in claim 1, wherein:
s200, obtaining the input predicted value according to the dynamic state of the prediction system and the real-time parameters comprises the following steps:
s201, acquiring a correction dynamic state expected value and a correction covariance value of the UKF of the system according to the real-time parameter at the moment t;
s202, estimating the corrected dynamic state and the corrected covariance of the system according to the expected value of the corrected dynamic state and the corrected covariance value;
s203 executes S202 until an optimal value of the revised dynamic state estimate is acquired, the optimal value being set as an input predicted value.
4. A method for estimating a tea leaf storage environment state as claimed in claim 1, wherein:
s300, correcting the input estimated value according to the real-time measurement state and the real-time parameter to obtain a final estimated value of the system comprises the following steps:
s301, acquiring equivalent substitution data of error data in the real-time parameters according to the real-time measurement state;
s302, correcting the input predicted value according to the equivalent substitution data to obtain the final estimated value.
5. A method for estimating a tea leaf storage environment state as claimed in claim 1, wherein:
the expression of UFK is as follows:
wherein ,is a state vector +.>，/>Is a parameter of the system->For observing vector, +.>For inputting vectors, ++>For process noise->For measuring noise +.>For the state transition function +.>For the state observation function +.>Andconsider the input vector->，/>Is->Covariance matrix of>Is->Covariance matrix of>Is->Time instant of (2)>Is the sampling period.
6. A method for estimating a tea leaf storage environment state as claimed in claim 2, wherein:
the solving process of the S100 for obtaining the dynamic state of the prediction system according to the UKF is as follows:
wherein ,representing the expected value; />Representing the covariance matrix of the image and,
wherein ,representing the prediction of the time of day,
when (when)Is a time step->When the vector is measured:
wherein ,is a state estimate; />Is a state estimate;
go to S102 until。
7. A method for estimating a tea leaf storage environment state as claimed in claim 3, wherein:
the step S200 is to obtain the solving process of the input predicted value according to the dynamic state of the prediction system and the real-time parameters as follows:
wherein ,involves estimating parameters and each step is updated to replace all known +.>A known input line of zero/non-zero;
when (when)Is a time step->When the vector is measured:
wherein ,is a state estimate; />Is a state estimate;
estimating for the final input;
go to the S202 until。
8. A method of estimating a state of a tea leaf storage environment as claimed in claim 4, wherein:
the step S300 is to correct the input estimated value according to the real-time measurement state and the real-time parameter to obtain the final estimated value of the system, wherein the solving process is as follows:
wherein in case of incorrect damping parametersRigidity parameter->Or input->In the case of (2), the equation of motion is modified to:
wherein ,
the linear equation of motion thus modified, assuming zero at the known input of the system, is modified as follows:
wherein use is made ofInstead of the true value, +.>A part representing an error;
this means that even if the error part of one unknown is minimized, there is a second error combination.
9. An estimation system of the state of a tea storage environment is characterized in that:
the prediction system dynamic state acquisition module is used for acquiring the prediction system dynamic state of the tea storage environment according to the extended Kalman filtering UKF;
the input pre-estimation value acquisition module is used for acquiring an input pre-estimation value of the system according to the dynamic state of the prediction system and the real-time parameters of the system;
the final estimation value acquisition module is used for correcting the input estimated value according to the real-time measurement state and the real-time parameter of the system to acquire the final estimation value of the system.
10. A computer readable medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the method of any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310817685.8A CN116796110A (en) | 2023-07-05 | 2023-07-05 | Tea leaf storage environment state estimation method, system and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310817685.8A CN116796110A (en) | 2023-07-05 | 2023-07-05 | Tea leaf storage environment state estimation method, system and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116796110A true CN116796110A (en) | 2023-09-22 |
Family
ID=88034366
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310817685.8A Pending CN116796110A (en) | 2023-07-05 | 2023-07-05 | Tea leaf storage environment state estimation method, system and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116796110A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105607009A (en) * | 2016-02-01 | 2016-05-25 | 深圳大学 | Power battery SOC estimation method and system based on dynamic parameter model |
| CN111223209A (en) * | 2020-01-10 | 2020-06-02 | 北京理工大学 | Vehicle driving parameter and state parameter joint estimation method based on parallel structure |
-
2023
- 2023-07-05 CN CN202310817685.8A patent/CN116796110A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105607009A (en) * | 2016-02-01 | 2016-05-25 | 深圳大学 | Power battery SOC estimation method and system based on dynamic parameter model |
| CN111223209A (en) * | 2020-01-10 | 2020-06-02 | 北京理工大学 | Vehicle driving parameter and state parameter joint estimation method based on parallel structure |
Non-Patent Citations (2)
| Title |
|---|
| MARIOS IMPRAIMAKIS等: "An unscented Kalman filter method for real time input-parameter-state estimation", MECHANICAL SYSTEMS AND SIGNAL PROCESSING, vol. 162, pages 1 - 14 * |
| XU BO等: "State Estimation of Bearingless Permanent Magnet Synchronous Motor Using Improved UKF", PROCEEDINGS OF THE 31ST CHINESE CONTROL CONFERENCE, pages 1 - 4 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Daescu | On the sensitivity equations of four-dimensional variational (4D-Var) data assimilation | |
| Zhou et al. | Differential feature based hierarchical PCA fault detection method for dynamic fault | |
| CN113886973B (en) | Ship navigational speed processing method, device and processing equipment based on virtual-real mapping | |
| CN103792910B (en) | A kind of network group control formula error dynamic compensation system of Digit Control Machine Tool | |
| CN111679035A (en) | Data compensation method, device, equipment and medium for gas analyzer | |
| CN110657806B (en) | Position resolving method based on CKF, chan resolving and Savitzky-Golay smooth filtering | |
| CN112713881A (en) | Synchronous clock maintaining system and method based on edge calculation | |
| Tian | Design and implementation of distributed measurement systems using fieldbus-based intelligent sensors | |
| Ge et al. | Observable degree analysis to match estimation performance for wireless tracking networks | |
| CN103356179B (en) | A kind of temperature taking analytical system and temperature taking analytical method | |
| CN116796110A (en) | Tea leaf storage environment state estimation method, system and storage medium | |
| CN105571666B (en) | Flow-compensated method and compensation device, flow sensor | |
| Peng et al. | Shrinkage estimation of varying covariate effects based on quantile regression | |
| CN106871905B (en) | Gaussian filtering substitution framework combined navigation method under non-ideal condition | |
| Yang et al. | Event-triggered consensus for second-order multi-agent systems via asynchronous periodic sampling control approach | |
| CN117490819A (en) | Error compensation system and method for electronic scale | |
| KR20230009484A (en) | Engine abnormality diagnosis method, engine abnormality diagnosis program, and engine abnormality diagnosis system | |
| Król et al. | Intelligent sensor platform with open architecture for monitoring and control of Industry 4.0 systems | |
| Zhang et al. | Optimal and Self-correcting Covariance Intersection Fusion Kalman Filters | |
| Haq | Recovering the sensitivity of an auxiliary information based EWMA mean chart | |
| Zhao et al. | Nonlinear state estimation based on continuous-discrete cubature kalman filter for fermentation process | |
| WO2014063633A1 (en) | Method and system for stable state processing during digital measurement or remote measurement of physical quantity | |
| CN116105731B (en) | Navigation method and device under sparse ranging condition, computer equipment and medium | |
| Paolucci et al. | Ad-Hoc Algorithm for Direct Data Processing on Board Each Single Node in the Structural Health Monitoring | |
| CN116482405A (en) | Accelerometer calibration method and system based on ADMM algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |