Industrial System Delay Control Method, Device and Non-Volatile Storage Medium

Abstract

The present application discloses an industrial system time-delay control method, device and non-volatile storage medium. Among them, the method includes: obtaining initial time-delay distribution data in a first format and performing format conversion processing to obtain first target time-delay distribution data in a second format, where the first target time-delay distribution data is characterized by a basis function and a weight vector corresponding to the basis function; comparing the first target time-delay distribution data with the expected time-delay distribution data to obtain a performance index, and generating control data based on the performance index; and performing iterative control on the first target time-delay distribution data according to the control data to determine second target time-delay distribution data. The present application solves the technical problem that due to the prior art only making the control system work properly by minimizing the time delay, without considering the impact of time-delay jitter on the performance of the industrial control system, the performance of the industrial control system is low and the stability is poor.

Description

Technical Field

[0001] The present application relates to the field of network communication technologies, and in particular, to an industrial system delay control method, apparatus, and non-volatile storage medium. Background Art

[0002] With the rapid development of modern industry, traditional wired control systems can no longer meet the needs of production development. Wireless networks with mobility and high flexibility can well make up for the deficiencies of wired control systems. Therefore, they have been widely tried and applied in the industrial field in recent years. However, due to the relatively complex environment where industrial equipment is located, the transmission process of wireless signals will be interfered by many unstable factors. Any change in the surrounding environment will be mapped and amplified in the perturbation of the channel transfer function, which will cause the radio wave to interact with the surrounding environment to form a multipath fading channel. Obviously, after the signal passes through the non-stationary fading channel, its transmission delay will be affected by the time-varying channel capacity and exhibit non-stationary time-varying characteristics, seriously reducing the security and reliability of the control system.

[0003] Currently, in the prior art, only by minimizing the delay can the control system work properly, without considering the impact of delay jitter on the performance of the industrial control system, resulting in the technical problems of low performance and poor stability of the industrial control system.

[0004] In view of the above problems, no effective solution has been proposed yet. Summary of the Invention

[0005] Embodiments of the present application provide an industrial system delay control method, apparatus, and non-volatile storage medium, so as to at least solve the technical problems of low performance and poor stability of the industrial control system caused by the prior art only making the control system work properly by minimizing the delay without considering the impact of delay jitter on the performance of the industrial control system.

[0006] According to one aspect of the embodiments of the present application, an industrial system delay control method is provided, including: obtaining initial delay distribution data in a first format and performing format conversion processing to obtain first target delay distribution data in a second format, where the first target delay distribution data is characterized by a basis function and a weight vector corresponding to the basis function; comparing the first target delay distribution data with expected delay distribution data to obtain a performance index, and generating control data based on the performance index; and performing iterative control on the first target delay distribution data according to the control data to determine second target delay distribution data.

[0007] Optionally, obtaining the initial delay distribution data in the first format and performing format conversion processing to obtain the first target delay distribution data in the second format includes: determining a target model, and determining the target parameters and the target function of the target model based on the initial delay distribution data; using the target parameters and the target function to characterize the initial delay distribution data to obtain the first target delay distribution data.

[0008] Optionally, the target parameters include the number of basis functions and the weight vector corresponding to the basis function, and the target function includes the basis function. Determining the target parameters and the target function of the target model based on the initial delay distribution data includes: determining the number of basis functions and the basis function based on the initial delay distribution data; performing arithmetic processing on the target model based on the initial delay distribution data and the constraint conditions to obtain a target correlation relationship, where the correlation relationship is used to characterize the dynamic relationship between the weight vector and the initial delay distribution data; determining the weight vector corresponding to the basis function based on the target correlation relationship.

[0009] Optionally, generating control data based on the performance index includes: determining a performance function; minimizing the performance index based on the performance function to obtain the control data.

[0010] Optionally, iteratively controlling the first target delay distribution data based on the control data to determine the second target delay distribution data includes: obtaining the control data, and controlling the transmit power based on the control data, where the transmit power is used to control the change of the delay distribution data; adjusting the mean vector of the basis function until the distribution position of the first target delay distribution data is consistent with the distribution position of the expected delay distribution data; adjusting the weight vector corresponding to the basis function until the performance index of the first target delay distribution data and the expected delay distribution data is within a preset threshold; determining the adjusted first target delay distribution data as the second target delay distribution data.

[0011] Optionally, the method further includes: obtaining the first control data in the first period; in the second period, adjusting the first target delay distribution data based on the first control data, where the second period is the next period immediately following the first period; comparing the adjusted first target delay distribution data with the expected delay distribution data to obtain a performance index, and generating second control data based on the performance index.

[0012] Optionally, the method further includes: determining the adjusted first target delay distribution data as the second target delay distribution data when the performance index meets the preset conditions.

[0013] According to another aspect of the embodiments of the present application, an industrial system delay control device is further provided, including: a data determination module, configured to obtain initial delay distribution data in a first format, perform format conversion processing, and obtain first target delay distribution data in a second format, where the first target delay distribution data is characterized by a basis function and a weight vector corresponding to the basis function; a data comparison module, configured to compare the first target delay distribution data with expected delay distribution data to obtain a performance index, and generate control data based on the performance index; an iterative control module, configured to perform iterative control on the first target delay distribution data according to the control data to determine second target delay distribution data.

[0014] According to still another aspect of the embodiments of the present application, an electronic device is further provided. The electronic device includes a processor, and the processor is used to run a program. When the program runs, it executes an industrial system delay control method.

[0015] According to yet another aspect of the embodiments of the present application, a non-volatile storage medium is further provided. The non-volatile storage medium includes a stored program. When the program runs, it controls the device where the non-volatile storage medium is located to execute an industrial system delay control method.

[0016] In the embodiments of the present application, by continuously performing iterative control to approach the expected delay distribution data, through obtaining initial delay distribution data in a first format and performing format conversion processing, first target delay distribution data in a second format is obtained, where the first target delay distribution data is characterized by a basis function and a weight vector corresponding to the basis function; comparing the first target delay distribution data with the expected delay distribution data to obtain a performance index, and generating control data based on the performance index; performing iterative control on the first target delay distribution data according to the control data to determine second target delay distribution data, the purpose of overall control of the delay distribution data is achieved, and further, the technical problem that the performance of the industrial control system is low and the stability is poor due to the fact that the existing technology only makes the control system work properly by minimizing the delay without considering the impact of delay jitter on the performance of the industrial control system is solved. Description of the Drawings

[0017] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation to the present application. In the drawings:

[0018] Figure 1 is a schematic diagram of the method flow of an industrial system delay control provided according to the embodiments of the present application;

[0019] Figure 2 is a schematic diagram of the method flow of a delay probability density distribution control based on iterative learning provided according to the embodiments of the present application;

[0020] Figure 3 It is a schematic diagram of the overall control process of a double - layer closed - loop feedback control structure provided according to an embodiment of the present application;

[0021] Figure 4 It is a schematic diagram including the initial, the time - delay probability density distribution after the first control cycle, and the target distribution curve provided according to an embodiment of the present application;

[0022] Figure 5 It is a schematic diagram of the change process of the performance index E1 in the mean - vector control stage provided according to an embodiment of the present application;

[0023] Figure 6 It is a schematic diagram of the curve comparison between the time - delay probability density distributions at the beginning of controlling the weight vector and after the control of the weight vector ends provided according to an embodiment of the present application;

[0024] Figure 7 It is a schematic diagram of the structure of an industrial - system time - delay control device provided according to an embodiment of the present invention;

[0025] Figure 8 It is a hardware structure block diagram of a computer terminal (or electronic device) for a method of implementing industrial - system time - delay control provided according to an embodiment of the present application. Detailed implementation manners

[0026] In order to enable the personnel in the technical field to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present application.

[0027] For the convenience of those skilled in the art to better understand the embodiments of the present application, some technical terms or noun explanations related to the embodiments of the present application are as follows:

[0028] MATLAB: A commercial mathematical software used in fields such as data analysis, wireless communication, and deep learning.

[0029] It should be noted that the terms "first", "second", etc. in the description, claims and the above-mentioned drawings of this application are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application described here can be implemented in an order different from those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0030] Embodiment 1

[0031] According to an embodiment of the present application, an embodiment of a method for industrial system delay control is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that here.

[0032] Figure 1 is a schematic diagram of a method flow for industrial system delay control provided according to an embodiment of the present application, as Figure 1 shown, the method includes the following steps:

[0033] Step S102, obtaining initial delay distribution data in a first format and performing format conversion processing to obtain first target delay distribution data in a second format, where the first target delay distribution data is characterized by a basis function and a weight vector corresponding to the basis function;

[0034] In some embodiments of the present application, obtaining initial delay distribution data in a first format and performing format conversion processing to obtain first target delay distribution data in a second format includes: determining a target model, and determining the target parameters and target function of the target model according to the initial delay distribution data; using the target parameters and target function to characterize the initial delay distribution data to obtain the first target delay distribution data.

[0035] Among them, the target parameters include the number of basis functions and the weight vector corresponding to the basis function, and the target function includes the basis function. Determining the target parameters and target function of the target model according to the initial delay distribution data includes: determining the number of basis functions and the basis function according to the initial delay distribution data; performing arithmetic processing on the target model according to the initial delay distribution data and the constraint conditions to obtain a target correlation relationship, where the correlation relationship is used to characterize the dynamic relationship between the weight vector and the initial delay distribution data; determining the weight vector corresponding to the basis function according to the target correlation relationship.

[0036] In this embodiment, the initial delay distribution data in the first format is the delay probability density distribution presenting a non-Gaussian distribution received in an actual industrial scenario. The first target delay distribution data in the second format is the delay probability density distribution after modeling the non-Gaussian distribution of the delay probability density distribution using a mixture Gaussian model. The target model is a mixture Gaussian model.

[0037] In this embodiment, a mechanism model is established for the delay probability density distribution in an industrial wireless network using a mixture Gaussian model. Specifically, assuming that the delay probability density distribution is continuous and bounded, the mathematical model constructed for the delay probability density distribution using a mixture Gaussian model is as shown in formula (1):

[0038]

[0039] In the above formula, γ k (τ, u(i)) represents the delay probability density distribution, where τ represents the transmission delay, u(i) ∈ R is a control variable representing the transmission power of a data packet; i represents the sampling time, e0(k) represents the fitting error between the modeling result and the actual delay distribution, C 1,k (τ) = [R 1,k (τ), R 2,k (τ), …, R N-1,k (τ)] represents the basis function vector composed of the first n - 1 Gaussian basis functions, where R n,k (τ) represents the nth Gaussian basis function; V k (i) = [ω 1,k (i), ω 2,k (i), …, ω N-1,k (i)] T represents the weight vector corresponding to the first n - 1 basis functions, ω n,k represents the weight of the nth Gaussian basis function; μ n,k and σ n,k are respectively the mean and variance of the nth Gaussian basis function.

[0040] It can be seen from formula (1) that in the process of establishing a mathematical model for the delay probability density distribution using a mixture Gaussian model, the number of basis functions included in the basis function vector C 1,k (τ) is determined, and the selection of this number needs to be reasonable: if the number of basis functions is too large, the computational workload in the iterative process will be heavy, resulting in low control efficiency; if the number is too small, the mixture Gaussian model will not accurately fit the delay probability density distribution, resulting in too large an error in the approximation result, and ultimately the control result will be difficult to meet the target requirements. Therefore, a reasonable number of basis functions is determined based on the initial delay distribution data;

[0041] When the number of the above basis functions is reasonable, the approximation error e0(k) is small enough to be negligible. At the same time, since the probability density distribution needs to satisfy the requirement that the integral within the coordinate interval is 1, the time-delay probability density distribution can be obtained which needs to satisfy the constraint conditions shown in formula (2):

[0042]

[0043] Substitute formula (1) into formula (2) and expand it to obtain formula (3), as shown below:

[0044] V k T (i)Σ0V k (i)+2Σ1V k (i)ω N (i)+Σ2ω N (i) 2 =1 (3)

[0045] where After performing mathematical operations, the Nth weight ω N (i) can be expressed as a non-linear function h(V k (i)) of the first N - 1 weights, as shown in formula (4):

[0046]

[0047] where Σ3 = Σ2Σ0 - Σ1 T Σ1

[0048] The above formula (4) shows that the Nth weight ω N,k (i) of the Gaussian mixture model can be represented by the non-linear function h(V k (i)) of the first N - 1 weights. Then, by rearranging formula (1), the matrix form of the time-delay probability density distribution can be obtained, as shown in formula (5):

[0049]

[0050] Next, multiply both sides of formula (5) on the left by [C 1,k (τ),R N,k (τ)] T and integrate over the time-delay domain [a, b] to obtain formula (6), as shown below:

[0051]

[0052] where When the matrix When it is non - singular, it can be transformed into the form shown in formula (7) by transposing terms.

[0053]

[0054] Formula (7) represents the dynamic relationship (i.e., the above - mentioned target association relationship) between the weight vector of the Gaussian mixture model and the time - delay probability density distribution. When the time - delay probability density distribution can be obtained through measurement and the Gaussian basis function is also determined, the corresponding weight vector can be obtained through formula (7).

[0055] In summary, the mathematical modeling of the time - delay probability density distribution in the industrial wireless network is completed, as shown in formula (8). In addition, since there will inevitably be errors in the approximation of the probability density distribution by the mathematical modeling method, the probability density distribution of the time - delay should use its estimated form to represent.

[0056]

[0057] Step S104: Compare the first - target time - delay distribution data with the expected time - delay distribution data to obtain a performance index, and generate control data based on the performance index;

[0058] In some embodiments of the present application, generating control data based on the performance index includes: determining a performance function; minimizing the performance index according to the performance function to obtain control data.

[0059] In this embodiment, the above - mentioned performance function can be written as shown in formula (9):

[0060] J=(N(y)-g(y)) 2 (9)

[0061] Among them, N(y) represents the time - delay probability density distribution of the industrial wireless network, and g(y) represents the target distribution. The performance function J is in the form of a quadratic form of the difference between the two distributions. By designing a control algorithm to minimize the performance function J, the control objective of approximating the time - delay distribution to the target distribution is realized.

[0062] To achieve the control objective of approximating the time - delay distribution to the target distribution, the generalized performance function shown in formula (9) can be specifically written as shown in formula (10). Among them, in order to maintain consistency with the time - delay distribution modeling method, the target distribution is also written as a square - root model, that is μ n and μ mean,g respectively represent the mean values of the basis functions of the time - delay probability density distribution and the target distribution:

[0063]

[0064] Specifically, formula (10) uses the form of a difference quadratic form. The first term shows the performance result of controlling the weight vector by further integrating the quadratic form in the time-delay domain [a, b]. The second term shows the control result of the time-delay distribution position (i.e., the above control data) by comparing the mean of the basis functions of the time-delay probability density distribution and the target distribution and then accumulating the difference between the two.

[0065] Step S106: Iteratively control the first target time-delay distribution data according to the control data to determine the second target time-delay distribution data.

[0066] Iteratively controlling the first target time-delay distribution data according to the control data to determine the second target time-delay distribution data includes: obtaining the control data and controlling the transmit power according to the control data, where the transmit power is used to control the change of the time-delay distribution data; adjusting the mean vector of the basis function until the distribution position of the first target time-delay distribution data is the same as that of the expected time-delay distribution data; adjusting the weight vector corresponding to the basis function until the performance index of the first target time-delay distribution data and the expected time-delay distribution data is within a preset threshold; and determining the adjusted first target time-delay distribution data as the second target time-delay distribution data.

[0067] In some embodiments of the present application, the method further includes: obtaining the first control data of the first period; in the second period, adjusting the first target time-delay distribution data according to the first control data, where the second period is the next period adjacent to the first period; comparing the adjusted first target time-delay distribution data with the expected time-delay distribution data to obtain a performance index, and generating second control data according to the performance index.

[0068] In some embodiments of the present application, the method further includes: when the performance index meets a preset condition, determining the adjusted first target time-delay distribution data as the second target time-delay distribution data.

[0069] It can be seen from the performance function shown in formula (10) that it is necessary to control the probability density distribution of the time delay from two aspects. First, since after any given target distribution, the time-delay distribution actually output by the system is unlikely to coincide with it, it is first necessary to control the distribution position of the time delay, which corresponds to the second term in the performance function. Otherwise, it is difficult to approximate the target distribution no matter how the control algorithm is designed later. Then, after the distribution position is adjusted to be basically the same as the target distribution, the method of reallocating the weight vector, which corresponds to the first term of the performance function, is used to finally achieve the approximation of the time-delay distribution to the target distribution.

[0070] Specifically, for a delay probability density distribution, the basis functions and the corresponding weight vectors determine its specific shape. Therefore, given any target distribution (i.e., the above-mentioned expected delay distribution data), it is necessary to start from two aspects to achieve the approximation of the delay probability density distribution to the target distribution. One aspect is the control of the Gaussian basis functions, and the other aspect is the control of the corresponding weight vectors. The method of controlling the basis functions lies in adjusting the mean and variance to change their positions and widths; the control of the weight vectors lies in adjusting the proportion of each weight. From the perspective of fitting accuracy, the requirement for variance control can be relaxed by increasing the number of basis functions. Therefore, after reasonably selecting the number of Gaussian basis functions, the control target for the Gaussian basis functions is to adjust the mean of each basis function to achieve the approximation of the distribution position to the target distribution. When the delay probability density distribution is basically the same as the target distribution in terms of distribution position, then adjust the weight ratio to achieve the further approximation effect of the delay probability density distribution to the target distribution.

[0071] In this embodiment, the industrial wireless network is modeled as a closed-loop feedback system, and the control variable is the transmit power. Therefore, first, adjust the distribution position of the delay by controlling the transmit power. Since the actual output delay distribution of the system is approximately represented by a mixture Gaussian model, for the adjustment of the delay distribution position, it is necessary to control the mean of each basis function separately. When the transmit power is controlled based on the mean of a certain basis function, due to the influence of the uncertain multipath fading channel on the transmitted signal, there will inevitably be an error between the controlled object and the target. To eliminate the error, the closed-loop feedback control process is divided into a series of cycles and used as the control unit, and the error between the delay distribution and the target distribution is continuously reduced through multiple cycles of iteration.

[0072] Figure 2 It is a schematic flowchart of a method for controlling the delay probability density distribution based on iterative learning according to an embodiment of the present application. As Figure 2 shown, in this embodiment, a double-layer closed-loop feedback control system is used, including an upper-layer structure S202 and a lower-layer structure S204.

[0073] The upper-layer structure S202 includes an iterative learning control process performed in each cycle. The inputs include the target distribution and the mathematical approximation of the actual output delay probability density distribution at the receiving end After the two enter the iterative learning controller, the control target for the next cycle will be obtained: μ k (mean vector) or ω k (weight vector);

[0074] The lower-layer structure S204 includes adjusting the transmission power according to the control target output by the upper-layer structure. After transmitting a certain number of data packets, the receiving end will obtain an estimate of the delay probability density distribution again through statistical calculation. And return it to the upper-layer structure.

[0075] Specifically, in each period, first in the upper-layer control structure, the iterative learning algorithm is used to control the mean of the Gaussian basis function first and then the weight vector. In the lower-layer control structure, the size of the delay is controlled by adjusting the transmission power. After transmitting a certain number of data packets, the delay probability density distribution of this period can be obtained. Figure 3 It is a schematic diagram of the overall control process of the double-layer closed-loop feedback control structure, as Figure 3 shown in the figure, where PDF represents the delay distribution data. Under the joint control of the upper and lower layers, the approximation of the target distribution is finally achieved, that is, the control of the delay probability density distribution is completed.

[0076] Next, the control algorithms in the upper-layer structure and the lower-layer structure will be further introduced respectively.

[0077] The upper-layer structure S202 includes the iterative learning control process carried out in each period, and the inputs include the target distribution and the mathematical approximation of the actual output delay probability density distribution at the receiving end After the two enter the iterative learning controller, the control target for the next period will be obtained: μ k (mean vector) or ω k (weight vector);

[0078] In this embodiment, in each period, the control process of the Gaussian basis function by the upper-layer structure using the iterative learning control method is divided into two parts: First, adjust the mean of the basis function; Second, adjust the weights corresponding to each basis function. In the upper-layer control structure, the estimated delay probability density distribution statistically obtained at the receiving end is used as the input quantity, and then it is compared with the target distribution, thereby forming the performance index of the closed-loop control. The performance function is used to minimize the performance index. The updated value of the control target (μ k+1 or ω k+1 ) obtained by minimizing the performance function will be used as the control input of the lower-layer structure, thereby guiding the control process of the transmission power at the transmitting end. By transmitting a number of data packets, the parameters of the delay probability density distribution obtained at the receiving end can continuously approach the control target value of this period. However, after each period has successively passed through the upper-layer control and the lower-layer control, there will inevitably still be an error between the mean vector or the weight vector of the delay probability density distribution and the expected target value. At this time, the iterative learning method is used to control this output error.

[0079] Specifically, in each cycle, by iteratively controlling the adjustment of the mean vector of the Gaussian basis function, the first-step control objective of the time-delay probability density distribution can be achieved, as shown in formula (11). This control mechanism indicates that the control objective of the k-th cycle will be determined by the control input of the (k-1)-th cycle and the correction term, and the control performance of the closed-loop feedback system can be gradually improved through continuous iteration between cycles:

[0080] μ k = μ k-1 + λ μ E 1,k (11)

[0081] In the above formula, λ μ represents the learning rate for iteratively controlling the mean vector; the correction term E 1,k represents the sum of the differences between the means of each Gaussian basis function and the mean of the target distribution after the control of the previous cycle, indicating the performance of the closed-loop control in the previous cycle, as shown in formula (12):

[0082]

[0083] To ensure that the effect of the iterative control is continuously improved, it is necessary to first adjust the position of the time-delay probability density distribution in the upper-layer control process. Otherwise, it will be difficult to approximate the target distribution in the subsequent control process of the weight vector. Therefore, E 1,k as the performance index of the upper-layer control process should show a decreasing characteristic, which means that the control algorithm needs to meet the following conditions, as shown in formula (13):

[0084] ΔE 1,k = E 1,k - E 1,k-1 ≤0 (13)

[0085] From formula (12), it can be seen that the performance index E 1,k is a quadratic function. According to the properties of the quadratic function, there is an extreme point in its domain; and because E 1,k is always greater than 0, so the performance index E 1,k has a minimum point. Select an appropriate learning rate λ μ , and after several cycles of iterative learning control, when the correction term E 1,k no longer decreases, that is, ΔE 1,k > 0, the control of the mean vector by the upper-layer structure ends, and then the control process of the weight vector enters the second stage.

[0086] After the control of the mean vector of the delay probability density distribution is completed, the positions of the basis functions are basically the same as the target distribution. Therefore, the next step in controlling the delay probability density distribution is to make the weight vector of the delay probability density distribution basically the same as the weight vector of the target distribution.

[0087] The control mechanism for the weight vector is shown in Equation (14):

[0088] ω k = ω k-1 + λ ω E 2,k (14)

[0089] where λ ω represents the learning rate for iterative control of the weight vector, and the sign of each element in λ ω indicates the control direction of the upper layer structure for the corresponding weight. E 2,k represents the difference between the delay probability density distribution at the receiving end and the target distribution, indicating the performance of the control of the weight vector in the previous cycle, as shown in Equation (15):

[0090]

[0091] Select the weight corresponding to the maximum term of the difference between the weight vector of the actual delay distribution and the target weight vector as the control target for the next cycle, and at the same time determine the control direction for this weight, from which the learning rate λ ω .

[0092] Similarly, the upper layer control algorithm should ensure that the performance index E 2,k keeps decreasing, so that the delay probability density distribution actually output by the system can gradually approach the target distribution.

[0093] The lower layer structure S204 includes adjusting the transmission power according to the control target output by the upper layer structure. After transmitting a certain number of data packets, the receiving end will obtain an estimate of the delay probability density distribution again through statistical calculation and return it to the upper layer structure.

[0094] In this embodiment, M represents the number of data packets transmitted; P t (i + 1) represents the transmission power, P d represents the path loss, P f (i + 1) represents the fading intensity of the industrial multipath fading channel, n represents the noise at the receiving end; μ represents the mean of the delay probability density distribution, and B represents the signal bandwidth.

[0095] In this embodiment, the coupling amount between the upper layer structure and the lower layer structure is the control target value for the next cycle: the mean vector μ of the Gaussian basis functionk and the weight vector ω k , which are uniformly defined as the control target value X k . They will be used as the target values of the controlled objects in the lower-layer structure of this cycle. By controlling the transmit power, the delay size formed at the receiving end after the data packet passes through the fading channel is adjusted, so that after the transmitter sends a certain number of data packets, the delay distribution of the basis functions in the mean vector, or the weight vector, can continuously approximate the target value X k .

[0096] Specifically, given a Rice fading channel, the state probability transition matrix P of the channel fading can be obtained by traversing the global information of the channel. According to the property that the future state probability of a first-order Markov chain only depends on the current state, after knowing the state probability p(i) of the fading channel at this moment, the state probability p(i + 1) at the next moment can be predicted using the state probability transition matrix information.

[0097] In each cycle, the lower-layer control structure compares the actual situation of the controlled object (mean vector or weight vector) in its current delay distribution with the control target value X k input by the upper-layer structure, and dynamically selects a specific control target value x k from the vector X * , thereby establishing an optimization equation for the transmission process Then, the optimal solution of the transmit power in this transmission process is calculated through an optimization algorithm, and the control target value of this cycle can be continuously approximated after multiple transmission processes.

[0098] Taking the control target value μ k as an example, after mathematically modeling the initial delay distribution using a mixture Gaussian model, the mean vector μ k =[μ 1,k , μ 2,k , …, μ N,k T formed by the means of each basis function in this distribution can be obtained. The optimization equation established for the transmission process at the next moment is shown in formula (16):

[0099]

[0100] where μ * is the control target of the delay mean at the next moment. After optimizing the above formula, the optimal solution of the transmit power P t (i + 1) can be obtained, as shown in formula (17), where P L (i + 1)=P d +Pf (i + 1), which includes two parts: path loss and Rice fading intensity.

[0101]

[0102] In this embodiment, the mathematical expectation of the channel fading is used to characterize the true channel fading intensity at the next moment, that is where p r (i + 1) represents the probability of the r-th channel state occurring at the next moment, and P f,r represents the Rice fading gain corresponding to the r-th channel state. Substituting P f ′(i + 1) into formula (17), the sub-optimal solution P t ′(i + 1) of the transmit power can be obtained, as shown in formula (18):

[0103]

[0104] where, represents the optimal solution of the transmit power obtained by optimizing for each channel fading state P f,r .

[0105] After one iteration control of the upper-layer structure, the updated value X k of the control target of the delay probability density distribution can be obtained. It is input into the lower-layer control structure and defined as the target value X ref of the lower-layer control process in this period; at the same time, the actual output value X k-1 of the controlled object in the previous period is also input into the lower-layer structure. Then the difference between the two can be defined as X d = X ref - X k-1 . By using the iterative method, the optimal atom selection strategy is adopted in turn, that is, in each iteration, the column vector corresponding to the maximum value in the inner product result of the measurement matrix Φ and the current difference matrix X d is used as the candidate control target. Since the difference matrix is one-dimensional, the measurement matrix can be simply designed as a one-dimensional vector [1, 1,..., 1]. Corresponding to each transmission process of the lower-layer control, the difference between the control target value μ ref of the mean vector output by the upper-layer structure and the mean vector μ k-1 of the delay distribution in the previous period is calculated to obtain the difference matrix μ d : μ d = μ ref - μ k-1 . The mean value corresponding to the maximum value in the difference matrix μ d is selected, that is, μ * = max(μ d), as the optimization equation in the next moment's transmission process of the control target value μ * .

[0106] Substitute the control target μ of the next moment * into the performance function of the lower-layer structure, that is, formula (16). Furthermore, the sub-optimal solution of this optimization equation is obtained and used as the transmission power of the next moment, as shown in formula (18). After each transmission process ends, at the receiving end, a mixture Gaussian model is used to mathematically model the delay distribution. After a single control process of the lower-layer structure, the updated value μ k ′(i) of the mean vector of the delay probability density distribution and the difference matrix μ d can be obtained. Continuously repeat the process of selecting the mean corresponding to the maximum term in the difference matrix as the control target value. After transmitting a certain number of data packets, the mean vector μ k of the delay probability density distribution can reach the target value μ ref .

[0107] Similarly, when the upper-layer control structure starts to control the weight vector of the delay probability density distribution, the lower-layer control structure adopts a greedy algorithm to determine the control target value of the system delay in the next moment. The updated value ω k of the weight vector output by the upper-layer structure is used as the control target value ω ref of the lower-layer structure. By comparing the control target value ω ref with the weight vector ω k ′(i) of the actual delay distribution, the difference matrix between the two can be obtained: ω d = ω ref - ω k ′(i). Then, by selecting the basis function corresponding to the maximum value in the difference matrix ω d , the mean value of the controlled delay in the next moment is determined.

[0108] Through the above steps, by obtaining the initial delay distribution data in the first format and performing format conversion processing, the first target delay distribution data in the second format is obtained, where the first target delay distribution data is characterized by basis functions and the weight vectors corresponding to the basis functions; comparing the first target delay distribution data with the expected delay distribution data to obtain performance indicators, and generating control data based on the performance indicators; according to the control data, performing iterative control on the first target delay distribution data to determine the second target delay distribution data, achieving the purpose of overall control of the delay distribution data, and further solving the technical problems of low performance and poor stability of the industrial control system caused by the fact that the existing technology only makes the control system work properly by minimizing the delay without considering the impact of delay jitter on the performance of the industrial control system.

[0109] Embodiment 2

[0110] According to an embodiment of the present invention, an embodiment of an industrial system delay control method is further provided. Specifically, numerical simulation is deployed based on MATLAB to verify the industrial system delay control method.

[0111] In this embodiment, the wireless fading channel adopts a classical Rice fading channel, which is obtained from the MATLAB standard communication toolbox. Its sampling time is 100 μs, the maximum Doppler shift is 100 Hz, and the Rice factor k = 2. The transmission power range of the data packet is between -12 dBm and 8 dBm, and the path loss during signal transmission is set to -26 dB. The length of the transmitted data packet is set to 800 bits.

[0112] In this embodiment, on the premise of giving a target distribution (i.e., the above-mentioned expected experimental distribution data), any initial delay probability density distribution (i.e., the above-mentioned initial delay distribution data) can approximate this target distribution after passing through a double-layer closed-loop feedback control system; for the actual industrial wireless network, it is hoped that the variance of the delay probability density distribution of the system can be as small as possible, so as to meet the requirements of wireless communication security and reliability.

[0113] Specifically, the simplest ideal Gaussian distribution is selected as the target distribution of the control system, denoted as Figure 4 According to an embodiment of the present application, a schematic diagram of the initial, delay probability density distribution after the first control cycle, and the target distribution curve is shown in Figure 4 the curve connected by triangles as shown.

[0114] After the transmitter sends a certain number of data packets, the initial delay probability density distribution will be obtained at the receiver, and its statistical output result is shown in Figure 4 the curve connected by circles as shown. When the delay probability density distribution is measurable, a mixture Gaussian model is used to estimate this distribution. First, the number of Gaussian basis functions is determined according to the shape of the initial delay probability density distribution. Then, three Gaussian basis functions are selected to fit this distribution according to the characteristics of the initial delay probability density distribution, and its mean vector is μ0 = [0.0243 s, 0.0307 s, 0.0362 s] T .

[0115] Next, the first control cycle is carried out. After the initial delay probability density distribution and the corresponding basis functions are determined, the weight vector corresponding to the basis function is then solved as [V0(i), h(V0(i))] T = [0.29, 0.23, 0.48] T. In the upper-layer control structure, the control target value of the mean vector for the first cycle (k = 1) can be obtained through the iterative learning algorithm, and the learning rate λ μ is [-0.18, -0.17, -0.16] T . The correction term E1 is obtained based on the difference between the mean vector of the delay distribution and the mean vector of the target distribution. Then, in the next cycle, the mean vector is updated to μ1 = [0.0230s, 0.0286s, 0.0345s] T .

[0116] . Next, the lower-layer structure performs specific control operations according to the control target μ1 of the mean vector output by the upper-layer structure, so that the delay probability density distribution at the receiving end can meet the control requirements of this cycle (k = 1). First, the mean vector μ1 input from the upper layer to the lower layer is used as the control target value μ of this cycle ref , and the difference matrix between the target value of the mean vector and the actual output at the receiving end is calculated as μ d = [0.0010s, 0.0018s, 0.0016s] T . The delay mean corresponding to the maximum value in the difference matrix μ d is selected as the control target of the optimization equation JPt, that is, μ * = 0.0286s. Then, the fading state of the fading channel at the next moment is predicted, and the transmit power P t '(i + 1) at the next moment is obtained. Transmitting the data packet at this power through the fading channel will result in a delay τ(i) at the receiving end. At the same time, the delay at the receiving end is statistically processed again, and the delay probability density distribution after this transmission process will be obtained

[0117] . After multiple transmission processes, the delay probability density distribution at the receiving end meets the control requirements of this cycle for the mean vector of the basis function, as Figure 4 shown. The curves connected by circles and the curves connected by short lines in the figure show the change of the delay probability density distribution after passing through the control of the upper-layer structure and the lower-layer structure in sequence in the first control cycle. Among them, the curve connected by circles is the initial delay probability density distribution; the curve connected by short lines is the delay probability density distribution after the first control cycle, and the mean vector corresponding to each basis function is μ1' = [0.0233s, 0.0289s, 0.0346s] T , which basically meets the requirements of the control target value μ1. Then, the lower-layer structure feeds back the mean vector of the current delay distribution to the upper-layer structure, and the upper-layer structure will obtain the control target value of the next cycle according to the control algorithm

[0118] . Repeat the control process within the above unit cycle. When E 1,kNo longer decreases, that is, ΔE 1,k > 0 indicates that the control of the mean vector has been completed at this time. Figure 5 It is a schematic diagram of the change process of the performance index E1 in the mean vector control stage provided by the embodiment of the present application. As Figure 5 shown, it can be seen that after the 10th cycle (k = 10), E1 will show an upward trend, which indicates that if the system continues to use the iterative learning algorithm to control the mean vector, the control algorithm will no longer satisfy the boundary conditions shown in formula (13). Therefore, after passing through this minimum point, the upper layer structure will start to iteratively control the weight vector of the delay probability density distribution.

[0119] Figure 6 It is a schematic diagram of the curve comparison between the delay probability density distribution at the beginning of controlling the weight vector and after the control of the weight vector ends, provided by the embodiment of the present application. As Figure 6 shown. The delay distribution at the beginning of controlling the weight vector is the same as the distribution after the control of the mean vector ends. Since only the movement of the basis function position in the delay distribution is considered in the mean vector control stage, the weight vector will show randomness characteristics. As shown by the curve connected in a star shape, the weight vector at this time is ω 10 = [0.27, 0.53, 0.20] T , and there is a large gap between it and the weight vector of the target distribution. After 57 cycles of iterative control process, the weight vector becomes ω 57 = [0.85, 0.12, 0.03] T , as shown by the curve connected in a circle shape. Among them, the two peaks are significantly suppressed, and the weight vector of the delay probability density distribution is basically the same as the weight vector of the target distribution, and the system control ends.

[0120] Embodiment 3

[0121] According to the embodiment of the present invention, an embodiment of an industrial system delay control device is further provided. Figure 7 It is a schematic structural diagram of an industrial system delay control device provided by the embodiment of the present invention. As Figure 7 shown, the device includes:

[0122] A data determination module 70, configured to obtain initial delay distribution data in a first format, perform format conversion processing, and obtain first target delay distribution data in a second format, where the first target delay distribution data is characterized by a basis function and a weight vector corresponding to the basis function;

[0123] A data comparison module 72, configured to compare the first target delay distribution data with the expected delay distribution data to obtain a performance index, and generate control data based on the performance index;

[0124] An iterative control module 74, configured to perform iterative control on the first target delay distribution data according to control data to determine second target delay distribution data.

[0125] It should be noted that the industrial system delay control device provided in this embodiment can be used to execute Figure 1 the industrial system delay control method shown. Therefore, the relevant explanations of the above industrial system delay control method also apply to the embodiments of this application and will not be elaborated here.

[0126] According to an embodiment of the present invention, there is also provided an embodiment of a computer terminal for implementing a method for industrial system delay control. Figure 8 is a hardware structure block diagram of a computer terminal (or electronic device) for implementing a method for industrial system delay control according to an embodiment of the present invention. As Figure 8 shown, the computer terminal 80 (or electronic device 80) may include one or more processors (processors may include, but are not limited to, processing devices such as a microprocessor MCU or a programmable logic device FPGA, shown as 802a, 802b,..., 802n in the figure), a memory 804 for storing data, and a transmission module 806 for communication functions. In addition, it may further include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, a power supply, and / or a camera. Those of ordinary skill in the art can understand that Figure 8 the structure shown is only illustrative and does not limit the structure of the above electronic device. For example, the computer terminal 80 may further include more or fewer components than Figure 8 shown, or have a different configuration from Figure 8 shown.

[0127] It should be noted that the above one or more processors and / or other data processing circuits are generally referred to as "data processing circuits" in this article. The data processing circuit may be embodied in whole or in part as software, hardware, firmware, or any combination thereof. In addition, the data processing circuit may be a single independent processing module, or be incorporated in whole or in part into any one of the other elements in the computer terminal 80 (or electronic device). As involved in the embodiments of this application, the data processing circuit is a processor control (such as the selection of a variable resistance terminal path connected to an interface).

[0128] The memory 804 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the industrial system delay control method in the embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 804, that is, implements the above-mentioned industrial system delay control method. The memory 804 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory 804 may further include a memory remotely provided with respect to the processor, and these remote memories can be connected to the computer terminal 80 through a network. Examples of the above network include but are not limited to the Internet, enterprise intranet, local area network, mobile communication network, and combinations thereof.

[0129] The transmission module 806 is used to receive or send data via a network. Specific examples of the above network may include a wireless network provided by the communication provider of the computer terminal 80. In one instance, the transmission device 806 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station and thus communicate with the Internet. In one instance, the transmission device 806 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0130] The display can be, for example, a touch-screen liquid crystal display (LCD), and this liquid crystal display enables a user to interact with the user interface of the computer terminal 80 (or electronic device).

[0131] It should be noted here that in some alternative embodiments, the above Figure 8 shown computer device (or electronic device) may include hardware elements (including circuits), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware elements and software elements. It should be pointed out that Figure 8 is only an example of a specific specific instance, and is intended to show the types of components that may exist in the above computer device (or electronic device).

[0132] It should be noted that Figure 8 the shown electronic device for industrial system delay control is used to execute Figure 1 the shown industrial system delay control method. Therefore, the relevant explanations in the above industrial system delay control method also apply to this industrial system delay control electronic device, and will not be elaborated here.

[0133] According to an embodiment of the present application, an embodiment of a non-volatile storage medium is further provided. The non-volatile storage medium includes a stored program. When the program runs, it controls the device where the non-volatile storage medium is located to execute the following method for industrial system delay control: obtaining initial delay distribution data in a first format and performing format conversion processing to obtain first target delay distribution data in a second format, where the first target delay distribution data is characterized by a basis function and a weight vector corresponding to the basis function; comparing the first target delay distribution data with the expected delay distribution data to obtain a performance index, and generating control data based on the performance index; and performing iterative control on the first target delay distribution data according to the control data to determine second target delay distribution data.

[0134] The serial numbers of the above embodiments of the present application are only for description and do not represent the advantages or disadvantages of the embodiments.

[0135] In the above embodiments of the present application, the descriptions of the respective embodiments have their own emphases. For parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0136] In several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of the units can be a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of units or modules can be in an electrical or other form.

[0137] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0138] In addition, the functional units in each embodiment of the present application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0139] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), mobile hard disks, magnetic disks, or optical discs.

[0140] The above are only the preferred embodiments of this application. It should be noted that for those of ordinary skill in the art of this technology, without departing from the principle of this application, several improvements and refinements can still be made, and these improvements and refinements should also be regarded as the protection scope of this application.

Claims

1. An industrial system delay control method, characterized in that, including: obtaining initial delay distribution data in a first format, performing format conversion processing on the initial delay distribution data to obtain first target delay distribution data in a second format, where the first target delay distribution data is characterized by a basis function and a weight vector corresponding to the basis function; comparing the first target delay distribution data with expected delay distribution data to obtain a performance metric, and generating control data based on the performance metric; performing iterative control on the first target delay distribution data according to the control data to determine second target delay distribution data; performing iterative control on the first target delay distribution data according to the control data to determine second target delay distribution data includes: obtaining the control data, and controlling the transmit power according to the control data, where the transmit power is used to control the change of the delay distribution data; adjusting the mean vector of the basis function until the distribution position of the first target delay distribution data is consistent with the distribution position of the expected delay distribution data; adjusting the weight vector corresponding to the basis function until the performance metric between the first target delay distribution data and the expected delay distribution data is within a preset threshold; determining the adjusted first target delay distribution data as the second target delay distribution data.

2. The industrial system delay control method according to claim 1, characterized in that, obtaining initial delay distribution data in a first format, performing format conversion processing on the initial delay distribution data to obtain first target delay distribution data in a second format includes: determining a target model, and determining target parameters and a target function of the target model according to the initial delay distribution data; using the target parameters and the target function to characterize the initial delay distribution data to obtain the first target delay distribution data.

3. The industrial system delay control method according to claim 2, characterized in that, the target parameters include the number of basis functions and a weight vector corresponding to the basis function, the target function includes a basis function, determining target parameters and a target function of the target model according to the initial delay distribution data includes: determining the number of basis functions and the basis function according to the initial delay distribution data; performing arithmetic processing on the target model according to the initial delay distribution data and constraint conditions to obtain a target correlation relationship, where the correlation relationship is used to characterize the dynamic relationship between the weight vector and the initial delay distribution data; determining the weight vector corresponding to the basis function according to the target correlation relationship.

4. The industrial system delay control method according to claim 1, characterized in that, generating control data based on the performance metric includes: determining a performance function; performing minimization processing on the performance metric according to the performance function to obtain the control data.

5. The industrial system delay control method according to claim 1, characterized in that, the method further includes: obtaining first control data in a first period; in a second period, adjusting the first target delay distribution data according to the first control data, where the second period is the next period immediately following the first period; comparing the adjusted first target delay distribution data with the expected delay distribution data to obtain a performance metric, and generating second control data based on the performance metric.

6. The industrial system delay control method according to claim 5, characterized in that, the method further includes: when the performance metric meets a preset condition, determining the adjusted first target delay distribution data as the second target delay distribution data.

7. An industrial system delay control device, characterized in that, including: A data determination module, configured to obtain initial delay distribution data in a first format, perform format conversion processing thereon, and obtain first target delay distribution data in a second format, wherein the first target delay distribution data is characterized by a basis function and a weight vector corresponding to the basis function; A data comparison module, configured to compare the first target delay distribution data with expected delay distribution data to obtain a performance index, and generate control data according to the performance index; An iterative control module, configured to perform iterative control on the first target delay distribution data according to the control data to determine second target delay distribution data; Performing iterative control on the first target delay distribution data according to the control data to determine second target delay distribution data includes: obtaining the control data, and controlling the transmit power according to the control data, wherein the transmit power is used to control the change of the delay distribution data; adjusting the mean vector of the basis function until the distribution position of the first target delay distribution data is consistent with the distribution position of the expected delay distribution data; adjusting the weight vector corresponding to the basis function until the performance index of the first target delay distribution data and the expected delay distribution data is within a preset threshold; determining the adjusted first target delay distribution data as the second target delay distribution data.

8. An electronic device, the electronic device includes a processor, characterized in that, The processor is configured to run a program, wherein when the program runs, it executes the industrial system delay control method according to any one of claims 1 to 6.

9. A non - volatile storage medium, characterized in that, The non-volatile storage medium includes a stored program, wherein when the program runs, it controls the device where the non-volatile storage medium is located to execute the industrial system delay control method according to any one of claims 1 to 6.

Images