Operation parameter prediction method and device, equipment and storage medium

Abstract

Embodiments of the invention provide an operation parameter prediction method and apparatus, a device and a storage medium. The method comprises the steps of obtaining a plurality of pre-trained prediction models and a gating neural network; inputting a current first operation parameter of the target object and a current control parameter for the target object into a plurality of prediction models and a gating neural network to obtain a specific initial second operation parameter of the target object in a plurality of different environment types and a specific probability of the target object in the different environment types; and predicting a current target second operation parameter of the target object according to the specific initial second operation parameter of the target object in the plurality of different environment types and the specific probability of the target object in the different environment types.

Description

Technical Field

[0001] The present disclosure relates to the field of IT application technology, and in particular to a method, device, equipment and storage medium for predicting operating parameters. Background Art

[0002] At present, a single, global machine learning model is usually used to characterize the dynamic characteristics of agricultural machinery in the entire agricultural environment, and based on the characterized dynamic characteristics of agricultural machinery, the running trajectory of agricultural machinery under the control of relevant personnel is predicted, so that relevant personnel can adjust the control amount of agricultural machinery according to the predicted running trajectory, so that agricultural machinery can run according to the trajectory expected by relevant personnel under appropriate control amount. However, due to the complexity and variability of the agricultural environment, agricultural machinery exhibits significantly different dynamic characteristics in different scenarios, such as flat roads, rugged mountain roads, soft soil, etc., and a single global model is difficult to accurately describe the diverse dynamic characteristics in different scenarios. Therefore, the global-based machine learning model performs poorly in the face of diverse agricultural scenarios, and the predicted running trajectory of agricultural machinery is not accurate, which makes it impossible for relevant personnel to determine the appropriate control amount, and thus cannot accurately control the agricultural machinery to run according to the expected trajectory. Summary of the invention

[0003] The embodiments of the present disclosure provide a method, device, equipment and storage medium for predicting operating parameters to solve the problem in the related art that the predicted operating trajectory of agricultural machinery in various agricultural scenarios is not accurate due to the use of a single global model that is difficult to accurately describe the diverse dynamic characteristics in different scenarios, making it impossible for relevant personnel to determine the appropriate control amount and thus unable to accurately control the agricultural machinery to run along the desired trajectory.

[0004] In a first aspect, an embodiment of the present disclosure provides a method for predicting an operating parameter, the method comprising: Acquire multiple pre-trained prediction models and gated neural networks; wherein the prediction model is used to predict the initial second operating parameters of the target object according to the control parameters for the target object and the first operating parameters of the target object; different prediction models are used to predict the initial second operating parameters of the target object in different environmental types; the gated neural network is used to determine the probability of the target object being in the different environmental types according to the first operating parameters of the target object; the initial second operating parameters predicted by different prediction models and the probability of the target object being in the different environmental types determined by the gated neural network are used to predict the target second operating parameters of the target object; Inputting the current first operating parameter of the target object and the current control parameter for the target object into the plurality of prediction models and the gated neural network to obtain specific initial second operating parameters of the target object under the plurality of different environmental types, and specific probabilities of the target object being in the different environmental types; According to the specific initial second operating parameters of the target object in the multiple different environment types and the specific probability that the target object is in the different environment types, the current target second operating parameters of the target object are predicted.

[0005] In a second aspect, an embodiment of the present disclosure provides a device for predicting operating parameters, the device comprising: A first acquisition module is used to acquire a plurality of pre-trained prediction models and a gated neural network; wherein the prediction model is used to predict the initial second operating parameter of the target object according to the control parameter for the target object and the first operating parameter of the target object; different prediction models are used to predict the initial second operating parameter of the target object in different environmental types; the gated neural network is used to determine the probability of the target object being in the different environmental types according to the first operating parameter of the target object; the initial second operating parameters predicted by different prediction models and the probability of the target object being in the different environmental types determined by the gated neural network are used to predict the target second operating parameter of the target object; A second acquisition model is used to input the current first operating parameter of the target object and the current control parameter for the target object into the plurality of prediction models and the gated neural network to obtain specific initial second operating parameters of the target object under the plurality of different environmental types, and specific probabilities of the target object being in the different environmental types; The prediction module is used to predict the current target second operating parameter of the target object according to the specific initial second operating parameters of the target object in the multiple different environment types and the specific probability that the target object is in the different environment types.

[0006] In a third aspect, an embodiment of the present disclosure provides a device for predicting operating parameters, comprising: a processor; and a memory configured to store computer-executable instructions, wherein the computer-executable instructions, when executed, enable the processor to implement the steps of the method described in the first aspect above.

[0007] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store computer-executable instructions, and the computer-executable instructions implement the steps of the method described in the first aspect when executed by a processor.

[0008] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, wherein the computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the method described in the first aspect are implemented.

[0009] At least one of the above technical solutions provided by the embodiments of the present invention can achieve the following technical effects: In an embodiment of the present invention, a plurality of pre-trained prediction models and a gated neural network may be first obtained, wherein the prediction model may be used to predict the initial second operating parameters of the target object based on the control parameters for the target object and the first operating parameters of the target object, different prediction models are used to predict the initial second operating parameters of the target object in different environmental types, and the gated neural network may be used to determine the probability of the target object being in different environmental types based on the first operating parameters of the target object, the initial second operating parameters predicted by the different prediction models, and the probability of the target object being in different environmental types determined by the gated neural network may be used to predict the target second operating parameters of the target object, and then, the current first operating parameters of the target object and the current control parameters for the target object may be input into the plurality of prediction models and the gated neural network to obtain specific initial second operating parameters of the target object in multiple different environmental types, and specific probabilities of the target object being in different environmental types, and the current target second operating parameters of the target object may be predicted based on the obtained specific initial second operating parameters and specific probabilities.

[0010] As can be seen from the above content, the embodiment of the present invention can obtain multiple prediction models for different environmental types, and can predict the initial second operating parameters of the target object under the control parameters based on the obtained prediction models, as well as the current operating parameters and control parameters of the target object. In addition, the embodiment of the present invention can also combine the gated neural network to determine the probability that the target object is currently in different environmental types based on the current operating parameters of the target object, and combine the probability of the target object currently being in different environmental types, and focus on combining the initial second operating parameters of the target object in different environmental types under the control of the input control parameters to predict the current target second operating parameters of the target object, so that it can maintain a high prediction accuracy in a variety of environments, effectively respond to complex and diverse environments, and effectively solve the problem in the related technology that the prediction of the operation trajectory of agricultural machinery in a variety of agricultural scenes is not accurate due to the use of a single global model that is difficult to accurately describe the diverse dynamic characteristics in different scenes, so that the relevant personnel cannot determine the appropriate control amount, and thus cannot accurately control the agricultural machinery to run according to the desired trajectory. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate one or more embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative labor. Figure 1 One of the flow charts of a method for predicting operating parameters provided by an embodiment of the present invention; Figure 2 A second flow chart of a method for predicting operating parameters provided by an embodiment of the present invention; Figure 3 A third flow chart of a method for predicting operating parameters provided by an embodiment of the present invention; Figure 4 A fourth flow chart of a method for predicting operating parameters provided by an embodiment of the present invention; Figure 5 A fifth flow chart of a method for predicting operating parameters provided by an embodiment of the present invention; Figure 6 A schematic diagram of the module composition of an operating parameter prediction device 600 provided in one embodiment of the present invention; Figure 7 A schematic diagram of the hardware structure of an operating parameter prediction device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0012] In order to make the technical parties in this technical field better understand the technical aspects of one or more embodiments of the present disclosure, and to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in combination with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0013] The technical solutions provided by various embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0014] See also Figure 1 , Figure 1 One of the flow charts of the method for predicting operating parameters provided by one embodiment of the present invention is as follows: Figure 1 As shown, the method comprises the following steps: Step 102: Acquire multiple pre-trained prediction models and gated neural networks; wherein the prediction model is used to predict the initial second operating parameters of the target object based on the control parameters for the target object and the first operating parameters of the target object; different prediction models are used to predict the initial second operating parameters of the target object in different environmental types; the gated neural network is used to determine the probability of the target object being in different environmental types based on the first operating parameters of the target object; the initial second operating parameters predicted by different prediction models and the probability of the target object being in different environmental types determined by the gated neural network are used to predict the target second operating parameters of the target object.

[0015] Step 104: Input the current first operating parameters of the target object and the current control parameters for the target object into multiple prediction models and gated neural networks to obtain specific initial second operating parameters of the target object under multiple different environmental types, as well as specific probabilities of the target object being in different environmental types.

[0016] Step 106: predicting the current target second operating parameter of the target object based on the specific initial second operating parameters of the target object in a plurality of different environment types and the specific probabilities of the target object being in different environment types.

[0017] In the embodiment of the present invention, a plurality of pre-trained prediction models may be obtained first, wherein the prediction model may be used to predict the initial second operating parameter of the target object according to the control parameter for the target object and the first operating parameter of the target object.

[0018] In one embodiment of the present invention, the operating parameters may include abscissa, ordinate, heading angle, speed, and front wheel steering angle.

[0019] In one example, the operating parameters can be expressed as , which can respectively represent the abscissa, ordinate, heading angle, speed and front wheel turning angle of the jth time step of the ith environment type, where the time step can be set by relevant personnel or use the default value, which is not limited in this example.

[0020] In the embodiment of the present invention, different prediction models can be used to predict the initial second operating parameters of the same target object in different environment types.

[0021] In one embodiment of the present invention, when acquiring multiple prediction models, the historical first operating parameters of the target object can be acquired first, and then the historical first operating parameters can be clustered to obtain the historical first operating parameters corresponding to multiple environmental types. When clustering, existing clustering methods can be used for clustering, such as using a GMM (Gaussian Mixture Model) clustering algorithm for clustering, and the embodiment of the present invention does not limit this.

[0022] In one example, you can get The first operating parameter points of agricultural machinery are collected together to form a data set. ,in, is the total number of data points, and Then, we can perform GMM clustering on the data set D to obtain K clusters. , where each data point can be viewed as a sample in a five-dimensional space.

[0023] Then, in an embodiment of the present invention, a specified number of induced operating parameters can be obtained from the historical operating parameters corresponding to any environment type by means of maximum mutual information, and an induced data set corresponding to any environment type can be generated based on each induced operating parameter and its corresponding control parameter.

[0024] In one example, for each cluster , we can select from the data points corresponding to each cluster by maximizing mutual information data points as induced operation parameters, and the selected The data points constitute the induced operation parameter set ,in, Can be input; is the range space of the entire input parameter; Can be the corresponding output; are the input control parameters corresponding to the induced operating parameters.

[0025] Then, the prediction model corresponding to any environment type can be trained based on the induced data set corresponding to the environment type.

[0026] In one example, after determining each cluster Corresponding induced operation parameter set After that, the induced operation parameter set corresponding to each cluster can be , to train the prediction model corresponding to each cluster.

[0027] In another example, the trained prediction model may be referred to as a SGPR (Sparse Gaussian Process Regression) model.

[0028] From the above content, it can be seen that the embodiment of the present invention can cluster the historical operating parameters of the target object according to the clustering algorithm to obtain multiple different environment types, and then select representative induced operating parameters from the historical operating parameters corresponding to each environment type, and train the prediction model corresponding to each environment type according to the selected induced operating parameters, so as to obtain a prediction model suitable for different environment types, improve the prediction accuracy in different environments, and effectively enhance the ability to cope with diversified environments.

[0029] In one embodiment of the present invention, a pre-trained gated neural network may be obtained. The gated neural network may be used to determine the probability of the target object being in different types of environments based on the first operating parameters of the target object. The probability of the target object being in different types of environments determined by the gated neural network may be used together with the initial second operating parameters predicted by different prediction models to predict the target second operating parameters of the target object.

[0030] In this scheme, the five-dimensional operating parameters can be obtained Then, the five-dimensional operating parameters are input into the gated neural network to obtain the output is the probability that x belongs to each cluster, where T can represent the transpose of the matrix.

[0031] In one example, the training loss function when training the gated neural network can be:

[0032] in, is the indicative function; if the data point Belongs to cluster , then the value of the indicative function is 1; if it does not belong to a cluster , then the value of the indicative function is 0.

[0033] like Figure 2 As shown, in an embodiment of the present invention, the prediction model and the gated neural network can be trained in advance, and during the training, offline training can be performed using the collected data. Specifically, the first operating parameters and corresponding control parameters of the original target object, such as agricultural machinery, can be obtained first, and the obtained data can be used as the original data set. Then, clustering processing is performed on the obtained data to obtain representative data sets corresponding to different environmental types. For the representative data sets, the gated neural network training and the prediction model (i.e. Figure 2When training the gated neural network, the training can be carried out according to the first operating parameter in the data set and the environment type to which the first operating parameter belongs, combined with the above loss function; when training the prediction model, the induction point operating parameters corresponding to each environment type can be determined by the maximum mutual information method, and then the prediction model can be trained according to the induction point operating parameters to optimize the hyperparameters of the prediction model and obtain the trained prediction model. After obtaining the trained gated neural network and prediction model, a local composite prediction model can be obtained according to the trained gated neural network and each prediction model. After obtaining all the local composite prediction models, the final prediction result can be determined based on the output results of all the local composite prediction models. Since the prediction results under different environment types can be determined with emphasis through the gated neural network, and the final prediction result can be determined by combining the prediction results under each environment type, the scenario adaptability of the solution and the prediction accuracy for diversified scenarios can be effectively improved.

[0034] After obtaining multiple pre-trained prediction models and gated neural networks, the current first operating parameters of the target object and the current control parameters for the target object can be input into the multiple prediction models and gated neural networks to obtain specific initial second operating parameters of the target object in multiple different environmental types, as well as specific probabilities of the target object being in different environmental types.

[0035] Then, the current target second operating parameter of the target object may be predicted based on the specific initial second operating parameters of the target object in a plurality of different environment types and the specific probabilities that the target object is in different environment types.

[0036] Among them, when predicting the target object, for any environment type, the specific initial second operating parameters of the target object under any environment type and the product of the specific probability that the target object is in any environment type can be obtained, and then, the current target second operating parameters of the target object can be predicted based on the sum of the multiple products obtained.

[0037] In one example, the specific probability that the target object is in any environment type can be used as the weight value of the specific initial second operating parameter, the specific probability is multiplied by the specific initial second operating parameter, and the multiplication values ​​corresponding to different environment types are summed, and the current target second operating parameter of the target object is predicted based on the obtained sum.

[0038] In an example, the following formula can be obtained:

[0039] in, represents the prediction model, that is, after z is input into the prediction model, the prediction model outputs the above-mentioned specific initial second operating parameter. In this example, Can be used with Figure 2 It corresponds to the local composite prediction model in, that is, The value of can correspond to Figure 2 Output of the local composite prediction model in .

[0040] like Figure 3 As shown, as can be seen from the above content, this scheme uses the specific probability of the target object being in any environment type as the weight value of the specific initial second operating parameter in any environment type, so that the target second operating parameter can be predicted in combination with the environment type in a focused manner, thereby maintaining a high prediction accuracy in a variety of environments and effectively responding to complex and diverse environments.

[0041] In an embodiment of the present invention, Figure 4 As shown, when the current first operating parameters of the target object and the current control parameters for the target object are input into multiple prediction models and gated neural networks, the specific initial second operating parameters of the target object under multiple different environmental types and the specific probabilities of the target object being in different environmental types are obtained, it can be determined whether to add the input data to the induced data set.

[0042] Specifically, the specific probability of the target object being in each environment type can be determined according to the gated neural network, and then the specific environment type corresponding to the maximum specific probability is determined, and the first posterior variance of the prediction model corresponding to the specific environment type is determined, and after the current first operating parameter of the target object is added to the induced data set corresponding to the specific environment type, the second posterior variance of the prediction model corresponding to the specific environment type is determined. When the difference between the first posterior variance and the second posterior variance is greater than the preset addition difference threshold, the current first operating parameter of the target object is added to the induced data set corresponding to the specific environment type.

[0043] In one example, when a new data point When it arrives, it first passes through the gating network to calculate the probability of belonging to each cluster Then, for the cluster with the highest probability Can you consider whether Add its induced operation parameter set .

[0044] Specifically, we can calculate the addition The reduction in posterior variance of the posterior prediction model at this point is:

[0045] in, and Clustering before and after joining The corresponding prediction model is The posterior variance at . Greater than the preset difference addition threshold , then join in .

[0046] In another embodiment of the present invention, in order to avoid too many induced data points in the induced data set, it is possible to consider removing points with redundant information and poor representativeness from the induced operation parameter set. When removing, the increase in the posterior variance of the prediction model at the point after removal can be calculated for each candidate removal point, that is, the difference between the posterior variance of the prediction model at the point after removal and the posterior variance of the prediction model at the point before removal. When the difference is less than the preset removal difference threshold, the point is removed.

[0047] In one example, to avoid too many induction points, consider Remove points with redundant information. For each candidate removal point , we can calculate the increase in the posterior variance of the prediction model at this point after removal:

[0048] in and The prediction models before and after removal are The posterior variance at . Less than the preset removal difference threshold , execute the removal.

[0049] like Figure 4 As shown in Figure 2, after updating the induced operation parameter set, the hyperparameters of the prediction model corresponding to each cluster can be updated by maximizing the corrected marginal likelihood logarithm. :

[0050] In an embodiment of the present invention, after predicting the current target second operating parameters of the target object based on the specific initial second operating parameters of the target object in multiple different environmental types and the specific probabilities of the target object being in different environmental types, the data distribution of the current first operating parameters can also be determined based on the current first operating parameters of the target object.

[0051] In one example, the Gaussian process model of the residual dynamics can be given by , approximating the propagation of state distributions through a composite system model:

[0052] in, It can be a state evolution function, and an existing state evolution function can be used; It can be a residual coefficient matrix, which can be determined based on the difference between the predicted value and the actual value of the predetermined prediction model; It is the noise value, which can be set by relevant personnel according to the current environment type.

[0053] Then, the prediction variance of the specific initial second operating parameter can be determined based on the data distribution of the current first operating parameter, a predetermined state evolution formula, a predetermined uncertainty propagation formula, and a predetermined residual mean and residual variance, and the constraint bound corresponding to the specific initial second operating parameter can be determined based on the diagonal elements of the prediction variance of the specific initial second operating parameter, a predetermined inverse cumulative distribution function of the standard normal distribution, and a predetermined confidence parameter.

[0054] In one example, assume that the initial state follows a Gaussian distribution , the first-order Taylor expansion approximate model predictive control near the mean trajectory can be used to predict the state prediction mean and variance of each time step k+i in the time domain:

[0055]

[0056] in, ; .

[0057] Then, using the Gaussian distribution of states, a tightened bound is calculated:

[0058] in, is the uncertainty corresponding to the lateral error diagonal element; It can be the inverse cumulative distribution function of the standard normal distribution; It is a parameter related to the confidence level or constraint condition, which is used to determine the strictness or confidence level when calculating the constraint limit. It can be a default value or set by relevant personnel.

[0059] Simplify the joint chance constraints to a set of deterministic constraints:

[0060] In the embodiment of the present invention, Figure 5 As shown, after determining the constraint bounds corresponding to the specific initial second operating parameters, the control strategy for the current first operating parameters of the target object can be determined through a predetermined stochastic model predictive control optimization algorithm based on the specific initial second operating parameters, the specific probability, and the constraint bounds corresponding to the specific initial second operating parameters.

[0061] In one example, the prediction results of the prediction model determined above, the approximate uncertainty propagation and the simplified chance constraints can be combined, and the optimization problem of the control strategy of the target object can be expressed as:

[0062]

[0063]

[0064]

[0065]

[0066]

[0067]

[0068]

[0069] in, , are the predicted mean and covariance, respectively; , are the state and control reference trajectories respectively; , , is the weighting matrix; , , , state and control boundaries; , is the posterior mean and covariance of j local GPRs; is the jth local GPR weight output by the hybrid expert gating network; is the lateral tracking error The chance-constrained confidence is .

[0070] In the embodiment of the present invention, the above optimization problem can be solved using nonlinear programming methods such as SQP (Sequential quadratic programming) or interior point method to obtain the optimal control sequence The target object can be controlled to run in a rolling horizon mode, executing at each time step And repeat the optimization.

[0071] In an embodiment of the present invention, a plurality of pre-trained prediction models and a gated neural network may be first obtained, wherein the prediction model may be used to predict the initial second operating parameters of the target object based on the control parameters for the target object and the first operating parameters of the target object, different prediction models are used to predict the initial second operating parameters of the target object in different environmental types, and the gated neural network may be used to determine the probability of the target object being in different environmental types based on the first operating parameters of the target object, the initial second operating parameters predicted by the different prediction models, and the probability of the target object being in different environmental types determined by the gated neural network may be used to predict the target second operating parameters of the target object, and then, the current first operating parameters of the target object and the current control parameters for the target object may be input into the plurality of prediction models and the gated neural network to obtain specific initial second operating parameters of the target object in multiple different environmental types, and specific probabilities of the target object being in different environmental types, and the current target second operating parameters of the target object may be predicted based on the obtained specific initial second operating parameters and specific probabilities.

[0072] As can be seen from the above content, the embodiment of the present invention can obtain multiple prediction models for different environmental types, and can predict the initial second operating parameters of the target object under the control parameters based on the obtained prediction models, as well as the current operating parameters and control parameters of the target object. In addition, the embodiment of the present invention can also combine the gated neural network to determine the probability that the target object is currently in different environmental types based on the current operating parameters of the target object, and combine the probability of the target object currently being in different environmental types, and focus on combining the initial second operating parameters of the target object in different environmental types under the control of the input control parameters to predict the current target second operating parameters of the target object, so that it can maintain a high prediction accuracy in a variety of environments, effectively respond to complex and diverse environments, and effectively solve the problem in the related technology that the prediction of the operation trajectory of agricultural machinery in a variety of agricultural scenes is not accurate due to the use of a single global model that is difficult to accurately describe the diverse dynamic characteristics in different scenes, so that the relevant personnel cannot determine the appropriate control amount, and thus cannot accurately control the agricultural machinery to run according to the desired trajectory.

[0073] Corresponding to the above-mentioned operation parameter prediction method, the embodiment of the present invention further provides an operation parameter prediction device, Figure 6 A schematic diagram of the module composition of the operating parameter prediction device 600 provided in an embodiment of the present invention, as shown in FIG. Figure 6 As shown, the operating parameter prediction device 600 includes: The first acquisition module 601 is used to acquire a plurality of pre-trained prediction models and gated neural networks; wherein the prediction model is used to predict the initial second operating parameters of the target object according to the control parameters for the target object and the first operating parameters of the target object; different prediction models are used to predict the initial second operating parameters of the target object in different environmental types; the gated neural network is used to determine the probability of the target object being in the different environmental types according to the first operating parameters of the target object; the initial second operating parameters predicted by different prediction models and the probability of the target object being in the different environmental types determined by the gated neural network are used to predict the target second operating parameters of the target object; A second acquisition module 602 is used to input the current first operating parameter of the target object and the current control parameter for the target object into the multiple prediction models and the gated neural network to obtain specific initial second operating parameters of the target object under the multiple different environment types, and specific probabilities of the target object being in the different environment types; The prediction module 603 is used to predict the current target second operating parameters of the target object according to the specific initial second operating parameters of the target object in the multiple different environment types and the specific probability that the target object is in the different environment types.

[0074] Optionally, the first obtaining module 601 is used to: Acquire a first historical operating parameter of the target object; Clustering the historical first operating parameters to obtain historical first operating parameters corresponding to multiple environment types; Obtaining a specified number of induced operating parameters from the historical operating parameters corresponding to any environment type by means of maximum mutual information, and generating an induced data set corresponding to any environment type according to each of the induced operating parameters and their corresponding control parameters; The prediction model corresponding to any environment type is trained according to the induced data set corresponding to any environment type.

[0075] Optionally, the prediction module 603 is used to: For any environment type, obtaining a product of a specific initial second operating parameter of the target object in the environment type and a specific probability that the target object is in the environment type; The current target second operating parameter of the target object is predicted according to the sum of the multiple products obtained.

[0076] Optionally, the device further comprises: A first determination module 604 is configured to determine a data distribution of a current first operating parameter according to the current first operating parameter of the target object after predicting the current target second operating parameter of the target object according to the specific initial second operating parameters of the target object in the multiple different environment types and the specific probabilities that the target object is in the different environment types; A second determination module 605 is used to determine the prediction variance of the specific initial second operating parameter according to the data distribution of the current first operating parameter, a predetermined state evolution formula, a predetermined uncertainty propagation formula, and a predetermined residual mean and residual variance; The third determination module 606 is used to determine the constraint bound corresponding to the specific initial second operating parameter according to the diagonal element of the prediction variance of the specific initial second operating parameter, the predetermined inverse cumulative distribution function of the standard normal distribution, and the predetermined confidence parameter.

[0077] Optionally, the device further comprises ( Figure 6 not shown): The fourth determination module 607 is used to determine the control strategy for the current first operating parameter of the target object through a predetermined stochastic model predictive control optimization algorithm based on the specific initial second operating parameter, the specific probability, and the constraint bound corresponding to the specific initial second operating parameter after determining the constraint bound corresponding to the specific initial second operating parameter.

[0078] Optionally, the second acquisition module 602 is used to: Determining, according to the gated neural network, a specific probability that the target object is in each type of environment; Determine the specific type of environment corresponding to the maximum specific probability; Determining a first a posteriori variance of the prediction model corresponding to the specific environment type, and a second a posteriori variance of the prediction model corresponding to the specific environment type after adding the current first operating parameter of the target object to the induced data set corresponding to the specific environment type; When the difference between the first a posteriori variance and the second a posteriori variance is greater than a preset addition difference threshold, the current first operating parameter of the target object is added to the induced data set corresponding to the specific environment type.

[0079] In an embodiment of the present invention, a plurality of pre-trained prediction models and a gated neural network may be first obtained, wherein the prediction model may be used to predict the initial second operating parameters of the target object based on the control parameters for the target object and the first operating parameters of the target object, different prediction models are used to predict the initial second operating parameters of the target object in different environmental types, and the gated neural network may be used to determine the probability of the target object being in different environmental types based on the first operating parameters of the target object, the initial second operating parameters predicted by the different prediction models, and the probability of the target object being in different environmental types determined by the gated neural network may be used to predict the target second operating parameters of the target object, and then, the current first operating parameters of the target object and the current control parameters for the target object may be input into the plurality of prediction models and the gated neural network to obtain specific initial second operating parameters of the target object in multiple different environmental types, and specific probabilities of the target object being in different environmental types, and the current target second operating parameters of the target object may be predicted based on the obtained specific initial second operating parameters and specific probabilities.

[0080] As can be seen from the above content, the embodiment of the present invention can obtain multiple prediction models for different environmental types, and can predict the initial second operating parameters of the target object under the control parameters based on the obtained prediction models, as well as the current operating parameters and control parameters of the target object. In addition, the embodiment of the present invention can also combine the gated neural network to determine the probability that the target object is currently in different environmental types based on the current operating parameters of the target object, and combine the probability of the target object currently being in different environmental types, and focus on combining the initial second operating parameters of the target object in different environmental types under the control of the input control parameters to predict the current target second operating parameters of the target object, so that it can maintain a high prediction accuracy in a variety of environments, effectively respond to complex and diverse environments, and effectively solve the problem in the related technology that the prediction of the operation trajectory of agricultural machinery in a variety of agricultural scenes is not accurate due to the use of a single global model that is difficult to accurately describe the diverse dynamic characteristics in different scenes, so that the relevant personnel cannot determine the appropriate control amount, and thus cannot accurately control the agricultural machinery to run according to the desired trajectory.

[0081] Corresponding to the above-mentioned operation parameter prediction method, the embodiment of the present invention further provides an operation parameter prediction device, Figure 7 A schematic diagram of the hardware structure of an operating parameter prediction device provided by an embodiment of the present invention.

[0082] The operating parameter prediction device may be the terminal device or server for predicting the trajectory provided in the above embodiment.

[0083] The prediction device of the operating parameters may have relatively large differences due to different configurations or performances, and may include one or more processors 701 and memory 702, and one or more storage applications or data may be stored in the memory 702. Among them, the memory 702 may be a short-term storage or a persistent storage. The application stored in the memory 702 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the prediction device of the operating parameters. Furthermore, the processor 701 may be configured to communicate with the memory 702, and execute a series of computer executable instructions in the memory 702 on the prediction device of the operating parameters. The prediction device of the operating parameters may also include one or more power supplies 703, one or more wired or wireless network interfaces 704, one or more input and output interfaces 705, and one or more keyboards 706.

[0084] Specifically in this embodiment, the prediction device for operating parameters includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions in the prediction device for operating parameters, and is configured to be executed by one or more processors to perform the above embodiments.

[0085] In an embodiment of the present invention, a plurality of pre-trained prediction models and a gated neural network may be first obtained, wherein the prediction model may be used to predict the initial second operating parameters of the target object based on the control parameters for the target object and the first operating parameters of the target object, different prediction models are used to predict the initial second operating parameters of the target object in different environmental types, and the gated neural network may be used to determine the probability of the target object being in different environmental types based on the first operating parameters of the target object, the initial second operating parameters predicted by the different prediction models, and the probability of the target object being in different environmental types determined by the gated neural network may be used to predict the target second operating parameters of the target object, and then, the current first operating parameters of the target object and the current control parameters for the target object may be input into the plurality of prediction models and the gated neural network to obtain specific initial second operating parameters of the target object in multiple different environmental types, and specific probabilities of the target object being in different environmental types, and the current target second operating parameters of the target object may be predicted based on the obtained specific initial second operating parameters and specific probabilities.

[0086] As can be seen from the above content, the embodiment of the present invention can obtain multiple prediction models for different environmental types, and can predict the initial second operating parameters of the target object under the control parameters based on the obtained prediction models, as well as the current operating parameters and control parameters of the target object. In addition, the embodiment of the present invention can also combine the gated neural network to determine the probability that the target object is currently in different environmental types based on the current operating parameters of the target object, and combine the probability of the target object currently being in different environmental types, and focus on combining the initial second operating parameters of the target object in different environmental types under the control of the input control parameters to predict the current target second operating parameters of the target object, so that it can maintain a high prediction accuracy in a variety of environments, effectively respond to complex and diverse environments, and effectively solve the problem in the related technology that the prediction of the operation trajectory of agricultural machinery in a variety of agricultural scenes is not accurate due to the use of a single global model that is difficult to accurately describe the diverse dynamic characteristics in different scenes, so that the relevant personnel cannot determine the appropriate control amount, and thus cannot accurately control the agricultural machinery to run according to the desired trajectory.

[0087] Another embodiment of the present disclosure further provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store computer-executable instructions, and the computer-executable instructions implement the above process when executed by a processor.

[0088] The storage medium in the embodiment of the present disclosure can implement each process of the above-mentioned operating parameter prediction method embodiment and achieve the same effect and function, which will not be repeated here.

[0089] Another embodiment of the present disclosure further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the above process is implemented.

[0090] The computer program product in the embodiments of the present disclosure can implement the various processes of the above-mentioned operating parameter prediction method embodiment and achieve the same effects and functions, which will not be repeated here.

[0091] In various embodiments of the present disclosure, the computer-readable storage medium includes a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0092] In the 1990s, it was very clear whether the improvement of a technology was hardware improvement (for example, improvement of the circuit structure of diodes, transistors, switches, etc.) or software improvement (improvement of the method flow). However, with the development of technology, many improvements of the method flow today can be regarded as direct improvements of the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that the improvement of a method flow cannot be implemented with a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming themselves, without having to ask chip manufacturers to design and make dedicated integrated circuit chips. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one kind of HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0093] The controller may be implemented in any suitable manner, for example, the controller may take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (e.g., software or firmware) executable by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, and the memory controller may also be implemented as part of the control logic of the memory. It is also known to those skilled in the art that, in addition to implementing the controller in a purely computer-readable program code manner, the controller may be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, such a controller may be considered as a hardware component, and the devices for implementing various functions included therein may also be considered as structures within the hardware component. Or even, the devices for implementing various functions may be considered as both software modules for implementing the method and structures within the hardware component.

[0094] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0095] For the convenience of description, the above devices are described in terms of functions and are divided into various units. Of course, when implementing the embodiments of the present disclosure, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0096] It should be understood by those skilled in the art that one or more embodiments of the present disclosure may be provided as a method, system or computer program product. Therefore, one or more embodiments of the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, one or more embodiments of the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0097] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0098] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0099] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0100] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0101] One or more embodiments of the present disclosure may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. One or more embodiments of the present disclosure may also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0102] Each embodiment in the present disclosure is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0103] The above description is only an embodiment of the present disclosure and is not intended to limit the present disclosure. For those skilled in the art, the present disclosure may have various changes and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the scope of the claims of the present disclosure.

Claims

1. A method for predicting operating parameters, characterized in that: The method comprises: Acquire multiple pre-trained prediction models and gated neural networks; wherein the prediction model is used to predict the initial second operating parameters of the target object according to the control parameters for the target object and the first operating parameters of the target object; different prediction models are used to predict the initial second operating parameters of the target object in different environmental types; the gated neural network is used to determine the probability of the target object being in the different environmental types according to the first operating parameters of the target object; the initial second operating parameters predicted by different prediction models and the probability of the target object being in the different environmental types determined by the gated neural network are used to predict the target second operating parameters of the target object; Inputting the current first operating parameter of the target object and the current control parameter for the target object into the plurality of prediction models and the gated neural network to obtain specific initial second operating parameters of the target object under the plurality of different environmental types, and specific probabilities of the target object being in the different environmental types; According to the specific initial second operating parameters of the target object in the multiple different environment types and the specific probability that the target object is in the different environment types, the current target second operating parameters of the target object are predicted.

2. The method according to claim 1, characterized in that The obtaining of multiple pre-trained prediction models includes: Acquire a first historical operating parameter of the target object; Clustering the historical first operating parameters to obtain historical first operating parameters corresponding to multiple environment types; Obtaining a specified number of induced operating parameters from the historical first operating parameters corresponding to any environment type by means of maximum mutual information, and generating an induced data set corresponding to any environment type according to each of the induced operating parameters and their corresponding control parameters; The prediction model corresponding to any environment type is trained according to the induced data set corresponding to any environment type.

3. The method according to claim 1, characterized in that The predicting of the current target second operating parameter of the target object according to the specific initial second operating parameters of the target object in the multiple different environment types and the specific probability that the target object is in the different environment types includes: For any environment type, obtaining a product of a specific initial second operating parameter of the target object in the environment type and a specific probability that the target object is in the environment type; The current target second operating parameter of the target object is predicted according to the sum of the multiple products obtained.

4. The method according to claim 1, characterized in that After predicting the current target second operating parameter of the target object according to the specific initial second operating parameters of the target object in the multiple different environment types and the specific probability that the target object is in the different environment types, the method further includes: Determining data distribution of the current first operating parameter according to the current first operating parameter of the target object; Determining the prediction variance of the specific initial second operating parameter according to the data distribution of the current first operating parameter, a predetermined state evolution formula, a predetermined uncertainty propagation formula, and a predetermined residual mean and residual variance; The constraint bound corresponding to the specific initial second operating parameter is determined according to the diagonal element of the prediction variance of the specific initial second operating parameter, a predetermined inverse cumulative distribution function of a standard normal distribution, and a predetermined confidence parameter.

5. The method according to claim 4, characterized in that After determining the constraint bound corresponding to the specific initial second operating parameter, the method further includes: According to the specific initial second operating parameter, the specific probability, and the constraint bound corresponding to the specific initial second operating parameter, a control strategy for the current first operating parameter of the target object is determined through a predetermined stochastic model predictive control optimization algorithm.

6. The method according to claim 2, characterized in that The step of inputting the current first operating parameter of the target object and the current control parameter for the target object into the plurality of prediction models and the gated neural network to obtain specific initial second operating parameters of the target object under the plurality of different environment types, and specific probabilities of the target object being in the different environment types, comprises: Determining, according to the gated neural network, a specific probability that the target object is in each type of environment; Determine the specific type of environment corresponding to the maximum specific probability; Determining a first a posteriori variance of the prediction model corresponding to the specific environment type, and a second a posteriori variance of the prediction model corresponding to the specific environment type after adding the current first operating parameter of the target object to the induced data set corresponding to the specific environment type; When the difference between the first a posteriori variance and the second a posteriori variance is greater than a preset addition difference threshold, the current first operating parameter of the target object is added to the induced data set corresponding to the specific environment type.

7. A device for predicting operating parameters, characterized in that: The device comprises: A first acquisition module is used to acquire a plurality of pre-trained prediction models and a gated neural network; wherein the prediction model is used to predict the initial second operating parameter of the target object according to the control parameter for the target object and the first operating parameter of the target object; different prediction models are used to predict the initial second operating parameter of the target object in different environmental types; the gated neural network is used to determine the probability of the target object being in the different environmental types according to the first operating parameter of the target object; the initial second operating parameters predicted by different prediction models and the probability of the target object being in the different environmental types determined by the gated neural network are used to predict the target second operating parameter of the target object; A second acquisition model is used to input the current first operating parameter of the target object and the current control parameter for the target object into the plurality of prediction models and the gated neural network to obtain specific initial second operating parameters of the target object under the plurality of different environmental types, and specific probabilities of the target object being in the different environmental types; The prediction module is used to predict the current target second operating parameter of the target object according to the specific initial second operating parameters of the target object in the multiple different environment types and the specific probability that the target object is in the different environment types.

8. A device for predicting operating parameters, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program implements the steps of the method according to any one of claims 1 to 6 when executed by the processor.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store computer-executable instructions, and when the computer-executable instructions are executed by a processor, the steps of the method described in any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images