Motion state estimation method of spatial non-cooperative target

Abstract

The invention relates to a motion state estimation method for a spatial non-cooperative target, and relates to the technical field of aerospace, and the method comprises the steps: inputting a current state estimation value into a disturbing force prediction model, and obtaining a current disturbing force prediction value of the spatial non-cooperative target at a current moment; determining an average observation value of the spatial non-cooperative target at a next moment corresponding to the current moment according to the current state estimation value and the current covariance estimation value, and determining a prediction state estimation value and a prediction covariance estimation value of the spatial non-cooperative target at the next moment according to the current interference force prediction value; and determining a target state estimation value of the spatial non-cooperative target at the next moment based on the average observation value, the prediction state estimation value and the prediction covariance estimation value, and estimating the motion state of the spatial non-cooperative target at the next moment based on the target state estimation value. According to the invention, the accuracy of the estimated motion state is improved.

Description

Background Art

[0002] During the flight of non-cooperative space targets, they are subject to multiple types of unmodeled, nonlinear external interference, including air resistance, solar pressure, geomagnetic field force, etc. This part of the dynamics makes it extremely difficult to accurately calculate the motion of non-cooperative targets. Traditionally, the unmodeled part of the target dynamics is regarded as generalized noise, which is filtered and eliminated by a low-pass filter. However, this method is only applicable to constant interference forces.

[0003] However, in practice, in most cases the interference force is related to the motion state of the non-cooperative target and is therefore time-varying. There is a phase difference between the estimated interference force obtained using a low-pass filter and the actual interference force. If the estimated interference force is substituted into the motion estimator, significant errors will be introduced, ultimately reducing the service satellite's estimation effect on the non-cooperative target state; that is, the accuracy of the estimated motion state is low.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0005] The purpose of the present disclosure is to provide a method for estimating the motion state of a spatial non-cooperative target, thereby overcoming the problem of low accuracy of the estimated motion state caused by the limitations and defects of related technologies, at least to a certain extent.

[0006] According to one aspect of the present disclosure, a method for estimating a motion state of a space non-cooperative target is provided, comprising:

[0007] Establishing a state space equation of a space non-cooperative target and an observation equation of a service satellite to the space non-cooperative target, and creating a state filter based on the state space equation and the observation equation;

[0008] Determine a current state estimation value and a current covariance estimation value of the spatial non-cooperative target at the current moment based on the state filter, and input the current state estimation value into the interference force prediction model to obtain a current interference force prediction value of the spatial non-cooperative target at the current moment;

[0009] Determine the average observation value of the spatial non-cooperative target at a next moment corresponding to the current moment according to the current state estimate and the current covariance estimate, and determine the predicted state estimate and the predicted covariance estimate of the spatial non-cooperative target at the next moment according to the current interference force prediction value;

[0010] Based on the average observation value, the predicted state estimate value and the predicted covariance estimate value, the target state estimate value of the spatial non-cooperative target at the next moment is determined, and based on the target state estimate value, the motion state of the spatial non-cooperative target at the next moment is estimated.

[0011] In an exemplary embodiment of the present disclosure, the interference force prediction model is obtained in the following manner:

[0012] Determine a first historical state estimate, a first historical covariance estimate, and a first historical residual value of the spatial non-cooperative target at time k-1 according to the state filter;

[0013] The first historical residual value is subjected to a limiting process and a recursive average filtering process to obtain a first deviation filtering value; wherein the timestamp of the middle value of the average filtering queue is M;

[0014] Reading a second historical state estimation value corresponding to the timestamp M in the state filter, and inputting the second historical state estimation value into the neural network model to be trained to obtain a deviation prediction result;

[0015] The parameters in the neural network model to be trained are adjusted according to the first deviation filtering value and the deviation prediction result to obtain the interference force prediction model.

[0016] In an exemplary embodiment of the present disclosure, the second historical state estimation value is input into a neural network model to be trained to obtain a deviation prediction result, including:

[0017] Randomly extracting a plurality of different samples from the second historical state estimation value as initial cluster centers, and randomly extracting training samples from the second historical state estimation value as input samples;

[0018] Determine, according to the input sample and the initial cluster center, the sample category to which the input sample belongs, and update the initial cluster center based on the sample category;

[0019] Based on the neural network model to be trained, the updated cluster center is updated again, and it is determined whether the change range of the updated cluster center is less than or equal to a preset threshold;

[0020] If the change range of the updated cluster center is less than or equal to the preset threshold, the deviation prediction result is obtained based on the updated cluster center.

[0021] In an exemplary embodiment of the present disclosure, adjusting the parameters in the neural network model to be trained according to the first deviation filter value and the deviation prediction result to obtain the interference force prediction model includes:

[0022] The standard deviation between the first deviation filter value and the deviation prediction result is calculated according to the Gaussian function in the neural network model to be trained, and the parameters in the neural network model to be trained are adjusted based on the standard deviation to obtain the interference force prediction model.

[0023] In an exemplary embodiment of the present disclosure, determining the average observation value of the space non-cooperative target at a next moment corresponding to the current moment according to the current state estimate and the current covariance estimate includes:

[0024] An unscented transformation is performed on the current state estimate and the current covariance estimate to obtain a current sampling point, and an average observation value of the non-cooperative target at a next moment is determined according to the current sampling point and the observation equation.

[0025] In an exemplary embodiment of the present disclosure, determining a predicted state estimate value and a predicted covariance estimate value of the space non-cooperative target at the next moment according to a current interference force prediction value includes:

[0026] Determining a predicted state estimate value of the spatial non-cooperative target at a next moment according to the current interference force prediction value and the current sampling point;

[0027] Determine a covariance estimate of the spatial non-cooperative target at a next moment according to the current sampling point and the predicted state estimate.

[0028] In an exemplary embodiment of the present disclosure, determining a target state estimate of the space non-cooperative target at the next moment based on the average observation value, the predicted state estimate, and the predicted covariance estimate includes:

[0029] Determining the information covariance and the cross covariance of the spatial non-cooperative target based on the average observation value, and determining the gain matrix of the spatial non-cooperative target according to the information covariance and the cross covariance;

[0030] According to the predicted state estimate value and the gain matrix, a target state estimate value of the spatial non-cooperative target at the next moment is determined.

[0031] In an exemplary embodiment of the present disclosure, determining the target state estimate value of the spatial non-cooperative target at the next moment according to the predicted state estimate value and the gain matrix includes:

[0032] Determine the target relative position between the space non-cooperative target and the service satellite at the next moment, and determine the target state estimate of the space non-cooperative target at the next moment based on the predicted state estimate, the gain matrix and the target relative position.

[0033] In an exemplary embodiment of the present disclosure, determining the relative position of the space non-cooperative target with respect to the serving satellite at the next moment includes:

[0034] Determine a current satellite position of the serving satellite in the orbital coordinate system at a current moment, and determine a current relative position between the space non-cooperative target and the serving satellite at a current moment according to the current state estimate;

[0035] The target relative position between the space non-cooperative target and the service satellite at the next moment is determined according to the current relative position and the current satellite position.

[0036] In an exemplary embodiment of the present disclosure, the state filter includes any one of an adaptive unscented Kalman filter, an adaptive extended Kalman filter and an adaptive cubature Kalman filter.

[0037] The present disclosure provides a method for estimating the motion state of a space non-cooperative target. On the one hand, a state space equation of the space non-cooperative target and an observation equation of the space non-cooperative target by a service satellite are established, and a state filter is created based on the state space equation and the observation equation; then, a current state estimation value and a current covariance estimation value of the space non-cooperative target at the current moment are determined based on the state filter, and the current state estimation value is input into an interference force prediction model to obtain a current interference force prediction value of the space non-cooperative target at the current moment; then, an average observation value of the space non-cooperative target at the next moment corresponding to the current moment is determined based on the current state estimation value and the current covariance estimation value, and a current interference force prediction value is obtained based on the current interference force. The predicted value determines the predicted state estimate and predicted covariance estimate of the spatial non-cooperative target at the next moment; finally, based on the average observation value, the predicted state estimate and the predicted covariance estimate, the target state estimate of the spatial non-cooperative target at the next moment is determined, and the motion state of the spatial non-cooperative target at the next moment is estimated based on the target state estimate. Since the interference force prediction value can be determined based on the interference force cloud measurement model, and then the motion state is estimated based on the interference force prediction value, the accuracy of the estimated motion state is improved; on the other hand, since the interference force prediction can be performed through the trained interference force prediction model, there is no need to train the model in real time, thereby improving the efficiency of motion state estimation.

[0038] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0040] Figure 1 A flowchart of a method for estimating the motion state of a spatial non-cooperative target according to an exemplary embodiment of the present disclosure is schematically shown.

[0041] Figure 2 A block diagram schematically illustrates the implementation principle of a method for estimating the motion state of a spatial non-cooperative target according to an exemplary embodiment of the present disclosure.

[0042] Figure 3 A diagram schematically shows an example structure of a radial basis function network according to an example embodiment of the present disclosure.

[0043] Figure 4 A flowchart of a radial basis function network training method according to an exemplary embodiment of the present disclosure is schematically shown.

[0044] Figure 5 An example diagram schematically illustrates a training scenario of a radial basis function network according to an example embodiment of the present disclosure.

[0045] Figure 6 An example diagram of a location scenario of an observation satellite and a space non-cooperative target according to an example embodiment of the present disclosure is schematically shown.

[0046] Figure 7 A block diagram schematically shows a motion state estimation device for a spatial non-cooperative target according to an exemplary embodiment of the present disclosure.

[0047] Figure 8 An electronic device for implementing a motion state estimation method for a spatial non-cooperative target according to an exemplary embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0048] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concepts of the example embodiments are fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0049] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0050] In the field of aerospace, on-orbit service missions for non-cooperative targets in space can ensure the safety of the space environment. Non-cooperative targets in space are uncontrollable and difficult to measure, so accurately measuring the motion state of non-cooperative targets in space is the basis for carrying out other types of on-orbit services. During the flight, non-cooperative targets in space are subject to multiple types of unmodeled, nonlinear external interference, including air resistance, solar pressure, geomagnetic field force, etc. This part of the dynamics brings great difficulties to the accurate calculation of the motion of non-cooperative targets. Traditionally, the unmodeled part of the target dynamics is regarded as generalized noise, which is filtered and eliminated by a low-pass filter. However, this method is only applicable to constant interference forces. In practice, in most cases, the interference force is related to the motion state of the non-cooperative target and is therefore time-varying. There is a phase difference between the estimated interference force obtained by the low-pass filter and the real interference force. If the estimated interference force is substituted into the motion estimator, significant errors will be introduced, which will eventually reduce the estimation effect of the service satellite on the state of the non-cooperative target.

[0051] The interference force is usually related to the state of the non-cooperative target. If a motion state-interference force model can be constructed using a neural network, the motion state estimated by the motion state estimator can be input into the interference force model, and the interference force currently suffered by the non-cooperative target output by the interference force model can be input into the motion state estimator, thereby making the dynamic integral part of the motion state estimator more accurate, and ultimately greatly improving the motion estimation effect. From this point of view, the residual of the offline training neural network calculation estimator proposed in the prior art cannot be applied to the aerospace field to estimate the state of non-cooperative targets; the specific reason is that: on the one hand, the satellite cannot obtain the data of external interference suffered by non-cooperative targets in space in advance, so it is impossible to realize offline training of neural networks; on the other hand, the quality of the residual data contains a lot of noise and wild values, and direct use will seriously pollute the training data. Therefore, it is extremely necessary to design a new algorithm for online and accurate calculation of environmental interference suffered by non-cooperative targets in space.

[0052] Therefore, compared with previous patents, the innovation of this patent lies in: 1) the idea of ​​"first using adaptive Kalman filtering to train the neural network online, and then using the neural network to improve the state estimation effect of the adaptive Kalman filter" is proposed, which solves the problem of state-related interference force in the motion state estimation of non-cooperative targets; 2) a method of using adaptive Kalman filtering to train the neural network online is proposed, in which the state quantity output by the adaptive Kalman filter is used as the known quantity of the neural network, and the residual quantity output is used as the expected value of the neural network; 3) it is proposed to perform low-pass filtering on the residual obtained by the adaptive Kalman filter calculation and time-align the obtained state, so as to obtain a more accurate state-interference force data pair, providing high-quality data support for training a reliable interference force model. At the same time, in the example embodiment of the present disclosure, the state estimator is basically equivalent to the adaptive Kalman filtering method, the residual value of the state estimator is essentially the interference force, and the motion state quantity includes the position and speed of the non-cooperative target. Furthermore, the state filter recorded in the example embodiments of the present disclosure adopts an adaptive unscented Kalman filter, and other adaptive Kalman filter schemes such as an adaptive extended Kalman filter or an adaptive cubature Kalman filter may also be adopted; similarly, the low-pass filter adopts a recursive averaging filter, and other low-pass digital filters may also be adopted, such as an averaging filter, a recursive filter, etc.; similarly, the neural network adopts a radial basis neural network, and in fact other schemes such as a multilayer perceptron neural network may also be adopted.

[0053] In an exemplary embodiment, the motion state estimation method of a spatial non-cooperative target described in the exemplary embodiment of the present disclosure may be run on a server, a server cluster, or a cloud server, etc. Of course, those skilled in the art may also run the method of the present disclosure on other platforms as required, and this exemplary embodiment does not specifically limit this. Figure 1 As shown, the motion state estimation method of the spatial non-cooperative target may include the following steps:

[0054] Step S110. Establishing a state space equation of a space non-cooperative target and an observation equation of the service satellite to the space non-cooperative target, and creating a state filter based on the state space equation and the observation equation;

[0055] Step S120. Determine the current state estimation value and the current covariance estimation value of the spatial non-cooperative target at the current moment based on the state filter, and input the current state estimation value into the interference force prediction model to obtain the current interference force prediction value of the spatial non-cooperative target at the current moment;

[0056] Step S130. Determine the average observation value of the spatial non-cooperative target at the next moment corresponding to the current moment according to the current state estimation value and the current covariance estimation value, and determine the predicted state estimation value and predicted covariance estimation value of the spatial non-cooperative target at the next moment according to the current interference force prediction value;

[0057] Step S140. Based on the average observation value, the predicted state estimate value and the predicted covariance estimate value, determine the target state estimate value of the spatial non-cooperative target at the next moment, and estimate the motion state of the spatial non-cooperative target at the next moment based on the target state estimate value.

[0058] In the above-mentioned motion state estimation method of the space non-cooperative target, on the one hand, by establishing the state space equation of the space non-cooperative target and the observation equation of the service satellite to the space non-cooperative target, and creating a state filter based on the state space equation and the observation equation; then, based on the state filter, the current state estimation value and the current covariance estimation value of the space non-cooperative target at the current moment are determined, and the current state estimation value is input into the interference force prediction model to obtain the current interference force prediction value of the space non-cooperative target at the current moment; then, according to the current state estimation value and the current covariance estimation value, the average observation value of the space non-cooperative target at the next moment corresponding to the current moment is determined, and according to the current interference force prediction value, the interference force prediction value is determined. Determine the predicted state estimate and predicted covariance estimate of the spatial non-cooperative target at the next moment; finally, based on the average observation value, the predicted state estimate and the predicted covariance estimate, determine the target state estimate of the spatial non-cooperative target at the next moment, and estimate the motion state of the spatial non-cooperative target at the next moment based on the target state estimate. Since the interference force prediction value can be determined based on the interference force cloud measurement model, and then the motion state is estimated based on the interference force prediction value, the accuracy of the estimated motion state is improved. On the other hand, since the interference force prediction can be performed through the trained interference force prediction model, there is no need to train the model in real time, thereby improving the efficiency of motion state estimation.

[0059] Hereinafter, the motion state estimation method of a spatial non-cooperative target recorded in an exemplary embodiment of the present disclosure will be explained and illustrated in detail with reference to the accompanying drawings.

[0060] First, the technical implementation principle of the exemplary embodiment of the present disclosure is explained and illustrated. Specifically, the motion state estimation method of the space non-cooperative target recorded in the exemplary embodiment of the present disclosure can accurately calculate the influence of environmental interference such as solar perturbation and air resistance on the motion of the non-cooperative target online; at the same time, it can also calculate the position of the space non-cooperative target and realize the online training of the neural network of the state prediction of the space non-cooperative target based on overcoming the inability of the existing space non-cooperative target motion estimation algorithm to handle state-related external interference forces.

[0061] Secondly, the motion state estimation method of the spatial non-cooperative target recorded in the exemplary embodiment of the present disclosure can be divided into a training phase and an operation phase. The purpose of the training phase is to train the radial basis neural network online and achieve more accurate motion estimation at the same time; the operation phase refers to the time period when the neural network training is completed, the normal operation is performed, and the motion estimation algorithm is corrected; wherein, the specific implementation process can be referred to Figure 2 As shown; at the same time, Figure 2 In the schematic diagram shown, the solid line and the dotted line represent the training phase and the running phase respectively.

[0062] In the actual application process, the main idea of ​​the training stage is: the state motion estimation is completed by using the framework of adaptive Kalman filtering. The residual and state generated by the adaptive Kalman filtering are low-pass filtered and time-aligned respectively, and then sent to the neural network for training. Finally, the state-disturbance model can be obtained. This model is an approximation of the real disturbance model. The disturbance model does not participate in the motion estimation task in the training stage; therefore, the entire calculation process can be divided into two modules in the training stage, namely the motion estimation module and the network training module. Specifically, the main process of motion estimation is: the position and velocity of the non-cooperative target in space are estimated by using the motion state estimator based on Kalman filtering, the residual in the state equation is obtained, and the residual is passed to the recursive average filter to complete the noise reduction and outlier removal tasks, so as to obtain the estimated disturbance force, which is transmitted to the Kalman filter to complete the correction of the process equation, and finally achieve a more accurate motion estimation; and the main process of neural network training is: considering the estimated disturbance force and motion estimation provided in the motion estimation, as mentioned above, the filter value generated by the recursive average filter has a certain phase difference, so the state estimation corresponding to the residual value after filtering in time is found. Then the state quantity and residual quantity are used as the known quantity and expected output of the neural network respectively, and input into the radial basis function neural network for parameter training to generate the interference model.

[0063] Furthermore, the main idea of ​​the operation phase is: no longer training the neural network, using the trained neural network as the interference model, inputting the current state, and outputting the interference force at the current moment. The main process is: the Kalman filter's current state estimation value is sent to the interference model to calculate the interference force on the current non-cooperative target, and then substitute it into the dynamic recursive formula of the Kalman filter in the next estimation cycle to complete the recursion of the state of motion at the next moment. The accuracy of the interference force greatly improves the accuracy of motion estimation, and at the same time, the amount of calculation is greatly reduced compared to the training phase, so that there is more sufficient calculation margin for other satellite payloads.

[0064] The following will be combined Figure 3 The interference force prediction model involved in the exemplary embodiment of the present disclosure is explained and illustrated. Specifically, refer to Figure 3 As shown, the interference force prediction model recorded here can be a RBFNN (Radial Basis Function Neural Network) model, which is a three-layer feedforward network, which can include an input layer 301, a hidden layer 302 and an output layer 303; wherein, the neurons of the RBFNN are nonlinear functions, so even if it is a shallow network, it can achieve an effect similar to that of a multilayer perceptron.

[0065] The following will be combined Figure 4The specific training process of the interference force prediction model is explained and illustrated. For details, refer to Figure 4 As shown, the specific training process of the interference force prediction model may include the following steps:

[0066] Step S410, determining a first historical state estimation value, a first historical covariance estimation value, and a first historical residual value of the spatial non-cooperative target at time k-1 according to the state filter;

[0067] Step S420, performing a limiting process and a recursive average filtering process on the first historical residual value to obtain a first deviation filtering value; wherein the timestamp of the middle value of the average filtering queue is M;

[0068] Step S430, reading a second historical state estimation value corresponding to the timestamp M in the state filter, and inputting the second historical state estimation value into the neural network model to be trained to obtain a deviation prediction result;

[0069] Step S440: adjusting the parameters in the neural network model to be trained according to the first deviation filtering value and the deviation prediction result to obtain the interference force prediction model.

[0070] In an example embodiment, the second historical state estimate is input into the neural network model to be trained to obtain a deviation prediction result, which can be achieved in the following manner: a plurality of different samples are randomly selected from the second historical state estimate as initial cluster centers, and training samples are randomly selected from the second historical state estimate as input samples; based on the input samples and the initial cluster centers, the sample category to which the input samples belong is determined, and the initial cluster centers are updated based on the sample categories; based on the neural network model to be trained, the updated cluster centers are updated again, and it is determined whether the change amplitude of the updated cluster centers is less than or equal to a preset threshold; if the change amplitude of the updated cluster centers is less than or equal to the preset threshold, a deviation prediction result is obtained based on the updated cluster centers.

[0071] In an exemplary embodiment, the parameters in the neural network model to be trained are adjusted according to the first deviation filter value and the deviation prediction result to obtain the interference force prediction model, which can be achieved in the following way: the standard deviation between the first deviation filter value and the deviation prediction result is calculated according to the Gaussian function in the neural network model to be trained, and the parameters in the neural network model to be trained are adjusted based on the standard deviation to obtain the interference force prediction model.

[0072] The following will further explain and illustrate the specific training process of the interference force prediction model. Specifically, in the training process of the interference force prediction model, it is first necessary to design a radial basis function neural network assisted adaptive unscented Kalman filter algorithm in the training phase. Among them, the main steps of the training phase are: using the framework of the adaptive Kalman filter to complete the state motion estimation and obtain the first historical state estimation value of k-1 steps and the first historical residual value The positive superscripts of the two variables represent estimated values ​​or predicted values, which are different from the true values ​​without positive superscripts. The right superscript of the plus sign represents the estimated value, which is different from the predicted value in the filter. The first historical residual generated by the adaptive Kalman filter and the first historical state estimated state quantity After low-pass filtering and time alignment, we can get k-1 steps and The upper right subscript represents the expected value and input of the RBFNN neural network, which is different from the output of the adaptive Kalman filter. and It is sent to the neural network for training, and finally the state quantity-interference force model can be obtained. This model is an approximation of the real interference force model. During the training stage, the interference force model does not participate in the motion estimation task. In addition to being used to train neural networks, it is also used in the extrapolation process of the state estimator. The state estimator uses the corrected prediction value to complete the estimation and ultimately achieve a more accurate motion estimation. Therefore, the specific implementation process of the unscented Kalman filter method for estimating the target motion state during the training phase is:

[0073] (1) Randomly initialize BRFNN to generate μ 0 , σ 0 , W 0 , which are all parameters of RBFNN. These three parameters will be described in detail below. Then, initialize AUKF (i.e., state filter) to generate (first historical residual value), (first historical state estimate), (first historical covariance estimate); further, let k=1 and start the loop;

[0074] (2) The limiting process is performed, and the recursive average filtering process is performed. The timestamp of the middle value of the average filtering queue is M, and the filtering generates a new deviation filtering value. (that is, the first deviation filter value);

[0075] (3) Retrieve the historical filter value of AUKF (i.e., the second historical state estimate), as mentioned above, can be written as Will Store it in memory for later retrieval;

[0076] (4) Use and Train RBFNN and update μ k-1 , σ k-1 , W k-1 ,The structure and training expressions of RBFNN are listed below;

[0077] Furthermore, in order to further iterate the training of RBFNN, it is necessary to cyclically generate historical residual values, historical state estimates, and historical covariance estimates. Under this premise, it can be implemented based on fearless transformation and state filter. Specifically, the specific implementation process of unscented transformation is:

[0078] (1) When the state estimate is known and covariance estimates Generate sampling points according to the unscented transform: Represents the sampling point. Unscented transformation is a common method in state estimation technology and will not be elaborated here.

[0079] (2) The sampling status Substituting into the process equation, all sampling points can be extrapolated, that is:

[0080] The right subscript of k|k-1 indicates the value at time k predicted based on the process equation or observation equation using the state at time k-1.

[0081] (3) The predicted state and predicted covariance are:

[0082]

[0083] Among them, Q k is the covariance matrix of process noise, It is the weight value of the unscented transformation. Its calculation method is also a common method of unscented Kalman filtering, which will not be described here.

[0084] (4) Substituting the predicted sample value into the observation equation, we have:

[0085] The average observed value is:

[0086] (5) The solutions for the innovation covariance, cross covariance, and gain matrix are:

[0087]

[0088] Among them, S k is the covariance matrix of the observation noise, which is usually easy to calculate in engineering and is a known quantity. The state vector, process disturbance and state covariance matrix, and the deviation value update method are:

[0089]

[0090] Where α is the forgetting factor, which ranges from 0 to 1.

[0091] Finally, determine whether the RBFNN has completed convergence. If not, set k = k + 1 and recursively perform the filtering process; if the convergence condition is met, stop training.

[0092] Furthermore, after obtaining the second historical state estimation value, the RBFNN can be trained. Figure 5 As shown in Figure 1, the neurons of RBFNN are nonlinear functions, so even though it is a shallow network, it can achieve similar effects to multilayer perceptrons. Figure 5 In the example scene shown, the input layer of the neuron is recorded as The hidden layer is represented as The output layer is denoted as At the same time, RBFNN is written as a function of d(x). The radial basis neurons of the hidden layer can map low-dimensional linearly inseparable input values ​​to high-dimensional linearly separable space. The activation function of the hidden layer neurons responds locally to the input information. When the input value is in the central area of ​​the basis function, the output of the hidden layer neurons is large, and when the output is far away from the center point, the input decays exponentially. The radial basis function RBFNN ensures that the function value is radially symmetric about the center point, and the function value is a non-negative real number and depends only on the distance from the input to the center point. You can choose Euclidean distance or Gaussian function, which can be recorded as Φ i (x) = L(||x-μ i || 2 ), where μ i is the center point of the i-th hidden layer neuron, L(·) represents a configurable scalar function; in this project, a nonlinear Gaussian function is used, then where σ i is the width parameter of the i-th hidden layer neuron. i The smaller the value, the more concentrated the distribution is and the narrower the central area of ​​the curve is. The larger the value, the smoother the curve is. On the one hand, the narrow Gaussian function increases the responsiveness of the RBFNN. On the other hand, if the overlap of the neuron function is limited due to the narrow central area of ​​the Gaussian function, it is difficult for the network to obtain a non-zero output. Output value of RBFNN The lth element value of Based on this equation, the hidden layer-output layer can be written as d(x)=W TΦ(x); where W is the weight coefficient matrix, W = [w il ], i=1,...,m, l=1,...,j. Φ(x)=[Φ 1 (x) Φ 2 (x) ...Φ m (x)] T .

[0093] As mentioned above, RBFNN has three types of unknown parameters, namely the center point matrix of the hidden layer And the weight coefficient matrix At the same time, RBFNN can use a self-organizing algorithm to achieve online training of all network parameters. The specific training process can be shown as follows:

[0094] (1) Use the K-means algorithm to select a reasonable center position and divide the input into several categories. The same category has similar characteristics and properties, so as to ensure that the selected center point is representative. Specifically, assume that the i-th cluster center of the k-th iteration is μ i,k , you can perform the following steps:

[0095] First, initialize and randomly select I different samples from the input sample data (that is, the second historical state estimate) as the initial cluster center μ i,k ; Secondly, input samples; that is, randomly extract training samples x from the training data k As input; then, matching; that is, we can calculate which cluster center the input sample is closest to, and classify it into the same category of the cluster center, that is, calculate i(x k ) = argmin||x k -μ i,k ||; find the corresponding i and classify xk as the i-th category; further, update the cluster center; that is, due to x k With the addition of , the cluster center of the i-th category will change. The new cluster center is updated as follows:

[0096] Among them, η is the learning step length, satisfying 0<η<1; only one cluster center is updated each time, and the others will not be updated; finally, judgment; that is, it is necessary to judge whether the algorithm converges; for example, a small threshold can be set, and if the change range of the cluster center is less than the threshold, the algorithm is considered to converge; otherwise, continue to iterate.

[0097] (2) Calculate the standard deviation; that is, when the algorithm converges, the standard deviation needs to be calculated to determine whether the model training is complete. Specifically, the exemplary embodiment of the present disclosure uses a Gaussian function as the basis function, and the standard deviation can be selected as:

[0098] Among them, max||μ i,k -μ j,k || represents the maximum distance between cluster centers. This method can avoid the radial basis function from being too concentrated or too dispersed.

[0099] (3) Learning weights; specifically, the least mean square algorithm (LMS) can be used directly. The input of LMS is the output generated by the hidden layer, that is, W = Φ + d; Φ + =(Φ T Φ) -1 Φ T .

[0100] Based on the above method, the Kalman filter algorithm continuously The data pairs are input into the training algorithm to generate the network parameters of μ, W and σ, thereby completing the construction of the radial basis function neural network.

[0101] The following will be combined Figure 2-Figure 5 right Figure 1 The motion state estimation method of the spatial non-cooperative target shown in is further explained and illustrated. Specifically:

[0102] In step S110, a state space equation of a space non-cooperative target and an observation equation of a service satellite to the space non-cooperative target are established, and a state filter is created based on the state space equation and the observation equation.

[0103] Specifically, first, refer to Figure 6 As shown, it is assumed that an orbital coordinate system can be set near the orbit of the space non-cooperative target 601, the x direction is the velocity direction, the y direction is the direction pointing to the center of the earth, and the z-axis direction is determined according to the right-hand coordinate system; then, according to the CW equation, the space non-cooperative target 601 gradually moves away from the origin of the orbital coordinate system. At the same time, considering that the environmental interference can be decomposed into non-zero components and zero-mean Gaussian noise terms, the differential motion equations of the position r and velocity v of the space non-cooperative target 601 can be shown as follows:

[0104]

[0105] Where d represents the non-zero environmental interference term (i.e., the interference value), which is difficult to be removed by the filter and is usually state-dependent. And ξ represents the zero-mean Gaussian noise term, which is regarded as process noise in the following text. The A and B matrices are in the form of:

[0106]

[0107] Where w represents the orbital angular velocity of the non-cooperative target, which is a known constant.

[0108] Furthermore, the differential motion formulas of position and velocity can be transformed into discrete motion formulas, namely:

[0109]

[0110] Among them, Δt is the discrete time step. In practical applications, the required step size can be selected according to actual needs. The subscript of the variable represents the number of discrete time steps. At the same time, the above equation is written in matrix form, then:

[0111]

[0112] Where I represents the 3*3 identity matrix. At the same time, for the convenience of the following text, the state variable can be simplified to a variable in matrix form, that is, the state variable. At the same time, the first term on the right side of the above equation can be simplified to, and the coefficients of the second and third terms on the right side can be written as C, so the above formula can be simplified to:

[0113] x k =f k-1 +Cd k-1 +Cξ k-1 ;

[0114] Furthermore, due to f k-1 ,d k-1 With x k-1 So the above formula can also be written as:

[0115] x k =f(x k-1 )+Cd(x k-1 )+Cξ k-1 ; The brackets in the formula represent functional relationships, not operation symbols. As mentioned above, the function f(x k-1 ) is known, the coefficient matrix C is known, and the process noise ξ k-1 The statistical properties of d(x k-1 ) has an unknown functional form.

[0116] Finally, continue to refer to Figure 6 As shown, based on the geometric relationship and taking into account the observation noise, the value of the position of the non-cooperative target measured by the service satellite is:

[0117] z k =r k-1 -r s,k-1 +θ k-1 ; where z k represents the relative position of the non-cooperative target measured by the serving satellite at time k, r s represents the known position of the service satellite in the orbital coordinate system, θ kis zero-mean observation noise with known statistical properties. Similarly, the above equation can be written as k = z(x k-1 ,r s,k-1 )+θ k-1 The brackets of z(·) represent functions, not operation symbols. This equation is the observation equation.

[0118] Finally, after obtaining the observation equation and the state space equation, the state filter, namely AUKF (Adaptive Unscented Kalman Filter), can be obtained based on the observation equation and the state space equation. In addition, the motion state estimation method of the non-cooperative target recorded in the exemplary embodiment of the present disclosure can finally be obtained in d(x k-1 ) is unknown, and there is interference from process noise and observation noise. k The accurate estimation of the motion state of non-cooperative targets is achieved.

[0119] In step S120, the current state estimation value and the current covariance estimation value of the spatial non-cooperative target at the current moment are determined based on the state filter, and the current state estimation value is input into the interference force prediction model to obtain the current interference force prediction value of the spatial non-cooperative target at the current moment.

[0120] In step S130, the average observation value of the spatial non-cooperative target at the next moment corresponding to the current moment is determined based on the current state estimate value and the current covariance estimate value, and the predicted state estimate value and predicted covariance estimate value of the spatial non-cooperative target at the next moment are determined based on the current interference force prediction value.

[0121] In an exemplary embodiment, determining the average observation value of the spatial non-cooperative target at the next moment corresponding to the current moment based on the current state estimate and the current covariance estimate can be achieved in the following way: performing an unscented transform on the current state estimate and the current covariance estimate to obtain the current sampling point, and determining the average observation value of the non-cooperative target at the next moment based on the current sampling point and the observation equation.

[0122] In an exemplary embodiment, determining the predicted state estimate and predicted covariance estimate of the spatial non-cooperative target at the next moment based on the current interference force prediction value can be achieved in the following manner: determining the predicted state estimate of the spatial non-cooperative target at the next moment based on the current interference force prediction value and the current sampling point; determining the covariance estimate of the spatial non-cooperative target at the next moment based on the current sampling point and the predicted state estimate.

[0123] In step S140, based on the average observation value, the predicted state estimate value and the predicted covariance estimate value, the target state estimate value of the spatial non-cooperative target at the next moment is determined, and based on the target state estimate value, the motion state of the spatial non-cooperative target at the next moment is estimated.

[0124] In an exemplary embodiment, determining the target state estimate of the spatial non-cooperative target at the next moment based on the average observation value, the predicted state estimate and the predicted covariance estimate can be achieved in the following manner: determining the information covariance and the cross covariance of the spatial non-cooperative target based on the average observation value, and determining the gain matrix of the spatial non-cooperative target based on the information covariance and the cross covariance; determining the target state estimate of the spatial non-cooperative target at the next moment based on the predicted state estimate and the gain matrix.

[0125] In an exemplary embodiment, determining the target state estimate value of the space non-cooperative target at the next moment based on the predicted state estimate value and the gain matrix can be achieved in the following manner: determining the target relative position between the space non-cooperative target and the service satellite at the next moment, and determining the target state estimate value of the space non-cooperative target at the next moment based on the predicted state estimate value, the gain matrix and the target relative position.

[0126] In an exemplary embodiment of the present disclosure, determining the target relative position between the space non-cooperative target and the service satellite at the next moment can be achieved in the following manner: determining the current satellite position of the service satellite in the orbital coordinate system at the current moment, and determining the current relative position between the space non-cooperative target and the service satellite at the current moment based on the current state estimate; determining the target relative position between the space non-cooperative target and the service satellite at the next moment based on the current relative position and the current satellite position.

[0127] In the following, steps S120 to S140 will be further explained and illustrated. Specifically, in the actual application process, the estimated value of the Mann filter can be (that is, the current state estimate) is sent to the interference model, and the interference value can be directly calculated (that is, the current interference force prediction value of the space non-cooperative target at the current moment), and then substitute it into the dynamic recursive formula of the filter in the next estimation cycle, which greatly improves the accuracy of motion estimation and greatly reduces the amount of calculation compared with the training stage, thereby providing more sufficient calculation margin for other satellite payloads.

[0128] Among them, the specific calculation process of RBFNN-AUKF for the target state estimation value of the spatial non-cooperative target at the next moment can be achieved as follows:

[0129] (1) Initialize AUKF, let are equal to the last time point of the training phase.

[0130] (2) Set k = 1 and start the loop;

[0131] (3) Imported into RBFNN, the predicted interference force is generated, that is, (i.e., the current interference force prediction value of the spatial non-cooperative target at the current moment);

[0132] (4) Known state estimate and covariance estimates Generate sampling points according to the unscented transform:

[0133] (5) The sampling status Substituting into the process equation, all sampling points can be extrapolated to obtain:

[0134] (6) The predicted state and predicted covariance are:

[0135]

[0136] (7), substituting the predicted sample value into the observation equation, we have:

[0137]

[0138] (8) The average observed value is:

[0139]

[0140] (9) The solutions for the innovation covariance, cross covariance, and gain matrix are:

[0141]

[0142] (10) The state vector (i.e., the estimated target state of the spatial non-cooperative target at the next moment), process disturbance and state covariance matrix, and the deviation value update method are:

[0143]

[0144]

[0145] Determine whether the task is completed; if not, set k = k + 1, return to the third step and repeat the calculation, and you can calculate the target state estimation value of the spatial non-cooperative target at multiple different times.

[0146] So far, the motion state estimation method of a non-cooperative target in space recorded in the exemplary embodiment of the present disclosure has been fully realized. Based on the above-mentioned contents, it can be known that the motion state estimation method of a non-cooperative target in space recorded in the exemplary embodiment of the present disclosure, on the one hand, can realize high-precision state estimation of a non-cooperative target subject to state-related interference forces; on the other hand, it does not require offline training of the neural network, and realizes online training while ensuring the normal operation of motion estimation.

[0147] The following are embodiments of the device disclosed herein, which can be used to execute the method embodiments disclosed herein. For details not disclosed in the device embodiments disclosed herein, please refer to the method embodiments disclosed herein.

[0148] The exemplary embodiment of the present disclosure also provides a motion state estimation device for a space non-cooperative target. Specifically, refer to Figure 7 As shown, the motion state estimation device of a space non-cooperative target may include a state filter creation module 710, a current interference force prediction value determination module 720, an average observation value determination module 730 and a state estimation module 740. Among them:

[0149] The state filter creation module 710 may be used to establish a state space equation of a space non-cooperative target and an observation equation of the service satellite to the space non-cooperative target, and to create a state filter based on the state space equation and the observation equation;

[0150] The current interference force prediction value determination module 720 can be used to determine the current state estimation value and the current covariance estimation value of the space non-cooperative target at the current moment based on the state filter, and input the current state estimation value into the interference force prediction model to obtain the current interference force prediction value of the space non-cooperative target at the current moment;

[0151] The average observation value determination module 730 may be used to determine the average observation value of the space non-cooperative target at the next moment corresponding to the current moment according to the current state estimation value and the current covariance estimation value, and determine the predicted state estimation value and the predicted covariance estimation value of the space non-cooperative target at the next moment according to the current interference force prediction value;

[0152] The state estimation module 740 can be used to determine the target state estimation value of the spatial non-cooperative target at the next moment based on the average observation value, the predicted state estimation value and the predicted covariance estimation value, and estimate the motion state of the spatial non-cooperative target at the next moment based on the target state estimation value.

[0153] In an exemplary embodiment of the present disclosure, the interference force prediction model is obtained in the following manner: determining the first historical state estimation value, the first historical covariance estimation value and the first historical residual value of the spatial non-cooperative target at time k-1 according to the state filter; performing limiting processing and recursive average filtering processing on the first historical residual value to obtain a first deviation filter value; wherein the timestamp of the middle value of the average filter queue is denoted as M; reading the second historical state estimation value corresponding to the timestamp M in the state filter, and inputting the second historical state estimation value into the neural network model to be trained to obtain a deviation prediction result; adjusting the parameters in the neural network model to be trained according to the first deviation filter value and the deviation prediction result to obtain the interference force prediction model.

[0154] In an exemplary embodiment of the present disclosure, the second historical state estimation value is input into the neural network model to be trained to obtain a deviation prediction result, including: randomly extracting a plurality of different samples from the second historical state estimation value as initial cluster centers, and randomly extracting training samples from the second historical state estimation value as input samples; determining the sample category to which the input sample belongs based on the input sample and the initial cluster center, and updating the initial cluster center based on the sample category; updating the updated cluster center again based on the neural network model to be trained, and determining whether the change amplitude of the updated cluster center is less than or equal to a preset threshold; if the change amplitude of the updated cluster center is less than or equal to the preset threshold, obtaining a deviation prediction result based on the updated cluster center.

[0155] In an exemplary embodiment of the present disclosure, the parameters in the neural network model to be trained are adjusted according to the first deviation filter value and the deviation prediction result to obtain the interference force prediction model, including: calculating the standard deviation between the first deviation filter value and the deviation prediction result according to the Gaussian function in the neural network model to be trained, and adjusting the parameters in the neural network model to be trained based on the standard deviation to obtain the interference force prediction model.

[0156] In an exemplary embodiment of the present disclosure, the average observation value of the spatial non-cooperative target at the next moment corresponding to the current moment is determined based on the current state estimate and the current covariance estimate, including: performing an unscented transform on the current state estimate and the current covariance estimate to obtain the current sampling point, and determining the average observation value of the non-cooperative target at the next moment based on the current sampling point and the observation equation.

[0157] In an exemplary embodiment of the present disclosure, a predicted state estimate and a predicted covariance estimate of the spatial non-cooperative target at the next moment are determined based on a current interference force prediction value, including: determining the predicted state estimate of the spatial non-cooperative target at the next moment based on the current interference force prediction value and the current sampling point; determining the covariance estimate of the spatial non-cooperative target at the next moment based on the current sampling point and the predicted state estimate.

[0158] In an exemplary embodiment of the present disclosure, the target state estimate of the spatial non-cooperative target at the next moment is determined based on the average observation value, the predicted state estimate value and the predicted covariance estimate value, including: determining the information covariance and the cross covariance of the spatial non-cooperative target based on the average observation value, and determining the gain matrix of the spatial non-cooperative target according to the information covariance and the cross covariance; determining the target state estimate of the spatial non-cooperative target at the next moment according to the predicted state estimate value and the gain matrix.

[0159] In an exemplary embodiment of the present disclosure, the target state estimate value of the space non-cooperative target at the next moment is determined according to the predicted state estimate value and the gain matrix, including: determining the target relative position between the space non-cooperative target and the service satellite at the next moment, and determining the target state estimate value of the space non-cooperative target at the next moment based on the predicted state estimate value, the gain matrix and the target relative position.

[0160] In an exemplary embodiment of the present disclosure, determining the target relative position between the space non-cooperative target and the service satellite at the next moment includes: determining the current satellite position of the service satellite in the orbital coordinate system at the current moment, and determining the current relative position between the space non-cooperative target and the service satellite at the current moment based on the current state estimate; determining the target relative position between the space non-cooperative target and the service satellite at the next moment based on the current relative position and the current satellite position.

[0161] In an exemplary embodiment of the present disclosure, the state filter includes any one of an adaptive unscented Kalman filter, an adaptive extended Kalman filter and an adaptive cubature Kalman filter.

[0162] The specific details of each module in the above-mentioned state estimation device for a spatial non-cooperative target have been described in detail in the corresponding state estimation method for a spatial non-cooperative target, and therefore will not be repeated here.

[0163] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0164] In addition, although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0165] In an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.

[0166] Those skilled in the art will appreciate that various aspects of the present disclosure may be implemented as systems, methods or program products. Therefore, various aspects of the present disclosure may be specifically implemented in the following forms, namely: complete hardware implementation, complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as "circuits", "modules" or "systems".

[0167] Refer to the following Figure 8 The electronic device 800 according to this embodiment of the present disclosure is described. Figure 8 The electronic device 800 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0168] like Figure 8 As shown, the electronic device 800 is in the form of a general computing device. The components of the electronic device 800 may include, but are not limited to: the at least one processing unit 810, the at least one storage unit 820, a bus 830 connecting different system components (including the storage unit 820 and the processing unit 810), and a display unit 840.

[0169] The storage unit stores program codes, which can be executed by the processing unit 810, so that the processing unit 810 performs the steps according to various exemplary embodiments of the present disclosure described in the above “Exemplary Method” section of this specification. For example, the processing unit 810 can perform the following steps: Figure 1Step S110 shown in: establish the state space equation of the space non-cooperative target and the observation equation of the service satellite to the space non-cooperative target, and create a state filter based on the state space equation and the observation equation; step S120: determine the current state estimate and the current covariance estimate of the space non-cooperative target at the current moment based on the state filter, and input the current state estimate into the interference force prediction model to obtain the current interference force prediction value of the space non-cooperative target at the current moment; step S130: determine the average observation value of the space non-cooperative target at the next moment corresponding to the current moment according to the current state estimate and the current covariance estimate, and determine the predicted state estimate and the predicted covariance estimate of the space non-cooperative target at the next moment according to the current interference force prediction value; step S140: determine the target state estimate of the space non-cooperative target at the next moment based on the average observation value, the predicted state estimate and the predicted covariance estimate, and estimate the motion state of the space non-cooperative target at the next moment based on the target state estimate.

[0170] The storage unit 820 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 8201 and / or a cache memory unit 8202 , and may further include a read-only memory unit (ROM) 8203 .

[0171] The storage unit 820 may also include a program / utility 8204 having a set (at least one) of program modules 8205, such program modules 8205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0172] Bus 830 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0173] The electronic device 800 may also communicate with one or more external devices 900 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 800, and / or communicate with any device that enables the electronic device 800 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 850. Furthermore, the electronic device 800 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 860. As shown, the network adapter 860 communicates with other modules of the electronic device 800 via a bus 830. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 800, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0174] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0175] In an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the above method of the present specification is stored. In some possible implementations, various aspects of the present disclosure may also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary implementations of the present disclosure described in the above "Exemplary Method" section of the present specification.

[0176] According to the program product for implementing the above method in the embodiment of the present disclosure, it can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0177] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0178] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0179] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0180] Program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0181] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present disclosure, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0182] Other embodiments of the present disclosure will be readily apparent to those skilled in the art after considering the specification and practicing the inventions invented herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not invented by the present disclosure. The specification and examples are to be considered merely exemplary, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A method for estimating the motion state of a non-cooperative target in space, characterized in that: include: Establishing a state space equation of a space non-cooperative target and an observation equation of a service satellite to the space non-cooperative target, and creating a state filter based on the state space equation and the observation equation; Determine a current state estimation value and a current covariance estimation value of the spatial non-cooperative target at the current moment based on the state filter, and input the current state estimation value into the interference force prediction model to obtain a current interference force prediction value of the spatial non-cooperative target at the current moment; Determine the average observation value of the spatial non-cooperative target at a next moment corresponding to the current moment according to the current state estimate and the current covariance estimate, and determine the predicted state estimate and the predicted covariance estimate of the spatial non-cooperative target at the next moment according to the current interference force prediction value; Based on the average observation value, the predicted state estimate value and the predicted covariance estimate value, the target state estimate value of the spatial non-cooperative target at the next moment is determined, and based on the target state estimate value, the motion state of the spatial non-cooperative target at the next moment is estimated.

2. The motion state estimation method of a space non-cooperative target according to claim 1, characterized in that: The interference force prediction model is obtained in the following way: Determine a first historical state estimate, a first historical covariance estimate, and a first historical residual value of the spatial non-cooperative target at time k-1 according to the state filter; The first historical residual value is subjected to a limiting process and a recursive average filtering process to obtain a first deviation filtering value; wherein the timestamp of the middle value of the average filtering queue is M; Reading a second historical state estimation value corresponding to the timestamp M in the state filter, and inputting the second historical state estimation value into the neural network model to be trained to obtain a deviation prediction result; The parameters in the neural network model to be trained are adjusted according to the first deviation filtering value and the deviation prediction result to obtain the interference force prediction model.

3. The motion state estimation method of a space non-cooperative target according to claim 2, characterized in that: Inputting the second historical state estimation value into the neural network model to be trained to obtain a deviation prediction result, including: Randomly extracting a plurality of different samples from the second historical state estimation value as initial cluster centers, and randomly extracting training samples from the second historical state estimation value as input samples; Determine, according to the input sample and the initial cluster center, the sample category to which the input sample belongs, and update the initial cluster center based on the sample category; Based on the neural network model to be trained, the updated cluster center is updated again, and it is determined whether the change range of the updated cluster center is less than or equal to a preset threshold; If the change range of the updated cluster center is less than or equal to the preset threshold, the deviation prediction result is obtained based on the updated cluster center.

4. The method for estimating the motion state of a space non-cooperative target according to claim 2, characterized in that: The parameters in the neural network model to be trained are adjusted according to the first deviation filtering value and the deviation prediction result to obtain the interference force prediction model, including: The standard deviation between the first deviation filter value and the deviation prediction result is calculated according to the Gaussian function in the neural network model to be trained, and the parameters in the neural network model to be trained are adjusted based on the standard deviation to obtain the interference force prediction model.

5. The method for estimating the motion state of a space non-cooperative target according to claim 1, characterized in that: Determining the average observation value of the spatial non-cooperative target at a next moment corresponding to the current moment according to the current state estimate and the current covariance estimate, including: An unscented transformation is performed on the current state estimate and the current covariance estimate to obtain a current sampling point, and an average observation value of the non-cooperative target at a next moment is determined according to the current sampling point and the observation equation.

6. The method for estimating the motion state of a space non-cooperative target according to claim 5, characterized in that: Determining a predicted state estimate value and a predicted covariance estimate value of the space non-cooperative target at the next moment according to the current interference force prediction value includes: Determining a predicted state estimate value of the spatial non-cooperative target at a next moment according to the current interference force prediction value and the current sampling point; Determine a covariance estimate of the spatial non-cooperative target at a next moment according to the current sampling point and the predicted state estimate.

7. The motion state estimation method of a space non-cooperative target according to claim 1, characterized in that: Determining a target state estimate of the space non-cooperative target at a next moment based on the average observation value, the predicted state estimate, and the predicted covariance estimate, including: Determining the information covariance and the cross covariance of the spatial non-cooperative target based on the average observation value, and determining the gain matrix of the spatial non-cooperative target according to the information covariance and the cross covariance; According to the predicted state estimate value and the gain matrix, a target state estimate value of the spatial non-cooperative target at the next moment is determined.

8. The method for estimating the motion state of a space non-cooperative target according to claim 7, characterized in that: Determining a target state estimate value of the spatial non-cooperative target at the next moment according to the predicted state estimate value and the gain matrix includes: Determine the target relative position between the space non-cooperative target and the service satellite at the next moment, and determine the target state estimate of the space non-cooperative target at the next moment based on the predicted state estimate, the gain matrix and the target relative position.

9. The method for estimating the motion state of a space non-cooperative target according to claim 8, characterized in that: Determining a relative position between the space non-cooperative target and the service satellite at a next moment includes: Determine a current satellite position of the serving satellite in the orbital coordinate system at a current moment, and determine a current relative position between the space non-cooperative target and the serving satellite at a current moment according to the current state estimate; The target relative position between the space non-cooperative target and the service satellite at the next moment is determined according to the current relative position and the current satellite position.

10. The motion state estimation method of a space non-cooperative target according to claim 1, characterized in that: The state filter includes any one of an adaptive unscented Kalman filter, an adaptive extended Kalman filter and an adaptive cubature Kalman filter.

Images