Optical equipment station distribution method and system based on BP neural network

Abstract

The invention discloses an optical equipment station distribution method and system based on a BP neural network, and relates to the technical field of flight external measurement, and the method comprises the specific steps: obtaining a to-be-distributed target task, and analyzing the to-be-distributed target task; carrying out track interception according to the analysis data, and carrying out feature extraction on the intercepted track and the analysis data to obtain input data; the input data comprises position data of the fixed optoelectronic device, parameter data of the fixed optoelectronic device, motion track feature data of the aircraft and parameter data of the target to be laid; and inputting the input data into a pre-trained layout strategy generation model, and outputting site position coordinates of the to-be-laid target. According to the method, measurement and control equipment capacity parameters, measured object motion trail characteristics and a BP neural network algorithm are combined, and a high-precision station distribution result can be quickly given by a layout strategy generation model obtained through final optimization training.

Description

Technical Field

[0001] The present invention relates to the technical field of external measurement of flight, and more specifically, to a method and system for arranging optical devices based on a BP neural network. Background Art

[0002] Currently, in the field of arranging measurement and control equipment, especially for equipment such as photoelectric theodolites that directly measure flight targets, traditional methods rely on rule-based arrangement and intelligent optimization algorithms such as genetic algorithms. The rule-based arrangement method performs well when dealing with stable and predictable flight paths, but when faced with unstable aircraft motion states or complex environmental changes, its flexibility is insufficient and it is difficult to provide an optimal arrangement plan. Although using intelligent optimization algorithms such as genetic algorithms can theoretically find the global optimal solution, these algorithms usually require a large amount of computing resources and the solution process takes a long time, unable to meet the requirement of real-time and rapid adjustment.

[0003] With the development of aircraft technology and the increasing complexity of the application environment, the requirements for arranging measurement and control equipment are gradually increasing. It is not only necessary to be able to adapt to dynamically changing flight conditions, but also to be able to complete arrangement optimization within a short time to ensure the accuracy and stability of measurement results. Due to their inherent limitations, traditional arrangement methods can no longer fully meet the needs of modern measurement and control tasks, especially when a rapid response is required in an emergency.

[0004] Therefore, developing an arrangement method that can quickly adapt to changes, efficiently solve problems, and has real-time performance has become an urgent problem for those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides a method and system for arranging optical devices based on a BP neural network, which overcomes the above defects.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A method for arranging optical devices based on a BP neural network, the specific steps are as follows:

[0008] Obtain the target task to be arranged, and analyze the target task to be arranged;

[0009] Perform track interception according to the analysis data, extract features from the intercepted track and the analysis data to obtain input data; the input data includes: position data of fixed optoelectronic devices, parameter data of the fixed optoelectronic devices, aircraft motion trajectory feature data, and parameter data of the target to be arranged;

[0010] Input the input data into a pre-trained arrangement strategy generation model, and output the site position coordinates of the target to be arranged.

[0011] Optionally, the aircraft motion trajectory feature data includes the endpoints, center point, and random points of the intersection arc between the coverage range of the fixed optoelectronic device and the intercepted track.

[0012] Optionally, there are multiple random points.

[0013] Optionally, the parameter data includes detection distance, pitch angle error, and azimuth angle error.

[0014] Optionally, the training steps of the layout strategy generation model are as follows:

[0015] Intercept multiple tracks of the aircraft, extract the aircraft motion trajectory feature data using the relevant parameters of the optical device, and construct an input data set based on the relevant parameters of the optical device and the aircraft motion trajectory feature data;

[0016] Extract the site position coordinates of the optical device based on the relevant parameters of the optical device, and optimize the site position coordinates of the optical device using the particle swarm algorithm to generate an output data set;

[0017] Construct a training sample set based on the input data set and the output data set;

[0018] Construct an initial layout strategy generation model, and perform iterative training using the training sample set until a preset target is reached to obtain the layout strategy generation model.

[0019] Optionally, during the construction of the input data set, the track points in multiple intercepted tracks of the aircraft are screened according to a preset rule, and the screening expression is:

[0020]

[0021] In the formula, X s1 , Y s1 , Z s1 are the coordinates of the fixed optoelectronic device; X, Y, and Z are the coordinates of the aircraft track.

[0022] Optionally, the initial layout strategy generation model is a four-layer BP neural network, including an input layer, a first hidden layer, a second hidden layer, and an output layer.

[0023] An optoelectronic device station layout system based on a BP neural network includes:

[0024] A data acquisition module, configured to acquire a target task to be laid out and standardize the target task to be laid out;

[0025] A feature extraction module, which is used to parse the standardized target task to be deployed, perform track selection on the parsed data, extract features from the selected track and the parsed data, and obtain input data; the input data includes: position data of a fixed optoelectronic device, parameter data of the fixed optoelectronic device, flight trajectory feature data of an aircraft, and parameter data of the target to be deployed.

[0026] A position prediction module, which is used to input the input data into a pre-trained deployment strategy generation model and output the site position coordinates of the target to be deployed.

[0027] As can be seen from the above technical solutions, compared with the prior art, the present invention provides an optical equipment station layout method and system based on a BP neural network. The BP neural network algorithm is introduced, and a four-layer neural network is designed. By selecting representative feature points of the flight trajectory of the aircraft to make an input data set, it can not only fully reflect the movement process of the measured object, but also control the data volume within a more appropriate range, which is convenient for the BP network to quickly adjust parameters during training and quickly converge to the optimal parameters; the present invention combines the measurement and control equipment capability parameters, the movement trajectory characteristics of the measured object, and the BP neural network algorithm. Finally, the optimized and trained deployment strategy generation model can quickly give a high-precision station layout result. Description of the Drawings

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the provided drawings.

[0029] Figure 1 It is a schematic diagram of the method flow provided by the present invention;

[0030] Figure 2 It is a schematic diagram of the selection of flight trajectory feature data of the optical equipment station layout method based on the BP neural network provided by the present invention;

[0031] Figure 3 It is a schematic diagram of the model training process provided by the present invention. Detailed Embodiments

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0033] An embodiment of the present invention discloses an optical equipment station layout method based on a BP neural network, as Figure 1 shown. The specific steps are as follows:

[0034] Step 1: Obtain the target task to be laid out and analyze the target task to be laid out;

[0035] Step 2: Perform track intercept selection according to the analysis data, extract features from the intercepted track and the analysis data to obtain input data; the input data includes: position data of fixed optoelectronic equipment, parameter data of fixed optoelectronic equipment, flight trajectory feature data of the aircraft, and parameter data of the target to be laid out;

[0036] Step 3: Input the input data into the pre-trained layout strategy generation model to output the site position coordinates of the target to be laid out.

[0037] In one embodiment, the flight trajectory feature data of the aircraft includes the endpoints, center point, and random points of the intersecting arc between the coverage range of the fixed optoelectronic equipment and the intercepted track.

[0038] In one embodiment, there are multiple random points.

[0039] In one embodiment, the parameter data includes detection distance, pitch angle error, and azimuth angle error.

[0040] Furthermore, the number of stations to be laid out can be set as required according to the actual situation. When the number of stations to be laid out is n, the input value is 19 + 3 * n. In this embodiment, the number of stations to be laid out n = 2.

[0041] Furthermore, each piece of input data is as Figure 2 shown, including the fixed optoelectronic equipment, that is, the position of the existing fixed equipment (X s1 , Y s1 , Z s1 ), the capabilities of this equipment: detection distance L s1 , equipment pitch angle error A s1 and azimuth angle error E s1 ; the two endpoints of the intersecting arc between the equipment coverage range and the track: starting point P1(0, Y p1 , 0), end point P3(X p3 , Y p3 , Z p3 ), the track coordinates corresponding to the center point of the arc P2(X p2 , Y p2 , Z p2 ); the random points within this arc segment. In this embodiment, the coordinates of the two endpoints of the key focus paragraph P4(X p4 , Y P4 , Z p4 ), P5(X p5 , Yp5 , Z p5 ), with the focus on generally selecting a part of the powered flight segment to the re - entry segment for the arc segment. The specific start and end points of the arc segment are selected according to different aircraft parameters and operator experience. For the parameter data of the target to be deployed, in this embodiment, it includes the measurement and control capabilities of two pieces of equipment to be deployed: the detection distances of the equipment to be deployed 2 are L s2 , and the pitch angle error and azimuth angle error are A s2 and E s2 respectively; the detection distance of the equipment to be deployed 3 is L s3 , and the pitch angle error and azimuth angle error are A s3 and E s3 respectively. In summary, a total of 25 input values are designed. The site coordinates S2(X s2 , 0, Z s2 ), S3(X s3 , 0, Z s3 ) of the two measurement and control equipment to be deployed are output.

[0042] In one embodiment, the training steps of the deployment strategy generation model are as Figure 3 shown, specifically:

[0043] Intercept various flight trajectories of the aircraft, extract the aircraft motion trajectory feature data using the relevant parameters of the optical equipment, and construct an input data set based on the relevant parameters of the optical equipment and the aircraft motion trajectory feature data;

[0044] Extract the site position coordinates of the optical equipment based on the relevant parameters of the optical equipment, and use the particle swarm algorithm to optimize the site position coordinates of the optical equipment to generate an output data set;

[0045] Construct a training sample set based on the input data set and the output data set;

[0046] Construct an initial deployment strategy generation model, and use the training sample set for iterative training until a preset goal is reached to obtain the deployment strategy generation model.

[0047] In one embodiment, during the construction of the input data set, the track points in the various intercepted flight trajectories of the aircraft are screened according to a preset rule, and the screening expression is:

[0048]

[0049] In the formula, X s1 , Y s1 , Z s1 are the coordinates of the fixed optoelectronic equipment; the detection capability of the fixed optoelectronic equipment is L s1 .

[0050] Further, the construction steps of the input data set are as follows: Randomly deploy fixed optoelectronic devices at a distance of 10 - 20 KM from the track to obtain the coordinates (X s1 , Y s1 , Z s1 ). Allocate the detection capabilities and errors L s1 , A s1 , E s1 , L s2 , A s2 , E s2 , L s3 , A s3 , E s3 according to the actual equipment performance. Traverse the track segment with ΔX as the step size. The observed track points satisfy: Screen the observed track points, select the endpoint close to the starting position of the track as P1, the other endpoint as P3, and the x-axis coordinate of P2 Obtain the corresponding position P2 of the track curve based on X p2 and determine the endpoints P4 and P5 of the concerned paragraph.

[0051] The construction steps of the output data set are as follows: Use the particle swarm algorithm to optimize the results, select the minimum fitness to obtain the output set, and the fitness function is GDOP is the error precision of 5 points, and ω is the weight of each point.

[0052] Use 95% of the samples in the training sample set as the training set and 5% of the samples as the test set.

[0053] In one embodiment, the initial layout strategy generation model is a four-layer BP neural network, including an input layer, a first hidden layer, a second hidden layer, and an output layer.

[0054] Further, establish a four-layer BP neural network with 25 input neurons, 64 neurons in the second layer, 64 neurons in the third layer, and 4 neurons in the output layer. The Leaky ReLU function is used in the hidden layer, and the purelin function is used in the output layer. It includes the weight matrix W1 of the first hidden layer, the weight matrix W2 of the second hidden layer, the weight matrix W3 of the output layer, and their bias vectors B1, B2, B3; the loss function is MSE; the optimizer is Adam, and the learning rate is 0.01.

[0055] Use the data in the test set to test after each round of training is completed. The output results are evaluated using the fitness function, and the error acceptance value is ±5%. If not satisfied, perform the next round of training.

[0056] On the other hand, this embodiment discloses an optical equipment layout system based on a BP neural network, including:

[0057] A data acquisition module, configured to acquire a target task to be deployed and standardize the target task to be deployed;

[0058] A feature extraction module, configured to parse the standardized target task to be deployed, perform track intercept selection on the parsed data, extract features from the intercepted track and the parsed data to obtain input data; the input data includes: position data of a fixed optoelectronic device, parameter data of the fixed optoelectronic device, flight track feature data of an aircraft, and parameter data of the target to be deployed;

[0059] A position prediction module, configured to input the input data into a pre-trained deployment strategy generation model and output the site position coordinates of the target to be deployed.

[0060] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0061] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for deploying optical equipment based on BP neural network, characterized in that: The specific steps are: Obtaining target tasks to be deployed, and parsing the target tasks to be deployed; Performing track interception according to the analyzed data, extracting features from the intercepted track and the analyzed data, and obtaining input data; the input data includes: position data of the fixed optoelectronic device, parameter data of the fixed optoelectronic device, characteristic data of the motion trajectory of the aircraft, and parameter data of the target to be deployed; The input data is input into a pre-trained deployment strategy generation model, and the site location coordinates of the target to be deployed are output.

2. According to the BP neural network-based optical equipment station layout method of claim 1, it is characterized in that: The aircraft motion trajectory characteristic data includes the end points, center points and random points of the arc where the coverage of the fixed optoelectronic device intersects the intercepted track.

3. The optical equipment station layout method based on BP neural network according to claim 2 is characterized in that: There are multiple random points.

4. The optical equipment station layout method based on BP neural network according to claim 1 is characterized in that: The parameter data includes detection distance, pitch angle error and azimuth angle error.

5. The optical equipment station layout method based on BP neural network according to claim 1 is characterized in that: The training steps of the deployment strategy generation model are: Intercepting a plurality of flight paths of an aircraft, extracting characteristic data of the aircraft motion trajectory using relevant parameters of an optical device, and constructing an input data set based on the relevant parameters of the optical device and the characteristic data of the aircraft motion trajectory; Extracting the site position coordinates of the optical device based on relevant parameters of the optical device, and optimizing the site position coordinates of the optical device using a particle swarm algorithm to generate an output data set; Constructing a training sample set based on the input data set and the output data set; An initial deployment strategy generation model is constructed, and iterative training is performed using the training sample set until a preset goal is achieved, thereby obtaining the deployment strategy generation model.

6. The optical equipment station layout method based on BP neural network according to claim 5 is characterized in that: During the construction of the input data set, the track points in the intercepted aircraft's various tracks are filtered according to the preset rules. The filtering expression is: Where, X s1 ,Y s1 ,Z s1 are the coordinates of the fixed optoelectronic equipment; X, Y, Z are the coordinates of the aircraft track.

7. The optical equipment station layout method based on BP neural network according to claim 5 is characterized in that: The initial deployment strategy generation model is a four-layer BP neural network, including an input layer, a first hidden layer, a second hidden layer and an output layer.

8. An optical equipment station layout system based on BP neural network, characterized in that: include: A data acquisition module, used for acquiring target tasks to be deployed and standardizing the target tasks to be deployed; A feature extraction module is used to parse the standardized target task to be deployed, perform track interception on the parsed data, perform feature extraction on the intercepted track and the parsed data, and obtain input data; the input data includes: position data of the fixed optoelectronic device, parameter data of the fixed optoelectronic device, feature data of the motion trajectory of the aircraft, and parameter data of the target to be deployed; The location prediction module is used to input the input data into a pre-trained deployment strategy generation model and output the site location coordinates of the target to be deployed.

Images