Track smoothing method and device in high disturbance environment, medium and product
By combining Seq2Seq and CAE structures, the problems of noise and missing data in track data under high disturbance environments are solved, and efficient track smoothing and accurate positioning are achieved in complex environments.
Patent Information
- Application Number
- CN202511025909.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-24
- Publication Date
- 2025-11-14
AI Technical Summary
In highly disturbed environments, noise disturbances and data gaps in track data lead to uneven distribution and fragmentation of track information, making it difficult for existing track smoothing algorithms to achieve real-time and accurate tracking of rapidly maneuvering or suddenly changing targets.
The trend semantic information of the trajectory sequence data is obtained by using a sequence-to-sequence (Seq2Seq) structure, and smooth reconstruction is performed by combining a one-dimensional convolutional autoencoder (CAE) structure. The smooth prediction results are generated by utilizing the attention mechanism and the denoising performance of convolution.
It reduces the parameter requirements for track smoothing in highly disturbed environments, improves track smoothing effect and positioning accuracy, and is suitable for track data processing in complex environments.
Smart Images
Figure CN120952057A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of sensing, and more specifically, to a method, apparatus, medium, and product for smoothing flight paths in highly disturbed environments. Background Technology
[0002] Track data typically refers to the sequence of coordinate points of a target's flight over a period of time. Normally, connecting these points sequentially yields a relatively smooth target trajectory curve. However, in complex real-world environments, due to the perception environment and the inherent characteristics of the detection methods, track data often contains significant noise disturbances and missing data. To address these track data issues, various track smoothing algorithms have been proposed to remove noise disturbances and obtain more complete and accurate track data, thereby reducing the difficulty of subsequent analysis tasks, including track pattern recognition and target detection. Current research on track smoothing algorithms mainly focuses on traditional methods such as Kalman filtering, intelligent methods based on deep learning, and combinations of both.
[0003] Traditional Kalman filtering correlation methods offer high real-time performance, but often require modeling the target's motion state to support corrections and updates between adjacent time steps. Traditional methods include trajectory prediction algorithms based on Kalman filtering, point coordinate smoothing and heading angle smoothing methods based on historical tracks, and track smoothing algorithms based on improved Kalman filtering. Traditional Kalman filtering correlation algorithms are computationally simple and offer good real-time performance, but they require modeling the target's motion state and are suitable for targets with simple and clear motion patterns, such as satellites. They struggle to track rapidly maneuvering or abruptly changing flight targets in a timely and accurate manner.
[0004] Since track smoothing under noisy conditions can also be viewed as a regression prediction task of time-series point data, deep learning-based intelligent methods mainly focus on using recurrent neural network structures adapted to predict time-series point data to discover the movement patterns of target tracks, and completing track data prediction and smoothing through various constraints. The introduction of intelligent methods has optimized the computational methods of traditional track data analysis and processing to some extent, but in most cases, accurate modeling of the target's motion state is still required. Therefore, it is necessary to consider new methods for handling special conditions under high disturbances while combining intelligent methods. Summary of the Invention
[0005] Traditional track smoothing or prediction methods often rely on target motion state information. Furthermore, in real-world scenarios, factors such as sensing methods and complex sensing environments can lead to missing or significantly inaccurate parameter information. After data sorting and merging, the track positioning information provided by different sensing points may also be difficult to align, resulting in uneven distribution and fragmentation of the merged track information. This invention provides a track smoothing method, device, medium, and product for high-disturbance environments.
[0006] In a first aspect, the present invention provides a method for track smoothing under high disturbance environments, comprising: By passing the normalized track latitude and longitude point sequence observation data through a sequence-to-sequence attention structure, the pre-estimated latitude and longitude point sequence data is obtained. The estimated latitude and longitude point sequence data and the normalized track latitude and longitude point sequence observation data are spliced together according to longitude and latitude data respectively, and each is smoothed and reconstructed using a multi-channel one-dimensional convolutional autoencoder structure. The reconstructed longitude sequence data and the reconstructed latitude sequence data obtained after smoothing and reconstruction are spliced together to obtain the smoothed estimated normalized latitude and longitude point sequence data.
[0007] In a preferred embodiment, the step of processing the normalized track latitude and longitude point sequence observation data through a sequence-to-sequence attention structure to obtain the pre-estimated latitude and longitude point sequence data includes: The normalized track latitude and longitude point sequence observation data is expanded in dimension by the embedding layer and then input into the encoding layer Encoder_Layer for encoding; The EncoderLayer outputs the hidden layer information and the encoded output value; The encoded hidden layer information is processed by the Attention Layer to calculate the assigned weights; Adjust the encoded output value using assigned weights and input it into the Decoder_Layer; The Decoder_Layer outputs the pre-estimated latitude and longitude point sequence data.
[0008] In a preferred embodiment, the smooth reconstruction using a multi-channel one-dimensional convolutional autoencoder structure includes: Each channel includes an encoding convolutional structure and a decoding convolutional structure. The encoding convolutional structure is a convolutional structure with downsampling function, and the decoding convolutional structure is a convolutional structure with corresponding upsampling function. Longitude or latitude stitched data is processed by an encoding convolutional structure to obtain multi-channel encoded convolutional results. The end of the encoding convolutional structure maps the multi-channel encoded convolutional results to a low-dimensional semantic space through tiling and full connection. Then, the decoding and reconstruction process of the decoding convolutional structure completes the smooth reconstruction, resulting in reconstructed longitude sequence data and reconstructed latitude sequence data.
[0009] In a preferred embodiment, the root mean square error of the normalized latitude and longitude point sequence data and the normalized latitude and longitude point sequence data is used as the loss function value of the sequence-to-sequence attention structure and the one-dimensional convolutional autoencoder structure.
[0010] In a preferred embodiment, the normalization calculation formula is:
[0011]
[0012] in, This is the offset coefficient. , This is the original latitude and longitude coordinate sequence data. These are the minimum and maximum longitude values in the original track latitude and longitude point sequence data. These are the minimum and maximum latitude values in the original track latitude and longitude point sequence data. , This is a normalized sequence of latitude and longitude points on the flight path.
[0013] In a second aspect, the present invention provides an electronic device, comprising: At least one processor; and a memory communicatively connected to said at least one processor; The memory stores instructions that can be executed by the at least one processor, and the at least one processor executes the instructions stored in the memory to perform the method described above.
[0014] Thirdly, the present invention provides a computer-readable storage medium for storing instructions that, when executed, enable the above-described method to be implemented.
[0015] Fourthly, the present invention provides a computer program product that, when invoked by a computer, causes the computer to execute the above-described method.
[0016] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: This invention utilizes a sequence-to-sequence (Seq2Seq) structure to obtain trend semantic information from historical trajectory sequence data, providing an estimated range for the actual trajectory. It also leverages the low-pass denoising performance of the Convolution Auto-Encoder (CAE) structure on time series data to smoothly reconstruct the predicted data, ultimately obtaining a smooth prediction result for the actual trajectory. This invention can reduce the parameter requirements for trajectory smoothing in highly disturbed environments, improve the trajectory smoothing effect, and enhance positioning accuracy, making it suitable for widespread application. Attached Figure Description
[0017] Figure 1 This is a flowchart of a trajectory smoothing method under high disturbance environment provided by an embodiment of the present invention.
[0018] Figure 2 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.
[0020] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.
[0021] Example The design principle of this invention is as follows: using a sequence-to-sequence (Seq2Seq) structure to obtain the trend semantic information of historical trajectory sequence data, giving the estimated range of the real trajectory, and using the low-pass denoising performance of the convolutional auto-encoder (CAE) structure on time series data to complete the smooth reconstruction of the prediction data, and finally obtain the smooth prediction result of the real trajectory.
[0022] In view of this, such as Figure 1 As shown, this embodiment of the invention provides a trajectory smoothing method under high disturbance environment, which is implemented as follows: First, the original latitude and longitude point sequence data of the flight path is preprocessed. The original latitude and longitude point sequence data of the flight path includes the observed data of the original latitude and longitude point sequence and the actual data of the original latitude and longitude point sequence.
[0023] In this embodiment, the data preprocessing is normalization. Let the original latitude and longitude point sequence data be... The normalized latitude and longitude point sequence data of the flight path is calculated using the following formula:
[0024]
[0025] in, This is the offset coefficient, which can be set according to needs and actual conditions. In this embodiment, it is taken as... . These are the minimum and maximum longitude values in the original track latitude and longitude point sequence data. These are the minimum and maximum latitude values in the original track latitude and longitude point sequence data. , This is a normalized sequence of latitude and longitude points on the flight path.
[0026] Therefore, after normalization, the original track latitude and longitude point sequence observation data yields the normalized track latitude and longitude point sequence observation data, represented as:
[0027] Furthermore, the original latitude and longitude coordinate sequence data is normalized to obtain the normalized latitude and longitude coordinate sequence data, which is represented as follows:
[0028] in, Long Represents longitude data. La This represents latitude data.
[0029] Secondly, an SSAC-1D model is constructed to achieve trajectory smoothing under high-disturbance environments. The SSAC-1D model includes a sequence-to-sequence attention structure (Seq2Seq-Attention structure) and a one-dimensional convolutional autoencoder structure (CAE-Conv1D structure). The process of achieving trajectory smoothing under high-disturbance environments based on the SSAC-1D model is as follows: The first step, processing with the Seq2Seq-Attention structure: The normalized track latitude and longitude point sequence observation data is processed through a sequence-to-sequence attention structure to obtain the pre-estimated latitude and longitude point sequence data. Specifically: The normalized track latitude and longitude point sequence observation data is expanded in dimension by the embedding layer and then input into the encoding layer Encoder_Layer for encoding; The EncoderLayer outputs the hidden layer information and the encoded output value; The encoded hidden layer information is processed by the Attention Layer to calculate the assigned weights; Adjust the encoded output value using assigned weights and input it into the Decoder_Layer; The Decoder_Layer outputs the pre-estimated latitude and longitude point sequence data.
[0030] The second step is processing the CAE-Conv1D structure: The estimated latitude and longitude point sequence data and the normalized track latitude and longitude point sequence observation data are spliced together separately according to longitude and latitude data, that is: The estimated longitude point sequence data is stitched together with the normalized track longitude point sequence observation data; The estimated latitude point sequence data is spliced with the normalized track latitude point sequence observation data; Both longitude and latitude stitched data are smoothly reconstructed using a multi-channel one-dimensional convolutional autoencoder structure. In the multi-channel one-dimensional convolutional autoencoder structure, each channel includes an encoding convolutional structure and a decoding convolutional structure. The encoding convolutional structure is a Conv-Down convolutional structure with downsampling function, and the decoding convolutional structure is a Conv-Up convolutional structure with corresponding upsampling function. The longitude or latitude stitched data is processed by the encoding convolutional structure to obtain multi-channel encoding convolution results. The end of the encoding convolutional structure maps the multi-channel encoding convolution results to a low-dimensional semantic space through tiling and full connection. Then, the decoding convolutional structure completes the smooth reconstruction process to obtain reconstructed longitude sequence data and reconstructed latitude sequence data.
[0031] Finally, the reconstructed longitude and latitude series data obtained after smooth reconstruction are concatenated to obtain the smoothed estimated normalized longitude and latitude point series data, represented as:
[0032] The root mean square error of the normalized latitude and longitude point sequence data and the normalized latitude and longitude point sequence data is used as the loss function value of the sequence-to-sequence attention structure and the one-dimensional convolutional autoencoder structure.
[0033] The entire processing procedure for the SSAC-1D model can be represented as follows:
[0034] This represents the output of the embedding layer; This represents the encoded hidden layer information. Indicates the encoded output value; Indicates the allocation of weights; This represents the estimated latitude and longitude point sequence data; This represents the CAE-Conv1D structure input data, including longitude stitched data and latitude stitched data; This represents a smoothed estimate of normalized latitude and longitude point sequence data; This represents the loss function value based on the root mean square error.
[0035] Based on the same technical concept, embodiments of the present invention also provide an electronic device that can implement the flight path smoothing method for high-disturbance environments provided in the above embodiments of the present invention. In one embodiment, the electronic device may be a server, a terminal device, or other electronic equipment. Figure 2 As shown, the electronic device may include: At least one processor and a memory connected to the at least one processor. In this embodiment of the invention, the specific connection medium between the processor and the memory is not limited. Figure 2 The example used is the connection between the processor and memory via a bus. The bus... Figure 2 The connections between other components are indicated by thick lines and are for illustrative purposes only, not as limiting information. Buses can be divided into address buses, data buses, control buses, etc., but for ease of representation, [the specific bus type is not shown here]. Figure 2 The processor is represented by a single thick line, but this does not imply that there is only one bus or one type of bus. Alternatively, a processor can also be called a controller; there are no restrictions on the name.
[0036] In this embodiment of the invention, the memory stores instructions that can be executed by at least one processor. By executing the instructions stored in the memory, at least one processor can execute a trajectory smoothing method under high disturbance environment as described above.
[0037] The processor is the control center of the device. It can connect to various parts of the control device through various interfaces and lines. By running or executing instructions stored in memory and calling data stored in memory, it can monitor the device's various functions and process data, thereby enabling overall monitoring of the device.
[0038] In an alternative design, the processor may include one or more processing units. The processor may integrate an application processor and a modem processor, wherein the application processor primarily handles the operating system, user interface, and applications, while the modem processor primarily handles wireless communication. It is understood that the modem processor may also not be integrated into the processor. In some embodiments, the processor and memory may be implemented on the same chip; in some embodiments, they may also be implemented separately on separate chips.
[0039] The processor can be a general-purpose processor, such as a CPU, digital signal processor, application-specific integrated circuit, field-programmable gate array or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the trajectory smoothing method under high disturbance environment disclosed in the embodiments of this invention can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor.
[0040] Memory, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. Memory can include at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type memory, random access memory (RAM), static random access memory (SRAM), programmable read-only memory (PROM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), magnetic memory, magnetic disk, optical disk, etc. Memory is any other medium capable of carrying or storing desired program code in the form of instructions or data structures, and accessible by a computer, but is not limited thereto. In embodiments of the present invention, memory can also be a circuit or any other device capable of implementing storage functions, used to store program instructions and / or data.
[0041] By designing and programming the processor, the code corresponding to the trajectory smoothing method under high disturbance environment described in the foregoing embodiments can be embedded into the chip, thereby enabling the chip to execute the steps of the method described in the foregoing embodiments during operation. How to design and program the processor is a technique well known to those skilled in the art, and will not be described in detail here.
[0042] Based on the same inventive concept, embodiments of the present invention also provide a storage medium storing computer instructions that, when executed on a computer, cause the computer to perform a trajectory smoothing method under high disturbance environment as described above.
[0043] In some alternative embodiments, the present invention also provides a method for smoothing a flight path in a highly disturbed environment, which can also be implemented as a program product including program code that, when the program product is run on a device, causes the control device to perform the steps in the method for smoothing a flight path in a highly disturbed environment according to various exemplary embodiments of the present invention as described above.
[0044] It should be noted that although several units or sub-units of the apparatus have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of the invention, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by multiple units. Furthermore, although the operation of the method of the invention is described in a specific order in the drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.
[0045] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0046] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a server, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0047] Program code for performing the operations of this invention can be written using any combination of one or more programming languages, including object-oriented programming languages such as Java and C++, as well as conventional procedural programming languages such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0048] In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).
[0049] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0050] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0051] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. For those skilled in the art, the present invention can be modified and varied in various ways; for example, the Seq2Seq structure can be replaced with other recurrent neural networks. In the convolutional autoencoder structure, in addition to the multi-channel one-dimensional convolutional autoencoder structure in the embodiments of the present invention, other high-dimensional encoders can also be applied. In application, besides the perceptual processing scenarios in which this method is applied, it can also be extended to other data processing scenarios under high-perturbation environments. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for smoothing flight paths under high disturbance environments, characterized in that, include: By passing the normalized track latitude and longitude point sequence observation data through a sequence-to-sequence attention structure, the pre-estimated latitude and longitude point sequence data is obtained. The estimated latitude and longitude point sequence data and the normalized track latitude and longitude point sequence observation data are spliced together according to longitude and latitude data respectively, and each is smoothed and reconstructed using a multi-channel one-dimensional convolutional autoencoder structure. The reconstructed longitude sequence data and the reconstructed latitude sequence data obtained after smoothing and reconstruction are spliced together to obtain the smoothed estimated normalized latitude and longitude point sequence data.
2. The trajectory smoothing method under high disturbance environment according to claim 1, characterized in that, The process of passing the normalized track latitude and longitude point sequence observation data through a sequence-to-sequence attention structure to obtain the pre-estimated latitude and longitude point sequence data includes: The normalized track latitude and longitude point sequence observation data is expanded in dimension by the embedding layer and then input into the encoding layer Encoder_Layer for encoding; The EncoderLayer outputs the hidden layer information and the encoded output value; The encoded hidden layer information is processed by the Attention Layer to calculate the assigned weights; Adjust the encoded output value using assigned weights and input it into the Decoder_Layer; The Decoder_Layer outputs the pre-estimated latitude and longitude point sequence data.
3. The trajectory smoothing method under high disturbance environment according to claim 1, characterized in that, The smooth reconstruction using a multi-channel one-dimensional convolutional autoencoder structure includes: Each channel includes an encoding convolutional structure and a decoding convolutional structure. The encoding convolutional structure is a convolutional structure with downsampling function, and the decoding convolutional structure is a convolutional structure with corresponding upsampling function. Longitude or latitude stitched data is processed by an encoding convolutional structure to obtain multi-channel encoded convolutional results. The end of the encoding convolutional structure maps the multi-channel encoded convolutional results to a low-dimensional semantic space through tiling and full connection. Then, the decoding and reconstruction process of the decoding convolutional structure completes the smooth reconstruction, resulting in reconstructed longitude sequence data and reconstructed latitude sequence data.
4. The trajectory smoothing method under high disturbance environment according to claim 1, characterized in that, The root mean square error of the normalized latitude and longitude point sequence data and the normalized latitude and longitude point sequence data is used as the loss function value of the sequence-to-sequence attention structure and the one-dimensional convolutional autoencoder structure.
5. The trajectory smoothing method under high disturbance environment according to claim 1, characterized in that, The normalization calculation formula is: in, This is the offset coefficient. , This is the original latitude and longitude coordinate sequence data of the flight path. These are the minimum and maximum longitude values in the original track latitude and longitude point sequence data. These are the minimum and maximum latitude values in the original track latitude and longitude point sequence data. , This is a normalized sequence of latitude and longitude points on the flight path.
6. An electronic device, characterized in that, include: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores instructions executable by the at least one processor, which executes the instructions stored in the memory to perform the method as described in any one of claims 1-5.
7. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store instructions that, when executed, cause the method as described in any one of claims 1-5 to be implemented.
8. A computer program product, characterized in that, When the computer program product is invoked by a computer, it causes the computer to perform the method as described in any one of claims 1-5.