Pulse phase estimation method based on photon TOA point cloudization

Abstract

The invention provides a pulse phase estimation method based on photon TOA point cloudization. The method comprises the following steps: step 1, cutting a photon TOA sequence into a plurality of sections by taking a pulsar pulse period as a segmentation interval, endowing each section with Y-axis coordinates at equal intervals, aligning starting points to realize point cloudization, and normalizing and converting a point cloud time axis into a phase axis; 2, performing point cloud registration on the photon TOA point cloud and the standard point cloud by adopting a kernel density estimation method, and outputting a translation amount as a pulse phase estimation value; compared with a conventional cross-correlation phase estimation method based on an epoch folding algorithm, the method can guarantee the estimation speed while greatly improving the estimation precision, is high in anti-noise capability, is not sensitive to observation duration, and is of great significance to real-time estimation of the pulse phase of an on-orbit spacecraft.

Description

Technical Field

[0001] The present invention belongs to the field of aerospace navigation technology, and in particular relates to a pulse phase estimation method based on photon TOA point clouding. Background Art

[0002] With the continuous development of deep space exploration technology, improving the autonomous navigation capabilities of deep space probes is imperative. Celestial navigation (CNS) is an autonomous navigation method that uses data obtained from observations of natural celestial bodies to perform navigation solutions. It can completely autonomously provide navigation parameters such as position and velocity to the probe. X-ray pulsar-based navigation (XPNAV) is one of many celestial navigation methods. It is an ideal autonomous navigation method for spacecraft, especially deep space probes, and provides a new and reliable navigation method for deep space exploration missions. Compared with traditional autonomous navigation methods, pulsar navigation technology can operate throughout the entire deep space exploration mission and can provide position, velocity, attitude, and timing services to meet all the navigation needs of deep space probes. With its unique characteristics and advantages, XPNAV has become a highly promising deep space autonomous navigation technology and has been a research hotspot in recent years.

[0003] The key to achieving high-precision pulsar navigation is to obtain high-precision pulse delay information. During the navigation solution process, the pulse delay estimate will be input into the navigation measurement model as a measurement, so the estimation accuracy will directly affect the accuracy of the spacecraft position estimation. ) for example, multiplying it by the speed of light will result in kilometer-level relative position errors. Therefore, improving estimation accuracy is a key research direction in pulsar navigation. Obtaining pulse delay information requires processing photon time of arrival (TOA) data. Therefore, pulse delay estimation methods based on TOA data processing are crucial to pulsar navigation performance. Pulse delay estimation includes both pulse period estimation and pulse phase estimation.

[0004] One mainstream pulse phase estimation method is based on the Epoch Folding (EF) algorithm, which calculates the phase difference through cross-correlation between the folded profile and the standard profile. Its main limitation lies in the limited bin width, resulting in insufficient estimation accuracy, which can lead to large position errors in actual navigation. Therefore, the current development focus of this type of method is still to further improve estimation accuracy, while maintaining or increasing estimation speed to ensure real-time performance. Another mainstream method is based on the Maximum Likelihood (ML) method, which directly processes the photon TOA to obtain a high-precision, asymptotically effective estimate of the pulse phase. However, this method requires a two-dimensional search of the maximum likelihood function, resulting in a considerable computational load and slow estimation speed. In addition, the algorithm's sensitivity to observation time and noise resistance, especially in low signal-to-noise ratio conditions, are also important criteria.

[0005] Considering the current research status and actual navigation needs, the accuracy of traditional pulse phase estimation methods is difficult to meet navigation needs. Therefore, a method that can improve the accuracy of pulse phase estimation is needed. Summary of the Invention

[0006] In order to solve the problems of insufficient estimation accuracy of the epoch folding method and slow estimation speed of the maximum likelihood method, the present invention proposes a pulse phase estimation method based on photon TOA point clouding. First, in order to avoid the limitation of bin width on phase estimation accuracy in epoch folding, the photon TOA point cloud is generated, and a standard point cloud is generated using a standard contour; secondly, in view of the large amount of computation required for cross-correlation operation, the kernel density estimation method is used to achieve the registration between the photon TOA point cloud and the standard point cloud. Compared with traditional methods, the point cloud pulse phase estimation method has higher noise resistance while being able to improve estimation accuracy and ensure estimation speed, and has a lower dependence on observation time. Therefore, the point cloud pulse phase estimation method is more suitable for the future development trend of pulsar navigation technology.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] A pulse phase estimation method based on photon TOA point clouding, the method comprising:

[0009] Step 1: Cut the photon TOA sequence into multiple segments using the pulsar pulse period as the segment interval, assign equally spaced Y-axis coordinates to each segment, align the starting points to achieve point cloud, and normalize the point cloud time axis to convert it into a phase axis;

[0010] Step 2: Use the kernel density estimation method to align the photon TOA point cloud with the standard point cloud, and output the translation as the pulse phase estimation value.

[0011] A pulse phase estimation device based on photon TOA point clouding, comprising:

[0012] The segmentation module divides the photon TOA sequence into multiple segments using the pulsar pulse period as the segment interval. Each segment is assigned an equally spaced Y-axis coordinate and the starting point is aligned to realize point cloud. The time axis of the point cloud is normalized and converted into a phase axis.

[0013] The pulse phase estimation value output module uses the kernel density estimation method to perform point cloud registration between the photon TOA point cloud and the standard point cloud, and outputs the translation as the pulse phase estimation value.

[0014] An electronic device comprises: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the described method.

[0015] A computer-readable storage medium stores executable instructions, which, when executed by a processor, enable the processor to implement the method described above.

[0016] The beneficial effects of the present invention are:

[0017] (1) The present invention does not require epoch folding operations, and the time / phase axis remains continuous, which can avoid the limitation of bin width on phase estimation accuracy, and has a faster estimation speed while greatly improving the pulse phase estimation accuracy;

[0018] (2) The method of the present invention has low dependence on observation time and strong anti-interference ability, which is of great significance for on-orbit real-time pulse phase estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of a pulse phase estimation method based on photon TOA point clouding of the present invention. DETAILED DESCRIPTION

[0020] In order to make the objectives, technical solutions, and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other. To achieve the above-mentioned objectives, the present invention adopts the following technical solutions.

[0021] like Figure 1As shown in the figure, the present invention provides a pulse phase estimation method based on photon TOA point clouding. Step 1: The photon TOA sequence is truncated into multiple segments. Each segment is assigned equally spaced Y-axis coordinates and the starting points are aligned to form a point cloud. The time axis of the point cloud is converted into a phase axis. Step 2: The photon TOA point cloud is aligned with a standard point cloud using kernel density estimation, and the translation value is output as the pulse phase estimate. This method can significantly improve estimation accuracy while maintaining a relatively fast estimation speed. In addition, it has strong anti-interference capabilities and low dependence on observation time.

[0022] The specific steps of step one are: using the pulse period as the segment interval to divide the photon TOA into multiple segments, aligning the starting point of each segment and assigning equidistant y-axis coordinates to obtain a set of two-dimensional point clouds, and normalizing the time axis of the point cloud to convert it into a phase axis. The peak part of the pulse signal is reflected in the point cloud as a straight line area with a higher density than other parts and a vertical posture. The specific steps of step two are: establishing kernel density estimation functions for the light TOA point cloud and the standard point cloud respectively, and obtaining the phase difference by calculating the similarity of the two functions, or directly calculating the phase difference corresponding to the maximum value of the function as the phase estimation output. After normalizing the time axis, align the photon TOA point cloud and the standard point cloud so that their high-density straight line areas are aligned, and the alignment translation is the estimated value of the pulse phase. The implementation process is described in detail below:

[0023] The point cloud process of photon TOA is as follows:

[0024] Assume that the pulse period of the pulsar is , starting from time 0 The photon TOA sequence in the observation period is divided into part:

[0025] (1)

[0026] Align the starting point of each sub-TOA sequence to time 0:

[0027] (2)

[0028] Assign equal spacing to the sub-TOA sequence Axis coordinates :

[0029] (3)

[0030] Represents a 2D point cloud composed of photon TOA data. In actual operation, the subsequence start point alignment and The axis coordinate assignment process can be done in any order. To ensure uniform distribution of point cloud and facilitate subsequent processing, Take the average of the photon TOA spacing:

[0031] (4)

[0032] Where, and represent the maximum and minimum photon TOA, respectively. represents the total number of photon TOA, is a hyperparameter, which represents the longitudinal spacing constant. The standard pulse profile photon TOA point cloud is obtained by .right 、 Normalize the time axis:

[0033] (5)

[0034] Where, 、 Respectively represent the time axis after normalization 、 , Represents the phase coordinates of point cloud data points, Represents the time coordinate of the point cloud data point, 、 Indicates the maximum and minimum time coordinates in the point cloud. The peak of the pulse signal appears as a dense, vertical straight line in the point cloud. The distance between the two straight lines is the phase difference.

[0035] The specific process of kernel density estimation point cloud registration is as follows:

[0036] Establish 、 Kernel density estimation function:

[0037] (6)

[0038] Where, 、 express 、 The kernel density estimation function, 、 express 、 The total number of data points in , 、 express 、 The data points in represents the translation amount to be registered, Represents the Gaussian kernel function, and its expression is:

[0039] (7)

[0040] in, is the kernel function variable, which represents the displacement vector in the present invention. Indicates bandwidth, which is selected based on the Silverman rule by default:

[0041] (8)

[0042] in, is the sample standard deviation, is the number of samples. In practice, cross-validation and other methods can be used for optimization and adjustment based on data characteristics. Kernel functions such as Epanechnikov and Quartic can also be used, also selected based on data characteristics. When the amount of photon TOA data is large, random sampling can be used to reduce computational complexity and improve estimation efficiency.

[0043] By maximizing and The similarity measure of the optimal registration translation can be solved :

[0044] (9)

[0045] Where, is the pulse phase estimate, express and The cross integral function of :

[0046] (10)

[0047] To solve equation (9), we can use optimization methods such as gradient ascent, and the gradient is:

[0048] (11)

[0049] In actual use, if the estimation accuracy requirement is relatively low and the estimation speed requirement is high, you can look for the kernel density estimation function 、 The pulse phase estimate can be obtained by directly performing the difference at the corresponding position of the highest peak in the pulse:

[0050] (12)

[0051] Any content not described in detail in this specification belongs to the prior art known to those skilled in the art. It will be readily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A pulse phase estimation method based on photon TOA point cloud, characterized in that: The method comprises: Step 1: Cut the photon TOA sequence into multiple segments using the pulsar pulse period as the segment interval, assign equally spaced Y-axis coordinates to each segment, align the starting points to achieve point cloud, and normalize the point cloud time axis to convert it into a phase axis; Step 2: Use the kernel density estimation method to align the photon TOA point cloud with the standard point cloud, and output the translation as the pulse phase estimation value.

2. The pulse phase estimation method based on photon TOA point clouding according to claim 1 is characterized in that: The photon TOA point clouding and normalization process includes: Assume that the pulse period of the pulsar is , starting from time 0 The photon TOA sequence in the observation period is divided into part: （1） Align the starting point of each sub-TOA sequence to time 0: （2） Assign equal spacing to the sub-TOA sequence Axis coordinates : （3） Represents a two-dimensional point cloud composed of photon TOA data. The standard pulse photon TOA point cloud is converted into a standard point cloud. ,right 、 Normalize the time axis: （5） Where, 、 Respectively represent the time axis after normalization 、 , Represents the phase coordinates of point cloud data points, Represents the time coordinate of the point cloud data point, 、 Indicates the maximum and minimum time coordinates in the point cloud.

3. The pulse phase estimation method based on photon TOA point clouding according to claim 2 is characterized in that: Take the average value of the photon TOA spacing and divide it by a constant: （4） Where, and represent the maximum and minimum photon TOA, respectively. represents the total number of photon TOA, is a hyperparameter, representing the vertical spacing constant.

4. The pulse phase estimation method based on photon TOA point clouding according to claim 1 is characterized in that: The kernel density estimation point cloud registration process includes: Establish 、 Kernel density estimation function: （6） Where, 、 express 、 The kernel density estimation function, 、 express 、 The total number of data points in , 、 express 、 The data points in represents the translation amount to be registered, Represents the Gaussian kernel function, and its expression is: （7） in, Indicates bandwidth, is the displacement vector, by maximizing and The similarity measure of the optimal registration translation is solved : （9） Where, is the pulse phase estimate, express and The cross integral function of : （10）。 5. The pulse phase estimation method based on photon TOA point clouding according to claim 4 is characterized in that: represents bandwidth, and the specific formula is: （8） in, is the sample standard deviation, is the sample size.

6. The pulse phase estimation method based on photon TOA point clouding according to claim 4, characterized in that: The solution of equation (9) uses optimization methods such as gradient ascent, and the gradient is: （11）。 7. The pulse phase estimation method based on photon TOA point clouding according to claim 4, characterized in that: Directly performing the difference also yields the pulse phase estimate: （12） and They are kernel density estimation functions 、 The highest peak position.

8. A pulse phase estimation device based on photon TOA point cloud, characterized in that: include: The segmentation module divides the photon TOA sequence into multiple segments using the pulsar pulse period as the segment interval. Each segment is assigned an equally spaced Y-axis coordinate and the starting point is aligned to realize point cloud. The time axis of the point cloud is normalized and converted into a phase axis. The pulse phase estimation value output module uses the kernel density estimation method to perform point cloud registration between the photon TOA point cloud and the standard point cloud, and outputs the translation as the pulse phase estimation value.

9. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that Executable instructions are stored thereon, and when the instructions are executed by a processor, the processor implements the method according to any one of claims 1 to 7.

Images