Autonomous avoidance maneuvering method for multiple interceptors by spacecraft based on reinforcement learning

A technology of reinforcement learning and spacecraft, applied in the field of anti-interception, can solve the problem of low success rate of evasion

Active Publication Date: 2020-11-27
HARBIN INST OF TECH
View PDF2 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to solve the problem of the low success rate of avoiding multi-interceptors in existing space

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Autonomous avoidance maneuvering method for multiple interceptors by spacecraft based on reinforcement learning
  • Autonomous avoidance maneuvering method for multiple interceptors by spacecraft based on reinforcement learning
  • Autonomous avoidance maneuvering method for multiple interceptors by spacecraft based on reinforcement learning

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0020] Specific implementation mode 1: A method for self-avoiding maneuvers of spacecraft to multiple interceptors based on reinforcement learning described in this implementation mode, the method is specifically implemented through the following steps:

[0021] Step 1: Establish the space dynamics models of the spacecraft and the interceptor respectively;

[0022] Step 2: Based on the space dynamics model of the spacecraft and interceptor established in Step 1, establish a true-scale guidance model for multiple interceptors;

[0023] Step 3: Take each engine of the spacecraft as an agent to establish a decision-making model for spacecraft evasion maneuvers;

[0024] Step 4: Establish a multi-agent autonomous decision-making training system based on reinforcement learning theory;

[0025] Step 5: Apply the model established in step 1, step 2 and step 3 to the system of step 4, and train the spacecraft evasive maneuver decision-making model offline;

[0026] Step 6: Apply the...

specific Embodiment approach 2

[0028] Specific embodiment two: the difference between this embodiment and specific embodiment one is: said step one establishes the space dynamics model of spacecraft and interceptor respectively, and its specific process is:

[0029] In the geocentric inertial coordinate system, the space dynamics model of the spacecraft is:

[0030]

[0031] in, is the space position vector of the spacecraft, r M for Corresponding scalar, m M is the instantaneous mass of the spacecraft, T M is the total thrust of the spacecraft engine, is the unit vector of the thrust direction of the spacecraft engine, μ is the gravitational constant of the earth, and the value is 3.986×10 5 km 3 / s 2 ; for The second derivative of , Be the perturbation acceleration vector, set as a constant value in the present invention;

[0032] The mass change rate of the spacecraft is:

[0033]

[0034] in, is the rate of change of spacecraft mass, I sp,M is the specific impulse of the space...

specific Embodiment approach 3

[0041] Specific embodiment three: the difference between this embodiment and specific embodiment two is that: in the step two, according to the space dynamics model of the spacecraft and the interceptor established in the step one, a multi-interceptor true-scale guidance model is established, which The specific process is:

[0042] According to the space dynamics model of the spacecraft and the interceptor established in step 1, the relative motion model of the spacecraft and the interceptor is obtained as:

[0043]

[0044] in, is the combined thrust vector of the spacecraft engine, is the combined thrust vector of the interceptor engine;

[0045] In order to simplify the calculation, only the maximum sight angle and saturated maneuvering overload are constrained, and the noise and other disturbance problems are not considered. Decompose equation (5) along the direction of the projectile line of sight and the direction perpendicular to the direction of projectile line...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses an autonomous avoidance maneuvering method for multiple interceptors by a spacecraft based on reinforcement learning, and belongs to the technical field of anti-interception. According to the method, the problem of low success rate of avoiding multiple interceptors by existing spacecraft programmed maneuver is solved. The invention provides an autonomous avoidance maneuvering method based on a deep neural network, which is not limited by the quality and material of a spacecraft, consists of two parts, namely an offline training system and an online decision-making network, uses less computing resources of the spacecraft, has real-time decision-making capability, and improves the avoidance success rate of the spacecraft for multiple interceptors. When the autonomousmaneuver avoidance method is adopted for the spacecraft, the average maneuver avoidance success rate is 49%, and the avoidance success rate is increased by 29%. According to the method, the engine switching time in the avoiding process can be effectively shortened, and more energy is saved. The method can be applied to autonomous avoidance of multiple interceptors by a spacecraft.

Description

technical field [0001] The invention belongs to the technical field of anti-interception, and in particular relates to a self-avoiding maneuver method for a spacecraft against multiple interceptors based on reinforcement learning. Background technique [0002] As early as the 1970s, research on maneuvering and avoiding technology has been carried out in foreign countries, mostly based on the analysis of simplified motion models, and maneuvering and avoiding strategies are only designed for special trajectory points. Early domestic research focused on simulation modeling, and a large number of interceptor evasion simulation systems were established based on kinematic constraints. On this basis, some scholars have proposed methods such as maneuver avoidance strategy based on differential game and impulsive avoidance strategy based on optimal control. These methods are offline planning methods based on mathematical models and do not have autonomy. During the orbital operation ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06F30/27G06N3/04G06N3/08
CPCG06F30/27G06N3/08G06N3/045
Inventor 白成超郭继峰郑红星赵毓
Owner HARBIN INST OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap