An Autonomous Avoidance Architecture Design and Analysis Method for Spacecraft Orbit Threats

Abstract

The present invention discloses an autonomous avoidance architecture design and analysis method for spacecraft orbit threats. The method includes: constructing a logical architecture of an autonomous avoidance system for spacecraft orbit threats based on a sensing module, a decision-making module, and an execution module; respectively establishing corresponding colored Petri net sub-models for the sensing module, the decision-making module, and the execution module based on colored Petri nets; establishing a hierarchical-colored Petri net for the entire system logical architecture; modeling the hierarchical-colored Petri net of the logical architecture of the autonomous avoidance system for spacecraft orbit threats; assigning time stamps to the transitions in the hierarchical-colored Petri net, building a hierarchical-time-colored Petri net of the logical architecture of the autonomous avoidance system for spacecraft orbit threats, and performing autonomous avoidance of spacecraft orbit threats based on the hierarchical-time-colored Petri net. The advantages of the present invention are: realizing the autonomous avoidance of spacecraft orbit threats.

Description

Technical Field

[0001] The present invention relates to the field of spacecraft system architecture modeling, and more particularly to an autonomous avoidance architecture design and analysis method for spacecraft orbit threats. Background Art

[0002] At present, due to the increasingly crowded orbital space, the surging collision risk, the intensifying space competition, and the increasing orbital harassment, the safe operation of spacecraft faces severe challenges. Facing orbital threats such as collisions between space debris and defunct satellites and harassment by hostile satellites in space, the current response method of "ground orbit determination + telemetry downlink → threat determination → decision-making and planning → uplink command → on-orbit execution" highly relies on the analysis and decision-making of ground personnel, and has problems such as many spatio-temporal constraints in windows and arcs, a long time chain in the space-ground loop, and many human factors in operation and maintenance control, resulting in untimely threat disposal, high operation and control pressure, over-dull or excessive responses to threats, seriously affecting the safety of spacecraft on-orbit operation and the continuity of services, and it is difficult to adapt to the space situation with a continuous increase in the number of threats and a deteriorating environment.

[0003] Spacecraft needs to achieve timely and appropriate autonomous responses to orbital threats under conditions such as a complex space environment and severely limited resources. Moreover, when spacecraft operates in orbit for a long time, it is difficult to replace and upgrade equipment, and it needs to adapt to "changing threats" with "unchanged hardware" for a long time. Resources such as sensors, computing, and storage are severely limited; accurate perception and autonomous decision-making have high resource requirements and face serious conflicts in time and resources. Therefore, when establishing the autonomous avoidance system architecture of "perception - decision - execution" for spacecraft orbit threats, it is necessary to fully consider the nested and coupled relationships among the three, and conduct overall modeling and analysis verification under an integrated framework. However, the prior art does not consider the nested and coupled relationships among "perception - decision - execution". For example, an intelligent spacecraft generalized control method and system disclosed in Chinese Patent Publication No. CN114967437A cannot achieve overall modeling and analysis verification, relies on the analysis and decision-making of ground personnel, and cannot autonomously avoid spacecraft orbit threats.

[0004] To model the autonomous avoidance system architecture of spacecraft orbits under an integrated framework, an appropriate model is needed to describe the relationships between the various components. The different levels of the integrated control system are nested with each other, have multi-loop coupling, and there are conflicts in time / event triggering. The system state includes multi-dimensional variables such as threat levels, morphological characteristics, and motion parameters, with both discrete and continuous characteristics, as well as both deterministic parameters and random variables. It is impossible to describe it using a single mathematical language and traditional modeling methods. Therefore, there are certain challenges in establishing a logically reasonable and clearly hierarchical model to depict the complex interaction relationships such as dependence, competition, and association between the components of the integrated control system. Petri nets are a theoretical method that can describe important behavioral attributes such as concurrency, conflict, and resource sharing of complex systems. They also have characteristics such as mathematical expression and graphical visualization, can effectively model and describe discrete event systems, and verify whether the model performance meets the requirements through simulation analysis. However, there is no relevant research on the autonomous avoidance of spacecraft orbit threats based on Petri nets in the existing technology. How to use the advantages of Petri nets to model the autonomous avoidance architecture of spacecraft orbit threats and thus achieve autonomous avoidance of spacecraft orbit threats has become a research hotspot. Summary of the Invention

[0005] The technical problem to be solved by the present invention is how to use Petri nets to model and design the autonomous avoidance architecture of spacecraft orbit threats and analyze it, which is conducive to realizing the autonomous avoidance of spacecraft orbit threats.

[0006] The present invention solves the above technical problems through the following technical means: A method for designing and analyzing the autonomous avoidance architecture of spacecraft orbit threats, the method comprising:

[0007] Step 1: Construct a logical architecture of the autonomous avoidance system for spacecraft orbit threats based on a perception module, a decision-making module, and an execution module;

[0008] Step 2: Based on the colored Petri net, establish corresponding colored Petri net sub-models for the perception module, the decision-making module, and the execution module respectively;

[0009] Step 3: Through the nested coupling relationship between the colored Petri net sub-models of the perception module, the decision-making module, and the execution module, set interfaces at the entrances and exits of each colored Petri net sub-model to realize the connection between the sub-models, so as to establish a hierarchical-colored Petri net of the entire system logical architecture, and conduct simulation analysis and verification on the avoidance process;

[0010] Step 4: Model the hierarchical-colored Petri net of the logical architecture of the autonomous avoidance system for spacecraft orbit threats, visually verify the correctness of the model through simulation, and analyze the dynamic characteristics of the model in the state space;

[0011] Step 5: Assign timestamps to the transitions in the hierarchical-colored Petri net, construct a hierarchical-time-colored Petri net for the logical architecture of the spacecraft orbit threat autonomous avoidance system, perform autonomous avoidance of the spacecraft orbit threat based on the hierarchical-time-colored Petri net, and analyze the process time of the model in response to threat events.

[0012] Beneficial effects: The present invention realizes the modeling and simulation of the perception module, decision-making module, and execution module of the spacecraft orbit threat autonomous avoidance system, respectively establishes corresponding colored Petri net sub-models, realizes the connection between the sub-models, constructs a hierarchical-colored Petri net for the logical architecture of the entire system, and then assigns timestamps to the transitions in the hierarchical-colored Petri net, makes full use of the advantages of the Petri net, constructs a hierarchical-time-colored Petri net based on the Petri net, can perform autonomous avoidance of the spacecraft orbit threat, and does not rely on the analysis and decision-making of ground personnel.

[0013] Further, the said Step 1 includes:

[0014] The perception model is used for the measurement of the spacecraft's own state and environmental scene, the identification of threat characteristics, the prediction of threat behaviors, and the determination of threat levels;

[0015] The decision-making module is used for parameter estimation of threat targets, prediction of threat behaviors, decision-making of avoidance behaviors, selecting the best solution among possible solutions, and forming a serialized sequence of behavioral actions;

[0016] The execution module is used to execute the action sequence through the attitude and orbit motion controller, control the actuators on the spacecraft to perform specific attitude and orbit change actions, and feed back some parameter data of the spacecraft to the perception module and the decision-making module.

[0017] Even further, the said perception module is also used for:

[0018] Using four sensing devices, namely an optical camera, an infrared camera, a microwave radar, and a laser point cloud, to obtain perception information of threat targets; fusing the information of the four sensors through a multi-layer parallel network and separating multi-layer image information and long-distance abnormal behavior information; based on the multi-layer image information, extracting the morphological characteristics of threat targets and payloads; further, based on the long-distance abnormal behavior information, the morphological characteristics of threat targets and payloads, extracting the abnormal behavior characteristics of threat targets and payloads; then storing the morphological characteristics and abnormal behavior characteristic data of threat targets and payloads as the historical behaviors of threat targets; finally, combining the morphological characteristics, abnormal behavior characteristics, historical behaviors, prior knowledge base, and the spacecraft's own parameter data collected by on-board sensors for fusion reasoning to obtain a quantitative evaluation of the threat category and threat level of the target.

[0019] Further, the decision-making module is further configured to:

[0020] Combine the threat category and threat level, long-distance abnormal behavior information, and the abnormal behavior characteristics of the threat target and payload to reason and make decisions on the specific avoidance behaviors that the spacecraft should take to deal with the threat target. The avoidance behaviors include five types: orbital maneuver, attitude maneuver, emergency avoidance, releasing interference, and normal operation. Secondly, combine the long-distance abnormal behavior information collected by the perception module at the current moment to estimate the future action behaviors of the threat target. Finally, comprehensively consider the predicted future actions of the threat target and the avoidance behaviors obtained from the reasoning and decision-making at the current moment, and solve the optimal action sequence that can minimize the avoidance time and fuel consumption.

[0021] Further, the execution module is further configured to:

[0022] Control the on-board actuators through the controller to execute the predetermined optimal action sequence, so that the spacecraft performs attitude and orbit changes to achieve threat avoidance. Secondly, the result of the avoidance will affect the on-orbit operation state of the spacecraft. Finally, the self-parameter data of the spacecraft measured by the on-board sensors are respectively fed back to the threat level reasoning part of the perception module and the action sequence planning algorithm solving part of the decision-making module to construct an integrated closed-loop system architecture for autonomous avoidance of spacecraft orbital threats of "perception - decision - execution".

[0023] Further, step two includes:

[0024] Define a colored Petri net model applicable to the spacecraft orbital threat autonomous avoidance system, and respectively establish colored Petri net sub-models corresponding to the perception module, decision-making module, and execution module according to the many behaviors involved in the perception module, decision-making module, and execution module, and explain the meanings and operation logics of the places and transitions in each colored Petri net sub-model.

[0025] Further, the definition of the colored Petri net model applicable to the spacecraft orbital threat autonomous avoidance system includes:

[0026] Define the colored Petri net model as a seven-tuple, expressed as:

[0027] Σ = (P, T, F, C, G, E, I)

[0028] Among them, (P, T, F) constitutes a basic Petri net. P is the set of all places in the Petri net, mapping various types of variable information and system states in the logical architecture of the spacecraft orbital threat autonomous avoidance system; T is the set of all transitions in the Petri net, mapping the many intelligent behaviors involved in the perception, decision-making, and execution modules in the logical architecture of the spacecraft orbital threat autonomous avoidance system; It is the set of input and output directed arcs of the Petri net, including the input mapping function on the input arc from place p to transition t and the output mapping function on the output arc from transition t to place p. F maps the occurrence rules of transitions and the input / output data types in the logical architecture of the autonomous orbit threat avoidance system; C is the color set associated with places and transitions, mapping different data and state types in the logical architecture of the autonomous orbit threat avoidance system; G is the guard function of transitions, mapping the guard expressions on each transition; E is the arc expression function, mapping the expressions on each input / output arc; I is the initialization function, mapping the token situation of each place in the initial state of the system.

[0029] Furthermore, the step three includes:

[0030] Build the top-level closed-loop model of the logical architecture of the spacecraft orbit threat autonomous avoidance system, which consists of the colored Petri net sub-model of the perception module, the colored Petri net sub-model of the decision-making module, the colored Petri net sub-model of the execution module, and the interface places between the sub-models. And construct the hierarchical-colored Petri net model of the logical architecture of the spacecraft orbit threat autonomous avoidance system through the port connections between the colored Petri net sub-models.

[0031] Furthermore, the step four includes:

[0032] Use the colored Petri net modeling tool CPNTools to model the hierarchical-colored Petri net model of the logical architecture of the spacecraft orbit threat autonomous avoidance system, and explain the meanings of the custom color sets and variables in the model. Use the simulation tool to simulate the established hierarchical-colored Petri net in the CPNTools environment. All reachable states in the model are the possible valid states of the spacecraft during the threat avoidance process.

[0033] Furthermore, the step five includes:

[0034] Estimate the possible delays in each link during the spacecraft orbit threat autonomous avoidance process to obtain the minimum and maximum delays of each transition, forming a delay interval. Conduct quantitative simulation analysis on the flow time of the entire autonomous avoidance system within the delay interval. On the basis of the original hierarchical-colored Petri net model of the logical architecture of the spacecraft orbit threat autonomous avoidance system, assign timestamps to all transitions in the model to obtain the hierarchical-time-colored Petri net of the logical architecture of the spacecraft orbit threat autonomous avoidance system. Use the CPNTools tool to simulate the model multiple times and adjust the timestamps in real time within the delay interval until the mean value of the obtained flow time meets the requirements of the predetermined index.

[0035] The advantages of the present invention are as follows: The present invention realizes the modeling and simulation of the perception module, decision-making module, and execution module of the spacecraft orbit threat autonomous avoidance system, and respectively establishes corresponding colored Petri net sub-models, realizes the connection between sub-models, establishes a hierarchical-colored Petri net of the entire system logic architecture, and then assigns time stamps to the transitions in the hierarchical-colored Petri net, making full use of the advantages of Petri nets to build a hierarchical-time-colored Petri net based on Petri nets, which can autonomously avoid spacecraft orbit threats without relying on the analysis and decision-making of ground personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a flowchart of a method for designing and analyzing a spacecraft orbit threat autonomous avoidance architecture disclosed in an embodiment of the present invention;

[0037] Figure 2 It is a logical architecture diagram of a method for designing and analyzing a spacecraft orbit threat autonomous avoidance architecture disclosed in an embodiment of the present invention;

[0038] Figure 3 It is a colored Petri net sub-model diagram of the perception module in a method for designing and analyzing a spacecraft orbit threat autonomous avoidance architecture disclosed in an embodiment of the present invention;

[0039] Figure 4 It is a colored Petri net sub-model diagram of the decision-making module in a method for designing and analyzing a spacecraft orbit threat autonomous avoidance architecture disclosed in an embodiment of the present invention;

[0040] Figure 5 It is a colored Petri net sub-model diagram of the execution module in a method for designing and analyzing a spacecraft orbit threat autonomous avoidance architecture disclosed in an embodiment of the present invention;

[0041] Figure 6 It is a top-level closed-loop model diagram of a hierarchical-colored Petri net in a method for designing and analyzing a spacecraft orbit threat autonomous avoidance architecture disclosed in an embodiment of the present invention;

[0042] Figure 7 It is a "perception-decision-execution" integrated colored Petri net model diagram in a method for designing and analyzing a spacecraft orbit threat autonomous avoidance architecture disclosed in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Apparently, the described embodiments are only a part rather than all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0044] A Petri net is a theoretical method that can describe important behavioral attributes such as concurrency, conflict, and resource sharing in complex systems. It also has characteristics such as mathematical expression and graphical visualization, and can effectively model and describe discrete event systems, and verify whether the model performance meets the requirements through simulation analysis. In addition to precisely defining the relationships such as sequence, concurrency, and asynchrony existing in the model process, a Petri net can also find out bad structures such as deadlocks and traps in the system through simulation analysis, and is an effective modeling and analysis tool for describing asynchronous, concurrent, and information processing systems with uncertainty. When modeling a relatively complex system, the model established by a basic Petri net is generally too large in scale, and the model cannot express attributes such as resources and time. Therefore, for different modeling requirements, the basic Petri net has been extended to colored Petri nets, timed Petri nets, hierarchical Petri nets, etc. A colored Petri net uses a color set to distinguish the tokens in the Petri net with different colors, and can fold a large net system to simplify the model structure; on the basis of the basic Petri net, a timed Petri net defines a mapping from transitions to real numbers or real number intervals for all transitions in the network, which respectively represent the specific delay when a transition occurs and the time interval consumed when a transition is enabled, so as to simulate the token migration time; a hierarchical Petri net can clearly describe a hierarchical complex system. By separately establishing subnets for independent subsystems in the system and building a top-level model to simplify the overall net structure, it not only realizes the encapsulation of the subnets, but also meets the repeated calls of the overall net to the subnets.

[0045] As Figure 1 shown, based on the above basic knowledge about Petri nets, the present invention provides a method for designing and analyzing an autonomous avoidance architecture for spacecraft orbit threats based on Petri nets, including the following steps:

[0046] S1: According to the autonomous avoidance task of spacecraft orbit threats, deeply and quantitatively describe the internal evolution law of the behaviors in the system, sort out the information flow logical relationship between nodes, and construct a logically clear autonomous avoidance system architecture for spacecraft orbit threats by establishing a perception module, a decision module, and an execution module.

[0047] Refer to Figure 2, An embodiment of the present invention provides a method for designing and analyzing the logical architecture of a spacecraft orbit threat autonomous avoidance system based on Petri nets. This method constructs a spacecraft orbit threat autonomous avoidance system by establishing a sensing module, a decision-making module, and an execution module. The ports of the execution module are respectively connected to the sensing module and the decision-making module to form an "awareness-decision-making-execution" integrated closed-loop system.

[0048] Continue to refer to Figure 2 , The sensing module is for measuring the state of the spacecraft itself and the environmental scenario, identifying threat features, predicting threat behaviors, and determining threat levels. First, complementary information such as images and point clouds is used for rapid segmentation and accurate identification of targets and characteristic components, giving the morphological features of threat targets and their expressions; secondly, under the condition of incomplete information in different spatio-temporal conditions, the abnormal behaviors of orbit threat targets are accurately detected, and the abnormal behavior features of threat targets and their expressions are given; finally, fusion reasoning is carried out by combining environmental features, morphological features, motion features, and historical behaviors to obtain a pre-judgment of the target behavior and a quantitative evaluation of the threat level. The specific introduction of the sensing module is as follows:

[0049] In terms of the sensing module, first, four types of sensing devices, namely optical cameras, infrared cameras, microwave radars, and laser point clouds, are used to obtain the sensing information of threat targets, and precise threat perception in complex space environments is achieved through the complementarity of information between different devices; secondly, the information of the four sensors is fused through a multi-layer parallel network to enhance some incomplete information and separate the multi-layer image information from the long-distance abnormal behavior information; based on the multi-layer image information, the morphological features of threat targets and payloads are extracted. Among them, the target morphological features include target category labels (enemy spacecraft, space debris), target line-of-sight information, target confidence, target timestamp, etc., and the payload morphological features include payload category labels (robotic arm, camera, antenna), payload line-of-sight information, payload confidence, payload timestamp, etc.; then, based on the long-distance abnormal behavior information and the morphological features of threat targets and payloads, the abnormal behavior features of threat targets and payloads are extracted, including threat target ID, payload quantity, six-dimensional attitude / orbit six orbital elements, current moment speed, behavior semantics (following flight, orbiting flight, accompanying flight, grazing flight, approaching rapidly), confidence, etc.; then, the morphological features and abnormal behavior feature data of threat targets and payloads are stored as the historical behaviors of threat targets; finally, fusion reasoning is carried out by combining morphological features, abnormal behavior features, historical behaviors, a priori knowledge base, and the spacecraft's own parameter data collected by on-board conventional sensors to obtain a quantitative evaluation of the threat category and threat level of the target. Among them, the threat categories are divided into three types: space debris with a collision threat (Type A), spacecraft with a close approach threat (Type B), and spacecraft with a normal patrol threat (Type C).

[0050] It should be noted that the main improvement of the present invention lies in constructing a perception - decision - execution integrated closed - loop system, constructing corresponding colored Petri net sub - models based on this system, then establishing a hierarchical - colored Petri net of the entire system logic architecture based on each colored Petri net sub - model, then conducting modeling and simulation, and finally assigning time stamps to the transitions in the hierarchical - colored Petri net to construct a hierarchical - time - colored Petri net. Based on this hierarchical - time - colored Petri net, autonomous avoidance of orbital threats is carried out. Therefore, the above - mentioned method of fusion reasoning by combining environmental characteristics, morphological characteristics, motion characteristics, and historical behaviors is not within the protection scope of the present invention. Any fusion reasoning method can be adopted. In practical applications, when the morphological characteristics, abnormal behavior characteristics, historical behaviors, prior knowledge bases, and spacecraft's own parameters collected by on - satellite conventional sensors reach different preset conditions, different threat categories can be set artificially.

[0051] Continue to refer to Figure 2 , the decision - making module estimates the parameters of the threat target, predicts the threat behavior, and makes decisions on avoidance behaviors, selects the best solution from possible solutions, and forms serialized action sequences. Considering complex multi - constraints such as the spacecraft's predetermined mission, orbit, maneuverability, computing power, and safety, rapid decision - making planning under limited computing resources is realized through online optimization of the threat avoidance strategy, and a feedback mechanism for online adjusting the parameters and structure of the decision - making module according to the situation characteristics is adopted to realize dynamic game decision - making under imperfect information. The specific introduction of the decision - making module is as follows:

[0052] In terms of the decision - making module, first, combined with the threat category, threat level, long - distance abnormal behavior information, and the abnormal behavior characteristics of the threat target and the payload, the specific avoidance behaviors that the spacecraft should take to deal with the threat target are inferred and decided. The avoidance behaviors include orbit maneuver, attitude maneuver, emergency avoidance, releasing interference, and normal operation, and multiple avoidance behaviors can be carried out simultaneously; second, combined with the long - distance abnormal behavior information collected by the perception module at the current moment, including speed, distance, azimuth angle, etc., the future action behaviors of the threat target are predicted; finally, based on the predicted future actions of the threat target and the avoidance behaviors inferred and decided at the current moment, the optimal action sequence that can minimize the avoidance time and fuel consumption is solved through a planning algorithm. It should be noted that what the present invention protects is the modeling idea based on Petri nets. Any existing relevant methods can be used for inferring and deciding the specific avoidance behaviors that the spacecraft should take to deal with the threat target, the inference of avoidance behaviors, and the planning algorithm, etc. The relevant inferences and planning algorithms or methods are not within the protection scope of the present invention.

[0053] Continue to refer to Figure 2, the execution module executes the action sequence through the attitude and orbit motion controller, controls the actuators on the spacecraft to perform specific attitude and orbit changes, and feeds back some parameter data of the spacecraft to the perception module and the decision-making module. The specific introduction of the execution module is as follows:

[0054] In terms of the execution module, first, the on-board actuators are controlled by the controller to execute a predetermined action sequence, enabling the spacecraft to perform attitude and orbit changes to avoid threats; second, the result of the avoidance will affect the on-orbit operation state of the spacecraft (including two types: normal operation and avoidance in progress); finally, the parameter data of the spacecraft itself measured by the on-board conventional sensors are respectively fed back to the threat level inference module of the perception module and the action sequence planning algorithm solving module of the decision-making module to construct an integrated closed-loop system architecture for autonomous avoidance of spacecraft orbit threats of "perception - decision - execution".

[0055] S2: By clarifying the operation process of the logical architecture of the spacecraft orbit threat autonomous avoidance system, using the Petri net modeling and simulation tool CPNTools, corresponding colored Petri net sub-models are established for the perception module, decision-making module, and execution module based on the colored Petri net to achieve modularization and generalization of model building. The basic principle and the specific process of constructing the Petri net simulation sub-model are as follows:

[0056] The colored Petri net is an advanced Petri net that can build compact and parameterized models, construct concurrent system models by modeling and specifying system behaviors, and can analyze the behavioral attributes of the models.

[0057] The present invention defines the colored Petri model as a seven-tuple, specifically expressed as:

[0058] Σ = (P, T, F, C, G, E, I)

[0059] Among them, (P, T, F) constitutes the basic Petri net. P is the set of all places in the Petri net, mapping various types of variable information and system states in the logical architecture of the spacecraft orbit threat autonomous avoidance system; T is the set of all transitions in the Petri net, mapping many intelligent behaviors involved in the perception, decision-making, and execution sub-models in the logical architecture of the spacecraft orbit threat autonomous avoidance system. And there are no instantaneous transitions in the Petri net model of this logical architecture, and all transitions are delay transitions; is the set of input and output directed arcs of the Petri net, including the input mapping function on the input arc from the place p to the transition t and the output mapping function on the output arc from the transition t to the place p. F maps the occurrence rules of transitions and the input / output data types in the logical architecture of the autonomous evasion system for orbital threats; C is the color set associated with places and transitions, mapping different data and state types in the logical architecture of the autonomous evasion system for orbital threats, and some places are described using a combined color set; G is the guard function of the transition, mapping the guard expression on each transition; E is the arc expression function, mapping the expression on each input / output arc; I is the initialization function, mapping the token situation of each place in the initial state of the system. The above clearly defines the constituent symbols of the colored Petri net model, indicating the elements contained in the network and their representations, and determining the symbol standard and model paradigm for the next step of characterizing the Petri net sub-model.

[0060] The colored Petri net sub-model of the sensing module (hereinafter referred to as the sensing sub-model) is as Figure 3 shown, and its specific description is as follows:

[0061] The sensing module starts from the port place P 17 . P 17 is a place with a combined color set, containing multiple information of threat targets and is also the input of the entire logical architecture of the autonomous evasion system for orbital threats. Since the field of view, detection distance, measurement accuracy, and imaging conditions of a single sensor device are relatively fixed, the perceivable information of threat targets is limited. To obtain multi-dimensional and rich perceivable information of threat targets, optical camera monitoring, infrared camera monitoring, microwave radar monitoring, and laser point cloud monitoring are used (corresponding to transitions T 11 , T 12 , T 13 , T 14 ) to obtain perceivable information under different sensor devices (corresponding to places P 1 , P 2 , P 3 , P 4 ); then, the above four types of perceivable information are used to execute the transition T 1 for multi-sensor information fusion to enhance some incomplete information. The output places after information fusion are: the multi-layer image information place P 5 , the long-distance abnormal behavior information place P 6 , and the long-distance abnormal behavior information cache place P 19 . Among them, the long-distance abnormal behavior information place P 6 includes information such as the speed, distance, and azimuth angle of threat targets, and the place P 19It is a cached copy of abnormal behavior information and one of the interface libraries between the perception module and the decision-making module. After obtaining multi-layer image information and abnormal behavior information respectively, first, for the multi-layer image information library P 5 Perform morphological feature extraction under the transition T 2 to obtain the morphological feature recognition result library P of the threat target and the payload 7 . Then, on this basis, combine the long-distance abnormal behavior information library P 6 Perform abnormal behavior feature extraction under the transition T 3 to obtain the abnormal behavior feature recognition result library P of the threat target and the payload 8 and its cached copy library P 18 . Among them, the library P 18 is also one of the interface libraries between the perception module and the decision-making module. Next, the perception module will combine the morphological feature recognition result library P of the threat target and the payload 7 , the abnormal behavior feature recognition result library P of the threat target and the payload 8 , the threat target historical behavior library P 9 , the prior knowledge base P 10 , and the spacecraft's own orbital parameter database library P fed back by the execution module 16 Execute T 5 to infer the threat level of the threat target and obtain the main output of the perception module, that is, the combined library P of threat category and threat level 11 . By setting the priority library priority, the threat level inference transition T 5 occurs before the perception information storage transition T 4 . Finally, store the above-mentioned morphological feature and abnormal behavior feature recognition results through the transition T 4 in the library P 9 as the historical behavior of the threat target to participate in the next threat level inference. The modeling and simulation of the perception module of the spacecraft orbit threat autonomous avoidance system are realized.

[0062] The colored Petri net sub-model of the decision-making module (hereinafter referred to as the decision sub-model) is as Figure 4 shown, and its specific description is as follows:

[0063] The decision-making module first executes the transition T 19 under the abnormal behavior information cached copy library P 7 to infer and estimate the future behavior of the threat target and output the future action library P of the threat target 13 . Secondly, the decision-making module synthesizes three interface libraries between the perception module and the decision-making module, including the abnormal behavior information cached copy library P 19 and the cached copy library P of the abnormal behavior feature recognition result of the threat target and the payload18 and the combined place P of threat category and threat level 11 and execute transition T 6 Make a decision on the specific avoidance behaviors that the spacecraft will take against the threat target, and output the place P of the avoidance behaviors that need to be taken at the current moment 12 Next, the decision-making module will synthesize the place P of the future actions of the threat target 13 and the place P of the avoidance behaviors that need to be taken at the current moment 12 and the place P of the spacecraft's own orbital parameter database fed back by the execution module 16 Execute transition T8 to solve the action sequence required for the avoidance behavior through the planning algorithm, so as to obtain the output of the decision-making module, that is, the optimal action sequence place P for the spacecraft to perform attitude and orbit changes and other actions 14 . The modeling and simulation of the decision-making module of the spacecraft orbit threat autonomous avoidance system are realized

[0064] The colored Petri net sub-model of the execution module (hereinafter referred to as the execution sub-model) is as follows Figure 5 shown, and its specific description is as follows

[0065] The input port place of the execution module is the output place P of the decision-making module 14 , and the execution module needs to execute transition T 9 Control the execution mechanism on the spacecraft through the controller to execute the action sequence instruction, complete the avoidance behaviors such as attitude and orbit changes of the spacecraft, and the execution result of this avoidance will affect the place P of the on-orbit operation state of the spacecraft 15 , where P 15 has two states: normal operation and avoidance. On this basis, execute transition T 10 Measure the own parameter data of the spacecraft in orbit by means of the conventional sensors equipped on the spacecraft, and store them in the place P 16 , and the place P 16 will participate in the threat level reasoning in the above-mentioned perception module and the solution of the action sequence planning algorithm in the decision-making module as a feedback link respectively. The modeling and simulation of the execution module of the spacecraft orbit threat autonomous avoidance system are realized

[0066] S3: Build the top-level closed-loop model of the spacecraft orbit threat autonomous avoidance system logic architecture, and according to the nested coupling relationship between the colored Petri net sub-models of the perception module, decision-making module and execution module, connect the sub-models by setting interfaces at the entrances and exits of each colored Petri net sub-model to establish the hierarchical-colored Petri net of the entire system logic architecture, and conduct simulation analysis and verification on the avoidance process. The detailed process is as follows

[0067] Refer to Figure 6, the present invention provides a top-level closed-loop model of the logical architecture of the autonomous avoidance system for spacecraft orbit threats, which consists of a perception sub-model, a decision-making sub-model, an execution sub-model, and an interface library between the sub-models. The input of this top-level model is the multi-layer information repository P of threat targets 17 , after the multi-sensor information fusion and feature extraction of the threat target multiple information by the perception sub-model, etc., the output is the cache copy repository P of abnormal behavior information 19 and the cache copy repository P of the recognition results of the abnormal behavior characteristics of the threat target and the payload 18 as well as the combined repository P of threat categories and threat levels 11 ; the above three interface repositories will participate in behaviors such as avoidance behavior decision-making and action planning algorithm solving as the input of the decision-making sub-model, and output the action sequence repository P 14 ; the execution sub-model executes the action sequence and finally outputs the database repository P of the parameters of the spacecraft's own in-orbit operation measured by the spacecraft's conventional sensors 16 , the repository P 16 will participate in the threat level reasoning in the above-mentioned perception module and the action sequence planning algorithm solving in the decision-making module respectively as the feedback link, and thus the top-level closed-loop model of "perception - decision-making - execution" is built.

[0068] At the same time, a hierarchical - colored Petri net model of the logical architecture of the autonomous avoidance system for spacecraft orbit threats is constructed through the port connections between the sub-models, as Figure 7 shown. Among them, the perception sub-model and the decision-making sub-model are connected through the port repositories P 11 , P 18 , P 19 , the decision-making sub-model and the execution sub-model are connected through the repository P 14 , and the execution sub-model and the perception sub-model and the decision-making sub-model are connected through the repository P 16 . The meanings of the repositories and transitions in the entire hierarchical - colored Petri net of the logical architecture of the autonomous avoidance system for spacecraft orbit threats are shown in Table 1.

[0069] Table 1 Meanings of repositories and transitions in the hierarchical - colored Petri net

[0070]

[0071] S4: Use the colored Petri net modeling tool CPNTools to model the hierarchical - colored Petri net of the logical architecture of the autonomous avoidance system for spacecraft orbit threats, visually verify the correctness of the model through simulation, and analyze the dynamic characteristics of the model in the state space. The detailed process is as follows:

[0072] The present invention uses the colored Petri net modeling tool CPNTools to model and simulate the hierarchical-colored Petri net model of the logic architecture of the spacecraft orbit threat autonomous avoidance system.

[0073] The CPNTools simulation tool supports the editing and simulation analysis of colored Petri net models, enables graphical modeling and dynamic simulation of colored Petri nets, and contains a large number of toolkits in CPNTools such as state space tools, model simulation tools, monitor tools, etc. With the help of these toolkits, the model can be dynamically simulated, and many behavioral attributes such as liveness, boundedness, fairness, etc. in the state space can be analyzed. It can also count the model process time and resource utilization.

[0074] Several basic types of color sets have been defined in the CPNTools simulation tool, and it supports users to customize color sets according to modeling needs. In addition to the basic color sets in CPNTools, some customized color sets in the present invention are as follows:

[0075] colset Single = with single_A | single_B | single_C | single_D; / / Declare the enumeration color set of four sensor signals

[0076] colset image = string; / / Declare the color set of multi-layer image information as String type

[0077] colset velocity = int; / / Declare the velocity information in the abnormal behavior information as int type

[0078] var v, v1: velocity;

[0079] colset distance = int; / / Declare the distance information in the abnormal behavior information as int type

[0080] var d, d1: distance;

[0081] colset azimuth = int; / / Declare the azimuth information in the abnormal behavior information as int type

[0082] var az, az1: azimuth;

[0083] colset behaviour = product velocity * distance * azimuth; / / Construct the color set of abnormal behavior information as a product color set containing velocity, time, and azimuth

[0084] colset information = union i:image + b:behaviour; / / Construct a combined color set of multi-layer image information and abnormal behavior information

[0085] colset target_type = with spacecraft|spacedebris|driftlessness; / / Declare an enumerated color set for threat target types, where the threat target types are enemy spacecraft, space debris, and no target

[0086] var r_t_t:target_type;

[0087] colset load_type = with mechanical_arm|camera|antennae|no_load; / / Declare an enumerated color set for load types, where the load types carried by threat targets are mechanical arm, camera, antenna, and no load

[0088] var r_l_t:load_type;

[0089] colset m_c = product target_type * load_type; / / Construct a product color set of threat target type and load type

[0090] colset target_behaviour = with follow_fly|around_fly|with_fly|Prey_fly|quick_close|target_no_action; / / Declare an enumerated color set for the abnormal behavior characteristics of threat targets, where the abnormal behaviors of threat targets are following flight, orbiting flight, accompanying flight, swooping flight, rapid approach, and no maneuver

[0091] var r_t_b:target_behaviour;

[0092] colset load_beaviour = with stick_out|in_focus|strike|load_no_action; / / Declare an enumerated color set for the abnormal behavior characteristics of loads, where the abnormal behaviors of loads carried by threat targets are sticking out, focusing, attacking, and no maneuver

[0093] var r_l_b:load_beaviour;

[0094] colset b_c = product target_behaviour * load_beahiour; / / Construct the product color set of the threat target abnormal behavior feature and the load abnormal behavior feature

[0095] colset result = product m_c * b_c; / / Construct the product color set of the morphological feature extraction result and the abnormal behavior feature extraction result

[0096] var res: result;

[0097] colset history_behaviour = list result; / / Declare the list color set for storing historical behaviors

[0098] var h_b: history_behaviour;

[0099] colset threat_type = with A|B|C; / / Declare the enumeration color set of threat types, where the threat types include space debris with collision threat (type A), spacecraft with approaching threat (type B), and spacecraft with regular inspection threat (type C)

[0100] var t_t: threat_type;

[0101] colset level_a = int with 1..3; / / Declare the index color set of threat levels, where the threat levels of type A threats are from level 1 to level 3

[0102] var l_a: level_a;

[0103] colset level_b = int with 4..7; / / Declare the index color set of threat levels, where the threat levels of type B threats are from level 4 to level 7

[0104] var l_b: level_b;

[0105] colset level_c = int with 8..11; / / Declare the index color set of threat levels, where the threat levels of type C threats are from level 8 to level 11

[0106] var l_c: level_c;

[0107] var l: INT;

[0108] colset t_l = product threat_type * INT; / / Construct the color set of the product of threat type and threat level

[0109] colset avoidance_behavior = with orbital_maneuver | attitude_maneuver

[0110] | emergency_avoidance | interference | normal_operation; Declare the color set of the product of avoidance behaviors, where avoidance behaviors include five types: orbital maneuver, attitude maneuver, emergency avoidance, releasing interference, and normal operation

[0111] var a_b: avoidance_behavior;

[0112] colset Status = bool with(evading, normal); / / Declare the Bool - type color set of the on - orbit operation status of the spacecraft

[0113] var s: Status;

[0114] colset Target = product target_type * load_type * behaviour; / / Declare the color set of the product of multiple information of threat targets;

[0115] The established hierarchical - colored Petri net is simulated using a simulation tool in the CPNTools environment. When running in a single - step mode, the Petri net model runs strictly according to the preset transition guard function and the arc expression functions on the input and output arcs. There is no dead - lock situation during the running process, and the feedback of the closed - loop model is effective, realizing the visual simulation of the operation of the logic architecture of the autonomous avoidance system for spacecraft orbital threats; to enable the model to continuously sense threat targets and take corresponding actions, simply modify the model structure to adjust the model to a closed - loop and recyclable running structure. At this time, the automatic - running simulation method in CPNTools can be adopted. The simulation results show that the model can sense threat targets multiple times and execute the corresponding actions in the sensing sub - model, decision - making sub - model, and execution sub - model, and take corresponding countermeasures against threat targets.

[0116] To analyze the behavioral properties such as reachability, liveness, and boundedness of the hierarchical - colored Petri net model for the autonomous avoidance system of spacecraft orbit threats, a state - space tool is used to generate the state - space report of this Petri net model. In terms of reachability, the state - space report lists all reachable states in the hierarchical - colored Petri net model of the autonomous avoidance system of spacecraft orbit threats. And through analysis, it can be seen that all the above - mentioned reachable states are valid states that may exist during the threat - avoidance process of the spacecraft. Therefore, this Petri model satisfies reachability; in terms of liveness, since every time the model loops, the perception information storage transition T 4 adds the morphological feature and abnormal behavior feature information to the historical behavior place P 9 In it, the number of tokens in the place P 9 will continue to increase, and the model will never return to the initial state. Therefore, this Petri model does not satisfy liveness; in terms of the liveness of the model, the state - space report shows that there are no live transitions and dead transitions in the model, and when the number of simulation steps of the model is set to 10,000, no deadlock occurs either. It can be inferred that there is no deadlock state in this model. Therefore, this Petri model satisfies liveness; in terms of boundedness, the state - space report shows that the model satisfies boundedness; in terms of fairness, because there is no infinitely occurring sequence in this Petri net model, the state - space report shows that the model does not satisfy fairness. All in all, this Petri net model satisfies basic behavioral properties such as reachability, liveness, and boundedness, and realizes the model simulation of the logical architecture of the autonomous avoidance system of spacecraft orbit threats.

[0117] S5: Assign time stamps to the transitions in the hierarchical - colored Petri net, add time attributes to the model, and build a hierarchical - time - colored Petri net for the logical architecture of the autonomous avoidance system of spacecraft orbit threats to analyze the process time when the model responds to threat events. The detailed process is as follows:

[0118] The autonomous avoidance task of spacecraft orbit threats requires the spacecraft to complete actions such as perception, decision - making, attitude and orbit change avoidance of threat targets in the shortest possible time. Therefore, it is necessary to use a timed Petri net to endow the system with time attributes and conduct quantitative simulation analysis on the process time of the entire autonomous avoidance system, so as to optimize the model structure to enable the spacecraft to quickly avoid threat targets under the premise of the least threat and return to the normal on - orbit operation state.

[0119] To conduct a quantitative simulation analysis of the process time of the entire autonomous avoidance system, it is necessary to assign timestamps to all transitions in the original hierarchical - colored Petri net model of the spacecraft orbit threat autonomous avoidance system logic architecture. Based on the estimated possible delays in each link during the spacecraft orbit threat autonomous avoidance process, the minimum and maximum delays of each transition are obtained as shown in Table 2. This time range is a random variable that satisfies a uniform distribution during the simulation process.

[0120] Table 2 Delay ranges of each transition in the model

[0121]

[0122] In the CPNTools simulation tool, the color sets involved in S4 are set as timed color sets by adding the "timed" suffix after the color sets, and a delay range is assigned to each transition, obtaining a hierarchical - time - colored Petri net of the spacecraft orbit threat autonomous avoidance system logic architecture. Use the CPNTools simulation tool to simulate the model 50 times, and use the data collection monitor in the built - in monitor function Monitors to collect the global time of the model at the end of each closed - loop model run. The single - run time of the spacecraft orbit threat autonomous avoidance system model is shown in Table 3. To obtain more accurate statistical results, use the CPNTools tool to simulate the model 500 times, and the average process time of a single run is 114.36 ms, meeting the predetermined index requirements.

[0123] Table 3 Single - run time of the spacecraft orbit threat autonomous avoidance system model

[0124]

[0125] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A design and analysis method for autonomous orbital threat avoidance architecture for spacecraft, It is characterized in that The method comprises: Step 1: Construct the logical architecture of the spacecraft orbit threat autonomous avoidance system based on the perception module, decision module, and execution module; Step 2: Based on the colored Petri net, corresponding colored Petri net sub-models are established for the perception module, decision module, and execution module respectively; Step 3: Through the nested coupling relationship between the colored Petri net sub-models of the perception module, decision module and execution module, interfaces are set at the entrance and exit of each colored Petri net sub-model to realize the connection between sub-models, so as to establish the hierarchical colored Petri net of the entire system logical architecture, and perform simulation analysis and verification on the avoidance process; Step 4: Model the hierarchical-colored Petri net of the logical architecture of the spacecraft orbital threat autonomous avoidance system, visually verify the correctness of the model through simulation, and analyze the dynamic characteristics of the model in the state space; Step 5: Assign timestamps to the transitions in the layered-colored Petri net, build the layered-time-colored Petri net of the logical architecture of the spacecraft orbit threat autonomous avoidance system, autonomously avoid spacecraft orbit threats based on the layered-time-colored Petri net, and analyze the process time of the model when responding to threat events.

2. A spacecraft orbital threat autonomous avoidance architecture design and analysis method according to claim 1, It is characterized in that The step one comprises: The perception model is used to measure the spacecraft's own state and environmental scene, identify threat characteristics, predict threat behavior, and determine threat levels; The decision module is used to estimate the parameters of the threat target, predict the threat behavior, make evasion decisions, select the best solution from possible solutions, and form a serialized action sequence; The execution module is used to execute the action sequence through the attitude and orbit action controller, control the actuators on the spacecraft to perform specific attitude and orbit change actions, and feed back some parameter data of the spacecraft to the perception module and decision module.

3. A spacecraft orbital threat autonomous avoidance architecture design and analysis method according to claim 2, It is characterized in that The perception module is also used for: Use optical camera, infrared camera, microwave radar and laser point cloud to obtain the perception information of threat targets; fuse the information of four sensors through multi-layer parallel network, and separate multi-layer image information and long-distance abnormal behavior information; extract the morphological characteristics of threat targets and payloads based on multi-layer image information; Then, based on the long-distance abnormal behavior information and the morphological characteristics of the threat target and the payload, the abnormal behavior characteristics of the threat target and the payload are extracted; then the morphological characteristics and abnormal behavior characteristic data of the threat target and the payload are stored as the historical behavior of the threat target; finally, the morphological characteristics, abnormal behavior characteristics, historical behavior, prior knowledge base and the spacecraft's own parameter data collected by the on-board sensors are combined for fusion reasoning to obtain a quantitative evaluation of the target's threat category and threat level.

4. A spacecraft orbital threat autonomous avoidance architecture design and analysis method according to claim 3, It is characterized in that the decision-making module is further configured to: reason and make decisions on the specific avoidance behaviors that the spacecraft should take to respond to threat targets by combining the threat category and threat level, long-distance abnormal behavior information, and the abnormal behavior characteristics of threat targets and payloads. The avoidance behaviors include five types: orbital maneuver, attitude maneuver, emergency avoidance, releasing interference, and normal operation. Secondly, combine the long-distance abnormal behavior information collected by the sensing module at the current moment to estimate the future action behaviors of the threat target. Finally, comprehensively consider the future actions of the threat target obtained by prediction and the avoidance behaviors obtained by reasoning and decision-making at the current moment, and solve the optimal action sequence that can minimize the avoidance time and fuel consumption.

5. The method for designing and analyzing an autonomous avoidance architecture for spacecraft orbital threats according to claim 4, It is characterized in that the execution module is further configured to: control the on-board actuators through the controller to execute the predetermined optimal action sequence, so that the spacecraft performs attitude and orbit change actions to achieve threat avoidance; secondly, the avoidance result will affect the on-orbit operation state of the spacecraft; finally, the self-parameter data of the spacecraft measured by the on-board sensors are respectively fed back to the threat level reasoning part of the sensing module and the action sequence planning algorithm solving part of the decision-making module to construct an integrated closed-loop system architecture for autonomous avoidance of spacecraft orbital threats of "sensing - decision-making - execution".

6. The method for designing and analyzing an autonomous avoidance architecture for spacecraft orbital threats according to claim 2, It is characterized in that Step two includes: Define a colored Petri net model applicable to the autonomous avoidance system for spacecraft orbital threats, and respectively establish colored Petri net sub-models corresponding to the sensing module, decision-making module, and execution module according to the many behaviors involved in the sensing module, decision-making module, and execution module, and explain the meanings and operation logics of places and transitions in each colored Petri net sub-model.

7. The method for designing and analyzing an autonomous avoidance architecture for spacecraft orbital threats according to claim 6, It is characterized in that The definition of the colored Petri net model applicable to the autonomous avoidance system for spacecraft orbital threats includes: Define the colored Petri net model as a seven-tuple, expressed as: Σ=(P,T,F,C,G,E,I) Among them, (P, T, F) constitutes a basic Petri net. P is the set of all places in the Petri net, mapping various types of variable information and system states in the logical architecture of the autonomous avoidance system for orbital threats; T is the set of all transitions in the Petri net, mapping many intelligent behaviors involved in the perception, decision-making, and execution modules in the logical architecture of the autonomous avoidance system for orbital threats. F is the set of input and output directed arcs of the Petri net, including the input mapping function on the input arc from place p to transition t and the output mapping function on the output arc from transition t to place p. F maps the occurrence rules of transitions and the input / output data types in the logical architecture of the autonomous avoidance system for orbital threats; C is the color set associated with places and transitions, mapping different data and state types in the logical architecture of the autonomous avoidance system for orbital threats; G is the guard function of the transition, mapping the guard expression on each transition; E is the arc expression function, mapping the expression on each input / output arc; I is the initialization function, mapping the token situation of each place in the initial state of the system.

8. The method for designing and analyzing an autonomous avoidance architecture for spacecraft orbital threats according to claim 6, It is characterized in that Step three includes: Build a top-level closed-loop model of the logical architecture of the spacecraft orbital threat autonomous avoidance system composed of the colored Petri net sub-model of the sensing module, the colored Petri net sub-model of the decision-making module, the colored Petri net sub-model of the execution module, and the interface places between the sub-models, and construct a hierarchical - colored Petri net model of the logical architecture of the spacecraft orbital threat autonomous avoidance system through the port connections between the colored Petri net sub-models.

9. The method for designing and analyzing an autonomous avoidance architecture for spacecraft orbital threats according to claim 8, It is characterized in that Step four includes: Use the colored Petri net modeling tool CPNTools to model the hierarchical - colored Petri net model of the logic architecture of the spacecraft orbit threat autonomous avoidance system, explain the meanings of the custom color sets and variables in the model, and use the simulation tool to simulate the established hierarchical - colored Petri net in the CPNTools environment. All reachable states in the model are valid states that may exist during the threat avoidance process of the spacecraft.

10. A method for designing and analyzing the autonomous avoidance architecture of a spacecraft orbit threat according to claim 9, characterized in that, the fifth step includes: Estimate the possible delays in each link during the autonomous avoidance process of the spacecraft orbit threat to obtain the minimum and maximum delays of each transition, forming a delay interval. Conduct quantitative simulation analysis on the process time of the entire autonomous avoidance system within the delay interval. Based on the original hierarchical - colored Petri net model of the logic architecture of the spacecraft orbit threat autonomous avoidance system, assign timestamps to all transitions in the model to obtain the hierarchical - time - colored Petri net of the logic architecture of the spacecraft orbit threat autonomous avoidance system. Use the CPNTools tool to simulate the model multiple times and adjust the timestamps in real - time within the delay interval until the mean value of the obtained process time meets the requirements of the predetermined index.

Images