A multi-node probe flexible attachment re-planning method

Abstract

The present application relates to a kind of multi-node probe flexible attachment re-planning method, belong to aerospace and artificial intelligence cross technical field.The system constraint conflict problem that may be generated by the task sequence of initial planning being executed in view of small asteroid exploration activities, due to insufficient cognition to small asteroid environment, each node of probe and possible fault in node subsystem, facing the flexible attachment task demand of multi-node probe, the present application establishes the flexible landing re-planning problem model of multi-node probe, creates the layered constraint graph model of re-planning, proposes the re-planning method of layered time delay satisfaction verification mechanism and layered neighborhood re-planning search algorithm based on layered constraint graph, to meet the demand of multi-node probe flexible landing to autonomous re-planning.The present application lays a technical foundation for the successful landing of multi-node probe and subsequent small asteroid resource collection, planet walk and other further exploration tasks.

Description

Technical Field

[0001] This invention relates to a flexible attachment replanning method for multi-node detectors, belonging to the interdisciplinary field of aerospace and artificial intelligence. Background Technology

[0002] Asteroid exploration is of great significance in studying the origin of life, developing deep space resources, and ensuring national security. Achieving a safe landing of a deep space probe is a necessary prerequisite for further exploration activities such as collecting asteroid surface samples and planetary walks.

[0003] Due to the weak gravity and unknown topography of asteroid surfaces, traditional rigid attachment methods are prone to causing probes to bounce or tumble. A multi-node cooperative flexible attachment method holds promise for solving the problem of unstable landings. During attachment, the multi-node flexible system can capture more information about the asteroid surface, and the flexible cooperation between nodes also contributes to a smooth landing. However, due to insufficient human understanding of the asteroid environment and potential malfunctions within the probe's nodes and subsystems, executing the initially planned mission sequence may lead to system constraint conflicts. In such cases, the multi-node probe system needs to activate its autonomous replanning function. Summary of the Invention

[0004] The purpose of this invention is to address the system constraint conflicts that may arise during the execution of the initially planned mission sequence in asteroid exploration activities due to insufficient understanding of the asteroid environment and potential failures within the nodes and subsystems of the probe. To meet the requirements of flexible attachment missions for multi-node probes, this invention creatively proposes a replanning method for flexible attachment of multi-node probes. This method lays the technical foundation for the successful landing of multi-node probes and subsequent exploration missions such as asteroid resource collection and planetary walks.

[0005] The innovations of this invention include: establishing a replanning problem model for flexible landing of multi-node detectors, creating a hierarchical constraint graph model for replanning, and proposing a replanning method based on a hierarchical delay satisfaction verification mechanism and a hierarchical neighborhood replanning search algorithm, thereby meeting the autonomous replanning requirements for flexible landing of multi-node detectors.

[0006] The present invention is achieved using the following technical solution.

[0007] A flexible attachment replanning method for multi-node detectors includes the following steps:

[0008] Step 1: Establish a model for the flexible landing replanning problem of a multi-node probe.

[0009] Specifically, the flexible landing replanning problem for a multi-node probe can be represented as a triple: , It is a time-ordered set of actions. any element , and These represent the activity start time object and the activity end time object, respectively. ,in , and These are all attributes of the activity, representing the node to which the activity belongs, the subsystem to which it belongs, and the activity number, respectively. This represents the point in time corresponding to the time object. This indicates fault information during the landing process of a multi-node probe; Constraints for multi-node detectors (including time, resources, etc.).

[0010] Multi-node deep space probes typically execute multiple activities concurrently, with time constraints existing between these activities. These constraints can be described by constraints between time objects associated with the activities. Resource constraints for multi-node probes include continuously consumed resources and exclusive resources. The consumption rate of continuously consumed resources by on-orbit activities should be limited to a certain range, while exclusive resources refer to resources that are only allowed to be used by one node at a time.

[0011] Step 2: Create a hierarchical constraint graph model for the flexible landing replanning of a multi-node detector.

[0012] In a multi-node detector system, a hierarchical constraint graph model for flexible landing replanning of a multi-node detector is created, with active time objects as vertices and constraints between time objects as constraint edges.

[0013] Specifically, construct a graph model G of the multi-node detector system.<V,E> V is the set of points representing the time objects of the activity, and E is the set of time constraint edges. From the points... Edge pointing to point j Represented as , , This represents the two time objects corresponding to a certain constraint edge.

[0014] Based on the time object attributes at both ends of the edge, the constraint edges are divided into intra-subsystem constraint edges, inter-subsystem constraint edges, and inter-node constraint edges. The constraint edge set of a multi-node detector system is represented as follows: ,in This represents the internal constraints of the subsystem. Represents the set of constraints between subsystems. This represents the set of constraints between nodes. Based on the constraint edge hierarchy, a hierarchical constraint graph model of the multi-node detector system is constructed.

[0015] Step 3: Within the subsystem to which fault information W belongs, if all unexecuted activities have the same delay, then when a delay exists... If the constraints of the activity layer of the faulty subsystem are satisfied, then the remaining unexecuted activity delay time is greater than the time. The activity layer constraints are also satisfied. Furthermore, when the remaining unexecuted activity delay time belongs to the activity delay time of the fault, the constraints of the activity layer of the subsystem to which the fault belongs will definitely be satisfied.

[0016] Based on the monotonically increasing time-delay property of the active layer of the subsystem to which the fault belongs, a binary search algorithm is used to find the minimum delay that satisfies the internal time constraints of the subsystem. .

[0017] Step 4: Within the node to which fault W belongs, consider the time constraints of on-orbit activities and adjacent activities of non-fault-related subsystems, the time constraints between remaining unexecuted activities of all subsystems, the time constraints between activities to which fault belongs and unexecuted activities, and the satisfaction of internal resource constraints within the node to which fault belongs. For the constraints between on-orbit activities and adjacent activities of non-fault-related subsystems, when delay exists... When this constraint is satisfied, the adjacency activity delay time is less than This also satisfies such constraints.

[0018] Based on the monotonically decreasing satisfaction of these constraints, a binary search algorithm is used to determine the maximum value of the subsystem delay. The delay of the faulty subsystem is ,like The overall delay of the faulty node If all resource constraints can be satisfied at this point, proceed to step 5. Otherwise, if all time constraints cannot be satisfied at the subsystem constraint layer, the hierarchical domain replanning search algorithm in steps 6 and 9 will be executed.

[0019] Step 5: In non-faulty nodes, the time constraints of unexecuted activities within the node, the time constraints of on-orbit activities and adjacent activities in the subsystem, the time constraints between unexecuted activities and the faulty subsystem, and the constraints of exclusive resources between nodes must also be considered.

[0020] For time constraints between on-orbit activities and adjacent activities, the maximum node delay under these constraints is determined using a binary search algorithm, based on the monotonicity of such constraints described in step 4. .

[0021] For the time constraints between the subsystem to which the fault belongs and the unexecuted activities, it is in It satisfies the property of monotonically increasing over the interval, and This is a feasible solution.

[0022] For exclusive resource constraints, any The exclusive resource usage range of the node to which the fault belongs is shifted to the right. At this point, the overall latency exhibits monotonically increasing satisfaction with respect to exclusive resources, and in [0, The optimal time delay solution for the non-faulty node can be obtained by executing the binary search algorithm within the node. .

[0023] Step 6: When all time constraints can no longer be satisfied at the subsystem level, execute the hierarchical neighborhood replanning search algorithm.

[0024] First, we search for the case that minimizes the violation of task-level time constraints using the following formula:

[0025] )

[0026] in, Indicates the preceding In the subsystem and the first Each subsystem is followed by a set of subsystems with time constraints. Indicates the preceding In each subsystem and the first The set of subsystems following each subsystem has time constraints. Indicates up to the The first subsystem Up to the last sub-activity, the time delay between searched activities and subsequent unsearched subsystems is recorded as the time constraint violation degree. , Table 1 The time delay of each activity is At that time, if the activity is In the middle, the status after the latency update. Indicates in In the case of the first Each activity and subsystem before The time constraint violation degree of each task Indicates in In this case, with the subsystem before Constraint violation degree of each subsystem.

[0027] To verify the satisfaction of resource constraints later, it is necessary to record as many possible latency scenarios as possible where the task layer achieves the minimum time constraint violation, that is, to record the latency scenarios where the time constraint violation is minimized. The path.

[0028] Step 7: Given the latency within the subsystem that minimizes the time constraint violation at the task level, update the time constraint violation at the subsystem level according to the following formula:

[0029] )

[0030] in, express The latency situation is as follows , for The time delay state, For the first Minimum time constraint violation at the task layer of each subsystem This represents the total number of activities or tasks within the subsystem. Indicates the first In each subsystem The latency of activities within.

[0031] To verify the final resource constraint satisfaction, it is necessary to record the delay of the minimum constraint violation of the subsystem layer of the previous subsystem when the time constraint violation of the subsystem layer is minimized, as well as the delay of the minimum constraint violation of the current subsystem task layer.

[0032] Step 8: Given the time delay condition that minimizes the time constraint violation degree at the known subsystem level, update the time constraint violation degree at the node level according to the following formula:

[0033]

[0034]

[0035] in, Indicates the preceding text The node with the first The subsystems following each node have a set of time constraints. Indicates the preceding text The node with the first The subsystems following each node have a set of time constraints; express; The latency situation, express The latency status needs to be recorded. To verify the availability of resources, the latency changes also need to be recorded.

[0036] Step 9: Through three-layer dynamic programming (task layer, subsystem layer, and node layer) from steps 6 to 8, obtain the optimal combination of time delays in the multi-node probe landing mission replanning scheme to minimize time constraint violation. Furthermore, the time delay status of the entire system can be inferred from the recorded time delays at the task layer and subsystem layer, thereby verifying whether the overall resource constraints of the probe system are met.

[0037] The queue size for each record is finite, and its value is a hyperparameter related to the system size. During hierarchical neighborhood search, the worst-case scenario is the time delay required for the overall detector system to fail (the time constraint violation is recorded as the maximum acceptable time constraint violation). The time constraint violation in the record list may be lower than this maximum. Therefore, among the delay schemes with lower than this maximum, the one with the smallest violation value that satisfies all resource constraints needs to be selected as the new planning sequence obtained from the final replanning.

[0038] Beneficial effects

[0039] This invention addresses the system constraint conflict problem that may arise during the flexible landing of a multi-node probe due to insufficient understanding of the asteroid environment and potential faults within the probe's nodes and subsystems. It proposes a replanning method based on a hierarchical delay satisfaction verification mechanism and a hierarchical neighborhood replanning search algorithm.

[0040] Compared with existing technologies, the present invention has the following technical advantages:

[0041] 1. Utilizing the hierarchical structure of multi-node detectors, constraint edges are divided into intra-subsystem constraint edges, inter-subsystem constraint edges, and inter-node constraint edges, achieving hierarchical storage at different granularities, thereby constructing a hierarchical constraint graph for the multi-node detector. The hierarchical constraint graph of the multi-node detector is a sparse graph with edge complexity approximately equal to vertex complexity, stored using an adjacency list. The average number of constraint edges related to the state of each subsystem is... The number of vertices is At that time, the overall storage complexity is This can significantly reduce memory usage and storage consumption during the modeling process.

[0042] 2. The designed flexible attachment replanning method for multi-node detectors fully utilizes the characteristics of hierarchical constraints to design related hierarchical verification and search algorithms, thereby reducing the time complexity of the multi-node detector replanning algorithm and improving the planning solution rate of replanning.

[0043] 3. The designed hierarchical time-delay verification mechanism and hierarchical neighborhood replanning search algorithm enhance the flexibility of autonomous replanning solutions. The hierarchical time-delay verification mechanism ensures that time constraints are met while obtaining new planning solutions quickly; the hierarchical neighborhood replanning search algorithm ensures that, when all time constraints cannot be satisfied, a replanning solution with minimal violation of resource and time constraints is found as much as possible. Attached Figure Description

[0044] Figure 1This is a schematic diagram of the hierarchical constraints of the multi-node detector system established in this invention.

[0045] Figure 2 The overall flowchart of the flexible attachment replanning method for multi-node detectors.

[0046] Figure 3 Flowchart for the layered delay verification mechanism.

[0047] Figure 4 Flowchart of the hierarchical neighborhood replanning search algorithm. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments.

[0049] Example

[0050] This invention establishes a replanning problem model for flexible landing of a multi-node detector and creates a hierarchical constraint graph model for the replanning process, such as... Figure 1 As shown, a replanning method based on a hierarchical delay satisfaction verification mechanism and a hierarchical neighborhood replanning search algorithm is proposed and designed, such as... Figure 2 As shown in the diagram. The flowchart of the layered delay satisfaction verification mechanism is as follows. Figure 3 As shown, the hierarchical neighborhood replanning search algorithm is as follows: Figure 4 As shown.

[0051] Step 1: Establish a model for the flexible landing replanning problem of a multi-node probe.

[0052] Step 2: Create a hierarchical constraint graph model for the flexible landing replanning of the multi-node detector, and realize hierarchical storage of constraints at different granularities. Based on the fault information, identify the node and subsystem to which the fault activity belongs.

[0053] Step 3: Within the subsystem to which the fault belongs, consider the constraint satisfaction of the remaining activities. The delay of the faulty activity is... ,exist Within the specified range, a right-first binary verification search algorithm is used to obtain the minimum delay of the faulty subsystem. .

[0054] Step 4: Considering the internal time and resource constraints of the node to which the fault belongs, and based on the constraints of on-orbit activities and adjacent activities of each subsystem, obtain the maximum delay of the subsystem. The delay of the faulty subsystem is ,like The overall delay of the faulty node If all resource constraints can be satisfied at this point, proceed to step 5. Otherwise, if all time constraints cannot be satisfied at the subsystem constraint layer, proceed to steps 6 through 9 of the hierarchical domain replanning search algorithm.

[0055] Step 5: Considering the time and resource constraints of the non-faulty nodes, and combining the exclusive resource characteristics between nodes, in The optimal time delay solution for the non-faulty node can be obtained by performing a binary search algorithm. .

[0056] Step 6: Execute the hierarchical neighborhood replanning search algorithm, searching for the case with the minimum time constraint violation at the task level according to the following formula, and update and record the case with the minimum time constraint violation at the task level. path.

[0057] )

[0058] Step 7: Update the time constraint violation status of the subsystem layer according to the following formula, and update and record the delay status with the minimum time constraint violation status of the subsystem layer.

[0059] )

[0060] Step 8: Update the time constraint violation status of the node layer according to the following formula, and update and record the delay status with the minimum time constraint violation status of the node layer.

[0061]

[0062]

[0063] Step 9: Select cases in the record queue that satisfy the time constraint violation degree below the highest violation degree, reverse reason to obtain a new multi-node detector planning sequence, and verify the resource satisfaction of the multi-node detector. If there are cases where all resource constraints are satisfied, select the scheme with the minimum constraint violation degree as the new replanning scheme. If not, use the overall delay fault subsystem time extension degree with the set highest violation degree as the new replanning scheme.

[0064] In this embodiment, the number of nodes of the multi-node detector is specifically set to 3. The node subsystem includes an imaging system, a lidar system, a GNC system, a navigation system, a guidance system, an attitude system, an internal communication system, a storage system, a thermal control system, a power supply system, a propulsion system, and an external communication system, which can autonomously replan the mission sequence.

Claims

1. A flexible attachment replanning method for multi-node detectors, characterized in that, Includes the following steps: Step 1: Establish a model for the flexible landing replanning problem of a multi-node probe; Step 2: Create a hierarchical constraint graph model for flexible landing replanning of multi-node detectors to achieve hierarchical storage of constraints at different granularities; Based on the fault information, identify the node and subsystem to which the fault activity belongs; Step 3: Within the subsystem to which the fault belongs, consider the constraint satisfaction of the remaining activities. The delay of the faulty activity is... ,exist Within the specified range, a right-first binary verification search algorithm is used to obtain the minimum delay of the faulty subsystem. ; Step 4: Considering the internal time and resource constraints of the node to which the fault belongs, and based on the constraints of on-orbit activities and adjacent activities of each subsystem, obtain the maximum delay of the subsystem. The delay of the faulty subsystem is ,like The overall delay of the faulty node If all resource constraints can be satisfied at this point, proceed to step 5; otherwise, if all time constraints cannot be satisfied at the subsystem constraint layer, proceed to the hierarchical domain replanning search algorithm in steps 6 to 9. Step 5: Considering the time and resource constraints of the non-faulty nodes, and combining the exclusive resource characteristics between nodes, in The optimal time delay solution for the non-faulty node can be obtained by performing a binary search algorithm. ; Step 6: Execute the hierarchical neighborhood replanning search algorithm, searching for the case with the minimum time constraint violation at the task level according to the following formula, and update and record the case with the minimum time constraint violation at the task level. path; First, we search for the case that minimizes the violation of task-level time constraints using the following formula: ) in, Indicates the preceding In the subsystem and the first Each subsystem is followed by a set of subsystems with time constraints. Indicates the preceding In each subsystem and the first The set of subsystems following each subsystem has time constraints. Indicates up to the The first subsystem Up to the last sub-activity, the time delay between searched activities and subsequent unsearched subsystems is recorded as the time constraint violation degree. , Table 1 The time delay of each activity is If the activity is in In the middle, the status after the latency update. Indicates in In the case of the first Each activity and subsystem before The time constraint violation degree of each task Indicates in In this case, with the subsystem before The degree of constraint violation of each subsystem; To verify the satisfaction of resource constraints later, it is necessary to record as many possible latency scenarios as possible at the task layer to achieve the minimum time constraint violation, that is, to record the latency scenarios where the time constraint violation is minimized. The path; Step 7: Given the latency within the subsystem that minimizes the time constraint violation at the task level, update the time constraint violation at the subsystem level according to the following formula: ) in, express The latency situation is as follows , for The time delay state, For the first Minimum time constraint violation at the task layer of each subsystem This represents the total number of activities or tasks within the subsystem. Indicates the first In each subsystem The time delay of the activities in the process; To verify the satisfaction of the final resource constraints, it is necessary to record the time delay of the minimum constraint violation of the subsystem layer of the previous subsystem when the time constraint violation of the subsystem layer is minimized, as well as the time delay of the minimum constraint violation of the current subsystem task layer. Step 8: Update the time constraint violation status of the node layer according to the following formula, and update and record the delay with the minimum time constraint violation status of the node layer: in, Indicates the preceding text The node with the first The subsystems following each node have a set of time constraints. Indicates the preceding text The node with the first The subsystems following each node have a set of time constraints; express; The latency situation, express The latency status; in order to verify the availability of resources, it is also necessary to record the changes in latency; Step 9: Select cases in the record queue that satisfy the time constraint violation degree below the highest violation degree, reverse reason to obtain a new multi-node detector planning sequence, and verify the resource satisfaction of the multi-node detector. If there are cases where all resource constraints are satisfied, select the scheme with the minimum constraint violation degree as the new replanning scheme. If not, use the overall delay fault subsystem time extension degree with the set highest violation degree as the new replanning scheme.

2. The flexible attachment and replanning method for multi-node detectors as described in claim 1, characterized in that, Step 1: Establish the model, as follows: The replanning problem for flexible landing of a multi-node probe can be represented as a triple: , It is a time-ordered set of actions. any element , and These represent the activity start time object and the activity end time object, respectively. ,in , and These are all attributes of the activity, representing the node to which the activity belongs, the subsystem to which it belongs, and the activity number, respectively. This represents the point in time corresponding to the time object. This indicates fault information during the landing process of a multi-node probe; Constraints for multi-node detectors; There are time constraints between the activities of multi-node deep space probes, and these constraints are described by the constraints between time objects related to the activities. Resource constraints for multi-node detectors include continuously consumed resources and exclusive resources. The consumption rate of continuously consumed resources by on-orbit activities should be limited to a certain range, while exclusive resources refer to resources that are only allowed to be used by one node at a time.

3. The flexible attachment replanning method for multi-node detectors as described in claim 1, characterized in that, In step 2, in the multi-node detector system, a hierarchical constraint graph model for flexible landing replanning of the multi-node detector is created, with the active time objects as vertices and the constraints between time objects as constraint edges. Constructing a graphical model of a multi-node detector system G=<V,E> V is the set of points representing the time objects of the activity, and E is the set of time constraint edges; from the points Edge pointing to point j Represented as , , This represents the two time objects corresponding to a given constraint edge. Based on the temporal object attributes at both ends of the edge, the constraint edges are divided into intra-subsystem constraint edges, inter-subsystem constraint edges, and inter-node constraint edges; the constraint edge set of a multi-node detector system is represented as follows: ,in, This represents the internal constraints of the subsystem. Represents the set of constraints between subsystems. Represent the constraint set between nodes; construct a hierarchical constraint graph model of the multi-node detector system based on the constraint edge hierarchy.

4. The flexible attachment replanning method for multi-node detectors as described in claim 1, characterized in that, In step 3, within the subsystem to which fault information W belongs, if all unexecuted activities have the same delay, then when a delay exists... If the constraints of the activity layer of the faulty subsystem are satisfied, then the remaining unexecuted activity delay time is greater than the time. The activity layer constraints are also satisfied; and when the remaining unexecuted activity delay time belongs to the activity delay time of the fault, the constraints of the activity layer of the subsystem to which the fault belongs must be satisfied. Based on the monotonically increasing time-delay property of the active layer of the subsystem to which the fault belongs, a binary search algorithm is used to find the minimum delay that satisfies the internal time constraints of the subsystem. .

5. The flexible attachment replanning method for multi-node detectors as described in claim 1, characterized in that, In step 4, within the node to which fault W belongs, the time constraints of on-orbit activities and adjacent activities of non-fault-related subsystems, the time constraints between remaining unexecuted activities of all subsystems, the time constraints between activities to which fault belongs and unexecuted activities, and the satisfaction of internal resource constraints of the node to which fault belongs are considered. For the constraints between on-orbit activities and adjacent activities of non-fault-related subsystems, when delays exist... When this constraint is satisfied, the adjacency activity delay time is less than This also satisfies such constraints.

6. The flexible attachment replanning method for multi-node detectors as described in claim 1, characterized in that, In step 5, for nodes that are not to which the fault belongs, it is necessary to consider the time constraints of unexecuted activities within the node, the time constraints of on-orbit activities and adjacent activities in the subsystem, the time constraints between unexecuted activities and the subsystem to which the fault belongs, and the constraints of exclusive resources between nodes. For time constraints between on-orbit activities and adjacent activities, the maximum node delay under these constraints is determined using a binary search algorithm, based on the monotonicity of such constraints described in step 4. ; For the time constraints between the subsystem to which the fault belongs and the unexecuted activities, it is in It satisfies the property of monotonically increasing over the interval, and This is a feasible solution; For exclusive resource constraints, any The exclusive resource usage range of the node to which the fault belongs is shifted to the right. At this point, the overall latency exhibits monotonically increasing satisfaction with respect to exclusive resources, and in [0, The optimal time delay solution for the non-faulty node can be obtained by executing the binary search algorithm within the node. .

7. The flexible attachment replanning method for multi-node detectors as described in claim 1, characterized in that, In step 9, the latency status of the entire system is inferred by reverse reasoning through the recorded latency status of the task layer and the latency status recorded at the subsystem layer, thereby verifying whether the overall resource constraints of the detector system are met. The queue size for each record is finite, and its value is a hyperparameter related to the system size. In the worst case of hierarchical neighborhood search, the time length required for the overall delay failure subsystem of the detector system is recorded as the maximum acceptable time constraint violation. The time constraint violation in the record list may be lower than this maximum value. Among the delay schemes with a value lower than this maximum value, the delay scheme with the smallest violation value that satisfies all resource constraints is selected as the new planning sequence obtained by the final replanning.

Images