A method for optimizing an extravehicular servicing configuration of a space manipulator

Abstract

The embodiment of the application provides a space mechanical arm extravehicular care configuration optimization method, and belongs to the field of mechanical arm configuration optimization, and comprises the following steps: according to the space station cabin section care task type, the performance evaluation indexes of the space mechanical arm extravehicular care configuration are constructed as static load capacity, flexibility and collision safety distance; a space mechanical arm extravehicular care configuration optimization model is constructed, the optimization model comprises constraint conditions and decision variables; according to the space mechanical arm extravehicular care configuration optimization model and the performance evaluation indexes, an adapter selection strategy and a benchmark configuration selection strategy are constructed; according to the NSGA-II algorithm and the adapter selection strategy and the benchmark configuration selection strategy, a space mechanical arm configuration optimization strategy is constructed, and an optimal mechanical arm care configuration and an optimal adapter connection scheme are obtained. According to the technical scheme provided by the embodiment of the application, the redundant space mechanical arm can obtain the optimal mechanical arm configuration and the optimal adapter connection scheme under the condition that the end pose is fixed.

Description

[Technical Field]

[0001] This invention belongs to the field of robotic arm configuration optimization, and relates to a method for optimizing the configuration of a space robotic arm for extravehicular care. [Background Technology]

[0002] The space station will have an on-orbit operational lifespan exceeding ten years, during which wear and tear is inevitable, requiring regular maintenance. However, the complex and harsh space environment makes long-term extravehicular activities (EVAs) both detrimental to astronauts' health and technically challenging. Therefore, robotic arms are more suitable for space station module maintenance. Module maintenance encompasses various types of tasks, such as surface condition checks, equipment installation, replacement, or repair, hovering spacecraft capture and docking, assisting astronauts with EVAs, transporting extravehicular loads, and assisting in module repositioning. When a robotic arm performs full-module maintenance tasks, typically only the position of the maintenance point is known, i.e., the position of the robotic arm's end effector is known. Since multiple robotic arm adapters are installed on the surfaces of each space station module, and most space robotic arms are redundant, the robotic arm configuration and adapter selection are not unique when the maintenance point position is fixed. Therefore, optimization research on the extravehicular maintenance configuration of space robotic arms is essential.

[0003] Most existing robotic arm configuration optimization methods optimize the robotic arm configuration based on one or more optimization indices using optimization algorithms or analytical kinematics methods to obtain the optimal configuration. However, the constraint condition for optimizing the configuration of a space robotic arm for extravehicular care is that the end effector pose of the robotic arm remains fixed. Therefore, it belongs to the zero-space configuration optimization problem of the robotic arm, and existing robotic arm configuration optimization methods are not applicable. [Summary of the Invention]

[0004] In view of this, the present invention provides a method for optimizing the configuration of a space robotic arm for external care, so as to obtain the optimal robotic arm configuration and the optimal adapter connection scheme when the end-effector pose is fixed.

[0005] This invention provides a method for optimizing the configuration of an extravehicular care system for a space robotic arm, comprising:

[0006] Based on the type of space station module care mission, the performance evaluation indicators for the space robotic arm external care configuration are static load capacity, flexibility, and collision safety distance.

[0007] An optimization model for the extravehicular care configuration of a space robotic arm is constructed, and the optimization model includes constraints and decision variables.

[0008] Based on the aforementioned space robotic arm extravehicular care configuration optimization model and performance evaluation indicators, an adapter selection strategy and a baseline configuration selection strategy are constructed.

[0009] Based on the NSGA-II algorithm and the adapter selection strategy and baseline configuration selection strategy, a space robotic arm configuration optimization strategy is constructed to obtain the optimal robotic arm extravehicular care configuration and the optimal adapter connection scheme.

[0010] In the above method, the performance evaluation indicators for constructing the space robotic arm's external care configuration based on the space station module care mission type include static load capacity, flexibility, and collision safety distance, including:

[0011] (1) Static load capacity

[0012] Define the force F acting on the end effector of the robotic arm. n and torque F n Composed of vectors

[0013]

[0014] This is called the end-effector generalized force vector. It defines the vector formed by the driving forces of each joint of the robotic arm.

[0015]

[0016] This is called the joint torque vector. Based on the principle of virtual work, it is known that the sum of the virtual work generated by any virtual displacement of the robotic arm under stable conditions is 0. Therefore, the virtual work generated by virtual displacement in joint space is equal to the virtual work generated by virtual displacement in Cartesian space, i.e.

[0017] τ τ ·δq=F T ·D

[0018] Here, δq represents the virtual displacement in joint space, and D represents the virtual displacement in Cartesian space. These two are not unrelated; they are subject to geometric constraints. These geometric constraints are determined by the Jacobian of the robotic arm, i.e.

[0019] D=Jδq

[0020] Substituting, we can obtain

[0021] τ=J T F

[0022] F = (J T ) -1 τ

[0023] J in the formula is the Jacobian matrix of the robotic arm.

[0024] Therefore, the first optimization objective is:

[0025]

[0026] (2) Flexibility

[0027] Operability can quantitatively assess the motion flexibility of a robotic arm, and it is defined as:

[0028]

[0029] In the formula: ω is the operability metric, det() is the determinant of the square matrix, J(q) is the Jacobian matrix, J T (q) is the transpose of the Jacobian matrix, λ i For matrix J(q)J T The eigenvalues ​​of (q) are λ1≥λ2≥...≥λ m Let i = 1, 2, ..., m, where m is a matrix J(q). T Number of rows of (q), σ i Let J(q) be the singular value of the Jacobian matrix. The operability index ω is between [0, 1]. When ω = 0, the robotic arm is in a singular state. The larger the value of ω, the better the flexibility.

[0030] For a given robotic arm, once the structural parameters are determined, only changes in the joint angles of the robotic arm will affect its dexterity. Therefore, the second optimization objective is:

[0031] g2 = max(ω1, ω2, ..., ω) n )

[0032] Where n is the number of all spatial robotic arm configurations corresponding to the care point pose.

[0033] (3) Collision safety distance

[0034] The collision detection algorithm uses the detection algorithm in the FCL library. The FCL library is an open source collision detection library. Its core idea is to process the object model to be detected into a bounding box shape, and then calculate whether there is a collision between the bounding boxes and find the collision distance.

[0035] The formula for calculating the collision distance in the FCL library is shown below, where A and B are two bounding box sets, and dist(A, B) is the shortest distance between A and B.

[0036] dist(A,B)=min{||xy||:x∈A,y∈B}

[0037] Optimization objective three is:

[0038] g1=max{dist(i, j), 1≤i≤8, 1≤j≤m}

[0039] Where dist(i,j) is the minimum safe collision distance between each link of the robotic arm and the surrounding environment in the extravehicular care configuration; i is the number of robotic arm link models, and j is the number of models in the environment surrounding the robotic arm, with a maximum of m.

[0040] In the above method, the construction of the space robotic arm extravehicular care configuration optimization model includes constraints and decision variables, including:

[0041] (1) Constraints

[0042] Constraint 1 requires the care point to be reachable. The prerequisite for optimizing the robotic arm's external care configuration is that the care point pose is fixed, meaning the robotic arm's end effector pose is fixed. Therefore, each robotic arm configuration participating in the optimization algorithm must satisfy the condition that the end effector pose equals the care point pose, i.e., configuration optimization is performed in the zero-space environment of the robotic arm. Zero-space configuration optimization is as follows... Figure 2 As shown.

[0043] The second constraint is that the robotic arm does not collide with the surrounding environment, i.e., Dist(i,j)>0.

[0044] Constraint 3 is the limit of the robot arm joint angle, where q i Let θ be the angle of each joint of the robotic arm. min θ is the minimum designed operating angle for the robotic arm joints. max This is the maximum designed operating angle for the robotic arm joints.

[0045] θ min ≤q i ≤θ max i = 1, 2, ..., n

[0046] (2) Decision variables

[0047] The joint angular velocity in zero space satisfies Therefore, we consider using φ as a decision variable, and perform selection, crossover, and mutation operations on φ using a genetic algorithm. Then, we decode φ to convert it into a joint angle q, and perform optimization operations such as fitness function calculation and non-dominated sorting based on q. The specific decoding method is as follows:

[0048] Step 1: Based on Convert φ to joint angular velocity

[0049] Step 2: Set joint angular velocity Integrate the joint angular velocity Converted to joint angle q.

[0050] In the above method, the step of constructing an adapter selection strategy and a baseline configuration selection strategy based on the space robotic arm's extravehicular care configuration optimization model and performance evaluation indicators includes:

[0051] (1) Adapter selection strategy

[0052] Step 1: Initial adapter selection strategy based on workspace analysis.

[0053] Calculate the Cartesian space distance between the care point pose and each adapter, denoted as dist. i If dist i <r, then endPE i The algorithm stores the data and performs a secondary selection in the subsequent fine-tuning of adapter selection; if dist i If the value is greater than r, it means the adapter is not within the workspace and should be discarded, where r is the workspace radius of the robotic arm. i This represents the pose of the i-th adapter on the space station.

[0054] Step 2: Fine-tuning the adapter selection strategy.

[0055] Set the end effector pose of the robotic arm to be equal to the pose of the care point, and the base pose of the robotic arm to be equal to the pose of the adapter obtained from the initial adapter selection strategy. Based on the inverse kinematics of the robotic arm, calculate the inverse kinematics algorithm result that satisfies the above conditions. Theoretically, the inverse kinematics algorithm must have a feasible solution. Therefore, collision detection is performed on all feasible solutions. If there is a set of collision-free feasible solutions, the adapter can be used as a usable adapter near the care point; if there is no collision-free feasible solution, the adapter is an unusable adapter.

[0056] (2) Baseline Configuration Selection Strategy

[0057] The baseline configuration is determined based on the position-level inverse kinematics algorithm of the robotic arm. The joint angle q1 is preset in the position-level inverse kinematics algorithm. q1 is iterated between 0° and 270°, increasing by x° each time, so that the baseline configuration can be traversed within the range of joint angle values, thus avoiding the configuration optimization algorithm from getting stuck in local optima.

[0058] In the above method, the step of constructing a space manipulator configuration optimization strategy based on the NSGA-II algorithm, the adapter selection strategy, and the baseline configuration selection strategy to obtain the optimal manipulator extravehicular care configuration and the optimal adapter connection scheme includes:

[0059] Step 1: The pose of the care point is the input variable of the algorithm. Based on the adapter selection strategy, the adapter information that satisfies the reachability of the care point is obtained.

[0060] Step 2: If the number of available adapters is not 0, analyze each adapter in turn and use the baseline configuration selection strategy to obtain the baseline configuration corresponding to the current adapter.

[0061] Step 3: Using the NSGA-II optimization algorithm, the decision variables are transformed into joint angles based on the baseline configuration, and optimization indicators such as fitness are analyzed to obtain the optimized configuration.

[0062] Step 4: To avoid the optimization algorithm getting stuck in local optima, the position-level inverse solution is iterated on the baseline configuration. The above optimization algorithm process (Step 3) is repeated. After multiple iterations, the optimal configuration of the current adapter is obtained.

[0063] Step 5: Each adapter corresponds to an optimal external care configuration. Select the optimal external care configuration from among them, and the corresponding adapter is the optimal adapter. [Attached Image Description]

[0064] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort or labor.

[0065] Figure 1 This is a flowchart illustrating the method for optimizing the extravehicular care configuration of a space robotic arm provided in an embodiment of the present invention.

[0066] Figure 2 This is a schematic diagram of zero-space configuration optimization in an embodiment of the present invention;

[0067] Figure 3 This is a flowchart of the robotic arm configuration optimization strategy in an embodiment of the present invention;

[0068] Figure 4 This is the convergence graph of the optimization objective after seqCount iterations of the NSGA-II algorithm in the simulation experiment of this embodiment of the invention;

[0069] Figure 5 This is the convergence graph of objective 2 after seqCount iterations of the NSGA-II algorithm in the simulation experiment of this embodiment of the invention;

[0070] Figure 6 This is the convergence graph of the optimization objective after seqCount iterations of the NSGA-II algorithm in the simulation experiment of this embodiment of the invention;

[0071] Figure 7 This is the total weight convergence graph after seqCount iterations of the NSGA-II algorithm in the simulation experiment of this embodiment of the invention;

[0072] Figure 8 This is the X-off map of the end-effector pose of the robotic arm configuration generated by the NSGA-II algorithm in the simulation experiment of this invention.

[0073] Figure 9 This is the Y-offset map of the end pose of the robotic arm configuration generated by the NSGA-II algorithm in the simulation experiment of this invention.

[0074] Figure 10 This is the Z-off map of the end-effector pose of the robotic arm configuration generated by the NSGA-II algorithm in the simulation experiment of this invention.

[0075] Figure 11 This is the Rz offset map of the end pose of the robotic arm configuration generated by the NSGA-II algorithm in the simulation experiment of this invention.

[0076] Figure 12 This is the Ry offset map of the end pose of the robotic arm configuration generated by the NSGA-II algorithm in the simulation experiment of this invention;

[0077] Figure 13 This is the Rx offset map of the end pose of the robotic arm configuration generated by the NSGA-II algorithm in the simulation experiment of this invention.

[0078] Figure 14 This is an experimental result diagram showing how the configuration optimization algorithm avoids getting trapped in local optima in an embodiment of the present invention. [Specific Implementation Examples]

[0079] To better understand the technical solution of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0080] It should be understood that the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0081] This invention provides an optimization method for the extravehicular activity (EVA) configuration of a space robotic arm for full-module care of a space station. Please refer to [reference needed]. Figure 1 This is a flowchart illustrating the space robotic arm extravehicular care configuration optimization method provided in this invention example, as shown below. Figure 1 As shown, the method includes the following steps:

[0082] Step 101: Based on the type of space station module care mission, construct the performance evaluation indicators for the space robotic arm's external care configuration as static load capacity, flexibility, and collision safety distance.

[0083] Specifically, space station care missions are divided into two main categories: on-orbit capture and assembly missions and on-orbit inspection missions. Supporting astronauts' extravehicular activities (EVAs), module repositioning, maintenance of large external equipment, and cargo handling fall under the category of on-orbit capture and assembly missions. When performing these missions, the robotic arm's load-bearing capacity at the external care configuration needs to be considered. External condition inspection falls under the category of on-orbit inspection missions. Since the robotic arm needs to use the end-effector's vision system to photograph and record the external environment, it must maintain high flexibility, meaning it must be able to change its attitude rapidly and significantly to adapt to different photographing postures. Therefore, this type of mission requires optimizing the robotic arm's flexibility. Simultaneously, the external care configurations for both on-orbit capture and assembly and on-orbit inspection missions must meet the goal of maximizing collision safety distances. This means maximizing the collision safety distance between each component of the robotic arm and other parts of the space station, ensuring that the robotic arm's external care configuration will not collide with the space station to the greatest extent possible.

[0084] Therefore, the performance evaluation indicators for the space robotic arm's external care configuration are static load capacity, flexibility, and collision safety distance.

[0085] (1) Static load capacity

[0086] Define the force F acting on the end effector of the robotic arm. n and torque F n Composed of vectors

[0087]

[0088] This is called the end-effector generalized force vector. It defines the vector formed by the driving forces of each joint of the robotic arm.

[0089]

[0090] This is called the joint torque vector. Based on the principle of virtual work, it is known that the sum of the virtual work generated by any virtual displacement of the robotic arm under stable conditions is 0. Therefore, the virtual work generated by virtual displacement in joint space is equal to the virtual work generated by virtual displacement in Cartesian space, i.e.

[0091] τ τ ·δq=F T ·D

[0092] Here, δq represents the virtual displacement in joint space, and D represents the virtual displacement in Cartesian space. These two are not unrelated; they are subject to geometric constraints. These geometric constraints are determined by the Jacobian of the robotic arm, i.e.

[0093] D=Jδq

[0094] Substituting, we can obtain

[0095] τ=J T F

[0096] F = (J T ) -1 τ

[0097] J in the formula is the Jacobian matrix of the robotic arm.

[0098] Therefore, the first optimization objective is:

[0099]

[0100] (2) Flexibility

[0101] Operability can quantitatively assess the motion flexibility of a robotic arm, and it is defined as:

[0102]

[0103] In the formula: ω is the operability metric, det() is the determinant of the square matrix, J(q) is the Jacobian matrix, J T (q) is the transpose of the Jacobian matrix, λ i For matrix J(q)J T The eigenvalues ​​of (q) are λ1≥λ2≥...≥λ m Let i = 1, 2, ..., m, where m is a matrix J(q). T Number of rows of (q), σ i Let J(q) be the singular value of the Jacobian matrix. The operability index ω is between [0, 1]. When ω = 0, the robotic arm is in a singular state. The larger the value of ω, the better the flexibility.

[0104] For a given robotic arm, once the structural parameters are determined, only changes in the joint angles of the robotic arm will affect its dexterity. Therefore, the second optimization objective is:

[0105] g2 = max(ω1, ω2, ..., ω) n )

[0106] Where n is the number of all spatial robotic arm configurations corresponding to the care point pose.

[0107] (3) Collision safety distance

[0108] The collision detection algorithm uses the detection algorithm in the FCL library. The FCL library is an open source collision detection library. Its core idea is to process the object model to be detected into a bounding box shape, and then calculate whether there is a collision between the bounding boxes and find the collision distance.

[0109] The formula for calculating the collision distance in the FCL library is shown below, where A and B are two bounding box sets, and dist(A, B) is the shortest distance between A and B.

[0110] dist(A,B)=min{||xy||:x∈A,y∈B}

[0111] Therefore, the third optimization objective is:

[0112] g1=max{dist(i, j), 1≤i≤8, 1≤j≤m}

[0113] Where dist(i,j) is the minimum safe collision distance between each link of the robotic arm and the surrounding environment in the extravehicular care configuration; i is the number of robotic arm link models, and j is the number of models in the environment surrounding the robotic arm, with a maximum of m.

[0114] Step 102: Construct an optimization model for the space robotic arm's external care configuration. The optimization model includes constraints and decision variables.

[0115] Specifically, constraint one is that the care point is reachable. The prerequisite for optimizing the robotic arm's external care configuration is that the care point pose is fixed, that is, the robotic arm's end effector pose is fixed. Therefore, each robotic arm configuration participating in the optimization algorithm must satisfy the end effector pose being equal to the care point pose, that is, configuration optimization is carried out in the robotic arm's null space.

[0116] The second constraint is that the robotic arm does not collide with the surrounding environment, that is, to ensure that each configuration participating in the optimization algorithm meets the requirement that the collision distance with the surrounding environment is greater than 0, where Dist(i,j) is the collision distance between each link of the robotic arm and the surrounding environment.

[0117] Dist(i,j)>0

[0118] Constraint 3 is the limit of the robot arm joint angle, where q i Let θ be the angle of each joint of the robotic arm. min θ is the minimum designed operating angle for the robotic arm joints. max This is the maximum designed operating angle for the robotic arm joints.

[0119] θ min ≤q i ≤θ max i = 1, 2, ..., n

[0120] The joint angular velocity in zero space satisfies Therefore, we consider using φ as a decision variable, and perform selection, crossover, and mutation operations on φ using a genetic algorithm. Then, we decode φ to convert it into a joint angle q, and perform optimization operations such as fitness function calculation and non-dominated sorting based on q. The specific decoding method is as follows:

[0121] Step 1: Based on Convert φ to joint angular velocity

[0122] Step 2: Set joint angular velocity Integrate the joint angular velocity Converted to joint angle q;

[0123] Step 103: Based on the space robotic arm extravehicular care configuration optimization model and performance evaluation index, construct the adapter selection strategy and the baseline configuration selection strategy.

[0124] Specifically, the adapter selection strategy consists of two steps:

[0125] Step 1: Initial adapter selection strategy based on workspace analysis.

[0126] Calculate the Cartesian space distance between the care point pose and each adapter, denoted as dist. i If dist i <r, then endPE i The algorithm stores the data and performs a secondary selection in the subsequent fine-tuning of adapter selection; if dist i If the value is greater than r, it means the adapter is not within the workspace and should be discarded, where r is the workspace radius of the robotic arm. i This represents the pose of the i-th adapter on the space station.

[0127] Step 2: Fine-grained adapter selection strategy.

[0128] Set the end effector pose of the robotic arm to be equal to the pose of the care point, and the base pose of the robotic arm to be equal to the pose of the adapter obtained from the initial adapter selection strategy. Based on the inverse kinematics of the robotic arm, calculate the inverse kinematics algorithm result that satisfies the above conditions. Theoretically, the inverse kinematics algorithm must have a feasible solution. Therefore, collision detection is performed on all feasible solutions. If there is a set of collision-free feasible solutions, the adapter can be used as a usable adapter near the care point; if there is no collision-free feasible solution, the adapter is an unusable adapter.

[0129] Baseline configuration selection strategy:

[0130] The baseline configuration is determined based on the position-level inverse kinematics algorithm of the robotic arm. The joint angle is pre-set as q1 in the position-level inverse kinematics algorithm. q1 is iterated between 0° and 270°, and x° is added each time. This allows the baseline configuration to be traversed within the range of joint angle values, thus avoiding the configuration optimization algorithm from getting stuck in local optima.

[0131] Step 104: Based on the NSGA-II algorithm and the adapter selection strategy and baseline configuration selection strategy, construct a space robotic arm configuration optimization strategy to obtain the optimal robotic arm extravehicular care configuration and the optimal adapter connection scheme.

[0132] Specifically, the spatial robotic arm configuration optimization strategy is implemented according to the following steps:

[0133] Step 1: The pose of the care point is the input variable of the algorithm. Based on the adapter selection strategy, the adapter information that satisfies the reachability of the care point is obtained.

[0134] Step 2: If the number of available adapters is not 0, analyze each adapter in turn and use the baseline configuration selection strategy to obtain the baseline configuration corresponding to the current adapter.

[0135] Step 3: Using the NSGA-II optimization algorithm, the decision variables are transformed into joint angles based on the baseline configuration, and optimization indicators such as fitness are analyzed to obtain the optimized configuration.

[0136] Step 4: To avoid the optimization algorithm getting stuck in local optima, the position-level inverse solution is iterated on the baseline configuration. The above optimization algorithm process (Step 3) is repeated. After multiple iterations, the optimal configuration of the current adapter is obtained.

[0137] Step 5: Each adapter corresponds to an optimal external care configuration. Select the optimal external care configuration from among them, and the corresponding adapter is the optimal adapter.

[0138] Space robotic arm configuration optimization strategy, such as Figure 3 As shown.

[0139] Based on the method provided in the embodiments of the present invention, a simulation experiment was conducted on the space robotic arm carried by the core module of my country's space station. With collision safety distance, static load capacity and flexibility as configuration optimization objectives, a numerical simulation experiment based on NSGA-II multi-objective configuration optimization was carried out using C++ language. The simulation experiment settings are shown in Table 1.

[0140] Table 1. Parameter values ​​for simulation experiments

[0141]

[0142] First, the effectiveness of the robotic arm external care configuration optimization method is verified. (From...) Figure 4 , Figure 5 , Figure 6 It can be seen that all three optimization objectives converge after iterating the NSGA-II algorithm seqCount times, and by Figure 7 It can be seen that optimizing the total weights exhibits an upward convergence effect. Since all three optimization objectives result in better performance with larger values, therefore... Figures 4-7 It can be proven that the optimization method can change the configuration of the robotic arm in a direction that results in better overall weight performance, and thus it is effective.

[0143] Next, we verify the validity of the method. Validity means that the method satisfies the constraints. The most important constraint of this method is that the caretaker point is reachable, meaning that the end-effector pose of the robotic arm participating in the optimization algorithm remains unchanged. Figures 8-1 The end-effector position offset error corresponding to the robotic arm configuration generated by the surface optimization algorithm is within 0.001 meters, and the end-effector attitude offset is within 0.1°. This proves that the algorithm satisfies the constraint of reachability of the care point during operation and is legitimate.

[0144] Finally, we verified whether this method can prevent the algorithm from getting trapped in local optima. Figure 14 To avoid the configuration optimization algorithm getting trapped in local optima, the experimental results are plotted, where the horizontal axis represents the number of iterations of the NSGA-II algorithm, and the vertical axis represents the value of the total objective. The NSGA-II algorithm uses the same initial baseline configuration for every seqCount (50) iterations. Therefore, it can be seen that the total objective within each seqCount data point shows an upward trend, while the value of the total objective varies greatly between each seqCount data point. This proves that the method effectively avoids the total objective getting trapped in local optima and can significantly change the value of the optimization objective by adjusting the baseline configuration, thus laying the foundation for obtaining the globally optimal configuration.

[0145] The technical solutions of the embodiments of the present invention have the following beneficial effects:

[0146] Based on the performance evaluation index of the robotic arm's external care configuration obtained from the space station's full-module care mission, constraints and decision variables for the optimal robotic arm external care configuration were constructed. A robotic arm adapter selection strategy and a baseline configuration selection strategy were designed. In addition, an optimal robotic arm external care configuration strategy was designed by combining the NSGA-II algorithm. This enables the redundant space robotic arm to obtain the optimal robotic arm configuration and the optimal adapter connection scheme when the end-effector pose is fixed. This optimization scheme can avoid the algorithm from getting trapped in local optima and has effectiveness and feasibility.

[0147] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0148] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. A method for optimizing the configuration of a space robotic arm for extravehicular care, characterized in that, The method includes: Based on the type of space station module care mission, the performance evaluation indicators for the space robotic arm external care configuration are static load capacity, flexibility, and collision safety distance. An optimization model for the extravehicular care configuration of a space robotic arm is constructed, and the optimization model includes constraints and decision variables. Based on the aforementioned space robotic arm extravehicular care configuration optimization model and performance evaluation indicators, an adapter selection strategy and a baseline configuration selection strategy are constructed. Based on the NSGA-II algorithm and the adapter selection strategy and baseline configuration selection strategy, a space robotic arm configuration optimization strategy is constructed to obtain the optimal robotic arm extravehicular care configuration and the optimal adapter connection method.

2. The method according to claim 1, characterized in that, Based on the type of space station module care mission, the performance evaluation indicators for the space robotic arm's external care configuration are static load capacity, flexibility, and collision safety distance, including: (1) Static load capacity Define the force acting on the end effector of the robotic arm. and torque Composed of vectors The generalized force vector at the end effector; the vector formed by the driving forces of each joint of the robotic arm. Let be the joint torque vector, where These correspond to the angular moments of the seven joints of the robotic arm; based on the principle of virtual work, it is known that the virtual work generated by the virtual displacement in joint space is equal to the virtual work generated by the virtual displacement in Cartesian space, that is... in, This represents virtual displacement in the joint space. For the virtual displacement in Cartesian space, there is a geometric constraint relationship between them, namely... Substituting, we can obtain In the formula It is the Jacobian matrix of the robotic arm; Therefore, the first optimization objective is: in, To assess the static load capacity of different robotic arm configurations corresponding to the position and pose of the care point; (2) Flexibility Operability can quantitatively assess the motion flexibility of a robotic arm, and it is defined as: In the formula: As a metric for operability, The determinant of a square matrix. For Jacobian matrices, This is the transpose of the Jacobian matrix. For matrix eigenvalues, , , For matrix the number of rows, Jacobian matrix The singular values, when At that time, the robotic arm was in a strange state. The higher the value, the better the flexibility; Therefore, the second optimization objective is: in, To enhance the flexibility of different robotic arm configurations corresponding to the position and pose of the care point; (3) Collision safety distance The collision detection algorithm uses the detection algorithm from the FCL library. The formula for calculating the collision distance in the FCL library is shown below. Where A and B are two bounding box sets. The shortest distance between A and B. Two bounding boxes, A and B, are set together as bounding boxes; Optimization objective three is: in, The minimum safe collision distance between each link of the robotic arm in the extravehicular activity (EVA) configuration and the surrounding environment; It is the number of robotic arm component models. This refers to the number of models in the environment surrounding the robotic arm, with a maximum of [number missing]. indivual.

3. The method according to claim 1, characterized in that, The aforementioned construction of an optimization model for the extravehicular care configuration of a space robotic arm includes constraints and decision variables, including: (1) Constraints Constraint 1 is that the care point is reachable; Constraint 2 is that the robotic arm does not collide with the surrounding environment, that is ; Constraint three is the limitation of the robot arm joint angle, among which, For the angle of each joint of the robotic arm, The minimum operating angle for the robotic arm joints, This represents the maximum designed operating angle of the robotic arm joints. (2) Decision variables Joint angular velocities in zero space satisfy Therefore, As decision variables, genetic algorithms are used to... Perform selection, crossover, and mutation operations, and then... Decode the values ​​and convert them into joint angles. Then based on The specific decoding method involves performing fitness function calculation and non-dominated sorting optimization operations as follows: Step 1: Based on Will Converted to joint angular velocity ; Step 2: Set joint angular velocity Integrate the joint angular velocity Converted to joint angle .

4. The method according to claim 1, characterized in that, Based on the space robotic arm's extravehicular care configuration optimization model and performance evaluation indicators, the adapter selection strategy and baseline configuration selection strategy are constructed, including: (1) Adapter selection strategy Step 1: Preliminary adapter selection strategy based on workspace analysis; Calculate the Cartesian space distance between the care point pose and each adapter, denoted as . ;like Then The data is stored in the algorithm for secondary filtering during subsequent adapter fine-tuning; if... If the adapter is not within the workspace, it should be discarded; The radius of the working space of the space robotic arm. For the space station Each adapter corresponds to a pose; Step 2: Fine-tuning the adapter selection strategy; Set the end effector pose of the robotic arm to be equal to the pose of the care point, and the base pose of the robotic arm to be equal to the pose of the adapter obtained by the initial adapter selection strategy. Based on the inverse kinematics of the robotic arm, calculate the inverse kinematics algorithm result that satisfies the above conditions. Perform collision detection on all feasible solutions. If there is a set of collision-free feasible solutions, the adapter can be used as a usable adapter near the care point; if there is no collision-free feasible solution, the adapter is an unusable adapter. (2) Baseline configuration selection strategy The baseline configuration is determined based on the position-level inverse kinematics algorithm of the robotic arm. The joint angles are pre-set in the position-level inverse kinematics algorithm. ,make exist Iterate between them, increasing each time. This allows the baseline configuration to be traversed within the range of joint angle values, thus preventing the configuration optimization algorithm from getting stuck in local optima.

5. The method according to claim 1, characterized in that, Based on the NSGA-II algorithm and the adapter selection strategy and baseline configuration selection strategy, a space robotic arm configuration optimization strategy is constructed to obtain the optimal robotic arm extravehicular care configuration and the optimal adapter connection method, including: Step 1: The pose of the care point is the input variable of the algorithm. Based on the adapter selection strategy, the adapter information that satisfies the reachability of the care point is obtained. Step 2: If the number of available adapters is not 0, analyze each adapter in turn and use the baseline configuration selection strategy to obtain the baseline configuration corresponding to the current adapter. Step 3: Using the NSGA-II optimization algorithm, the decision variables are transformed into joint angles based on the baseline configuration, and optimization indicators such as fitness are analyzed to obtain the optimized configuration. Step 4: To avoid the optimization algorithm getting stuck in local optima, iterate the position-level inverse solution of the baseline configuration. Repeat step 3. After multiple iterations, the optimal configuration of the current adapter is obtained. Step 5: Each adapter corresponds to an optimal external care configuration. Select the optimal external care configuration from among them, and the corresponding adapter is the optimal adapter.

Images