Floating target iteration position prediction method and system for on-orbit satellite service

Abstract

The invention discloses a floating target iterative position prediction method and system for on-orbit satellite service, the system comprises three components: a position estimator, a trajectory planner and a compliance controller, dynamic data of a satellite interface are collected in real time by using a force-torque sensor, and the dynamic data of the satellite interface are calculated by combining with modeling of satellite target behaviors. By iteratively optimizing a prediction result and integrating a compliant control algorithm, oscillation can be effectively reduced during satellite on-orbit service, so that the system can stably converge to target force, the system stability is enhanced, accurate docking of satellite interfaces is realized, the convergence speed is high, docking tasks can be completed more quickly, and the method is suitable for large-scale popularization and application. The method is crucial to the task of on-orbit satellite fuel supplement with high time requirement.

Description

Technical Field

[0001] The present invention belongs to the technical field of on-orbit satellite fuel filling, and particularly relates to a method and system for predicting the iterative position of a floating target for on-orbit satellite services. Background Art

[0002] Satellite on-orbit service is a general term for various maintenance and repair operations carried out in the space Earth orbit, usually involving operations such as fuel filling, fault repair, and replacement of damaged parts; in this scenario, it is particularly important to accurately insert the fuel filling device into the predetermined interface; in the field of robotics, this problem is called the peg-in-hole problem.

[0003] The traditional peg-in-hole problem plays a fundamental role in satellite on-orbit service, and its core lies in accurately inserting the fuel filling device or repair tool into the satellite interface; however, due to the satellite being in a microgravity and free-floating state in the space environment, the position of its interface is usually not fixed and may be displaced or rotated due to contact force, external interference, or its own movement, forming a "floating target"; this dynamic uncertainty greatly increases the complexity of the operation, especially when facing the high-precision and high-reliability satellite docking requirements; the original technology cannot quickly handle the dynamic drift of the satellite interface and has limited effectiveness in static error handling. Summary of the Invention

[0004] To solve the above problems, the present invention discloses a method for predicting the iterative position of a floating target for on-orbit satellite services, which uses a force-torque sensor to collect real-time dynamic data of the satellite interface, combines the modeling of the satellite target behavior, iteratively optimizes the prediction result and incorporates a compliant control algorithm, and can effectively reduce oscillations when processing the shaft-hole assembly task of the floating target for on-orbit satellite fuel replenishment, enabling the system to stably converge to the target force, enhancing the system stability, and thus achieving accurate docking of the satellite interface.

[0005] To achieve the above object, the technical solution of the present invention is as follows:

[0006] A system for predicting the iterative position of a floating target for on-orbit satellite services includes three components: a position estimator, a trajectory planner, and a compliant controller.

[0007] The six-axis force sensor is responsible for measuring the forces and torques acting on the end of the robotic arm and transmitting this force feedback information to the position estimator and the compliance controller. The position estimator uses this force feedback information and possible visual or encoder feedback to calculate the accurate current position and orientation of the robotic arm. The trajectory planner generates a smooth motion trajectory from the current position to the target position based on the externally input target position and the information provided by the position estimator. The compliance controller receives the target trajectory generated by the trajectory planner and the actual position data from the position estimator, calculates the pose error, and uses a specific algorithm based on this error to determine the forces or torques applied to the robotic arm joints to achieve the desired compliant behavior. The control torques output by the compliance controller are sent to the servo drives of the robotic arm, which convert these torques into electrical signals to drive the motors to rotate the corresponding joints, causing the robotic arm to move along the predetermined trajectory and with the predetermined force. Throughout this process, the position estimator continuously updates the actual position information of the robotic arm and compares it with the target trajectory generated by the trajectory planner. Once a deviation is detected, the compliance controller immediately adjusts the output torque to ensure that the robotic arm can accurately track the predetermined trajectory until the task is completed.

[0008] Its usage method is as follows:

[0009] First, the position estimator estimates the target displacement d based on the force feedback data f from the six-axis force sensor; during the estimation process, by repeatedly reading the force f i , iteratively update and calculate the offset d i , and then append the target offset d obtained after accumulating the offsets to the initial position p estimated a priori, and the position under the action of the contact force can be estimated. During the mapping process from the contact force to the target displacement, directly using a proportional function for mapping will cause problems such as oscillation and overshoot; to solve the problems (such as oscillation, overshoot, etc.) brought by the above simple proportional mapping model, the present invention proposes two strategies: the non-linear error mapping function and the anchor point limiter.

[0010] In the non-linear error mapping function, a non-linear mapping g(f δ ) is introduced to suppress the influence of noise near the sensor zero point on the target position estimation. g(f δ ) is obtained by interpolating the key points of a piecewise linear mapping g 1 , k 2 , k 3 , f l and f h defined by parameters k 0 (f δ ) with a piecewise cubic Hermite polynomial.

[0011] g 0 (f δ) The expression is as follows:

[0012]

[0013] In the anchor limiter strategy, the displacement is constrained within a cubic region centered at the anchor point with side length r in the i-th iteration, so as to prevent overshoot of the position caused by excessive displacement. At the same time, the anchor point is also adjusted in each iteration, so as to be able to more flexibly and adaptively estimate the floating target:

[0014]

[0015] After completing the target prediction, for the predicted target position, a trajectory planner is used to calculate the trajectory in the Cartesian space. Then, the compliant controller continuously monitors the force and position states of the end effector, moves towards the currently estimated target position according to the planned trajectory, and avoids hard contact based on the end-effector compliance in the Cartesian space during contact, thereby further ensuring the stable insertion of the fuel filler.

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

[0017] (1) The present invention proposes an iterative target position prediction strategy. Through continuous iterative optimization, it overcomes the static error caused by the floating target and significantly improves the prediction accuracy of the target position.

[0018] (2) The present invention uses a non-linear error mapping. By introducing a piecewise linear mapping, when dealing with the shaft-hole assembly task of floating targets for in-orbit satellite fuel replenishment, it can effectively reduce oscillations, enable the system to stably converge to the target force, and enhance the system stability.

[0019] (3) The present invention introduces an anchor limiter. Through displacement constraint, it effectively reduces the overshoot and static error.

[0020] (4) The present invention performs excellently in terms of convergence speed and can complete the shaft-hole assembly task more quickly, which is crucial for tasks with high time requirements such as in-orbit satellite fuel replenishment. Description of the Drawings

[0021] Figure 1 is the structural diagram of the control strategy in the embodiment of the present invention.

[0022] Figure 2 is the schematic diagram of the iterative strategy for target position estimation in the embodiment of the present invention.

[0023] Figure 3 is the graph of the non-linear error mapping function in the embodiment of the present invention.

[0024] Figure 4Schematic diagram of the iterative update process of the position and anchor point after the anchor point limiter is introduced in an embodiment of the present invention.

[0025] Figure 5 It is the configuration of the assembly system in the embodiment of the present invention.

[0026] Figure 6 It is a schematic diagram of the results of the system experiment in the embodiment of the present invention. DETAILED DESCRIPTION

[0027] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention.

[0028] Figure 1 is the control strategy structure diagram of the present invention. The whole system consists of three components: position estimator, trajectory planner and compliance controller. The force feedback from the six-dimensional force sensor estimates the displacement of the hole and updates the pre-planned position to obtain the corrected target position ph. Based on this, the trajectory planner adjusts the path of the manipulator to obtain the corrected trajectory; then, the compliance controller continuously monitors the state of the manipulator and moves to the currently estimated target position following the optimal trajectory to ensure stable insertion of the nail.

[0029] Figure 2 : is a schematic diagram of the target position estimation iteration strategy in an embodiment of the present invention. The contact force data from the six-dimensional force sensor is used to iteratively estimate and accumulate the target displacement caused by the applied contact force. Since the contact force is sensed within the frame of the end effector, the mapping between displacement and force is also calculated in this frame; if a simple linear mapping is used, it will bring large static errors and overshoots, so nonlinear error mapping and anchor limiter are used for mapping. Finally, the target position is calculated through continuous updating and iteration.

[0030] Figure 3 is a nonlinear error mapping function diagram in an embodiment of the present invention. The present invention introduces a piecewise linear mapping, and different parameters k1, k2, and k3 control the slope of the linear segment. Using this piecewise linear formula, if the error is close to zero, the sensor noise effect can be suppressed to be close to zero to reduce oscillation; on the contrary, if the error is too large, a smaller slope k3 is used to suppress the gain to prevent excessive displacement estimation; in order to improve the smoothness of the mapping, a piecewise cubic Hermite polynomial is used to interpolate the waypoints on the piecewise linear mapping, thereby producing a smoother mapping, such as Figure 3 As shown by the grey dashed line in .

[0031] Figure 4It is a schematic diagram of the iterative update process of the position and the anchor point after introducing the anchor point limiter in the embodiments of the present invention. Even when using non-linear mapping, if the error is large, the estimated displacement may still be large. To further limit excessive movement, the displacement at the i-th step is restricted within a predefined cubic region centered on ph. This mechanism can prevent excessive movement that may cause the final estimate to deviate significantly from the initial estimate ph, thereby effectively reducing position overshoot. Although the static anchor point limiter effectively constrains excessive displacement, it may hinder the identification of the optimal insertion point when the trajectory is highly inaccurate or the insertion motion is unpredictable. To overcome this limitation, dynamic adjustment of the anchor point is introduced, enabling a more flexible and adaptive estimation of the floating target. Specifically, a1 is defined as the anchor point, and ph is replaced with the center of the cube. Initially, a1 = ph is set, and during the estimation process, it is continuously updated with a fixed step size s. Figure 4 The iterative update of the shown anchor point enables the nail to gradually reach the target position while ensuring that the displacement at each step remains controlled.

[0032] Figure 5 It is a configuration diagram of the assembly system in the embodiments of the present invention. The ATI Mini45 FT six-axis force sensor is installed on the end effector of the JakaZu12 robotic manipulator. To simulate the floating target platform, two cylindrical rods are inserted into a box, and at the same time, two parallel rods restrict the translational movement of the box along its direction, preventing it from moving perpendicular to the rods. Due to the circular through-holes, the box can slide within a certain range along the direction of the rods and may undergo slight rotation between the rods, with the rotation angle limited by the difference between the rod diameter and the through-hole.

[0033] Figure 6 It is a schematic diagram of the results of the system experiment in the embodiments of the present invention. The experimental process is to compare three methods with the solution of the present invention. The three methods are: force control (FC), direct compliance control (DCC), and artificial potential field (APF) strategy. In terms of the convergence speed, FC requires 16.21 seconds, DCC requires 6.92 seconds, APF requires 9.18 seconds, while the method of the present invention converges in only 5.34 seconds, showing faster performance. In terms of stability, when using all the above methods, the target axial force does not show large oscillations. However, regarding the radial force, the three baselines show different degrees of oscillation, while the method of the present invention still maintains the radial force stable.

[0034] The present invention shows obvious advantages in terms of performance compared with other methods. First, in terms of the convergence speed, the present invention is significantly superior to the existing methods, showing higher efficiency. Second, in terms of stability, the present invention can not only effectively maintain the smoothness of the target axial force, but also show excellent stability in the radial force control, avoiding the common oscillation problems in other methods. These characteristics fully demonstrate the unique advantages of the present invention in improving the control performance and optimizing the system response.

Claims

1. An iterative position prediction system for floating targets served by an on-orbit satellite, characterized in that: It includes a position estimator, a trajectory planner and a compliance controller. The six-dimensional force sensor is responsible for measuring the force and torque on the end of the robot arm, and transmits these force feedback information to the position estimator and the compliance controller. The position estimator uses these force feedback information and possible visual or encoder feedback to calculate the current accurate position and posture of the robot arm. The trajectory planner generates a smooth motion trajectory from the current position to the target position based on the external input target position and the information provided by the position estimator. The compliance controller receives the target trajectory generated by the trajectory planner and the actual position data of the position estimator, calculates the posture error, and uses a specific algorithm to determine the force or torque applied to the joint of the robot arm based on this error to achieve the established compliant behavior. The control torque output by the compliance controller is sent to the servo driver of the robot arm, which converts these torques into electrical signals, drives the motor to rotate the corresponding joints, and makes the robot arm move according to the predetermined trajectory and force. During the entire process, the position estimator continuously updates the actual position information of the robot arm and compares it with the target trajectory generated by the trajectory planner. Once a deviation is found, the compliant controller immediately adjusts the output torque to ensure that the robot arm can accurately track the predetermined trajectory until the task is completed.

2. The method for using the floating target iterative position prediction system for on-orbit satellite service according to claim 1, characterized in that: First, the six-dimensional force sensor is installed on the end effector of the robot manipulator. The position estimator estimates the target displacement d based on the force feedback data f of the six-dimensional force sensor. In the estimation process, the force f is read multiple times. i , iteratively update and calculate the offset d i , and then add the target offset d obtained after the offset accumulation to the initial position p estimated a priori, and the position under the contact force can be estimated After completing the target prediction, the trajectory planner calculates the trajectory in Cartesian space for the predicted target position; then, the compliant controller continuously monitors the force and position state of the end and moves to the current estimated target position according to the planned trajectory.

3. The method of use according to claim 2, characterized in that: In the process of mapping contact force to target displacement, two strategies are used: nonlinear error mapping function and anchor point limiter: In the nonlinear error mapping function, the nonlinear mapping g(f δ ) to suppress the influence of noise near the sensor zero point on the target position estimation; g(f δ ) is composed of a number of parameters k1, k2, k3, f l and f h The piecewise linear map g0(f δ ) is obtained by interpolating its key points with piecewise cubic Hermite polynomials; g0(f δ ) is as follows: In the anchor limiter strategy, the displacement is constrained to the anchor point at the i-th iteration. The center is defined in a cube area with a side length of r, so as to prevent over-adjustment of the position caused by excessive displacement and at the same time Adjustments are also made in each iteration to enable a more flexible and adaptive estimation of the floating target:

Images