CN113665841B - Method for testing steering wheel operation of aircraft cockpit based on cooperative robot - Google Patents

Abstract

The invention discloses an aircraft cockpit steering wheel manipulation test method based on a collaborative robot. The collaborative robot is used to replace a human to enter the cockpit for operation, and the robot's own sensor is used to measure the manipulation curve. The test device includes a collaborative robot, a two-finger gripper, a raised platform, and an aircraft cockpit. The two-fingered gripper is installed at the end of the collaborative robot. On the elevated platform, adjust the accessibility of its workspace to various operations in the cockpit; under the instructions of the operator, the robot completes the calibration of the coordinate system in the collaborative robot cabin, the programming movement of the collaborative robot, The steering wheel manipulation force and displacement data are collected, and finally the drawing of the external robot manipulation curve is completed. Through the conversion relationship of the curve, the selection of the impedance mode coefficient and the comparison with the internal sensor curve, the installation deviation of the steering wheel is obtained.

Description

Test method of aircraft cockpit steering wheel manipulation based on collaborative robot

technical field

The invention belongs to the application field of collaborative robots, and in particular relates to a method for using a robot to complete the operation of an aircraft cockpit and to realize the steering wheel manipulation test of the aircraft cockpit through a torque sensor and an angle sensor inside the robot.

Background technique

The aircraft assembly test stage needs to test the displacement and force characteristics of the aircraft driving and control components (such as steering wheel, steering column and pedals, etc.). The displacement and force sensor data on the aircraft driving and control parts, and the external manipulation data should theoretically also be measured and obtained by an external measuring device with displacement and force sensors, and then compared to see if it is within the tolerance range. However, the current domestic testing situation is that testers use experience perception, or use tools such as spring scales and rulers to detect limit position values. They have also tried to develop external measuring devices, but they have not been popularized and applied due to inconvenient use.

With the advancement of robot technology and the transformation of manufacturing models, collaborative robots have attracted more and more attention from all over the world in recent years. According to the definition in ISO 10218-2, a collaborative robot refers to a robot that can directly interact with humans in a designated collaborative area. At present, typical collaborative robots include KUKA iiwa, ABB Yumi, Sawyer, Baxter, Franka, etc. Compared with traditional industrial robots, collaborative robots have the advantages of high safety, good versatility, sensitivity, precision, ease of use, and human-machine collaboration. The above advantages make collaborative robots not only applied in the manufacturing field, but also have potential application value in the fields of home service and rehabilitation medicine.

Since the collaborative robot is equipped with torque sensors on each joint, it can sense the external situation, so it can perform feedback control based on the measured value of the joint torque, and can detect the contact force with the outside world in real time, so as to achieve emergency stop in dangerous situations. The use of collaborative robots, its high precision can ensure the reliability of the test and achieve the consistency of different aircraft tests. The 7-joint redundancy can adjust its own attitude according to the actual obstacles in the aircraft cabin, so as to complete the path planning in a narrow space.

Therefore, the present invention uses collaborative robots instead of testers to carry out aircraft driving and manipulation. Taking the steering wheel as an example, the displacement and force sensors of the robot itself are used to generate the external displacement and force manipulation curves of the driving and manipulation components.

Contents of the invention

The purpose of the present invention is to solve the problem that it is difficult to manually obtain the displacement/force curve of the aircraft driving control from the outside during the aircraft assembly test process. Its method of manipulating the curve.

The present invention adopts the following technical scheme, the aircraft cockpit steering wheel manipulation test method based on the collaborative robot, including the collaborative robot, the two-finger gripper, the raised platform, the aircraft cockpit, the two-fingered gripper is installed at the end of the collaborative robot, and the raised platform It is fixedly connected to the pilot's seat in the main cockpit of the aircraft. The seat is not installed in the final assembly stage. The seat is installed on the floor of the cockpit to fix the raised platform. The base of the collaborative robot is fixed on the raised platform. Adjust its working space for Accessibility of various operations in the cockpit.

Further, the collaborative robot is a 7-axis robot, and the impedance mode can generate a virtual spring at the end to solve the problem of suffocation on the rigid trajectory of the robot turning the steering wheel.

Further, torque sensors are installed on the seven axes of the collaborative robot, which can detect the torque of each axis, and calculate the force at the end through the Jacobian matrix.

Further, in order not to damage the ground of the aircraft cockpit, the mounting holes of the raised platform should be designed to match the seat mounting holes.

Further, in order to ensure the stability of the gripper holding the steering wheel, the fingers of the gripper are designed to wrap around the steering wheel extension bar in an arc-shaped structure.

The present invention provides a method for calibrating the collaborative robot test device for the above-mentioned aircraft automation test, the steps of which are as follows: calibration of the coordinate system in the cabin; programming motion of the collaborative robot; conversion relationship of the steering wheel manipulation curve; selection of impedance mode coefficients and external robot , Internal sensor curve comparison.

The calibration of the coordinate system in the cabin refers to taking the center of the base of the robot as the origin, the forward direction of the aircraft as the x-axis, the direction of the wing as the y-axis, and the vertical direction as the z-axis to establish a base coordinate system, and then rotate with the steering wheel The center establishes the local coordinate system. The specific method of calibration is to install an industrial 3D camera on the beam of the cabin, and obtain its 3D point cloud by shooting the environment in the cockpit, so as to realize the measurement of the spatial distance.

The collaborative robot programming motion refers to: through the previously established base coordinate system, local coordinate system, and the set position coordinates of the steering wheel, the designated point can be reached through PTP (Point-To-Point) motion, taking into account Avoid collisions and precise docking between the robot and the steering wheel tooling, adopt compliant control near the target point, and achieve compliant docking through the preset impedance coefficient.

The conversion relationship of the steering wheel steering curve refers to: for the measurement of the steering wheel rotation angle, it is first necessary to calculate the real-time Cartesian position of the end through the rotation angle of each axis, and then obtain the steering wheel angle by inverting the trigonometric function for the center of the circle; for the torque of the steering wheel, it is necessary to combine the robot The terminal attitude calculates the current Jacobian matrix, then calculates the force of the terminal according to the torque of each axis, and finally calculates the torque of the steering wheel according to the torque at the actual position and the force of the terminal.

The selection of the impedance mode coefficient refers to: for the trajectory of the programming movement and the comparison of the predetermined arc trajectory of the steering wheel, the precision error brought by the impedance mode and the radial force causing damage to the mechanism are calculated, and the precision error and the radial force are calculated. It is synthesized into an evaluation index, and the most suitable impedance mode stiffness coefficient k is obtained by solving the index minimum.

The comparison of the external robot and internal sensor curves refers to: by reading the steering wheel steering curve measured by the aircraft internal sensor and the ratio of the steering wheel torque and the rotation angle measured by the robot, the curves of the two are respectively made in matlab, and observed The difference between the two is to adjust the steering wheel.

The beneficial effect that the present invention has:

1) Compared with the traditional manual testing method that consumes a lot of manpower, the present invention uses a robot to enter the cockpit to simulate driving, which greatly improves the testing efficiency. At present, aircraft digital manufacturing and digital assembly have been realized, but after the installation of structural parts and finished products, the test of the whole machine system still stays in the state of discrete and manual testing. The collaborative robot testing method provided by the present invention not only It can obtain system closed-loop data, improve test efficiency and accuracy, increase test coverage, and accurately analyze data and test conclusions.

2) The present invention improves the objectivity and repeatability of the test. Through robot programming, it can ensure that the operating force and displacement are basically unchanged each time. Compared with manual operation, its repeatability is better. This is illustrated by the comparison of the steering curves produced by the robot turning the steering wheel experiments with the curves of the internal sensors.

Description of drawings

Fig. 1: The structural diagram of the aircraft simulation cockpit of the application environment of the present invention.

Figure 2: Robot-assisted operating system employed by the present invention.

Figure 3: Finger grip design used in the present invention.

Figure 4: The invention is programmed to achieve rhythmic motion.

Fig. 5: Analysis of the steering wheel operation process of the present invention.

Figure 6: Impedance mode force analysis of the present invention.

Fig. 7: Flowchart of robot-assisted detection adopted by the present invention.

Fig. 8: The steering curve measured by the internal sensor of the present invention.

Fig. 9: the no-load maneuvering curve measured by the robot in the present invention.

Fig. 10: The load manipulation curve measured by the robot in the present invention.

Fig. 11: The pure steering wheel manipulation curve measured by the robot in the present invention.

Figure 12: Comparison of the manipulation curves obtained by the internal sensor of the present invention and the robot.

Figure 13: Comparison of the programming curve and the actual displacement curve of the present invention.

Description of numbers in the figure: 1-main steering wheel; 2-main steering column; 3-main pedals; 4-central large screen; 5-button; 51-knob; 52-dial button; ; 7-copilot column; 8-pair of pedals; 9-KUKA robot; 10-handle; 11-raised platform.

Detailed ways

The present invention will be described in further detail below through the accompanying drawings and specific embodiments. The following examples are only descriptive and not restrictive, and cannot be used to limit the protection scope of the present invention.

As shown in Figures 1 and 2, the collaborative robot test device based on aircraft automation testing, the structure of the main pilot and the co-pilot is basically similar, the main steering wheel 1, the main steering column 2, the main pedal 3 are located on the left side, the auxiliary steering wheel 6, the auxiliary Steering column 7, auxiliary pedal 8 are positioned at the right side, and the current status information of the aircraft is displayed on the central large screen 4, and the button area 5 includes various buttons such as knob 51, toggle button 52, ship type switch 53, and the like. The gripper 10 is installed at the end of the KUKA robot 9 and connected by screws. The raised platform 11 is fixed to a certain place in the main cockpit of the aircraft, and the KUKA robot 9 is fixed to the raised platform 11. Accessibility of each operation.

In this embodiment, the test object is the internal sensor calibration of the main steering wheel in the cockpit of a certain type of aircraft. As shown in Figure 2, the steps of the internal sensor calibration method in this embodiment are as follows: calibration of the coordinate system in the cabin; programming motion of the collaborative robot; The conversion relationship of the steering wheel steering curve; the selection of impedance mode coefficients and the comparison of external robot and internal sensor curves.

1. Calibration of the coordinate system in the cabin

Install an industrial camera on the beam behind the cockpit to measure the absolute coordinates of each component in the camera coordinate system. Then measure the conversion relationship between the camera coordinate system and the robot base coordinate system, so as to obtain the coordinates of each component in the base coordinate system. The base coordinate system takes the center of the base of the KUKA robot 9 as the origin, and the forward direction of the aircraft is the x-axis , the wing direction is the y-axis, and the vertical direction is the z-axis to establish the base frame, where: the steering wheel rotation center is located at the (880, 169, 795) position of the base frame, and the steering wheel clamping point is located at the base frame (880, -90, 885) position.

2. Programmed movement of collaborative robots

In order to avoid the collision between the gripper and the steering wheel and suffocation during motion caused by the positioning error of the KUKA robot 9, a Cartesian impedance controller is used for programming, and the setStiffness(...) function is used to set the impedance value of the xyz axis. Under impedance control, KUKA The behavior of robot 9 is obedient, and the steering wheel will not be suffocated due to motion errors. During the movement, pay attention to the rhythmic nature of the work (as shown in Figure 4), move the KUKA robot 9 from the initial position PTP to the vicinity of the steering wheel, LIN to the steering wheel, the jaws to close, the arc to rotate, the jaws to open, LIN to move out and PTP Move to the starting point and program it into motion batch.

KUKA robot 9 works in the human-machine cooperation mode throughout the whole process. KUKA robot 9 is equipped with torque sensors at 7 joints to detect the force of each joint in real time. When KUKA robot 9 collides with objects in the cockpit or someone enters the driving When in the cabin, it can stop the movement to protect the safety of KUKA robot 9 and people. By reading the torque value of each joint and specifying a sensitivity coefficient, if the torque of each joint is between the normal torque minus the sensitivity coefficient and the normal torque plus the sensitivity coefficient, then the KUKA robot 9 operates normally, as long as one torque exceeds the limit range, the KUKA robot 9 will stop. Experiments can react to external influences such as obstacles or process forces. In the design of the gripper finger, a stable connection has been found in combination with the steering wheel tooling. After many tests, the gripper at the end of the KUKA robot 9 can grasp the steering wheel. In terms of operation, the trajectory of the steering wheel is arc-shaped, and the CIRC function is used to program according to the coordinates of the center of the circle and the radius.

3. Conversion relationship of steering wheel control curve

The collaborative robot is equipped with a displacement sensor and a torque sensor on each axis, and the angle and torque of each axis of the KUKA robot 9 can be obtained by programming at a certain time. The controller of the KUKA robot 9 is also programmed with a matrix conversion relationship, which can calculate the position of the end flange. Forces and Cartesian positions of . Since the end flange of the KUKA robot 9 is equipped with grippers, it is necessary to calculate the force and displacement at the actual point of action (jaws) between the KUKA robot 9 and the steering wheel for subsequent steering wheel torque and rotation angle calculations. The known D-H parameter table of KUKA robot 9 is shown in Table 1.

Table 1 KUKA robot DH parameter list

The distance between the action point of the KUKA robot 9 jaws and the end flange is 156mm, so when calculating the displacement and force at the action point, you can change 152 in the DH model to 308. First, it is necessary to derive from the controller the rotation angles and torques of the 7 axes of the KUKA robot 9 at each moment during the process of turning the steering wheel. Here, the TCP/IP protocol is used to transmit data: a server is written on the external receiving PC, which runs before the KUKA robot 9 program and waits for the connection of the client. Write a client in the KUKA robot 9 program, the client establishes a connection with the service segment before the KUKA robot 9 turns, and then sends data every 100ms until the turning operation ends and the connection is disconnected.

For a series of data received at the PC, the string needs to be intercepted according to the characteristic symbols such as “[]”, and then the valid information is transferred to matlab for calculation and curve drawing.

For the calculation of the steering wheel angle, it is the forward solution problem of KUKA robot 9. According to the DH model, the coordinate transformation relationship between each two adjacent joints can be obtained:

in It is a 4*4 matrix containing 7 joint angle variables, and the Cartesian coordinates (Px, Py, Pz) of the end can be obtained by bringing in the 7 angle values.

The coordinates of the steering wheel rotation center are known to be (x ₀ , y ₀ , z ₀ ), so the steering wheel rotation angle at this time can be calculated according to the arc tangent formula:

The schematic diagram for calculating the torque of the steering wheel from the force at the jaws to the rotational torque of the steering wheel is shown in Figure 5. First, the force at the end is calculated according to the torque of each axis, and then according to the analysis, only Fy and Fz have an effect on the torque of the steering wheel, and Fx is because It coincides with the direction of the steering wheel rotation axis, so it has no effect. The torque in the three axis directions and the reaction torque of the steering wheel rotation center are offset, so it does not need to be included. The specific calculation process is as follows. First, the Jacobian matrix J ₀ at this time can be calculated in matlab according to the angles of each axis of the KUKA robot 9. Then the force at the end can be calculated according to the moment of each axis:

For the calculated end force, the actual torque of the steering wheel can be calculated using the following formula:

M＝-F _y ×(z ₀ -z)+F _Z ×(y ₀ -y) (6)

Thereby, the relationship between the steering wheel torque measured by the KUKA robot 9 and the rotation angle can be calculated.

4. Selection of impedance mode coefficients

After determining to use the impedance control mode, it is necessary to select the stiffness coefficients in each direction. If the stiffness coefficient is too large, the trajectory of KUKA robot 9 and the steering wheel will be rigid, which will cause the steering wheel to suffocate; if the stiffness coefficient is too small, the distance between the programmed position and the actual position of KUKA robot 9 will become larger, resulting in poor motion accuracy and lack of precise control The ability to turn the steering wheel angle. The force analysis of impedance mode is shown in Figure 6. The goal of using impedance control is to reduce the stiffness coefficient as much as possible without losing accuracy, so that the internal restoring force is reduced, thereby reducing damage to the mechanism.

As shown in Figure 6, the trajectory of the solid line is the established trajectory of programming, and the trajectory of the dotted line is the theoretical trajectory of the steering wheel. The radius of rotation of the gripping point is r, and the radius of the pulley is R. After the gripper clamps the steering wheel, the predetermined track of the dotted line of the steering wheel is determined, so the two arcs coincide at the starting point. Subsequent curves are separated due to the inaccurate measurement of the center position and radius.

The force analysis of the upper jaw is carried out below. Since the posture of the KUKA robot 9 changes continuously during the motion, an isotropic stiffness coefficient kz is used. Under the limitation of the predetermined arc trajectory of the steering wheel, the P1 point of the source program can only reach the P2 point in the end. The error can be divided into radial direction and tangential direction. The error in the radial direction is caused by surface friction and the limit mechanism, which will cause damage to the mechanism. The radial force is recorded as F _r ; The tangential error is recorded as Δx.

The position measurement error of the center of the steering wheel is a circle with a diameter of 1mm, and the measurement error of the radius of gyration is ±1mm, so the maximum radial error of the two arcs is 2mm, namely:

F _r =2k _z (7)

The pressure experienced in the direction perpendicular to the steering wheel extension is related to the internal loading, so:

The goal is to reduce the radial force F _r and reduce the tangential error Δx, so the minimum value of the comprehensive index W=F _r +Δx is obtained, so as to obtain the most suitable impedance kz.

Therefore, the programming curve cannot be directly used as the output displacement curve like a rigid robot, but the terminal displacement output curve needs to be calculated in real time by reading the angle values of the 7 joints of the KUKA robot 9. Figures 9 and 10 are the terminal displacement output curves measured under no load (the steering wheel is not clamped) and load (the steering wheel is clamped). Since the programmed trajectory and the predetermined trajectory of the steering wheel are both circles, the circle equation is used for fitting, and the simulated The fitted load curve is: (y-186.19) ² +(z-767.72) ² =292.67 ² , and the fitted no-load curve is (y-187.42) ² +(z-748.22) ² =300.16 ² . During programming, due to the measurement error of the center and radius, the center of the circle has an error of 1.23mm in the y direction, an error of 19.5mm in the z direction, and a radius error of 7.49mm. Since the deviation is solved by the impedance control mode, the KUKA robot 9 can Following the established trajectory of the steering wheel to complete the turning motion of the steering wheel, it also shows that the output displacement curve of the end of the KUKA robot 9 must be re-acquired and calculated in real time, and cannot be directly obtained from the programming theoretical curve.

5. Comparison of external robot and internal sensor curves

Put the KUKA robot 9 into the preset position for operation, the operation process is shown in Figure 13, the KUKA robot 9 runs smoothly and can rotate to the limit position. Figure 8 is the steering wheel steering curve measured by Beckhoff. It can be seen that the control curve takes 20.5° as the boundary, and the linearity on both sides is very good. The slope of the first segment is approximately twice that of the second segment.

For the manipulation curve measured by KUKA robot 9, since the self-weight of the end gripper will also cause the change of the torque of each axis, and the torque value changes with the position of the end, so the KUKA robot 9 operating steering wheel and KUKA robot 9 are respectively measured. The no-load control curve, the subtraction of the two is the steering wheel control curve. Fig. 9 is the manipulation curve converted from the no-load curve motion of the KUKA robot 9 to the steering wheel, Fig. 10 is the manipulation curve obtained by the KUKA robot 9 manipulating the steering wheel, and Fig. 11 is the actual curve of the steering wheel calculated by the difference method. Comparing the manipulation curves measured by Beckhoff and KUKA robot 9, Figure 12 is obtained.

The theoretical control curve is two straight lines with unequal slopes. The slope of the first segment is twice that of the second segment. The dividing point between the two segments is the moment when one side of the spring is disengaged. It can be seen from the results that the curves measured by Beckhoff and KUKA robot 9 are two straight lines with different slopes. Among them, the curve measured by Beckhoff has good linearity, and the slope of the first section is approximately twice that of the second section. The source of the error is the inherent resistance of the steering wheel mechanism. In the curve measured by KUKA robot 9, the slope difference between the two straight lines is relatively small, mainly because the rubber pulley at the upper end of the rope has a certain degree of expansion and contraction, thus weakening the gap between the two straight lines. The steering curve measured by the KUKA robot 9 has a certain torque at the initial angle. This is mainly because the steering wheel and other mechanisms also have a certain initial torque, which can be processed by subtracting this value in the entire section, while the Beckhoff measured is The actual load and linear displacement of the spring, therefore a straight line through the origin. In addition, the measured value of the rotation angle of Beckhoff is 0.67° larger than that measured by the KUKA robot 9. This is because the rope is not tight and stretchable, and the rotation angle of the KUKA robot 9 in the initial stage is converted into the tension of the rope. By adjusting the intercept of the KUKA robot 9 curve, the relative error between the torque measured by the KUKA robot 9 and the system torque measured by Beckhoff is less than 1.75%, which is within the acceptable range. The comparison of the test results is shown in Table 2.

Table 2 Comparison of test results

Claims (3)

1. The aircraft cockpit steering wheel manipulation test method based on the collaborative robot is characterized in that the steering wheel manipulation curve test is carried out in the aircraft cockpit using a collaborative robot, a two-finger gripper, and a raised platform, and the two-finger gripper is installed at the end of the collaborative robot. The raised platform is fixedly connected to the driver's seat in the main cockpit of the aircraft. The seat is not installed in the final assembly stage, and the raised platform is fixed by using the hole where the seat is installed on the cockpit floor. The collaborative robot is fixed on the raised platform. And there is a space for the collaborative robot to complete various test operations in the cockpit. The collaborative robot is a 7-axis robot, and the impedance mode generates a virtual spring at the end. In order to avoid the collision between the gripper and the steering wheel caused by the positioning error of the collaborative robot and To suffocate in motion, use the Cartesian impedance controller for programming, and use the setStiffness(...) function to set the impedance value of the xyz axis. Under impedance control, the behavior of the collaborative robot is obedient, and the steering wheel will not be suffocated due to motion errors Death; the collaborative robot works in the human-machine cooperation mode throughout the whole process, and the collaborative robot is equipped with torque sensors at 7 joints to detect the force of each joint in real time; when the collaborative robot collides with objects in the cockpit or someone enters the cockpit It can stop the movement to protect the safety of the collaborative robot and humans. By reading the torque value of each joint and specifying a sensitivity coefficient, if the torque of each joint is between the normal torque minus the sensitivity coefficient and the normal torque plus the sensitivity coefficient , then the collaborative robot operates normally, as long as a torque exceeds the limited range, the collaborative robot will stop. 2. The aircraft cockpit steering wheel manipulation test method based on collaborative robot according to claim 1, characterized in that the pincers of the two-finger gripper are arc-shaped structures wrapping the steering wheel extension rod. 3. the aircraft cockpit steering wheel manipulation test method based on collaborative robot according to claim 1, is characterized in that its steps are successively: 3-1 Calibration of the coordinate system in the cabin: it refers to the center of the base of the collaborative robot as the origin, the forward direction of the aircraft as the x-axis, the direction of the wings as the y-axis, and the vertical direction as the z-axis to establish a base coordinate system, and then use the steering wheel The center of rotation establishes a local coordinate system, and the specific method of calibration is to install an industrial 3D camera on the beam of the cabin, and obtain its 3D point cloud by shooting the environment in the cockpit, so as to realize the measurement of the spatial distance; 3-2 The programming movement of the collaborative robot: through the previously established base coordinate system, local coordinate system, and the set position coordinates of the operating steering wheel, the Point-To-Point movement reaches the designated point, and the soft control is adopted when approaching the target point , through the preset impedance coefficient to achieve flexible docking; 3-3 The conversion relationship of the steering wheel manipulation curve: For the measurement of the steering wheel rotation angle, it is first necessary to calculate the real-time Cartesian position of the end through the rotation angle of each axis, and then obtain the steering wheel angle by inverting the trigonometric function for the center of the circle; for the torque of the steering wheel, first combine the end posture of the robot Calculate the current Jacobian matrix, then calculate the force at the end according to the torque of each axis, and finally calculate the torque of the steering wheel according to the torque at the actual position and the force at the end; 3-4 Selection of impedance mode coefficients: Comparing the trajectory of the programmed motion with the established arc trajectory of the steering wheel, calculate the precision error brought by the impedance mode and the radial force that causes damage to the mechanism, and combine the precision error and radial force into An evaluation index, the most suitable impedance mode stiffness coefficient k is obtained by solving the index minimum; 3-5 Comparison of external collaborative robot and internal sensor curves: By reading the ratio of the steering wheel steering curve measured by the internal sensor of the aircraft and the steering wheel torque to the rotation angle measured by the collaborative robot, draw the curves of the two in matlab and observe The difference between the two is to adjust the steering wheel.