Tension tracking for single robotic wire management
By adopting a single robot system combined with cable tension feedback and nonlinear dynamic modeling, the problem of expensive and complex wiring of wire harnesses in existing technologies is solved, and efficient and low-cost wire management is achieved.
Patent Information
- Application Number
- CN202510296350.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2025-02-26
- Filing Date
- 2025-03-13
- Publication Date
- 2025-09-16
AI Technical Summary
Prior art systems using two-handed robotic wire management systems are expensive and complex, while single robotic wire management systems use expensive and fragile tactile sensors, making it difficult to efficiently route wire harnesses to fixtures in a vehicle.
A single robot system is used, combined with cable tension feedback, using a fixture and camera on the robot to capture cable images. By generating a nonlinear cable dynamics model, the robot motion is controlled to achieve wiring of the wire harness, reducing system cost and improving robustness.
This enables efficient and cost-effective routing of wire harnesses to vehicle fixtures using a single robot, reducing system complexity and improving the robustness of wire management.
Smart Images
Figure CN120645185A_ABST
Abstract
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims the benefit of the filing date of U.S. Provisional Application Serial No. 63 / 565,196, filed on March 14, 2024, and entitled TENSION TRACKING FOR SINGLE ROBOTIC WIRE MANAGEMENT. Technical Field
[0002] The present disclosure relates generally to a robotic system for manipulating cables, and more particularly to a robotic system for routing wires or cables to and through various fixtures, wherein the system employs a single robot using cable tension feedback. Background Art
[0003] Various industries, such as automotive, aviation, medical, and telecommunications, often require the use of numerous wires, cables, and wiring harnesses to provide signals and power to various electrical devices and systems. For the automotive industry, a wiring harness is a collection of cables or wires that connect electrical and electronic components within a vehicle, such as sensors, electronic control units, batteries, and actuators. These harnesses transmit power and information to provide primary vehicle functions such as steering and braking, as well as auxiliary functions such as ventilation and infotainment. These harnesses are routed throughout the vehicle and connected to fixed devices during vehicle manufacturing.
[0004] Modern vehicle manufacturing is highly automated, and robots are often used to route wire harnesses and connect them to fixtures. Using robots to route wire harnesses creates many challenges, such as kinematic challenges due to the unlimited degrees of freedom of wires and cables, dynamic challenges due to the deformability of wires and cables under contact, and vision challenges due to the thinness and long length of wires and cables. There are currently two general technologies in the art for using robots to route cables and wire harnesses throughout a vehicle, namely two-hand robotic wire management and single-robot wire management. Two-hand robotic wire management uses two robotic arms to grasp the wires and cables and then place them into fixtures. However, this technology is expensive and introduces complexity into the system design. Single-robot wire management requires tactile sensors on the robot's fingertips to detect wire posture and grasping / shearing force. However, such tactile sensors are expensive and fragile. Therefore, improvements can be made. Summary of the Invention
[0005] The following discussion discloses and describes a robotic system for routing wires or cables to and through various fixtures, wherein the system employs a single robot that uses cable tension feedback. The robotic system employs a process that includes attaching the ends of the cables to fixed endpoints, mapping the positions and poses of various fixtures relative to the robot, and capturing images of the cables using a camera. The process also includes grasping the cable using a gripper on the robot and an image of the vicinity of the fixed endpoint, twisting the cable using the gripper so that tension exists on the cable between the gripper and the fixed endpoint, and sliding the gripper along the twisted cable away from the fixed endpoint and toward the target fixture while the cable is under tension. The process uses a learning-based algorithm to generate a nonlinear cable dynamics model, uses the cable dynamics model to generate robot motion command signals, and uses the robot motion command signals and robot pose and force measurements to control the robot's motion to route and secure the cable to the target fixture.
[0006] Additional features of the present disclosure will become apparent from the following description and appended claims, taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0007] Figure 1 is an isometric view of a system including a robot for routing cables to and through a support fixture;
[0008] Figure 2 is a flow chart of a method for routing and connecting cables to a series of fixtures;
[0009] Figure 3 is an overall block diagram of a single robotic wire harness manipulator system with force feedback;
[0010] Figure 4 is a block diagram of a robotic system including robotic visual perception for grasping a cable;
[0011] Figure 5 is a block diagram of a robotic system illustrating cable routing using learning-based model predictive control (MPC);
[0012] Figure 6 is a diagram of a force feedback system illustrating a robotic gripper applying force tension to a cable coupled to a fixed point;
[0013] Figure 7 is a block diagram of a learning phase system for a single robotic wire harness manipulator system; and
[0014] Figure 8 is a block diagram of the mission phase system for a single robotic wire harness manipulator system. DETAILED DESCRIPTION
[0015] The following discussion of embodiments of the present disclosure, which relates to a robotic system for routing wires or cables to and through various fixtures, wherein the system employs a single robot using cable tension feedback, is merely exemplary in nature and is in no way intended to limit the present disclosure or its applications or uses.
[0016] As discussed in detail below, the present disclosure proposes a robotic wire harness manipulator system that employs a single robot using cable tension feedback. Compared to the known two-handed robotic wire management techniques mentioned above, the proposed robotic wire harness manipulator system utilizes only a single robotic arm, which reduces system cost and complexity. Furthermore, compared to the known single-robot wire management techniques based on tactile sensing, the proposed robotic wire harness manipulator system uses built-in force sensors to maintain tension in the cables, more effectively solving the management problem with improved robustness.
[0017] Figure 1 is an isometric view of a robotic system 10, which includes a six-axis robot 12 controlled by a motion controller 8 and configured to route wires or cables 14 to and through various support fixtures 16, such as a C-shaped fixture 54 and a U-shaped fixture 56, secured to a structure 20. One end of the cable 14 is secured to a fixture 22, such as a cable connector. Robot 12 is intended to represent any robot suitable for the purposes discussed herein, and structure 20 is intended to represent any structure, such as a vehicle structure, that could benefit from a robot that provides wiring harnesses. Robot 12 includes a base 24 rotatably mounted to a support 26 via a joint 28, a first inner arm 30 coupled to base 24 via a joint 32, and a second inner arm 34 coupled to first inner arm 30 via a joint 36. A first outer arm 40 is coupled to second inner arm 32 via a joint 42, a second outer arm 44 is coupled to first outer arm 40 via a joint 46, and an end effector 48 having a gripper 50 is coupled to second outer arm 44. The camera 52 is coupled to the end effector 48 and captures images of the cable 14 , the fixture 16 , and the structure 20 .
[0018] The robot 12 knows the location and orientation or pose of the fixtures 16 on the structure 20 by using, for example, the augmented reality tag 58 in a manner well known to those skilled in the art. Other technologies are also available, such as vision sensing, to determine the location and orientation or pose of the fixtures 16 on the structure 20. The robot 12 also knows to which fixtures 16 the cables 14 will be secured and the route the robot 12 will take to move the cables 14 from one fixture 16 to the next. Each time a cable 14 is secured to a fixture 16, that fixture 16 becomes the fixation point for routing the cable 14 to the next fixture 16, where the next fixture 16 from the designated fixation point fixture 16 is the target fixture 16.
[0019] Figure 2 1 is a flow chart 150 of a method for routing and connecting cables 14 to fixtures 16. At block 152, the poses of all fixtures 16 are provided. At block 154, harness or cable motion planning is performed, which gives a complete path for routing the cables 14 to the fixtures 16 and uses the waypoints of the fixtures 16 on the structure 20 provided at block 156. The C-shaped fixture 54 includes three waypoints, while the U-shaped fixture 56 includes two waypoints. Cable motion planning includes aligning the waypoints of each fixture 16 along the path and merging waypoints that are too close. Cable following is then performed at block 158 using model predictive control (MPC).
[0020] Once the cable 14 is secured to a particular fixture 16, a fixation point switch is performed at block 160, where the last fixture 16 is now the connection point. At block 162, the cable following process may cause cables and assembly primitives to be inserted into the waypoint to better align the cable 14 with the fixture 16. A more detailed discussion of the cable routing process is provided below.
[0021] Figure 3 is an overall block diagram of a single robotic wire harness manipulator system 60 utilizing force or tension feedback, discussed in detail below, operable to route and insert a cable 14 into a fixture 16. At block 62, an algorithm provides an initial visual perception that identifies the location where the gripper 50 grasps the cable 14, which is typically a point on the cable 14 proximate to the fixture 16, which is the current fixation point fixture.
[0022] Figure 4is a block diagram illustrating a robotic system 64 that includes robotic vision perception for this purpose. The robotic system 64 includes a wiring computer 66, a robot 68 representing the robot 12, and a camera 70 representing the camera 52. The camera 70 sends an image of the fixture 16 to the computer 66, which determines a target pose for the cable 14 based on the image and sends a target pose signal to the robot 68. The robot 68 sends a robot pose signal to the computer 66, which sends a signal to the camera 70.
[0023] At block 80, the system 60 provides waypoint planning, which includes assigning a plurality of waypoints 78 to each fixture 16 that will receive the cable 14, where a waypoint is an intermediate point around and typically on the fixture 16. The C-shaped fixture 54 will include one type of waypoint location, and the U-shaped fixture 56 will require another type of waypoint location. Any suitable algorithm, such as an AprilTag, may be used to obtain the waypoints.
[0024] At block 82, the system 60 employs cable routing with a learning-based MPC, such as a Koopman operator-based MPC, to fit a nonlinear dynamic model representing the variation in tension on the cable 14 relative to the robot's motion. Learning-based MPC is an operator that provides a data-driven approach to constructing control-oriented models of nonlinear systems. The system 60 then employs the nonlinear dynamic model to follow the cable 14 while tracking the tension on the cable 14. While this discussion relates to MPC to provide cable routing, other control techniques, such as proportional-derivative (PD) control, may be employed. That process employs a dynamic model of the generative system, as discussed below.
[0025] The following equations can also be used to study and implement cable dynamics, especially the Koopman operator dynamics model.
[0026] The nonlinear equation (1) can be linear in the embedding space. S t+1 =f(S t ,u t ) (1) It gives the lifting function:
[0027] The approximate linear dynamics in the embedding space is: g(s t+1 )=Kg(s t )+Lu t (3)
[0028] Model fit is provided by: [K, L] = PG t (6)
[0029] Data was collected during real-world wire following trajectories, and the potential for improvement was: A second-order polynomial
[0030] The MPC equation is: g(s t+1 )=Kg(s t )+Lu t b l ≤Ag(s t+1 )≤b u
[0031] Figure 5 FIG6 is a block diagram of a robotic system 84 illustrating a learning dynamic model to be employed in cable routing with MPC, wherein similar elements to system 64 are identified by the same reference numerals. At block 86, the routing computer 66 provides predetermined motion signals to the robot 68, and the robot 68 provides gripper pose and tension signals to the offline data. The computer 66 uses the offline data to fit the cable dynamics.
[0032] Figure 6 FIG2 is a diagram of a force feedback system 90, showing a robotic gripper 92, representing the gripper 50, applying a torsional force and tension to a cable 94, representing the cable 14, coupled to a fixed point 96. Gripper 92 grasps cable 94 and applies tension to it by twisting it, causing the angle of cable 94 to be set between gripper 92 and fixed point 96. The greater the twist angle, the greater the tension on cable 94. The tension on cable 94 can be measured by six joint torque sensors on the robot 12, where the joint torques are mapped to Cartesian space to obtain tension. Tension can also be measured in other ways, such as by using force sensors. By applying such tension to cable 94, it is known that there is no slack in cable 94, and it can then be routed directly to the target fixture 16. When gripper 92 grasps cable 94 near fixed point 96 and applies tension to it through a twisting motion, gripper 92 will then slide along cable 94 while cable 94 is under tension as the robot 68 moves gripper 92 toward the target fixture 16.
[0033] The following equations can also be used to learn and implement cable dynamics. Equation (11) is the state s of the model at time tt , Equation (12) is the input u to the module at time t t , and equation (13) is the cable dynamics f(s t ,u t ) State s at time t+1 t+1 , where |f t | is the tension on cable 94, (x t ,y t ) is the position of the fixture 92 at time t relative to the position (0, 0) of the fixed point 96, and θ t is the angle of the fixture 92 at time t. The model converts the state s t is mapped to a latent space (s, z), where z contains the second-order terms of state s. For example, Equation (14) fits the linear dynamics i+1 with 40 real-world trajectories to obtain model learning, where K is a weighting function. s t =(x t ,y t ,θ t |f t |) (11) u t =(Δx t , Δy t , Δθ t ) (12) s t+1 =f(s t ,u t ) (13)
[0034] The linear model of Equation (14) obtained from the collected data is then used as a constraint in the optimization formula of Equation (15) to route the cable 14 to the next target fixture 16, where s i is the current state, s d is the desired state at the target fixture 16 , R and Q are weighting factors for balancing force tracking and robot pose tracking, and H is the predicted horizon and is 5, for example.
[0035] At block 100, system 60 uses two primitive motions: an assembly primitive for inserting cable 14 into fixture 16 while maintaining tension on the cable 14, and a cable primitive for collecting data. The assembly primitive involves pulling clamp 50 along cable 14 while maintaining tension on the cable 14 as it is guided to the target fixture 16 and twisting the angle of the cable 14 as the clamp 50 is moved. The forces and angles on cable 14, as well as the position of clamp 50 relative to the fixture, are measured and calculated to understand the dynamics of cable 14. The assembly primitive for C-shaped fixture 54 is used as follows. In one embodiment, for the first side waypoint, robot 12 moves downward 30 mm to lower its position for cable routing. After reaching the second side waypoint, robot 12 moves upward 30 mm to avoid collision with structure 20 and complete cable routing. The assembly primitive for U-shaped fixture 56 is used as follows. In one embodiment, after reaching the second side waypoint, robot 12 first moves downward 30 mm for insertion while twisting cable 14 to maintain cable tension. The robot 12 then moves 20 mm in both directions along the edge of the fixture 56 to further secure the cable 14. Finally, the robot 12 moves upward 30 mm to complete the insertion of the cable 14 into the fixture 56. For the cable primitive, in one embodiment, forty real-world wire-following trajectories [τi=(s0,u0,...,sT,uT),i=1,...,40] are collected as a dataset using scripted twisting and stretching motions of the cable 14 with random initial states.
[0036] Figure 7 FIG1 is a block diagram of a learning phase system 110 for a single robotic wire harness manipulator system, such as system 60, for learning cable dynamics before the robot 12 is used to route the cable 14 through the fixture 16. The learning phase system 110 includes a training robot 112 and a training computer 114. A motion primitive command module 116 provides motion command signals to the robot 112 to cause the robot 112 to grip and twist the cable 14 as described above to obtain a desired number of motion primitives, wherein these motion primitives are provided to a data collection module 118, and wherein the robot 112 provides external force robot pose signals, including measurement signals from joint torque sensors, to the data collection module 118. The data collection module 118 provides the collected data to a model learning module 120 in the computer 114 to learn the cable dynamics using, for example, the equations described above, and the model learning module 120 provides the model learning information to a cable dynamics model module 122.
[0037] Figure 8FIG1 is a block diagram of a task phase system 130 for a single robotic wire harness manipulator system that performs the actual routing of the cable 14 to the fixture 16 based on the learned cable dynamics provided by the system 110. The task phase system 130 includes a robot 132 and a computer 134. The cable dynamics generated by the cable dynamics model module 122 are provided to a cable dynamics model module 136 in the computer 134. The cable dynamics model module 136 provides the cable dynamics to a model predictive control module 138 in the computer 134, which, as discussed above, provides motion command signals to a motion controller module 140 in the robot 132, which controls the motion of the robot 132 to route the cable 14 to the fixture 16. Motor encoders 142 in the robot 132 provide robot pose and gripper position signals to the cable dynamics model module 136, and force / torque sensors 144 on the robot 132, such as joint torque sensors, provide force or tension signals to the cable dynamics model module 136.
[0038] The foregoing discussion discloses and describes only exemplary embodiments of the present disclosure. Those skilled in the art will readily recognize from such discussion and from the accompanying drawings and claims that various changes, modifications and variations can be made therein without departing from the spirit and scope of the present disclosure as defined in the appended claims.
Claims
1. A method for routing and securing cables to a plurality of fixtures mounted to a structure using a robot controlled by a motion controller and including a cable clamp and a camera, the method comprising: connecting one end of the cable to a fixed terminal; mapping the positions and poses of the plurality of fixtures relative to the robot; capturing an image of the cable using the camera; grasping the cable using the clamp and the image of the vicinity of the fixed endpoint; twisting the cable using the clamp to provide tension on the cable between the clamp and the fixed end point; while the cable is under tension, sliding the clamp along the twisted cable away from the fixed end point and toward a target fixture; Generate nonlinear cable dynamics models using a learning-based algorithm; generating a robot motion command signal using the cable dynamics model; controlling the motion of the robot using the robot motion command signals and robot pose and force measurements to route and secure the cable to the target fixture; as well as Switching to routing the cable to the next fixture occurs after the cable is connected to the target fixture. 2 . The method of claim 1 , further comprising sequentially routing and securing the cable to each fixture in the same manner until the cable is secured to all of the plurality of fixtures.
3. The method according to claim 1, wherein Generating robot motion command signals includes using a model predictive control (MPC) algorithm.
4. The method according to claim 1, wherein Generating a nonlinear cable dynamics model using a learning-based algorithm includes modeling the cable dynamics as a Koopman operator model fitted by the learning-based algorithm.
5. The method of claim 1, further comprising determining two primitive motions, including an assembly primitive that inserts the cable into the plurality of fixtures while maintaining tension on the cable and a cable primitive that collects data.
6. The method according to claim 1, wherein Grasping the cable using the gripper and the image includes determining a target pose for the cable based on the image, sending a target pose signal to the robot, and sending a robot pose signal from the robot controller to a wiring computer.
7. The method according to claim 1, wherein Mapping the positions and poses of the plurality of fixtures may include using a vision sensor.
8. The method according to claim 1, wherein Mapping the positions and attitudes of the plurality of fixed devices includes providing a waypoint plan including assigning a waypoint to each of the plurality of fixed devices.
9. The method according to claim 8, wherein The plurality of fixtures include C-shaped fixtures and U-shaped fixtures, and one type of waypoint is assigned to the C-shaped fixture and another type of waypoint is assigned to the U-shaped fixture.
10. The method according to claim 1, wherein The force measurements are obtained by measuring the tension provided by twisting the cables using joint torque sensors or force sensors on the robot.
11. A method for routing and securing cables to a plurality of fixtures mounted to a structure using a robot, the method comprising: Grab the cable; twisting the cable to provide tension on the cable; sliding the clamp along the twisted cable while the cable is under tension; Generate nonlinear cable dynamics models using a learning-based algorithm; generating a robot motion command signal using the cable dynamics model; controlling the motion of the robot using the robot motion command signals and the robot pose and force measurements to route and secure the cables to the plurality of fixtures; as well as Switching to routing the cable to the next fixture occurs after the cable is connected to the target fixture.
12. The method according to claim 11, wherein Generating robot motion command signals includes using a model predictive control (MPC) algorithm.
13. The method according to claim 11, wherein Generating a nonlinear cable dynamics model using a learning-based algorithm includes modeling the cable dynamics as a Koopman operator model fitted by the learning-based algorithm.
14. The method of claim 11, further comprising determining two primitive motions, including an assembly primitive that inserts the cable into the plurality of fixtures while maintaining tension on the cable and a cable primitive that collects data.
15. A system for routing and securing cables to a plurality of fixtures mounted to a structure using a robot controlled by a motion controller and including a cable clamp and a camera, the system comprising: means for mapping the positions and poses of the plurality of fixtures relative to the robot; means for capturing an image of the cable using the camera; means for grasping said cable using said clamp and said image near a fixed end point; means for twisting the cable using the clamp to provide tension on the cable between the clamp and the fixed end point; means for sliding the clamp along the twisted cable away from the fixed end point and toward a target fixture while the cable is under tension; means for generating a nonlinear cable dynamics model using a learning-based algorithm; means for generating robot motion command signals using the cable dynamics model; means for controlling the motion of the robot using the robot motion command signals and robot pose and force measurements to route and secure the cable to the target fixture; as well as Means for switching to routing the cable to a next fixture after the cable is connected to the target fixture.
16. The system of claim 15, further comprising means for sequentially routing and securing the cable to each fixture in the same manner until the cable is secured to all of the plurality of fixtures.
17. The system according to claim 15, wherein: The device for generating a robot motion command signal uses a model predictive control (MPC) algorithm.
18. The system according to claim 15, wherein: The means for generating a nonlinear cable dynamics model using a learning-based algorithm models the cable dynamics as a Koopman operator model fitted by a learning-based algorithm.
19. The system of claim 15, further comprising means for determining two primitive motions, the two primitive motions comprising an assembly primitive for inserting the cable into the plurality of fixtures while maintaining tension on the cable and a cable primitive for collecting data.
20. The system of claim 15, wherein: The means for grasping the cable using the gripper and the image determines a target pose of the cable based on the image, sends a target pose signal to the robot, and sends a robot pose signal from a robot controller to a wiring computer.