Tension-tracking for single robot wire harnessing
A single-robot system with cable tension feedback addresses kinematic and dynamic challenges in wire harness routing by employing a learning-based algorithm to efficiently route and secure cables to fixtures, reducing costs and enhancing robustness.
Patent Information
- Application Number
- JP2025039845
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-02-26
- Filing Date
- 2025-03-13
- Publication Date
- 2025-09-29
AI Technical Summary
Current robotic systems for routing wire harnesses in vehicles face challenges due to kinematic, dynamic, and vision-based issues, with bimanual robotic systems being expensive and single-robot systems relying on fragile tactile sensors.
A robotic system using a single robot with cable tension feedback, employing a learning-based algorithm to generate a nonlinear cable dynamics model, and controlling robot motion with force measurements to route and secure cables to fixtures.
Reduces system cost and complexity while improving robustness and efficiency in routing wire harnesses by using built-in force sensors to maintain cable tension.
Smart Images
Figure 2025141920000001_ABST
Abstract
Description
[Technical Field]
[0001] (CROSS-REFERENCE TO RELATED APPLICATIONS) This application claims the benefit of priority to U.S. Provisional Application No. 63 / 565,196, filed March 14, 2024, and entitled "Tension-Tracking For Single Robot Wire Harnessing."
[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 using a single robot with cable tension feedback. [Background technology]
[0003] Various industries, such as automotive, aerospace, medical, telecommunications, etc., frequently require the use of many wires, cables, and wire harnesses to provide signals and power to various electrical devices and systems. In the automotive industry, a wire harness can be a collection of electrical cables or wires that connect electrical and electronic components within a vehicle, such as sensors, electronic control units, batteries, actuators, etc. Wire harnesses route power and information to provide primary vehicle functions, such as steering and braking, and secondary vehicle functions, such as ventilation and infotainment. These wire harnesses must be routed throughout the vehicle and joined to fixtures when the vehicle is manufactured. Summary of the Invention [Problem to be solved by the invention]
[0004] Modern vehicle manufacturing is highly automated, often using robots to route and connect wire harnesses to fixtures. Using robots to route wire harnesses creates many challenges, including kinematic challenges due to the infinite degrees of freedom of wires and cables, dynamic challenges due to the deformability of wires and cables through contact, and vision-based challenges due to the thinness and length of wires and cables. Currently, two common technologies exist in the art for using robots to route cables and wire harnesses throughout a vehicle: bimanual robotic wire harnesses and single-robot wire harnesses. Bimanual robotic wire harnesses use two robotic arms to grasp and place wires and cables into fixtures. However, this technology is expensive and complicates system structure. Single-robot wire harnesses require tactile sensors on the robot's fingertips to detect the pose and grip / shear force of the wires. However, such tactile sensors are expensive and fragile. Therefore, there is room for improvement. [Means for solving the problem]
[0005] The following discussion discloses and describes a robotic system for routing wires or cables to and through various fixtures. The system uses a single robot with cable tension feedback. The robotic system employs a process that includes connecting the end of a cable to a fixed endpoint, mapping the position and pose of the fixture relative to the robot, and acquiring images of the cable using a camera. The process also includes gripping the cable using the robot's gripper and an image near the fixed endpoint, twisting the cable with the gripper so that the cable is under tension between the gripper and the fixed endpoint, and sliding the gripper along the twisted cable away from the fixed endpoint and toward a target fixture while the cable is under tension. The process uses a learning-based algorithm to generate a nonlinear cable dynamics model, generates robot motion command signals using the cable dynamics model, and controls the robot's motion using the robot motion command signals, as well as robot pose and force measurements, 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 claims, taken in conjunction with the accompanying drawings. [Brief explanation of the drawings]
[0007] [Figure 1] FIG. 1 is an isometric view of a robotic system including a robot for routing cables to and through a support fixture.
[0008] [Figure 2] FIG. 1 is a flow diagram of a method for routing and connecting cables to a series of fixtures.
[0009] [Figure 3] FIG. 1 is a general block diagram of a single robot wire harness manipulator system with force feedback.
[0010] [Figure 4] FIG. 1 is a block diagram of a robotic system including robotic visual recognition for gripping a cable.
[0011] [Figure 5] FIG. 1 is a block diagram of a robotic system illustrating cable routing using learning-based model predictive control (MPC).
[0012] [Figure 6] FIG. 1 is a diagram of a force feedback system showing a robotic gripper applying a surface tension force to a cable attached to a fixed point.
[0013] [Figure 7] FIG. 1 is a block diagram of a learning phase system for a single robot wire harness manipulator system.
[0014] [Figure 8] FIG. 1 is a block diagram of a task phase system for a single robot wire harness manipulator system. DETAILED DESCRIPTION OF THE INVENTION
[0015] The following description of embodiments of the present disclosure directed to a robotic system for routing wires or cables to and through various fixtures using a single robot with cable tension feedback is merely exemplary in nature and is not intended to limit the invention or its applications or uses.
[0016] As described in detail below, the present disclosure proposes a robotic wire harness manipulator system using a single robot with cable tension feedback. Compared to the known bi-manual robotic wire harness technology described above, the proposed robotic wire harness manipulator system uses only one robot arm, which reduces both system cost and complexity. Furthermore, compared to the known tactile sensing-based single-robot wire harness manipulator system described above, the proposed robotic wire harness manipulator system uses built-in force sensors to maintain cable tension, improving robustness and solving harness problems more efficiently.
[0017] FIG. 1 is an isometric view of a robotic system 10 including a six-axis robot 12 controlled by a motion controller 18. The robotic system 10 is configured to route wires or cables 14 to various support fixtures 16, such as C-shaped fixtures 54 and U-shaped fixtures 56, secured to a structure 20, with one end of the cable 14 secured to a fixture 22, such as a cable connector. The robot 12 is intended to represent any robot suitable for the purposes discussed herein, and the structure 20 is intended to represent any structure that would benefit from a robot performing wire harness routing, such as a vehicle structure. The robot 12 includes a base 24 rotatably mounted to a stand 26 by a joint 28, a first inner arm 30 connected to the base 24 by a joint 32, and a second inner arm 34 connected to the first inner arm 30 by a joint 36. A first outer arm 40 is connected to the second inner arm 34 by a joint 42, a second outer arm 44 is connected to the first outer arm 40 by a joint 46, and an end effector 48 having a gripper 50 is connected to the second outer arm 44. A camera 52 is connected to the end effector 48 to capture images of the cables 14, fixture 16 and structure 20.
[0018] The robot 12 knows the location and orientation or pose of the fixtures 16 on the structure 20, for example, by using augmented reality tags 58 in a manner well understood by those skilled in the art. Other techniques, such as visual sensing, can be used to determine the location and orientation or pose of the fixtures 16 on the structure 20. The robot 12 also knows which fixtures 16 the cable 14 is secured to and further knows the path the robot 12 will take to move the cable 14 from one fixture 16 to the next. Each time the cable 14 is secured to a fixture 16, that fixture 16 becomes a fixation point for routing the cable 14 to the next fixture 16, and the next fixture 16 from the specified fixation point 16 is the target fixture 16.
[0019] 2 is a flow diagram 150 of a method for routing cables 14 and connecting them to fixtures 16. The poses of all fixtures 16 are provided in box 152. Harness or cable motion planning is performed in box 154, providing a complete path for routing cables 14 to fixtures 16, using waypoints for fixtures 16 on structure 20, provided in box 156. The C-shaped fixture 54 includes three waypoints, and the U-shaped fixture 56 includes two waypoints. The cable motion planning includes aligning the waypoints for each fixture 16 along the path and merging waypoints that are too close. Cable following is then performed in box 158 using model predictive control (MPC).
[0020] Once the cable 14 is fixed to a particular fixture 16, a fix point switch is performed in box 160, where the last fixture 16 is now the connection point. Subsequent processing of the cable can insert cable and assembly primitives at waypoints in box 162 to better align the cable 14 with the fixture 16. A more detailed description of the cable routing process is provided below.
[0021] 3 is a schematic block diagram of a single robot wire harness manipulator system 60 that uses force or tension feedback operable to route and insert cable 14 into fixture 16, as described in more detail below. In box 62, an algorithm performs an initial visual recognition to identify a location for gripper 50 to grasp cable 14. This location is typically a point on cable 14 near fixture 16, which is a fixed-point fixture.
[0022] 4 is a block diagram of a robot system 64 illustrating robot visual recognition for this purpose. The robot system 64 includes a routing computer 66, a robot 68 representing the robot 12, and a camera 70 representing the camera 52. The camera 70 sends images of the fixture 16 to the computer 66, which determines a target pose for the cable 14 based on the images and sends a target pose signal to the robot 68. The robot 68 sends a robot pose signal to the computer 66, which sends the signal to the camera 70.
[0023] The system 60 provides a waypoint plan in box 80, which includes assigning a number of waypoints 78 to each fixture 16 that receives a cable 14, where the waypoints are midpoints around, typically above, the fixture 16. A C-shaped fixture 54 includes one type of waypoint location, while a U-shaped fixture 56 requires another type of waypoint location. Any suitable algorithm, such as Apriltags, can be used to obtain the waypoints.
[0024] To fit a nonlinear dynamics model that describes the changes in tension in cable 14 in response to the robot's motion, system 60 employs cable routing in box 82 using a learning-based MPC, such as a Koopman operator-based MPC, which is an operator that provides a data-driven method for building control-oriented models of nonlinear systems. System 60 then tracks the tension in cable 14 while following cable 14 using the nonlinear dynamics model. While this description refers to MPC for cable routing, other control techniques, such as proportional-derivative (PD) control, can also be used. This process generates a dynamics model of the system, as described below.
[0025] The following formulas can be used to learn and implement cable dynamics, specifically for the Koopman operator dynamics model:
[0026] JPEG2025141920000002.jpg33166
[0027] JPEG2025141920000003.jpg17163
[0028] JPEG2025141920000004.jpg68163
[0029] JPEG2025141920000005.jpg38166
[0030] JPEG2025141920000006.jpg52167
[0031] 5 is a block diagram of a robotic system 84 illustrating the learning of the dynamics model used for cable routing using MPC, with like elements to system 64 identified by the same reference numerals. A routing computer 66 provides predetermined motion signals to a robot 68, which provides gripper pose and tension signals to offline data in box 86. Computer 66 uses the offline data to adapt the cable dynamics.
[0032] FIG. 6 is an illustration of a force feedback system 90 showing a robot gripper 92, representing gripper 50, applying a twisting force and tension to a cable 94, representing cable 14, connected to a fixed point 96. Gripper 92 applies tension to cable 94 by grasping and twisting cable 94, thereby providing an angle of cable 94 between gripper 92 and fixed point 96; the greater the twist angle, the greater the tension on cable 94. The tension in cable 94 can be measured by six joint torque sensors on robot 12, and the joint torques are mapped into Cartesian space to obtain the tension. Tension can also be measured in other ways, such as using force sensors. It is known that applying such tension to cable 94 eliminates slack in cable 94, allowing it to be threaded directly through target fixture 16. Gripper 92 grasps cable 94 proximate fixation point 96 and applies surface tension to cable 94 through a twisting action, and then robot 68 moves gripper 92 toward target fixture 16, causing gripper 92 to slide along cable 94 while cable 94 is under tension.
[0033] JPEG2025141920000007.jpg91169
[0034] JPEG2025141920000008.jpg38168
[0035] The system 60 uses two primitive operations, including an assembly primitive that inserts the cable 14 into the fixture 16 while maintaining tension on the cable 14, and a cable primitive that collects data in the box 100. The assembly primitive involves pulling the gripper 50 along the cable 14 while the cable 14 is under tension and aimed at the target fixture 16, and measuring the twist angle in the cable 14 as the gripper 50 is moved. To learn the dynamics of the cable 14, the forces and angles on the cable 14 and the position of the gripper 50 relative to a fixed point are measured and calculated. The assembly primitive for the C-shaped fixture 54 is used as follows: In one embodiment, for the first side waypoint, the robot 12 moves down 30 mm to lower the robot 12's position for cable routing. After reaching the second side waypoint, the robot 12 moves up 30 mm to avoid a collision with the structure 20 and finish routing the cable. The assembly primitive for the U-shaped fixture 56 is used as follows: In one embodiment, after reaching the second side waypoint, the robot 12 first descends 30 mm for insertion while twisting the cable 14 to maintain tension on the cable. Next, the robot 12 moves 20 mm in two directions along the edges of the fixture 56 to further secure the cable 14. Finally, the robot 12 moves 30 mm upward to finish inserting the cable 14 into the fixture 56. For the cable primitive, in one embodiment, 40 real-world wire-following trajectories [τi=(s0, u0,..., sT, uT),i=1,...,40] are collected as a dataset using scripted twisting and stretching motion of the cable 14 with random initial conditions.
[0036] FIG. 7 is a block diagram of a learning phase system 110 for a single-robot wire harness manipulator system such as system 60. This system is used to learn the 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, causing the robot 12 to grasp and twist the cable 14 as described above to obtain a desired number of motion primitives. These motion primitives are provided to a data collection module 118, which in turn 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, for example, using the equations described above. The model learning module 120 provides the model learning information to a cable dynamics model module 122.
[0037] 8 is a block diagram of a task phase system 130 of a single-robot wire harness manipulator system that performs the actual routing of cable 14 to fixture 16 based on the learned cable dynamics provided by system 110. Task phase system 130 includes a robot 132 and a computer 134. The cable dynamics generated by cable dynamics model module 122 are provided to a cable dynamics model module 136 in computer 134. Cable dynamics model module 136 provides the cable dynamics to a model predictive control module 138 in computer 134, as described above, which provides motion command signals to a motion control module 140 in robot 132 that control the motion of robot 132 to route cable 14 to 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, such as joint torque sensors in the robot 132, provide force or tension signals to the cable dynamics model module 136.
[0038] The foregoing discussion discloses and describes merely exemplary embodiments of the present disclosure. Those skilled in the art will readily appreciate from such description, the accompanying drawings, and the claims that various changes, modifications, and variations can be made without departing from the spirit and scope of the present disclosure, as defined in the following claims.
Claims
1. 1. A method for routing and securing cables to a plurality of fixtures attached to a structure using a robot including a cable gripper and a camera, the robot being controlled by a motion controller, the method comprising: connecting an end of the cable to a fixed termination point; mapping the position and pose of the fixture relative to the robot; capturing an image of the cable using the camera; gripping the cable using the gripper and an image of the vicinity of the end point; twisting the cable with the gripper to apply tension to the cable between the gripper and the fixed endpoint; sliding the gripper along the twisted cable away from the fixed endpoint and toward a target fixture while the cable is under tension; generating a nonlinear cable dynamics model using a learning-based algorithm; generating a robot movement command signal using the cable dynamics model; controlling the robot's motion using the robot motion command signals and robot pose and force measurements to route and secure the cable to the target fixture; after the cable is connected to the target fixture, switching to routing the cable to a next fixture; A method comprising:
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 of claim 1 , wherein generating the robot motion command signals includes using a model predictive control (MPC) algorithm.
4. 2. The method of claim 1, wherein generating a nonlinear cable dynamics model using a learning-based algorithm comprises modeling the cable dynamics as a Koopman operator model adapted to the learning-based algorithm.
5. The method of claim 1 , further comprising determining two primitive operations including an assembly primitive that inserts the cable into the fixture while maintaining tension on the cable, and a cable primitive that collects data.
6. 2. The method of claim 1, wherein gripping the cable with the gripper and the image comprises determining a target pose of the cable based on the image, sending a target pose signal to the robot, and sending a robot pose signal from a robot controller to a routing computer.
7. The method of claim 1 , wherein mapping the position and pose of the fixture includes using a visual sensor.
8. The method of claim 1 , wherein mapping the positions and poses of the fixtures includes a waypoint plan that assigns waypoints to each of the fixtures.
9. 9. The method of claim 8, wherein the fixtures include C-shaped fixtures and U-shaped fixtures, the C-shaped fixtures being assigned one type of waypoint and the U-shaped fixtures being assigned another type of waypoint.
10. The method of claim 1 , wherein the force measurements are obtained by measuring tension caused by twisting the cable using joint torque or force sensors of the robot.
11. 1. A method for using a robot to route and secure cables to a plurality of fixtures attached to a structure, comprising: gripping the cable; twisting the cable to apply tension to the cable; sliding the gripper along the twisted cable while the cable is under tension; generating a nonlinear cable dynamics model using a learning-based algorithm; generating a robot movement command signal using the cable dynamics model; controlling the robot's motion using the robot motion command signals and robot pose and force measurements, and routing and securing the cables to the plurality of fixtures; after the cable is connected to the target fixture, switching to routing the cable to a next fixture; A method comprising:
12. The method of claim 11 , wherein generating the robot motion command signals includes using a model predictive control (MPC) algorithm.
13. 12. The method of claim 11, wherein generating a nonlinear cable dynamics model using a learning-based algorithm comprises modeling the cable dynamics as a Koopman operator model adapted to the learning-based algorithm.
14. The method of claim 11 , further comprising determining two primitive operations including an assembly primitive that inserts the cable into the fixture while maintaining tension on the cable, and a cable primitive that collects data.
15. 1. A system for routing and securing cables to a plurality of fixtures attached to a structure using a robot including a cable gripper and a camera, the robot being controlled by a motion controller, the system comprising: means for mapping the position and pose of the fixture relative to the robot; means for capturing an image of the cable using the camera; means for gripping the cable using an image of the gripper and the vicinity of the fixed end point; means for twisting the cable with the gripper to apply tension to the cable between the gripper and the fixed endpoint; means for sliding the gripper along the twisted cable away from the fixed endpoint 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 a robot operation command signal using the cable dynamics model; means for controlling the robot's motion using the robot motion command signals and robot pose and force measurements, and for routing and securing the cable to the target fixture; means for switching the routing of the cable to a next fixture after the cable has been connected to the target fixture; Including, the system.
16. 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. 16. The system of claim 15, wherein the means for generating the robot motion command signals uses a model predictive control (MPC) algorithm.
18. 16. The system of 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 adapted to the learning-based algorithm.
19. 16. The system of claim 15, further comprising means for determining two primitive operations including an assembly primitive that inserts the cable into the fixture while maintaining tension on the cable, and a cable primitive that collects data.
20. 16. The system of claim 15, wherein the means for gripping 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 routing computer.