Intelligent force control includes payload dynamics features such as torque-based collision detection, drag teaching, and torque feedforward.
Identification
For ease of use, the dynamics parameters in iNexBot control systems are generated automatically through identification. Simply run the preset trajectory automatically and the robot's dynamics parameters are obtained.
Collision Detection
As industrial robot technology advances, robots are required to perform increasingly complex functions. During robot motion, operators may need to enter the robot's working area, and their safety must be guaranteed. Collision detection must therefore identify in real time whether a collision has occurred between the robot and an operator, protecting operator safety — or preventing two collaborating robots from colliding during operation. The collision detection feature works with any six-axis robot. Even without external sensors, it remains highly sensitive while carrying a load, greatly reducing the likelihood of production accidents.
Torque Feedforward
Once dynamics identification is complete, enable the torque feedforward switch and it will be applied whenever job programs are run.
Torque feedforward anticipates how the arm will move before it actually does, and uses that information to adjust for and optimize deviations that may occur during motion — such as arm vibration in certain postures or after a load is attached. Because torque feedforward knows in advance what torque each joint should apply, it reduces overshoot during motion, which manifests macroscopically as suppressed arm vibration. Beyond suppressing vibration, torque feedforward also increases arm speed, reduces position tracking error, and makes arm motion smoother and more compliant.
To use it, simply enter the load's mass and offset and the payload dynamics features can be applied in real production. During production, the load at the arm's end effector can also be updated via commands, greatly increasing the flexibility of the feature.
Combining load-capable dynamics with torque feedforward effectively eliminates vibration when the arm carries heavy loads, and also allows the arm to move faster.
Drag Teaching
Drag teaching in iNexBot control systems supports two dragging modes: torque drag and 3D mouse drag.
Torque Drag
The robot is dragged by external force, and the dragged trajectory is recorded and can be replayed. Torque drag offers three modes: free drag, position drag, and posture drag. The smoothness of each axis during dragging can be adjusted via the joint friction coefficient; however, the coefficient must not be set too high, or the robot will sag when powered on, endangering both operators and the robot.
Dragged trajectories are recorded in the Human-Robot Collaboration > Drag Teaching screen, with trajectory names saved in the Trajectory Management screen. Dragged trajectories are stored permanently and can be replayed at any time.
Trajectory recording during dragging can also be performed in the Monitor > Shortcuts > Trajectory Replay screen, with dragged trajectories saved to the Trajectory Management screen in real time.
Drag teaching commands fall under the motion control command category.
Drag teaching commands also support trajectory replay, and you can change the trajectory name to replay at any time.
3D Mouse Drag
In certain human-robot interaction tasks, the operator needs to drag the robot along a straight line, which torque drag cannot achieve. Our 3D mouse enables dragging along the end-effector XYZ axes and the ABC posture axes, offering higher precision than torque drag.
Adaptive Acceleration and Deceleration
To use it, enable the switch after identification and enter the load parameters; the robot will then automatically adjust acceleration and deceleration during operation.
