Industrial Robot Deflection Correction via Torque Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for correcting mechanical deflection in industrial robots, such as those described in patent JP H04-233602 A, are inadequate as they require additional components like force sensors and increased rigidity, which can increase cost and size, and are ineffective when deflection is caused by external forces.
Innovation Solution
A method that calculates the moment applied to a robot's axis in a non-rotation direction based on load torque and self-weight, allowing for deflection correction without additional components, thereby maintaining size and weight reduction while ensuring accurate alignment and force estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a force sensor is attached to the tip end to detect external force for deflection calculation, then deflection correction accuracy is improved, but device complexity and cost increase
Solution Approach 1:
The robot system uses its own existing sensors (torque sensors in joints, encoders for position detection) and control data to calculate deflection, rather than requiring external force sensors. The deflection calculation leverages the robot's self-measured load torque and geometric parameters to determine and correct deflection without additional measurement components.
Solution Approach 2:
The patent introduces an intermediary calculation process that uses the relationship between load torque, robot geometry, and deflection mechanics to bridge the gap between measurable quantities (torque, position) and the unmeasured quantity (deflection). This mathematical intermediary eliminates the need for direct force measurement at the tip.
2Stability of the object's composition
If each axis is ensured to have high rigidity so as not to develop deflection, then deflection is reduced, but size and weight of the arm increase
Solution Approach 1:
Instead of changing the physical parameter of rigidity (which would increase weight), the patent changes the control parameter by introducing deflection correction calculations. The system accepts the actual rigidity of the arm and compensates for deflection through computational correction of position and orientation data, rather than attempting to eliminate deflection through increased structural rigidity.
Solution Approach 2:
The patent replaces the mechanical approach (increasing physical rigidity to prevent deflection) with a computational approach (calculating and correcting deflection based on measured load torque and geometric parameters). This substitution allows the use of lighter, more flexible arms while maintaining positioning accuracy through software-based compensation.
3Measurement precision
If calculated joint torque is different from actual torque, then end effector position correction accuracy deteriorates, but adding force sensors increases cost
Solution Approach 1:
The system uses feedback from the robot's existing torque sensors and position encoders to continuously monitor load conditions and calculate deflection. The control system feeds back corrected position and orientation data based on the calculated deflection, creating a closed-loop system that compensates for torque estimation errors without requiring additional force measurement devices.
Data Source
AI summary
In an industrial robot, correction is made for change in position and attitude of an arm distal end due to mechanical deflection of the robot. In the robot, a moment applied to the first axis in its non-rotation direction opposite to its rotation direction is calculated from a load torque applied to the second axis in its rotation direction, a moment due to a second-axis-side self-weight, and a ratio of a distance between the rotation centers of the first and second axes, to a distance between the rotation centers of the second axis and a tool. A deflection amount indicating an angle of the first axis tilting in the non-rotation direction is calculated from the moment applied to the first axis and the rigidity of the first axis in the non-rotation direction. A control value is corrected based on the deflection amount to control the robot.