Why industrial robots struggle with absolute accuracy, how laser trackers and iGPS close the gap, and what published trials have achieved.
Industrial robots are very good at returning to the same place. They are much less good at going to a place they have never been taught. For aerospace drilling, machining and assembly, that difference decides whether a robot cell works.
Repeatability is not accuracy
Robots have traditionally been programmed by teaching: an operator jogs the robot to a position and it repeats that path. This relies on good repeatability, but it is slow and impractical for low-volume, high-mix work. Offline programming from CAD removes the teaching step, but it demands good absolute accuracy, which standard robots lack [1].
A study from Loughborough University and partners described the underlying problem plainly: articulated robots are attractive for drilling, trimming and machining because they are flexible, cheap and have a large reach, but low stiffness and low positional accuracy have limited them to non-critical parts with loose tolerances [2].
How external metrology closes the gap
The established answer in aerospace is to add an external measurement system, such as a laser tracker or indoor GPS, and feed its data back to the robot. As early as 2011, researchers noted that most such systems still required the robot to stop and wait for a position update, with work under way on real-time control [3].
Published results show what this can achieve:
- Laser tracker guidance reduced the average absolute position error of an aviation drilling robot from 0.72 mm to 0.16 mm, as reported in a 2024 review of the field [1].
- A calibration approach combining kinematic modelling with a learned error model, measured with a laser tracker, improved a six-axis robot's average positioning accuracy from about 0.85 mm to about 0.05 mm [4].
- The Loughborough work showed that even a lower-cost three-degree-of-freedom tracker, which cannot correct tool orientation errors, improved performance enough to justify the investment in drilling and machining trials [2].
Choosing an approach
- Offline calibration suits cells where loads and temperatures are stable. It is cheaper to run but degrades as conditions change.
- Online compensation keeps correcting during the process. It costs more and needs clear lines of sight, but holds accuracy under changing loads.
- In both cases, the measurement uncertainty of the metrology system must be comfortably smaller than the accuracy you need from the robot.
References
- Positioning accuracy improvement for target point tracking of robots based on Extended Kalman Filter with an optical tracking system. Robotics and Autonomous Systems, 2024.
- Real-Time Laser Tracker Compensation of Robotic Drilling and Machining. Journal of Manufacturing and Materials Processing, 4(3), 79, 2020.
- Experimental comparison of dynamic tracking performance of iGPS and laser tracker. International Journal of Advanced Manufacturing Technology, 2011.
- Absolute Positioning Accuracy Improvement in an Industrial Robot. Sensors, 2020.
