IMPROVING ACCURACY AND ROBUSTNESS IN ROBOT KINEMATIC CALIBRATION USING A SINGLE DRAW-WIRE ENCODER: A COMPARATIVE ANALYSIS

The demand for high-accuracy industrial manipulators has driven the development of affordable robot calibration procedures. This paper presents a critical comparison of kinematic calibration methods employing a single draw-wire encoder, focusing on the impact of measurement noise and encoder placement on calibration performance. We compare two distinct approaches: a standard optimization-based procedure, incorporating both fixed and estimated encoder location, and a recently proposed step-by-step method. Through simulation, we evaluate the sensitivity of each method to varying levels of measurement noise, demonstrating that the step-by-step method exhibits superior robustness to noise compared to the standard method when encoder location is estimated. Furthermore, we assess the influence of encoder location estimation on overall calibration accuracy, revealing that inaccurate encoder location estimation in the standard method can significantly degrade calibration results, particularly in noisy environments. Our findings provide valuable insights into the practical implementation of draw-wire encoder-based calibration, highlighting the trade-offs between computational complexity, robustness, and the required accuracy of the measurement system. Specifically, we demonstrate that achieving sub-millimeter accuracy requires encoder accuracy below 0.1 mm, regardless of the calibration method employed. These results contribute to the advancement of cost-effective and reliable robot calibration techniques for industrial applications.