Freeform surface profiling by iterative learning-extremum seeking control

• Published in: International journal of advanced manufacturing technology Vol. 132; no. 11-12; pp. 5633 - 5646
• Main Authors: Nguyen, Phuc, Lee, ChaBum
• Format: Journal Article
• Language: English
• Published: London Springer London 01-06-2024; Springer Nature B.V
• Subjects: Accuracy; Advanced manufacturing technologies; Algorithms; CAE) and Design; Computer-Aided Engineering (CAD; Control algorithms; Control methods; Engineering; Error analysis; Industrial and Production Engineering; Lasers; Machine learning; Manufacturing; Mechanical Engineering; Media Management; Metrology; Misalignment; Optics; Original Article; Path planning; Planning; Reference signals; Sensors; Surface geometry; Tracks (paths); Uncertainty; Metrology tool path planning; Cosine error; Surface metrology; Extremum seeking control; Freeform surface; Iterative learning control
• ISSN: 0268-3768; 1433-3015
• DOI: 10.1007/s00170-024-13698-2

This paper presents a metrology tool path planning algorithm using an iterative learning-extremum seeking control (ILESC) scheme that combines iterative learning control (ILC) and extremum seeking control (ESC) schemes. In many surface metrology systems, it is inevitable to avoid cosine error that introduces extra uncertainty to the measurement. Cosine error results from the angular misalignment between the measurement probe and the target surface. The proposed ILESC scheme is suitable to address this problem because, unlike the classical control method where the system tracks a given reference signal, ILESC simultaneously compensates for cosine error by simultaneously regenerating the metrology tool path while scanning the target surface, even in case of initial profile unknown. Since no initial geometry profile is given, ESC is deployed to adaptively track the shortest measured distance to the target, which approximately aligns the measurement probe in the perpendicular path to the target. Then, ILC is deployed to each set of measurements to ensure the algorithm converges to an acceptable error threshold given the system’s uncertainty. The proposed scheme was implemented to measure the surface profile of a reference cylinder (ϕ25.4 mm) and freeform lens mold, and the effectiveness of ILESC was evaluated by comparing the measurement results and reference data. As a result, the implemented ILESC scheme showed a 0.32% deviation at full scale.