LearningBoost

AI-based algorithm improves motion control performance of high-dynamic precision motion systems

LearningBoost is an AI-driven servo control algorithm that is purpose‑built for the most demanding, high‑throughput multi‑axis motion systems.  This algorithm delivers dramatic improvements in position error during motion, move & settle times for aggressive motion, and cross‑axis coupling performance across semiconductor, electronics assembly, laser processing, biomedical next generation sequencing and many other precision applications. This state-of-the-art control algorithm delivers a new level of dynamic control capability by achieving significantly higher control bandwidth than traditional servo algorithms, enabling superior performance in applications with rapid motions, compliant mechanics, and reaction‑force‑coupled axes.

Key Features and Benefits

• Multiple operating modes (Robust, Recipe, Cyclic)
• Applicable for both non-repetitive (Robust) and repetitive (Recipe, Cyclic) motion profiles
• Flexible learning and compensation options
• Dynamic online learning modes for in-process improvement
• High-frequency compensation for mechanical resonances and their harmonics
• Reduce non-cyclic errors, including disturbances during motion and reaction-force-coupled axes
• Reduce cyclic errors, including encoder misalignment, and bearing-induced errors
• LearningBoost Wizard graphical user interface simplifies algorithm setup and optimization

Targeted for Demanding Applications

• Semiconductor Advanced Packaging: Die attach, LED bonding, and hybrid/TC bonding processes where fast XY motion creates cross‑axis disturbances.
• Electronics & PCB Assembly: SMT pick‑and‑place operations requiring >50 rapid XY moves per second with minimal following error.
• Precision Inspection & Metrology: Wafer defect inspection and spiral scanning applications needing aggressive disturbance rejection in X, Z, and Theta stages.
• Biomedical:  Next generation sequencing in genomics requires very fast move & settle to minimize overall cycle time
• Laser Microprocessing / Additive Manufacturing, where at least 2 coordinated axes are in continuous motion and following error must be minimized during motion
• Other:
  • Industrial roll-to-roll or raster scan printing