System and method for machine learning-based position estimation used in microassembly control using digital computers

Abstract

This invention provides a method and system for using a hybrid model that includes at least one novel physics-based model and a machine learning model. [Solution] A method for predicting the position of moving micro-objects that takes control loop latency into account, coupled using gradient boosting, wherein the model created during at least one of the stages is fitted based on residuals calculated during the previous stage, based on comparison with training data. The loss function for each stage is selected based on the model being created. The hybrid model is evaluated on extrapolated and interpolated data from the training data and can take into account both deterministic and probabilistic components involved in the movement of micro-objects by including both physics-based and machine learning models. Thus, the accuracy and throughput of microassemblies can be increased.

Images