Guided Policy Search Based Control of a High Dimensional Advanced Manufacturing Process

TL;DR

This paper demonstrates the application of guided policy search reinforcement learning to optimize control in a high-dimensional additive manufacturing process, achieving promising real-time control performance.

Contribution

It introduces a GPS-based reinforcement learning approach for controlling complex manufacturing processes, integrating simulation and real-time feedback for improved outcomes.

Findings

Successful training of neural network policy using GPS in simulation

Effective closed-loop control with in-situ measurements

Promising experimental results in process control

Abstract

In this paper we apply guided policy search (GPS) based reinforcement learning framework for a high dimensional optimal control problem arising in an additive manufacturing process. The problem comprises of controlling the process parameters so that layer-wise deposition of material leads to desired geometric characteristics of the resulting part surface while minimizing the material deposited. A realistic simulation model of the deposition process along with carefully selected set of guiding distributions generated based on iterative Linear Quadratic Regulator is used to train a neural network policy using GPS. A closed loop control based on the trained policy and in-situ measurement of the deposition profile is tested experimentally, and shows promising performance.