# Optimization of robotic spray painting trajectories using machine learning for improved surface quality

## TL;DR

This paper uses machine learning and experiments to optimize robotic spray painting for better surface quality and efficiency in manufacturing.

## Contribution

A novel predictive model is developed using machine learning and statistical analysis to optimize robotic spray painting parameters.

## Key findings

- Viscosity and temperature are the main factors affecting thickness deviation in spray painting.
- Speed and temperature together influence surface roughness outcomes.
- The predictive model achieved high accuracy with R² values of 0.9224 and 0.9707 for surface roughness and thickness variation.

## Abstract

The production process needs spray painting particularly within automobile manufacturing since product painting accuracy establishes product quality. The combination of hand spray techniques produces intricate designs as well as small quantity needs yet industrial robots excel at painting large industrial product orders. Taguchi Design of Experiments (DoE) is used to investigate the effect of six process variables which included spray distance along with pressure, temperature, humidity level, speed and viscosity rate. Experiments were conducted via industrial robotic spraying with subsequent statistical evaluation through ANOVA tests and regression calculations. The research shows that viscosity together with temperature stands as primary influential factors for thickness deviation, yet speed and temperature jointly determine surface roughness outcomes. The predictive model performed with substantial accuracy based on its ability to achieve R² values of 0.9224for surface roughness measurements and 0.9707 for thickness variation determination. The study offers clear guidelines for practitioners to enhance their processes to produce high-quality products and time efficiency.