| KEYWORD |
Application of Interpretable Machine Learning to the Analysis of Errors in Robot Assembly
keywords ARTIFICIAL INTELLIGENCE IN MANUFACTURING, COLLABORATIVE ROBOTICS, DEFECT ANALYSIS, INTERPRETABLE MACHINE LEARNING
Reference persons DARIO ANTONELLI
Research Groups Quality Engineering and Management Group
Thesis type EXPERIMENTAL APPLIED
Description Objective: The goal of this thesis is to develop an interpretable machine learning model capable of identifying factors that cause defects in robotic assembly and proposing modifications to process parameters to eliminate them.
Methodology:
Collaborative Robot Programming: A collaborative robot will be programmed to perform a realistic industrial assembly process.
Data Collection: The process will be executed multiple times, varying the process parameters (speed, force, position, etc.) and recording data related to any defects encountered.
Model Training: The collected data will be used to train an interpretable machine learning algorithm (e.g., decision trees, linear models). The interpretability of the model will allow for identifying which factors are most relevant to the presence of defects.
Modification Proposal: The trained model will be used to propose modifications to process parameters, aiming to reduce or eliminate defects.
Experimental Verification: The proposed modifications will be tested experimentally to verify their effectiveness.
Expected Results:
An interpretable machine learning model capable of diagnosing defects in robotic assembly.
Concrete proposals for modifications to process parameters to improve assembly quality.
A contribution to research in the field of machine learning applied to industrial robotics.
Required skills Statistics
Machine Learning
Notes Thesis will take place in the collaborative robotics laboratory at DIGEP
Deadline 22/07/2025
PROPONI LA TUA CANDIDATURA