Logo of Acerta
Book a demo

Case Study

Automating end-of-line tests with AI

Objectives

Improve EOL testing for electric power steering systems to catch more defects and reduce warranties.

Challenge

To perform as well or better than its human EOL testers, which meant a false negative rate of 0% and a false positive rate of <1%.

Key Results

  • identified electric power steering systems most likely to fail end of line testing with almost 100% accuracy.
  • Achieved the objective of a 0% false negative rate and <1% false positive rate
Electric Power Steering Systems

99.99%

Faulty systems identified

0%

False negative rate

<1%

False positive rate

Background

A leading Tier-1 supplier of driveline components was looking to improve its end-of-line (EOL) testing for electric power steering systems. The goals were to reduce warranties by identifying more defective systems and to replace their manual EOL test with an automated solution.

Problem

  • The existing end of line testing regime used data gathered from four vibration sensors over the course of 8 tests, for a total of 96 signals per unit tested. The results were then manually evaluated by a team of engineers.
  • The customer requested a machine learning model that would perform as well or better than its human EOL testers, which meant a false negative rate of 0% and a false positive rate of <1%.
  • Acerta’s training dataset consisted of roughly 700 electric power steering systems, none of which failed end of line testing. Our data scientists used 5-fold and 10-fold cross-validation plus 27 failed units to test their classification model.

Solution

Targeted Machine Learning Approach
  • Acerta’s data scientists began by analyzing the customer’s manufacturing process and existing data collection, focusing specifically on their use of signal processing for EOL vibration data.
  • Building on this foundation, Acerta engineered additional features and applied machine learning models optimized for handling complex vibration patterns.
Accelerated Model Development
  • Leveraging experience from previous projects involving vibration data, Acerta reconfigured proven machine learning models.
  • The approach accelerated deployment by avoiding the need to develop new models from scratch, ensuring rapid time-to-value without sacrificing accuracy.
Cross-Application Insights
  • By transferring learnings from a successful gearbox failure prediction project, Acerta enhanced model performance for electric power steering systems.
  • This cross-application insight enabled Acerta to deliver a solution that not only met but exceeded the client’s expectations for predictive accuracy and reliability.

Results

  • Acerta succeeded in identifying electric power steering systems most likely to fail end of line testing with almost 100% accuracy.
  • Acerta achieved the objective of a 0% false negative rate and <1% false positive rate
  • The result was an automated EOL test that performed comparably to (or possibly even better than) the client’s human testers.

Leveraging insights from similar projects, we delivered near-perfect accuracy

"Meeting the 0% false negative target meant deeply understanding the vibration data and the system. We engineered features from 96 signals, applied advanced signal processing, and fine-tuned classifiers using cross-validation. Leveraging insights from similar projects, we delivered near-perfect accuracy with minimal false positives."
Ning Jai
Principal Data Scientist

Case Studies

Proven Impact