Bachelor and Master Theses

Title: Automatic log analysis in industrial robot automation systems
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description: Introduction

The department PRVS within ABB Robotics Products is responsible for software customer support and development validation of the industrial robot systems. These tasks are important regarding quality assurance and customer satisfaction. The purpose with the validation is, by finding weaknesses in the system early in the product cycle, to increase the quality of robot system and the compatibility between the different products within the ABB Robotics family. This results in reduced number of faults in at customer installations, increased possibility of proactive quality measures and fewer problems for the customers. For ABB Robotics it means lower cost of poor quality (COPQ), a stronger trademark regarding quality and a more competitive product. The validation team tests the system from a customer perspective in an environment similar that of the customer, securing system quality in new releases of software and hardware.

The software in the controller part of the robot system is written in C/C++ on top of VxWorks (real-time operating system). The software for the human interface unit is written in C# and is handheld with a touch panel and a joystick. The software that handles robot motion and interaction with other units, such as IO units, are written in RAPID, an ABB Robotics developed high level language for robot automation control. For testing and validation ABB Robotics use a pearl based framework, WebAST, for parsing test cases and executing command towards the external robot interface, RobAPI.

Description

Several different logs can be extracted from an industrial robot system. Some of the information can be hard to understand and it takes a lot of time to manually sort out the important information needed for troubleshooting or analysis. Therefore a tool that automates this work saves a lot of time and increases the product quality by automatically finding abnormal events and data in the system. Statistical analysis of the log data can be used for trend analysis, showing the status of a robot system for a specific time period as well as other statistical analysis. This information should then work as a basis for proactive fault prevention when developing new features or for troubleshooting.

The thesis consists of identifying key information in the logs extracted from the robot system, such as memory usage etc. A framework for automatic log analysis and storage should be established and implemented in the existing validation framework. The log analysis tool used can be a commercial product on the market or an internal product developed for this purpose. The thesis also consists of developing validation scenarios, using RAPID, or test cases, using WebAST, for both long-term and short-term statistical analysis.

Scope: D-level for 1 person (or C-level for 2 persons)

Expected result

• A report discussing the log analysis concept and methods as well as evaluating and identifying important/interesting/unnecessary information in the existing robot system logs.
• Implement and define a log analysis concept within the existing validation framework that should extract, store, evaluate and present the information using a commercial or internal log analysis tool.
• Implement validation test scenarios using log analysis, according to previous work, as a component in the validation framework when identifying and finding system anomalies and problems.
Company: ABB Robotics, kontaktperson: Daniel Gustavsson
Proposed: 2008-01-21
IDT supervisor: Ingemar Reyier
ingemar.reyier@mdh.se, U3-045
Examinator: Bjrn Lisper
Bjrn Lisper
bjorn.lisper@mdh.se, +46-21-151709

Rapport och bilagor

Size

Senaste uppdatering


  • Mälardalen University |
  • Box 883 |
  • 721 23 Västerås/Eskilstuna |
  • 021-101300, 016-153600 |
  • webmaster |
  • Latest update: 2017.03.25