Test Generation for Robotized Paint Systems Using Constraint Programming in a Continuous Integration Environment

Abstract

Advanced industrial robots usually consist of several independent control systems. Particularly, robots that perform process-intensive tasks like painting, gluing, and sealing have dedicated process control systems that are more or less loosely coupled with the motion control system. Testing the software for such systems is challenging because physical systems are necessary to test many of their characteristics. This paper proposes a method for automated testing of such robot systems. Our approach draws on previous work on continuous integration, combined with constraint programming techniques for test sequence generation. In ABB Robotics' process control system for robotized painting, many tests are only conducted every six months, during the release test. With our automated test approach, we expect to reduce the round-trip time, from code change to test completion, to less than one day.