Testing robot controllers using constraint programming and continuous integration

作     者：Mossige, Morten Gotlieb, Arnaud Meling, Hein 

作者机构：ABB Robot N-4349 Bryme Norway Simula Res Lab Lysaker Norway Univ Stavanger N-4036 Stavanger Norway 

出 版 物：《INFORMATION AND SOFTWARE TECHNOLOGY》 (信息与软件技术)

年 卷 期：2015年第57卷

页      面：169-185页

核心收录：

学科分类：08[工学] 0835[工学-软件工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Norwegian Research Council under the Industrial PhD Program Certus SFI grant ABB Robotics 

主　　题：Constraint programming Continuous integration Robotized painting Software testing Distributed real time systems Agile development 

摘      要：Context: Testing complex industrial robots (CIRs) requires testing several interacting control systems. This is challenging, especially for robots performing process-intensive tasks such as painting or gluing, since their dedicated process control systems can be loosely coupled with the robot s motion control. Objective: Current practices for validating CIRs involve manual test case design and execution. To reduce testing costs and improve quality assurance, a trend is to automate the generation of test cases. Our work aims to define a cost-effective automated testing technique to validate CIR control systems in an industrial context. Method: This paper reports on a methodology, developed at ABB Robotics in collaboration with SIMULA, for the fully automated testing of CIRs control systems. Our approach draws on continuous integration principles and well-established constraint-based testing techniques. It is based on a novel constraint-based model for automatically generating test sequences where test sequences are both generated and executed as part of a continuous integration process. Results: By performing a detailed analysis of experimental results over a simplified version of our constraint model, we determine the most appropriate parameterization of the operational version of the constraint model. This version is now being deployed at ABB Robotics s CIR testing facilities and used on a permanent basis. This paper presents the empirical results obtained when automatically generating test sequences for CIRs at ABB Robotics. In a real industrial setting, the results show that our methodology is not only able to detect reintroduced known faults, but also to spot completely new faults. Conclusion: Our empirical evaluation shows that constraint-based testing is appropriate for automatically generating test sequences for CIRs and can be faithfully deployed in an industrial context. (C) 2014 Elsevier B.V. All rights reserved.