Integration testing for robotic systems
- PDF / 2,741,191 Bytes
- 33 Pages / 439.37 x 666.142 pts Page_size
- 72 Downloads / 248 Views
Integration testing for robotic systems Maria A. S. Brito1 · Simone R. S. Souza2 · Paulo S. L. Souza2 Accepted: 7 October 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Conventional approaches for the validation of robotic systems include simulations and functional software testing. Although these approaches are interesting, they do not offer coverage information to guide the testing activity. In this sense, the introduction of coverage testing to the validation of mobile robotic systems seems to promise. This paper proposes a integration testing approach for robotic systems in a simulation environment. The approach is composed of test scenarios, constructed based on functional testing, and structural testing towards revealing defects and providing a better systematization of the testing activity in this application domain. The proposed approach focuses on systems developed in ROS, the robot operating system, in which communication can be established through a publish/subscribe interaction schema. The paper provides an example of the way the integration testing approach can be applied to robotic systems. An experimental study conducted with developers of mobile robotic systems evaluated the gains and proved the applicability of the approach to the industry. The results confirmed its advantages for the integration testing of mobile robotic systems. It can check functional behavior (test scenarios), or reach a high structural coverage (coverage criteria), and reveal common defects in mobile robotic systems. Keywords Software testing · Integration testing · Test scenarios · Testing criteria · Robotic system · Experimental study
1 Introduction Mobile robots use locomotion mechanisms to move unboundedly and autonomously in their environment, without assistance from external human operators Tzafestas (2014). Their mobility makes them suitable to applications in several environments, as ground * Maria A. S. Brito [email protected]; [email protected] Simone R. S. Souza [email protected] Paulo S. L. Souza [email protected] 1
Departamento de Ciência da Computação, Universidade Federal de Lavras, Lavras, Brazil
2
Instituto de Ciências Matemáticas e de Computação, Universidade de Sao Paulo, Sao Carlos, Brazil
13
Vol.:(0123456789)
Software Quality Journal
(wheeled mobile robots—WMRs, and legged mobile robots—LMRs), air (unmanned aerial vehicles—UAVs), and water (autonomous underwater vehicles—AUVs). The design and development of robotic systems are complex tasks. Researchers have created a wide variety of middlewares to manage complexity and facilitate the prototyping of software for experiments, which have resulted in many robotic software systems currently used in both academia and industry Kramer and Scheutz (2007). Communication middlewares can provide information for a distributed network or a single device. Among such middlewares, ROS (robot operating system) has been widely used as a framework that supports distributed computing and component-based infrastructure
Data Loading...
