Specification and Execution of Robotic Acceptance Tests for Object Sorting
摘要
Ensuring reliable performance in tasks with high variability is an increasingly important task in the engineering process of robotics applications. This paper introduces an extended toolchain leveraging Behaviour-Driven Development to specify acceptance criteria and high-quality simulations to execute acceptance tests. We demonstrate the effectiveness of the testing approach in a dynamic sorting task with objects picked from a moving conveyor using a Franka Panda robot system. The contributions include an extended domain-specific language for specifying acceptance criteria and an implementation for automated testing of sorting scenarios using NVIDIA IsaacSim. We conclude with a discussion on the current state and future work on acceptance testing for robotics.