Abstract

Robots working in real-world applications need to be robust and reliable. However, ensuring robust software in an academic development environment with dozens of developers poses a significant challenge. This work presents a testing framework, successfully employed in a large-scale integrated robotics project, based on continuous integration and the fork-and-pull model of software development, implementing automated system regression testing for robot navigation. It presents a framework suitable for both regression testing and also providing processes for parameter optimisation and benchmarking.