The On-Ground Autonomy Test Environment: OGATE

This paper presents an ongoing work focused on providing a general testbench to assess the performance of autonomous software controllers for robotics platforms. Currently, it is not possible to objectively analyze the performance of such systems due to: (i) the lack of well defined metrics to compare the performance and (ii) a framework that allows to automatically perform large testbenchs of different controllers under the same platform and conditions. For these reasons we are working on the OGATE framework that tries to overcome such deficiencies providing a general and customizable testbench environment. In this regard, OGATE allows to monitor and control the execution of different planning and scheduling systems over a common robotic platform while collecting relevant metrics of the execution. Also, OGATE not only provides this capability in nominal conditions; it is possible to define different scenarios with dynamic goal injection or with execution failures to stress autonomous controllers, allowing to verify the behaviour in such conditions. In this paper we present the OGATE functionality and methodology, and the results obtained after testing a particular deployment of the GOAC controller for a set of increasing complexity problems of the planetary exploration mission case study.