An Evaluation Framework to Assess Autonomous Navigation Linked to Environment Complexity

Abstract

The evaluation of mobile robots performing autonomous navigation has recently gained significant attention, not only by the research community, but also by companies wanting to deploy robotic services into the market. In this work we present a framework for the assessment of autonomous navigation performance referenced to the environmental complexity in which the robot operates. The framework consists in selecting a set of metrics, proposing guidelines and comparing in the same test scenario theoretical optimal and measured trajectory performances while linking them to the environment complexity. Supporting the framework, two innovative algorithms are employed: one for trajectory planning and the other to assess the environment complexity. Initial results show the applicability of this approach.