Semantic Navigation & Explanations (SENS)

Problem Context

There is a need for safer and more explainable autonomous robot systems in the greenhouse domain in which potentially multiple robots and humans collaborate in a tight space. In order to collaborate effectively there needs to be interaction by means of communication and navigation. The future perspective in the greenhouse sector is that these robots each perform their missions in the greenhouse and can navigate and interact with each other and humans using traffic priority rules.

Solution

The objectives of the SENS project are 1) to enable semantic communication for autonomous systems in the greenhouse, and 2) to support semantic explanations to the human operator. Together it forms the ‘Semantic Explanation and Navigation System’ (SENS) that can be integrated into autonomous greenhouse systems. SENS is being developed as a system that is agnostic of a specific robot platform such that integration of the SENS system on the platforms of various robot manufacturers can be easily realised.

Results

In 2023 SENS has been realised as a reusable software tool that generates semantic translations of the communication of events from an autonomous robot to the operator. Based on this communication the operator can judge the reported urgency of an unexpected but foreseeable situation a robot encounters as well as the need to take action.