The SP-Freiburg TechX Challenge Entry: Performance Evaluation of a Fully Autonomous All-terrain Robot

Alex Kleiner

The TechX Challenge is a competition firstly held this year by DSTA, Singapore's national authority for defense science and technology. It aims to inspire and engage the development of a system that is capable of one hour autonomous navigation in unknown terrain without any kind of human intervention. It will have to navigate its way through a number of different settings, from an outdoor environment to an indoor urban location. Along the way, the system has to overcome obstacles, negotiate stairs, operate an elevator to reach a particular floor in the building, and to search for and engage designated static targets with a manipulator before proceeding back to the starting point. The robot that accomplishes the mission most efficiently will be rewarded with 470,000 Euros. In this talk I describe the performance metrics that have been applied by DTSA to qualify robots for this mission, as well as the approach of the SP Freiburg team that passed together with four others teams the qualification round. We focus particularly on Multi-Level Surface (MLS) maps- based SLAM, obstacle negotiation and path planning on behavior maps, object recognition from 3D range readings, and manipulator trajectory planning using a probabilistic roadmap planner. For this purpose, we extended the commercial Telemax robot, originally developed for tele-operated bomb disposal, by several sensors for facilitating autonomous operation.