Tobias Föhst

{RAVON} -- The Robust Autonomous Vehicle for Off-road Navigation

Abstract: This chapter describes the work of the Robotics Research Lab at the University of Kaiserslautern in the field of autonomous off-road robotics. It introduces concepts developed for hazard detection, terrain classification, and collision-free autonomous navigation. As an example of a system implementing the described techniques, the mobile off-road robot RAVON is presented. Experiments have been carried out to prove the effectiveness of the approaches.

A fuzzy approach to low level sensor fusion with limited system knowledge

When facing the need to perform low-level sensor fusion with only very limited knowledge available one has to come up with an alternative to the well known and proven Kalman filter. A very interesting candidate for such applications is the so called fuzzy (or soft) voter. This algorithm makes use of fuzzy logic principles to fuse signals in an efficient way and provides a figure of merit as well as sensor monitoring capabilities with very moderate demand for computation performance and memory. In this paper a computational efficient alternative implementation of soft voting for embedded applications is described. Furthermore, its performance is examined using scenarios typical to harsh operations environments using simulation.

Detection of Field Structures for Agricultural Vehicle Guidance

This paper introduces a model based detection approach for typical structures in the agricultural environment. In contrast to other existing approaches in this area it exclusively relies on distance data information generated by a laser scanner which makes the detection robust against varying illumination. Additionally, the method was designed to be easily adaptable to different agricultural structures as well as to minimize computation power. To test the performance and the capabilities of the presented approach a prototypical baling assistance system was implemented where a tractor had to follow a straw windrow while the implement created round bales.

An overview on framework design for autonomous robots

Abstract Robotic software frameworks have major impact on development effort and quality of robot control systems. This paper provides a condensed overview on the complex topic of robotic framework design. Important areas of design are discussed – together with design principles applied in state-of-the-art solutions. They are related to software quality attributes with a brief discussion on their impact. Based on this analysis, the approaches taken in the framework Finroc are briefly presented.

Perception of Environment Properties Relevant for Off-road Navigation

In this paper a set of physical properties is presented that can be utilized for save and efficient navigation in unstructured terrain. This set contains properties of positive obstacles, i.e. flexibility, shape, dimensions, etc. as well as properties of negative obstacles and ground, i.e. slope, carrying capacity, slippage, etc. By means of these properties a classifier is developed that supports the discrimination from traversable to non-traversable areas. Furthermore, an overview of different sensor systems, that can be employed to determine some these properties, is given.

Principles in Framework Design applied in Networked Robotics

Abstract In recent years, we developed a considerable range of networked robotic applications based on the software framework Finroc. Such systems often require integrating diverse communication technologies, protocols, and computing platforms. In this context, beneficial concepts and approaches applied in the framework are introduced — and their positive impact on flexibility, interoperability, and development effort of resulting applications is discussed. Furthermore, applications are presented and tool support is illustrated.

The Application of Design Schemata in Off-Road Robotics

This paper presents a novel concept that combines action-oriented and perception-oriented concepts for the development of robot control systems with the aim to close the semantic gap between the two approaches. Furthermore, the paper describes a representation scheme for a generic modeling of environment data from arbitrary sensors. Formal abstraction and fusion methods are introduced, and their application in the control system of a real robot is demonstrated using the example of the autonomous off-road vehicle RAVON.