Localisation and Navigation of Autonomous Service Robots

Service robots autonomously cut the lawn and weed in large parkways or sports sites. Tourists are comfortably driven to places of interest or from one place to the other on large fairgrounds. Handicapped and elderly people are supported with their everyday problems. Examples like these are numerous where a service robot is a possibility to support the human being. To solve these tasks efficiently, the robot has to know at any time where it is situated. It is therefore searching for the answer to the question: "Where am I?"

To find an answer to this question, the RTS works on a development of a new navigation method of autonomous systems. A special focus is set on the navigation in unknown and natural environment here. In opposition to indoor navigation, where the environment is structured through straight walls, doors and even floors, the obstacles and objects in outdoor areas are much more complex. The form and surface of the objects (trees, bushes, houses, etc.) is irregular and changeable during the time. Influences such as wind are difficult to be described by discrete mathematical models. However, with the consolidation of different sensors such as GPS, compass, laser scanner, wheel sensors and inertial circles, the RTS is capable of generating independent surrounding models of the working space. With the help of these maps, a localisation is possible to the accuracy of a few centimetres.

Nevertheless, the adaptability of this method is restricted to plane working spaces where there are sufficient natural landmarks. In order to make navigation possible in hilly regions where there are no definite obstacles it is necessary to collect 3D data.

Contact person: Matthias Hentschel

This project aims at the abstraction of the large data amounts during the gathering of three-dimensional data, so that the robot can accomplish a 3D localisation in real time during its journey. Mathematical algorithms as well as the sensor technology have to be developed by the RTS as there are no commercial products for these kinds of requirements yet.

Contact person: Marko Reimer

RTS-HANNA: Based on Kawasaki's Mule 3010 4x4

HANNA's four-wheel-drive and differential locks make it possible to use the vehicle in difficult terrain. HANNA is capable of driving autonomously as well as being controlled manually. Autonomous navigation is based on several 3D and 2D laser range scanners. Additionally, optical und infrared camera systems are utilised. HANNA's positioning is carried out by combining precise DGPS and the ambient data of the 3D laser range scanners. Real-time data acquisition and processing is performed by four on-board computers. [more information]

RTS-STILL – the Autonomous Forklift Truck

The essential innovation of the autonomous forklift truck can be found in the localization equipment:A 3D ceiling localization system enables navigation in highly dynamic environments. Furtherimportant achievements of our project have been the interoperability of the system with the human driver and its flexibility to perform pallet handling also when the pallet position is not precisely known in advance.
[more information]

ANNICA - Autonomous intelligent vehicle for Navigation

Annica is the smallest autonomous system that exists at the RTS. The drive results from two independently controllable wheels. A third supporting wheel gives the system the necessary stability. In addition, there are wheel coders attached to each axis with which the covered distance can be measured. The applications are programmed in C and are uploaded via a serial interface on the micro controller MC68HC11 by Motorola. As an additional sensor, Annica contains an infrared distance sensor, a bumper, photographic resistances and a microphone.

BRITTA - Scrabbling Robot for Applications in Trained Ambit

Britta is an autonomous scrabbling/crawling robot. The six independently moveable legs are controlled by twelve servo motors altogether. Each leg is steered by a motor for back-and-forth movements and by a motor for up-and-down movements. The heart of the system is a micro controller 68332 by Motorola. It processes the applications which were developed in gnu-C beforehand. For the navigation tasks, a customary web cam is attached. Furthermore, Britta is equipped with an infra red distance sensor, a microphone and tactile sensors at its legs.

CARLA - Computer based Robot for problems in navigation

Carla is the first robot that is used especially for navigation tasks in the outdoor area. Equipped with a standard GPS receiver, two correcting data receivers (DGPS), a 360° laser scanner, an odometer and a radio LAN, this system shall be capable of navigating in an unknown environment independently. Carla is driven be four electric motors while two are attached to each other respectively. This results in the same steering method as with chain vehicles so that even turnaround manoeuvres are possible around its own axis. The data processing is done by a low power pc with 233 MHz and a Linux operating system.

DORA - Outdoor robot for navigation

Dora fulfils navigation tasks in outdoor areas as well. The specialty of this system is the possibility of passenger transportation. At a maximum loading weight of 300 kg, Carla can carry up to two persons next a range of sensors.