“Autonomous vehicles will gain ground in agriculture and increase efficiency and productivity. To make autonomous vehicles operate safely, it must be provided with the ability to perceive and interpret the elements of danger in the concerned environment”, assesses Anders Petersen, robot consultant of the Danish Technological Institute's Centre for Robot Technology.
Petersen is leader of the European project QUAD-AV (Ambient Awareness for Autonomous Agricultural Vehicles), which the Danish Food Industry Agency under the Ministry of Food, Agriculture and Fisheries has granted funding for. Partners are the research institutions Fraunhofer (Germany), Cemagref (France), and the University of Salento (Italy) as well as the Danish Technological Institute and the tractor and harvester manufacturer Claas.
The project focuses on developing sensors and methods for data processing, allowing vehicles to become aware of their surroundings by detecting obstacles such as differences in the landscape, physical installations on the field and humans and animals that are present in the environment.
The project partners combine selected sensor technologies to achieve combined and enhanced effect. In order to enable the vehicles to find their ways autonomously, four types of existing sensor technologies are combined: stereo vision, thermography, radar and LADAR (aka LIDAR). The sensors are mounted on an autonomous tractor, and the technologies are tested in relevant scenarios. The gathered data forms the foundation for development of new techniques to process and merge sensor data with an eye to detection of obstacles in the field.
"In the project we examine which sensors and data processing techniques are best suited to detect different kinds of obstacles such as different soil characteristics, objects such as fences, humans and animals", explains Rainer Worst, project supervisor at Fraunhofer IAIS The institute contributes its LADAR 3DLS sensor to the project. The laser-based sensor scans the surroundings of the vehicle as generates a three-dimensional scatter diagram.