The idea of this project is that of using different sensor modalities and multi-algorithm approaches to detect the various kinds of obstacles and to build an obstacle database that can be used for vehicle control. For instance, bearing and distance to the nearest collision can be estimated and used by the path planner to change route or to lower the speed if an obstacle is in close proximity to the vehicle's planned path. Road and cliff edges should be handled as special cases since the consequences to the vehicle of breaching a cliff edge are very severe.
The project investigates the potential of four sensor technologies: (stereo) vision, radar, ladar and thermography. Existing state-of-the-art sensors, some previously developed by the partners themselves, will be modified and interfaced in such a way that they can be demonstrated in an agricultural context. The sensors will be mounted on an autonomous tractor and a data acquisition campaign will collect sensor data from a previously selected set of agriculturally relevant test scenarios. These data will provide the basic dataset for the development of novel sensor processing and sensor fusion techniques aiming to detect and classify obstacles in an agricultural environment. Research will focus on feature extraction and selection issues, and on classifier design, using the different sensorial inputs, in order to improve state-of-the art approaches or to introduce innovative methods accounting for real time processing constraints and challenging environmental conditions, such as lighting variations, smoke, dust or fog, presence of living beings. The proposed methods and systems will aim at increasing the overall level of safety of an autonomous agricultural vehicle with respect to itself, to people and animals as well as to property.
Project related results will be disseminated in conference and journal papers and video footage will be produced showing the various sensor types mounted on the vehicle. Also the consortium will participate in fair and conference activities with the purpose of presenting ideas and results from the project.