Field Robot Event 2014
One of the winners of the Field Robot Event 2014, a team from the University of Prague, used in their Eduro Maxi HD robot CANopen as in-vehicle network. The robots of the 23 partipating teams had to master five different tasks.
During the annual Field Robot Event the robots have to perform five tasks, which are related to agriculture applications. For the first two tasks (basic and advanced navigation) the robots require proper sensing in order to achieve accurate and fast navigation between crop rows. In task 3 the agricultural application will be visible by letting the robots detect weed plants and create weed maps using positional information from a GNSS. For task 4 always two teams will be asked to let their robots work together to show cooperative abilities. With regards to last contests the team or robot cooperation was highly appreciated. In 2014, the popular discipline Freestyle was conducted as task 5. In task 4 and 5 the teams are totally free in to present a robot performance based on their own innovative ideas. As during previous contests the general use of a GNSS system was not allowed, because the focus shall be on relative positioning and sensor-based behaviors. However, in 2014 GNSS was used in task 3 for weed mapping (absolute positioning) and on team request in task 4 (Collaboration) and 5 (Freestyle).
This year, the event saw two overall winners: The Eduro Maxi HD from the University of Life Science in Prague (Czech Republic) and the Phaeton from the University in Siegen (Germany). The third place went to the Great Cornholio robot from the Highschool of Osnabrueck (Germany). The Eduro Maxi HD robot is a three-wheeler robot with a differential drive. The modular robotic platform weighs about 15 kg and measures 38 cm x 50 cm x 56 cm. It uses a stepper motor with adaptive control and belt transmission. The power supply is provide by two Pb batteries. The central control unit is a single-board computer based on AMD’s Geode processor. The EIA 232 port is used to connect an adapter to the CANopen backbone network, which links the motors, the optional compass, and several I/Os to the Linux-based host computer. The more bandwidth requiring sensors (IP camera with fish-eye lens and the LMS100 laser scanner from Sick) are connected directly to the PC via Ethernet. The main application software is running on the singe-board computer under Linux and the Xenomai real-time extension. The motor control is provided locally in the drive units. The host computer just sends speed commands via the CANopen network.
Also other participating robots in the Field Robot Event 2014 are based on CAN networks. The Idefix robot by the pupils research center of Ueberlingen (Germany) uses two Maxon motors connected via a CAN network to a Linux-based PC. The Finnferno robot by the Aalto university implements an Isobus-derived CAN network. It links the Windows-based host controller with the units, which controls the steering and driving individually for each wheel. Additionally, a separate trailer communicates the stop request at detected weed plants via the CAN network. By the way, the Finnish team achieved the second place in the freestyle task.
News and reports