The joint team of The University of Electro-Communications,
Tokyo Institute of Technology
[Rover name]
MAICO(Motion Adjustment by intelligent Computing)
[member]
Supervisor: Keiki Takadama
Leader: Takehiro Okawa
Member: Yoshitaka Fujii, Eiki Tokita, Takashi Miura,
Taro Nyuwa, Keiji Sato, Ken Nonami
Observer: Tohru Takada
[Introduction of rover]
Major feature of the rover of our team is to employ PDA which provides a high calculation capability, and various controls. In particular the diagnosis and maintenance of emergency of the rover are possible by using the liquid crystal touch panel of PDA. When some troubles happen, the program on PDA is update by replacing the SD card. In contrast to the elaborate software features provide by PDA, the hardware of our rover is composed of the very simple and low-spec elements. For example the GPS loaded on our over has a normal capability on GPS data precision (which means that GPS data includes the large error). To overcome this problem, our mission aims at implementing the correction program of GPS error, and validation the effectiveness of the program.
[Approach and solution]
Since our team consists of students who have information science background we tackled the problem from the software viewpoint not the hardware viewpoint. In our mission case, it was necessary to remove the the ±8m GPS error. Inspire of employing a new high-spec GPS to overcome this problem, we address a correction of the GPS error with software technique.
[Result]
Unfortunately our rover in this come back competition resulted in no control in both 1st and 2nd try. The 1st try failure is mainly caused the rover has landed in the direction to which the sensor cannot be activate. The 2nd try failure is caused universal joint was broken due to the impact of landing. In the 2nd try, in particular, the big impact of landing occurred because we narrowed down a parachute to anti-strong wind. Our rover strayed in the same place by using raw GPS data. In contrast to the comeback competition, rover moved toward the goal by using our program in ground experiment. Our rover did not move in a try, and we are regret that we could not verify it enough.
[Next challenge]
Although our rover resulted in no control, we confirm behavior of our rover in the ground experiment in black rock desert. Next challenges are improvement of the motor shaft, state mechanism, learning mechanism. We could not take over well in this year. We did not have enough preparation time. Next year, we start by addressing to overcome these problems.
|
