Datasheet

9. Tests and Results
The FROG robot is currently being tested by the UPO team. IDM and UPO have been testing the
robot to detect and correct problems that can arise in the integration of new components and
systems.
After the Seville integration meeting the robot was left in Seville to allow the UPO team to integrate
their navigation software, construct the test bed navigation map and to test the robot hardware.
Then the robot was brought to the Lisbon Zoo for the second integration meeting and for the Data
collection on the Zoo premises. After the end of the Zoo meeting the robot went again to UPO, so that
they could continue with their tests with the platform in Seville.
The next chapter shows some of the detected problems that are being solved at the current time of
writing.
9.1. Detected Problems and Solutions
UPO and IDM have detected some problems that are being corrected. They are now discussed.
1. Motor, drivers and board lines overheat - UPO and IDM have detected some problems
between the control algorithms and the motor controllers that were producing quick
overheating of the motors, board lines and motor drivers. In the Lisbon Zoo this problem was
studied and UPO introduced smoother acceleration ramps in the manual navigation control
algorithms. With this change in the navigation controllers, it was possible to proceed with the
data collection to create a map of the Zoo. Due to the Zoo ground structure and high
inclination of the paths, the robot had to stop from time to time due to overheating of the
motors. Now the heating problem is being corrected by the introduction of heat sinks with
fans. The change of the robot wheels material and the introduction of very light weight
LiFePO4 batteries is also being considered.
2. Laser Range Finders Position - During the UPO tests, it was detected that the lasers
mounted on the robot platform were too low. The lasers were detecting the floor. This is
being corrected and the lasers will be raised 10 cm from the original position.
3. People tracking computer autonomy - It was detected that the People Tracking computer
autonomy, when running the code at full speed, was too low, about 1:30 hours of battery.
The problem was that the battery that was powering this computer was also powering all the
electronics aboard the robot. The solution was to use the PC3 Battery to power only the
People Tracking computer, the Dalsa stereo cameras and the sound system.
4. Sound system volume - During the Zoo meeting it was determined that the sound system
installed on the robot was too weak and it was impossible to hear the sounds produced by the
interaction system when the robot was outdoors. The solution will be to change the sound
system for a more powerful one.
5. FROG robot platform weight - The weight of the robot is a very important factor that should
be taken into account when designing the robot. It will affect the autonomy, the center of
mass of the robot, the way the robot can travel on different surfaces, the way it can rotate, the
maximum inclination slopes that the robot is allowed to pass or the brake capability. The
FROG robot is a very complex machine with lots of components, computers, sensors and
actuators; to power all the equipment aboard it will need a total of 8 batteries. The normal 12V
17Ah lead acid batteries, that are reliable for this kind of robots, weight about 5.5 Kg to 6 Kg
FROG – FP7 STREP nr. 288235
Deliverable: D1.4 – Platform User and Developer Manual 46