The scout robots are a simple differential drive base equipped with sonar and bumpers. (Note: the current exemplar we have is not yet fully functional --- you cannot access the bumpers and the odometry is not correct. ) The robot is attached directly to Plutonium which I have moved to the 5th floor of Dunham labs. I'll arrange to give you 24 hour access to this area for the remainder of the term.
Using real robots requires some care and attention to detail. While playing with devices that move and spin is fun, they are generally hard, heavy, and fast. This mean they can hurn you as well as themselves and their environment. Here are a few rules for your own safety and that of the robots:
As noted above, the robots are directly connected to Plutonium via a radio ethernet link. You'll notice that there is a modem attached to Plutonium and another atop the robot. Generally, you should be able to switch on the robot and run programs without any trouble. You might notice that occasionally the first couple of commands get dropped --- this is due to some unavoidable connection delay in the modems. Hence, it is a good idea to include a couple of "null" commands (e.g. tell the robot to move at zero velocity) at the start of the program to make sure everything is up and going before you commence your real work.
In order to have your progam communicate with the scout, you merely
need to recompile using -lrscout instead of -lscout. This library is compatible
with the simulation library in every respect except it'll make the real
robot move instead of the simulated one. Note that you DO NOT NEED NSERVER
to run the Scout. In fact, there is currently no way to use Nserver with
the Scouts.
Occasionally, the robot will not respond even after a lengthy period of time. Generally, this is because of a complete breakdown of the ethernet protocol used by the robots. In this case, do the following:
If this doesn't rescue your program, then something more disastrous has happened and you should contact me or the TA.
If you have comments or suggestions, email me at hager@cs.yale.edu .