Software
Linux 2.0.30 Installed-1/20/98
BTTV
module installed and functional 1/23/98 (The
first image!)
I have been poking with the VisLib package from Activemedia Robotics. It's fantastic, and the demo programs are relatively easy to modify for alternative uses. I recommend it to anyone experimenting with Linux-based robot vision systems. Great work guys! 2-10-2000
I have a basic version of the WebBICS fron-end working, but noticed a few problems. The control system is far too crude without sensors to feel comfortable using it anywhere than the lab. 1-30-2000
I have begun work on the WebBICS front-end. This version is intended to be simple enough to drive the robot with a simple client-side imagemap. The robot-side NPH CGI programs initially will be a front-end to the blue-move library, and Bt848 framegrabber library. As more sensors and software intelligence layers are added, the CGI will need to be modified to interface with the top-level control system. 1-8-2000
After several days of poking with the individual subroutines that send the serial commands I tried setting all of the variables that were of type byte to type char (these were set as integers before this experiment). The result is a slew of compiler warnings about type conversion, and the ability to cause the motors to turn as requested! The dissapearing verbose logging is still happening, but the commands are being sent. 11-29-1999
The cause of the read errors ended up being a simple transposition of the read success flag, the the read bit. It is now possible to query both PIC-SERVO boards, and load the servo gain, velocity, and acceleration profiles onto the boards. Something appears to be amiss with the start_move() subroutine as it is called by the blue_move-pstest demo program. Simple verbose logging just stops working for some commands, weird... 11-11-1999
I had limited success with the blue-move Linux PIC-SERVO control library last night, and was able to detect both of the PIC-SERVO boards on the Blue Cube. A few checksum errors are popping up, and the board type is being reported incorrectly. This is probably due to incorrect data type conversion with the status packet. Getting closer...11-8-1999
Last Updated: 1/8/2004 Brian Rudy