Robot Brain

Robot Brain Project

Jeff and Glenn working on The GNU/Linux Rack

The NetBSD Stack

About

The Robot Brain: The Robot Brain project is large project with the goal of providing a platform to test various robot control theories. It is in it's early stages but conisists of an ever growing number of GNU/Linux and NetBSD based computers networked together. The collection of rack mount computers are strapped together via ethernet to form the fixed brain for various mobile robot bases. We are currently working on getting "agent" based software running on this "Society" of machines. The brain team is convinced that interesting things can be done with the concept of "agents" introduced by Marvin Minsky in his book "The Society of Mind". We need all the help we can get and many friends and members of The Robot Group have already contributed greatly to the project and we owe them our thanks.

GlennCurrie is the project leader.

BrainAgentFrameworks for discussion of our experiences with agent architectures and frameworks. BrainAgentSoftware for software designs, ideas and related topics. BrainAgentJade for JADE specific links and discussion. BrainOrocos for Orocos Open Realtime Control Services/Open Robot Control Software discussion and links. BrainArch for architecture for other-than-agent discussions, including IMA. BrainPlayer for Player/Stage/Gazebo.

March 20, 2003 - Glenn added a 100 Base-T 24 port hub to the Robot Brain network. It is briged to the the 10 Base-T hub that has been linking the net. All computers with network cards capable of running the the increased speed have been moved to the new hub. The link to the Sun/NetBSD systems as well as the wireless link remain on the origninal hub. Glenn is preparing a laptop PC with 802.11b wireless capability to his robot platform. Intial tests were run with the laptop serial port connected to the platform motor control computer. The platform wheels moved as directed. The motor control P.I.D. control software had to be freshly loaded into the motor control computer as the battery backed RAM had not retained the program. The intergal RAM backup batteries seem to have died. Sibyl6 has been the main development machine for the Robot Brain project so far, but seems to have developed some intermittant hardware problems. Copies of user files have been moved to the recently installed Sibyl7 but Jeff continues to have problems getting the agent software to run properly on the new machine. He continues to try to track down the new "null pointer" problem. Sibly1 is on line as well and we are discussing moving the multiport serial card to this machine from Sibyl6. This card is used as part of the system to drive the Babbling Head. The current Sibyl1 is a second generation system in that the rack mount case remains the same but the mother board etc. has been upgraded. This system holds the archives from the origianl Mobile Platform control computer BamBam. As resouces allow, the video link and X-Window / Joystick control code for the Mobile Platform will be ported from machine BamBam to one of the Sibyl machines.

February 27, 2003 - Glenn added new server Siby7 to the Brain. It will require considerable configuration, but will eventaully take over from Sibyl6 as the main server in the system. Thanks to Jeff and Eric for their help in configuration. Many thanks to Ed for systems administration help in testing donated computer parts and his fabrication of custom panels using his nifty LASER cutter.

February 13, 2003 - Haven't posted an update in a while. Jeff has been doing battle with Java flakiness that is causing FIPAOS to crash. We installed another 2 port serial card into Sybil6. All of the PCI slots are now used in Sybil6. Future I/O expansion will now have to occur on other nodes like Sybil1 and be distributed across nodes. Eric is working on getting drivers working for a DataQ DI-194 4 channel A/D converter. Jeff has come up with a new high level data flow diagram for the brain that describes the high level message flow between agents. The team is also investigating the incorporation of GNU Talk Filters to add more "personality" to the brain speech.

January 16, 2003 - The BabblingHead babbles in sync with the DecTalk pretty well now.

January 15, 2003 - Jeff modified the Emotion Agent to only change voice parameters per sentence instead of per word. This resulted in much smoother speech.

January 10, 2003 - Jeff cleaned up some of the Java code to make it easier to test remotely. Glenn found a dual rs-232 PCI card for $5.00 at a local discount store that he installed on Sibyl6. We now have enough serial ports to drive the DecTalk, read the phonetic logging from the host port, and to drive the servo board for on the Babbling Head. This was much easier than trying to modify the cabling for the 8-port card.

December 11, 2002 - CD-ROM burner was installed and baseline backup of project software was placed on CD-ROM. NFS and CVS have been configured and are operational. Work was done on several of the multi-port serial port cards. Some success was acheived but it was determined that the custom (78 pin to 8 DB-25 pin) cable we have, was not designed to work with the card we have installed. The pin out of the cable is available but fabricating the cable would involve hours of work. Other multi-port cards that have been donated are being tested. Most are very old and data on them is hard to find on the net - lots of data on much newer cards is available. Jeff suggests installing Java test and verification utilities as the next major software project. The agent software makes extensive use of Java and therefore being able to verify that Java is working as required will help developers trace and fix bugs.

December 4, 2002 - Every project needs a plan, so the team members each took an action item to work on. Glenn will be getting the HP CD burner installed on the cluster and helping to set up NFS up on the nodes. Jeff will be setting up CVS and working on NFS. Eric will be installing the 8-port serial board.

November 20, 2002 - The Goodwill computer store had a barrel full of free power cords. This was our lucky day! We picked up enough of them to allow all of the Sparc NetBSD systems to all be powered up at the same time.

November 14, 2002 - Jeff was able to iron out the difficulties with the speech agent. It now freely speaks whatever is entered via the agent GUI.

November 13, 2002 - Jeff got the first agent program running under the FIPA-OS! It successfully loaded up the dectalk Java library and was able to make the DecTalk say "foo" before it became unresponsive, but it is a great start! Eric got 8 servos on the Babbling Head working and hooked up a terminal on the monitor port of the DecTalk to watch the phoneme speech coming out. Glenn was able to repair the Mini SSC with some fancy solder work and successfully got his ruggedized laptop patched into the mobile platform drive electronics.

November 10, 2002 - EricLundquist worked on the Babbling Head. Replaced 4 burned out servos. Reworked the Mini SSC II servo controller. Hopefully the Babbling Head will put a face to the voice of the DecTalk.

November 7, 2002 - JeffHurst got the FIPA-OS installed on Sybil6.

November 6, 2002 - EricLundquist got the DecTalk Java Linux interface working on Sybil6.

October 28, 2002 - GlennCurrie brought up Sybil2 and got the DecTalk mounted on the Linux rack. He also got the monitors, keyboards, power, etc all nicely mounted on the main rack. JeffHurst got Java 1.4.1 loaded on Sybil6.

October 27, 2002 - GlennCurrie got Sybil1 up and running with Red Hat Linux 8.0.

October 23, 2002 - GlennCurrie got the two UPS' for the Linux rack rebuilt and hooked up. We now have clean power!

FIPA The Foundation for Intelligent Physical Agents.

Professor Michael Wooldridge's homepage http://www.csc.liv.ac.uk/~mjw/ Especially 'My WWW links'. It is extensive.

Professor Nick Jennings's homepage http://www.ecs.soton.ac.uk/~nrj/

Intelligent Agents: Theory and Practice, Michael Wooldridge & Nicholas R. Jennings. Knowledge Engineering Review 10(2), 1995. pdf or html.

Marvin Minsky Bio http://education.yahoo.com/search/be?lb=t&p=url%3Am/minsky__marvin

The Society of Mind from Amazon http://www.amazon.com/exec/obidos/ISBN%3D0671657135/102-1116262-3243342

This is a place to gather some related page links. Ones for which we do not have enough to say so as to warrant another page.

The World-Wide-Mind project: "researchers construct their agent minds and their agent worlds as servers on the Internet. 3rd parties will use these servers as components in larger systems". also Mark Humphry's site, and SOML: Society of Mind Markup Language.

ABLE: A toolkit for building multiagent autonomic systems from IBM Research. "conforms to the FIPA 97 specifications" "...agents cooperate and compete to take control of the intelligent system and make the appropriate response, much like Minsky's The Society of Mind. The base agent architecture is shown in Figure 16. The architecture contains a central pool of subsumption agents similar to that of Brooks with sensors and effectors..."