Talk:Subsumption architecture
From Wikipedia, the free encyclopedia
I've just deleted this:
- This AI model may have been invented by Alexandre Parodi, and was first used in a real robot as part of the FMC/CEL (now United Defense) AVTB/AGVT program in 1984[citation needed].
-
- J. J. Nitao and A. M. Parodi (1986) "A Real-Time Reflexive Pilot for an Autonomous Land Vehicle", IEEE Control Systems, vol. 6, no. 1, February, pp.14-23.
Here is the actual quote about Nitao and Parodi's architecture from that paper. Anyone who reads any paper actually on Subsumption Architecture will see how far off this is immediately.
- Currently the FMC architecture constis of a Planner, Observer, Mapmaker and Pilot, and the vehicle control subsystems. Included in the Observer is the sensor subsystem... The system is now mainly hierarchical, but will become more hetrarchical as each expert shares information from its area of resposibility with the other subsystems. (p. 14)
This sounds like a cross between GOFAI and blackboard architectures. Arguably, SA is like the latter. But all the main points of the SA are missed except for those shared with any modular system. Also, the publications came out the same year so presumably the systems were both in development at the same time. There is no evidence this work was known in 1984.--Jaibe 22:47, 30 October 2006 (UTC)
[edit] disambiguating modules
Someone deleted the modules link & called it disambiguating. I looked at the ambiguous modules page, and at least half of the links there are relevant to the use of modularity in AI, so I think it makes sense to link to the whole page. Unless someone wants to start a special page on modules in AI, but I don't know what that would have beyond modularity in programming, maths and mind.--Jaibe 22:26, 29 November 2006 (UTC)
