Hello DaveX, thanks for the post.
I've not been on that forum for many
years. But I guess what I said is
still relevant.
Reactive subsumption robotics is really
a deadend, and mobile robots need to
incorporate a lot more intelligence.
The big gains in AI in the past 10 years
seems to have settled on Deep Learning,
which is ok for pattern recognition, but
not intelligence in general.
For your own projects I would recommend
looking at ...
- ROS (lots of available software).
- lidar using something like a Benewake
TF-Mini or Luna. - the 3-Layer Architecture work by Gatt,
Bonasso, Firby, Simmons, Kortenkamp, where
they add layers of intelligence onto
basic subsumption robots.