How the system will "see"

My plan for the system to see consists of a small matrix of cameras looking down into the area, with enough cameras to cover all blind spots. Places such as supermarkets will be split into de-centralised processing units about the size of an aisle where a server would "control" all units within this area but no other area, the reason for this is two-fold, it offers better performance over a centralised processing approach in a situation where performance is vital and its modular design means that breakdowns are not so detrimental to the system in the area as a whole compared to a breakdown of a central server. The servers would be linked though, not for "seeing" purposes but for information purposes, the system would setup as a distributed system where each server informed its neighbours what features/interests/products where within its visual boundary so the systems could know where you desired destination is and devise a route through each area to your destination, much like how a network routes packets to their destination.

Exact details of how this works will be presented later

Other features the system will have for seeing will include a link up with pedestrian crossings where the system can inform users when it is safe to cross and also a link up with home computers where users can set their plan for the day and the system could make route plans for that, etc