Thursday, August 11, 2011

On computational sustainability

Every once in a while I learn a new phrase - the one I have learned today is computational sustainability, in AAAI there is a track dedicated to it, it turns out that there is a community and even an organization (originated in Cornell), their site has a rotating logo, this is one of their logos.  They define computational sustainability as the way that computer scientists can help in improve the ecology, economics and society - pretty large.    They have a "green" orientation and various environmental projects.   The session I attended today in the AAAI track was dedicated to transportation management --  every speaker emphasized the possible impact on the environment -- if people stay less in cars, the amount of CO2 will be reduced,  but the presentations themselves were quite interesting about intersection control in autonomic cars, avoiding starvation and gridlocks, and another talk was about a system deployed in several cities in Oregon, taking reports from the city (in Portland -- periodic every second, in Eugene - event-driven every time that a traffic light is changed) about the traffic light status, displaying the result on mobile phones for subscribers (sounds dangerous to watch while driving) and advises about desired speed to get the green light and advice to go in alternative road,  if they'll have enough subscribers they'll be able to load balance the traffic.  Cool stuff -- they actually process events.  


BTW - I sat in the conference reception in a table with two robotics guys (they are a lot of them here) that deal with autonomic cars,  they said that autonomic cars cannot be trusted to react quickly enough to dangers, so there is actually a human in every car that takes the control when needed.   So the Google driverless car is not entirely driverless nowadays;   this stems from the current limitations in perception and action speed in robotic technology that I've written about before.      


We can label some of the things we are doing as computational sustainability -- a trend to follow!

No comments: