Rise of the Machines

I gave this talk at Quanda Convergence on Mar 4th. I was pleasantly surprised with the response the talk received. It was a very good gauge to ascertain if people in general were ready for having more robots in their everyday lives and the answer is an overwhelming YES!

The talk focused on my work in robotics and the kind of questions people asked me when I told that I build robots. While most of them were paranoid as to when we should welcome our robot overlords, others were skeptical if robotics was there yet. We’re somewhere in between today and this talk examined the topic in a lighthearted manner. It touched the various movies that had robots as a centerstage and linked them to real robots or intelligent systems that we have today.

So, when will the robots rise?

 


DAvinCi’s Revival

Aside

I came across this article on IEEE Spectrum today and I was pleasantly surprised to see DAvinCi mentioned in Prof.Kuffner’s presentation (slide 17).

Cloud Computing and Robotics

Prof.Kuffner’s Presentation -
Cloud Enabled Robots

DAvinCi started with the grand vision of providing intelligence cheaply to a horde of heterogeneous service robots and we made good progress in identifying the underlying technologies and in deploying a SLAM demonstration on a single node Hadoop cluster. We were the only group that existed publicly at that point and I believe we were one of the first to publish the concept of using Cloud-computing for Service Robots. My interest in DAvinCi has been resparked and I’m glad that players like Google have finally taken notice. I look forward to more efficient use of computing ahead.

LEO10 in the news

A hearty congratulations to my students for making the media with their recent accomplishment at the Intelligent Ground Vehicle Competition held in Michigan in June 2010. They placed 8th out of 48 teams in the design challenge during their debut.Their robot is featured in the front, while my robot Rio is in the background. You can also see our humanoid, HOAP-3 in the edge of the picture. The team members are Dev Chandan (team lead, mechanical), Hitesh Dhiman (software, ROS expert) and Ankit Sachdev (electronics, embedded systems). The picture was taken in my lab.

This is impressive on two counts -

1. The LEO10 team consisted of just 3 rookie members. Most teams have an average of 8 to 20 members.

2. This is the team’s first robot. Most teams have existed for atleast half a decade.

Given this precedence, I’m excited about the upcoming year’s design and I wish the team the very best.

To read the full article, click on – NUS team’s Robotic Turtle is no slouch

Alternative link

Rio 2 – The Holonomic Telepresence Robot

Rio has been constantly evolving in its application and is currently been outfitted as a Telepresence robot for remote surveillance, patient care, elderly and child monitoring applications. While the fundamental technology behind Telepresence robots are equivalent to web-conferencing, the real challenge lies in delivering enterprise level QoS. We are now evaluating methods for data compression on the fly to minimize latency, increase video resolution and to improve the controller frequency.

A robot’s birthday

Quite possibly my fastest fabrication to implementation for a full fledged autonomous robot. Rio, our holonomic robot, works; flawlessly too.

It was fabricated (mechanical and PCBs), wired, tested and ready to go in less than 12 days, quite a miracle given the failure rate in robotics, the lead time for procuring parts in Singapore and the paperwork involved.

Here’s Rio all wired up and good to go.

Don’t stare at the wheels too long, they make your head spin. (http://www.youtube.com/watch?v=Upio7feAY6w)

The Evolution of Rio

I never imagined that I would be going through so many iterations of Rio. And these are just from the Alpha prototype to the Beta. Also, I could not accommodate the countless other versions that superseded the Alpha. I need to employ Agile methods in my design planning and fix elements of the design in short iterations.

Adapt for Seniors

Aside

It just amazes me sometimes the kind of attention we get at Cornell MineSweeper, sometimes cause of the work we do and other times, well see for yourself.

I am a  73 year old woman, and my husband is 75 and we really like minesweeper for keeping our ability to focus fine-tuned, but we need more time in the learning phase of the game.  Right now, we don’t make errors, but we run out of time before we can win a game.  I wonder if you could set it up  so that we could adjust the time allowed.  That way, we could make progress in two dimensions:  time and accuracy…  Of course, if not, we can always just keep at it, until we can do it in the time that you allot.
Learning new things and doing things that require intense focus is so important as we get older, that it would be great to be able to do it in a game that we like —and win once in a while.
Thanks for considering the suggestion.
Judy & Phil

Maybe it’s time we changed our name?