Continuing with the basics, I took a TIG welding class at the TechShop.
I, along with three others, learned how make a basic weld on steel sheet metal.
After donning the appropriate protective gear (safety glasses and gloves), we proceeded to watch our instructor show us the ropes using the TIG welder.
The TIG welder had the following:
the power base (we used -75KV DC), plus foot pedal (like your typical sewing machine pedal but for serious power)
argon gas cylinder
electrode tip (tungsten)
grounded workbench
The argon gas is used to create micro cleanroom where the weld is being made. This protects the molten metal from getting oxidized. The hose is attached to the electrode assembly, which makes it a bit unwieldy (_that’s_ why it was so hard for me…)
Welding takes very fine hand movements to do it right. My first attempt ended with mixed results as I rushed through one section due to hand fatigue (I’m thinking RSI is probably common in the welding field). My second attempt with filler rod (think solder) scored low in my book. To nail the procedure down, I should go back and do a bunch of practice welds.
I won’t be welding any titanium bike frames, but next time I make a weld, I’ll be sure to photograph the results.
After listening to a number of Churchill Club events broadcasted on public radio, I finally attended an event in person at the Grand Hyatt in SF.
The event, ‘Best of the Best in a Growing Robotics Industry’, was a panel with four leaders in robotics:
Colin Angle, CEO & Co-founder, iRobot Corporation
Caleb Chung, board member & co-founder UGOBE, Inc.
Soren Lund, Senior Director, LEGO Group
Sebastian Thrun, Professor of Computer Science, Stanford University
Most of us probably know one or more of the companies, but how many of us actually have a robot at home? That was one of the questions posed to the audience by the moderator Elizabeth Corcoran, Sr. Editor, SV Bureau, Forbes Magazine. Only a handful of people raised their hands.
The answer to that question is very enlightening. Why don’t we have more robots in our lives? Let’s take Roomba’s history. Colin stated that iRobot was founded in 1990, but it took cycling through 18 business plans before hitting upon the Roomba model. Not to mention that iRobot also sells robots to the military which, if I recall correctly from an earlier chat last year with an iRobot employee, was (still is?) their bread and butter.
Eighteen business plans later, iRobot managed to find a problem with a solution that people will pay for. What makes the iRobot commercially-viable? Throughout the evening, Colin (as well as Caleb) mentioned that a significant design constraint is that the product has to cost less than X dollars. It’s no wonder then that most of us don’t have robots roaming around ours homes, handling our day-to-day chores we’d love to offload.The cost of the hardware has been dropping dramatically, and the software has gotten better, but it takes a special combination of cost, function, and emotional appeal to get a successful robot launch.
Emotional appeal?
Many of the robots I know in popular culture are emotional beings (e.g., Marvin the Paranoid Andriod,C3PO, Rosie, …), and so why it is surprising that emotional appeal plays such a strong role in designing and/or using a robot? Colin said that around 60% of Roomba owners have named their vacs. He went on to relate the story of one Roomba owner. who named her Roomba Russie(sp?). When Russie failed to run, she phoned the iRobot support line in a panic and was upset to find out that in order to get Russie back on its wheels, it had to be sent back in. She couldn’t fathom the idea that she’d have to pack Russie in a box and ship it away.
Caleb described the Cleo (the playful, AI-based dinosaur that he designed) as having the heart of toy. Caleb thinks of the Cleo as a robotics platform rather than a toy, but the Cleo has toy written all over it. One look at it and you can’t help but think of it as that pet dinosaur you have always wanted.
(Heard in background is Caleb Chung)
Sebastian, the leader of the Stanley team that won the DARPA competition, was on the opposite end of emotional spectrum when it came to robots, at least those that have your life in their hands. He said that he didn’t want to get to his car and have it throw a temper tantrum because he hasn’t seen it in five days. I couldn’t agree with him more.
All panelists agreed that a successful robot has to do one thing well—whether it be cleaning, being a platform to express creativity, providing some emotional connection, or to chauffeur one around.
There are so many varied applications of robotics that one cannot simply categorize the industry. The important thing to keep in mind is that each will have its own strengths, limitations, and hurdles. For example, consider Stanley. Sebastian mentioned that the concept of a self-driving vehicle is not a new. Most airline passengers have all been flown by an autopilot system. But getting people use to the idea that a car can drive them from one place to another has a number of hurdles. Sebastian pointed to incremental features in place today, such as the self-parking Lexus, or lane departure warning systems. More features such as these will come over time, and at some point we may be able to wrap them all together into one neat package and call it a robotic car.
Another idea that came up was that the robotics industry is where the PC industry was 20-25 years ago. However, the panelists added that the economics are much more difficult for a couple of reasons: platform incompatibilities and that Moore’s law doesn’t apply to robotics.
Key Advances in Robotics
Elizabeth asked what each of the panelists what advance has been key.
Colin: a good business model
Caleb: meshing of art and science
Sebastian the ability to build a robot that can learn
Soren: price component
Killer App in Robotics?
Colin: Assisted Living
Sebastian: Driving
Caleb: Robots as surrogates for caregivers
Soren: Robotics woven into curriculum
One of the questions posed by an audience member brought the panelists back full-circle: Why is it that I don’t have a fully-autonomous robot that can clean my room? Shouldn’t this be possible now in this day and age, asked one early 20s member. Another audience member characterized three of the presenters’ products as ‘gadgets’ that weren’t tackling the hardcore issues such as computer science.
The panelists responded by saying that these aren’t merely gadgets and each product has its own set of really hard problems. For example, in the case of the Roomba, is it as smart as it can be? Colin responded by saying that obviously it is not, but if it were as smart as they could make it, it would be prohibitively expensive. For the Cleo, Caleb said that just getting a child-proof layer of skin to cover the robotic assembly is a big challenge, not to mention adding the software/processors and electro-mechanics underneath. For Lego Mindstorms, it’s about making the platform as open to tinkering as possible. In the case of Stanley, getting across the dessert without manual intervention was a very hard problem (the next challenge is to do it in an urban environment).
I must admit that robotics did look easy to me at one point. But once you begin to put your thoughts into action, you’ll run into all sorts of issues. That’s also what makes it a lot of fun. I hope to capture my experiences with the issues, and the fun, here.
Building the Akrobot requires more power tools that I have available at home, so I signed up for a class at the TechShop in Menlo Park.
I took the Milling Machine Safety and Basics Usage class Tuesday evening. The goal of the class? To mill a block of aluminum from 0.5″x0.5″x1.25″ down to 0.400” x 0.400” x 1.000”, +/- 0.005”, all while keeping your limbs intact.
My instructor, Alec, started off with some of the safety basics such as:
Always wear safety goggles!
Don’t wear gloves.
No loose clothing.
etc…
If you are planning to take this class, I highly recommend that you pick up an apron to wear over your clothing. I wore a t-shirt and jeans, and wouldn’t have minded an apron to wipe the oil off of my hands that I picked up from using various tools.
