Science Ain't So Bad

COLLISION DENIAL?

INVISIBLE HAND

Say you’re driving and a brick wall pops into existence right in front of your speeding car. Do you drive right into it?  Or do you jam on the brakes and twist the wheel to avoid a collision? While you’re scratching your head over that tough one, let’s look at one way to  reduce your chances of an accident.

You’ve read about cars that drive themselves right? In two or three years, your car’s gonna be twice as smart as you are. Twice. That’s SOMETHING, huh? While you’re re-reading Fifty Shades Of Grey, your car will take you from point A to point B without crashing into vehicle C .

Sterling Anderson (Mechanical Engineering at MIT) has been concerned that we’re going about this all wrong. He says self-driving cars that follow preprogrammed paths don’t make sense. After all, it’s certainly not what you do in real life.  You’re constantly reinventing how you’re gonna get to the office as you drive . Get squeezed into the left lane? Traffic’s crawling to your right? Might as well stay left till you’re under the bridge (ASSUMING Mister Testosterone will just back off of your bumper). Anderson’s idea is to put in place an electronic “copilot” that doesn’t interfere except when needed. This arrangement lets you do all the driving unless you get out of your “safe zone”‘ If that happens,  Anderson’s “friendly ghost” grabs the wheel. He and his associate from the MIT Mobility Lab, Karl Iagnemma, think you will barely notice. You may even give yourself undeserved credit for an expert maneuver that you couldn’t have accomplished without that virtual tug on the wheel.

Anderson and Iagnemma also point to the cost advantages of their system over a fully autonomous vehicle. Lots cheaper. Less parts. Less complex.

Mister ScienceAintSoBad wishes Sterling Anderson good luck. His impressive new system will, no doubt, win him a well deserved PhD. But he may be underestimating the appeal and robustness of the “competition”. Google’s self driving cars have gone further and with less problems. And the marketing momentum seems to favor Google’s more comprehensive approach.

ScienceAintSoBadRating = 5 . Technological tour de force. Commercial acceptance? Time will tell.

– – – – – – – – –

ADD ANOTHER STEP

ENGINEERING?

Some things are just TOO simple. They should be much more complicated. The Purdue Society of Professional Engineers understands this. You’re gonna blow up a balloon and pop it? In two steps? How about seven? If you’re REALLY good, maybe you can do it in 300 really ridiculous steps.

Don’t forget the sink.

Zach Umperovitch (geology student) was team leader in Purdue’s Rube Goldberg Machine Contest . To his team’s eternal credit, it found the hardest way to bust a balloon.  So proud. Back in the day, East German engineers used to be pretty good at this stuff. It’s how they made cars.

Here’s another project.

Nuts?

See, here’s the thing, A good way to understand the principles of simplification is to de-simplify. A good way to understand how to optimize is to de-optimize. Believe it or not, this is a good way to learn.

Seriously.

Long as you don’t think this is what your next employer is looking for.

– – – – – –

Image credits: The image is courtesy of Purdue University.