I'm not sure if its just a frustrated childhood desire to live with R2-D2, but I'm a firm believer that autonomous cars will prove to be the first real, personally-interactive robots that humans will deploy in large numbers. Which means we need to stop thinking of them just as cars. There seems to be at least one engineer who agrees, and he's building his own experimental car called "Atticus."

Paulo Camasmie is that engineer, and he's a very impressive guy. Camasmie came to the U.S. from Brazil in 2000 with almost nothing but his engineering skills and experience. He's been an automotive engineer working for Chrysler in Detroit, before later moving back to Brazil.

He had an idea for a sort of recumbent tricycle, and started building them in his apartment. He's still building recumbent trikes, but now has an actual company, Catrike, which happens to be the biggest maker of this particular kind of trike.

I'm sure if you're part of the recumbent tricycle subculture, you'll know exactly what I'm talking about. And there must be some of you out there, because his company is cranking out about 4,000 of these things a year.

Camasmie was successful with his people-powered vehicles, but as a gearhead, he was always interested in making cars. His first "good" car, he told me, was a gold Volkswagen GTI, a limited edition one (one of 2000) he had to write a letter to be allowed to buy. So, it makes sense that a car would be the next step.

Camasmie was inspired to create a primarily autonomous car because of his memories of a car accident his sister was involved in in 1984:

Simone had a terrible car accident and stayed a month in coma. She is fine now, but not without paying a huge price her whole life dealing with her disability. A lot is said about traffic crash fatalities, but little is said about injuries that can also be incredibly devastating to individuals and families, and a huge cost to society.

As a result, active safety technology and car-to-car communication are key to the Atticus' design. And while there's no arguing safety's great and wildly important, but it's not that exciting to talk about. That video up there of ordering a car around is much more compelling.

Physically, Atticus is still in a very early state of development. It's all electric and has a very practical one-box design. It's using two off-the-shelf motors with a custom-designed electronic differential with reduction gears. The combination is good enough to get the car to 65 mph or so.

Also, note the central location of the battery pack; it's designed to be able to be slid out of the back of the car for rapid replacement. This concept has been suggested before, and so far hasn't met with much success, but it's such a potentially good solution to long battery charge times and concerns about battery longevity that I'm glad to see it incorporated here.

But, of course, that's not the cool part of that video. What's cool is that he's talking to the Atticus, and it's responding.

Now, pretty much any major auto company could have done something roughly like this. But they haven't because they're still thinking of their future autonomous cars as cars that drive themselves, as opposed to robots that happen to be able to drive you places.

The Atticus is being designed differently. Once you eliminate the need to drive, why would we need to interact with our cars the same old way at all? Just talking to a car and telling it what you want it to do seems natural and so basic I can't believe we haven't seen a major manufacturer concept car to do it yet.

I still think I'd want a more flexible interior design that wouldn't be locked into the usual four-seats-facing-forward restrictions of current cars, but that's me.

"Take me to the horse fighting arena," you'd tell your car. When you arrive, you'd say "Go pick up food from that Dutch place on Cadaver Street, then get my medicated salve from the drugstore, then come pick me up at 4:30." And off it'd go. Because it's your robot.

Also novel are the array of deployable racks for bikes and other leisure equipment that Atticus can deploy. A flexible transport robot like this absolutely should be able to handle all these sorts of things.

The name, by the way, comes from here:

This is how it happened. Since I grew up in Brazil, I had never read "To Kill a Mockingbird". However, as my daughter was reading it here in the US, she turned to me and said that I should read it. When I asked her why, she answered that it was because Atticus Finch reminded her of me so much. I asked, how so? She said that he was a man of his time, fair, and highly ethical. So than I read it, and obviously fell in love with the character.

Camasmie's plans include selling the Atticus for around $25,000. That seems wildly optimistic to me, at least in the near future. And, it's clear there's a lot of development yet to happen. Right now, Atticus is a bare-bones framework for a lot of good ideas. It's a proof of concept, and there's a long way for it to go before it can be considered to be something plausible for mass street-legal use.

This isn't news to Camasmie, and he has plenty of drawings and plans for what a production-ready Atticus might look like. He's not seeking money or investments yet, but is rather looking for an established manufacturing partner.

I'd hook him up with someone at Apple, since this is quite close to my vision of a future Apple car, but they never return my calls.

Atticus is a remarkable project and vision for what an autonomous car could be, without even considering that it's something a guy put together in his garage in his spare time. I have no idea how far along this project may get, but that little voice-activated demonstration is very provocative.