This type of thing could actually spark a robotics revolution. One of the large barrier to entry to robotics is that it is generally built on costly specialty hardware running costly specialty software.
Some people have used laptops as 'brains' for robots but they are bulky and often still quite costly.
Because of their size, long battery life, gps support, camera, touchscreen, microphone, accelerometer and potential lower price, smartphones are surprisingly well suited as commodity platform for robotics.
I disagree on "sparking a robotics revolution". It is akin to suggesting that Songsmith sparked a music revolution. Just because something is easier to do does not make it revolutionary, nor does it necessarily advance the field.
The barriers to entry that you speak of do not exist, unless you consider an Arduino, a motor shield, some sensors, cables and a breadboard costly speciality hardware. Take a look at www.sparkfun.com or www.pololu.com, all the things that come with an iPhone you can get in form of easily programmable modules. Need GPS? Buy a gps module. Need wireless communication? Get BT or XBEE.
Sure, building a robot this way is not as simple as plugging in your smartphone and dragging and dropping some blocks thorough the GUI, but it is not that difficult either, you learn more, and it does not get boring quickly.
Smartphones have orders of magnitudes more computing power than Arduino. Arduino is a nice learning platform but you will reach its limit rather quickly and have to start over with a new more expensive platform. Smartphones have enough power for your robot to evolve to using advanced control algorithms, advanced kinematics and mechanical modeling, advanced image processing (computer vision), speech recognition, text to speech etc...
I agree that maybe this 'Romo' project is too simple to do much. I am not a fan of these 'graphical programming' platforms but it's just a start. I think using mobile phones for robotics has much more potential to advance the state of the art than Arduino except maybe as part of that Android-Arduino combination Google did.
Revolution may be overstating the case, but taking a technology that is accessible to enthusiasts with the technical ability to program an Arduino and construct a robot out of components and making it accessible to users with the technical ability to plug in a smartphone and use a GUI opens up the technology to a whole new audience. Expanding the number of people interested in robotics is a good thing.
Just as I'm sure there are many people doing serious work in robotics now that got their start with Mindstorms, if this comes to fruition and is marketed successfully, I'm sure there will be many people who get their start with a simpler-to-understand smartphone robot now and later move on to bigger and better things.
I understand the sentiment for a getaway robot, however, I believe there are better alternatives for introducing people to robotics such as Mindstorms that you mentioned or Arduino. To clarify, by "better" I mean such that actually get you building and making things, rather than what to me at least boils down to playing with a toy.
It may be just the matter of personal preference, but to me this project is closer to the RoboSapien than it is to, say, making things with LEGOs or Arduino. It does not aim to teach anything about the mechanical aspect of robot building, connecting wires, soldering. Simplifying robotics for the sake of attracting everyone to it may be diminishing returns because at the point where a huge crowd is attracted not much of actual 'robotics' is left.
For those really interested in building robots, picking up a micro board and some parts online would be much more educational. Also, ability to program micros, connect wires and put parts together is something that robotics enthusiast should strive to learn, it's a big part of the hobby!
This being said, I still wish them best and hope they will succeed with their project. After all it is a cool ting to build! :)
> To clarify, by "better" I mean such that actually get you building and making things, rather than what to me at least boils down to playing with a toy.
To me, this is one of the attitudes that relegates robotics to the background. Tinkering together hardware is one small part of robotics, but even if you're interested mainly in other aspects you are forced into it.
I'm very interested in the autonomy and programming aspects of robotics rather than in soldering. A ready-to-roll platform that allows me to write my own code is exactly what I want. I've learned enough to design and build my own robot (motor drivers and the rest of it) and have done so, and I realize that this part is something I could hand over to pretty much anyone with an EE degree, or better yet just buy the hardware platform pre-assembled. I can do it but I'd rather not, because this isn't where I personally can do anything truly interesting.
There's room for hardware tinkerers and software people, and people in between. There's no reason for a one-size-fits all attitude that mandates that you need to assemble hardware to be "into" robotics.
That's actually an awesome point. A robot construction kit that you could connect to a phone would be a cooler way to really learn about robotics than a delivered, finished robot.
Even if it were modular -- so that you learned nothing about connecting wires, soldering, etc. -- having the experience of getting to choose how you build your robot and what sensors and drive train and such that you use would make a huge part of the experience. It's definitely fundamental to what makes Mindstorms so valuable.
This also true for other areas of embedded computing.It certainly offers a costs reduction and a lot less work on the hardware.
But the biggest impact would be on the software side. By offering a standard platform with alot of performance to spare at low cost, Embedded software could be written in higher level languages(lua/lua-jit seems interesting), use much more standard libraries, and higher level libraries.
So we might see a creativity boom in embedded devices, just as we saw when scripting languages started being used for web development.
I already got to play with it, its super fun and works just like you want it to. It's versatile enough to go over cracks and stuff in doorways and the video streaming is very smooth. This is an awesome project to back.
They are using the audio jack to send command data to the robot base. Notice how the iphone is not connected through a dock or 30-pin connector. Very clever indeed!
Making a "real" iphone accessory through the 30-pin connector is simply not possible for hobbyists/small outfits. This also makes their communication interface with the robot somewhat portable. I like :D
From what I've heard, negotiations with Apple took a lot of time and were quite "complex". Also, the cable unfortunately costs 80$ (because of the custom, "secret" DRM chip enabling the interface). Which goes to prove that the audio jack goes a long way!
The hardware required to prototype the Bilibot is a lot more expensive than to prototype one of these Romos, so I don't get what they need 6 times the funding of the Bilibot project for. $78 per Romo seems fine, but the overall goal seems a bit high.
This was my first thought as well. I built a similar robot in terms of physical design and motorized capabilities a couple years ago for ~$400. It wasn't controlled by a phone, which will admittedly adds some complexity, but I really can't see their overall hardware costs, including prototypes, getting into the 10's of thousands. Is the expectation here that this is also covering a living wage for the developers? It seems like they would be more successful with a lower funding request with the plan being to make a profit once the device hits market.
EDIT: It looks like they aren't even applying a solder mask to the board, which is a really cheap way to improve quality and reliability. For an extensible product that they want people hacking, I'd be spending the few dollars to help minimize short circuits. From my experience, everyone who has skimped and not gotten a solder mask applied to their board ends up regretting it down the line.
HNer, Matt Might, did a quick hack (weekend-type) to control a scribbler robot via BlueTooth via smartphone a while back . Given that ROS (Robot Operating System) is being ported to Android, the capabilities for these systems are going to explode! Soon enough, you'll have phone-based robots that have all the accouterments: 3D perception, localization, etc.
Done! I've made several half starts at robotics projects, mostly based on Arduino but I always felt exasperated at how weak they were compared to the smartphone in my pocket, which I can already program over a zillion standard interfaces. Getting an Arduino kit with the equivalent of this functionality would cost hundreds of dollars ... this is a steal.
seriously, I don't get it. There are tons of off-the-shelf electronic kits that already do this (ARM, PIC, AVAR, etc)... And most of this is just simple logic. Follow a light beam, follow color, bla, bla, bla... Anything really useful, to me, appears one would need a bit more than a smart phone. Honda's robot comes to mind and that quadraped that supposedly is suppose to help military carry heavy things through tough terrain.
It's not about price, it's about complexity. Plugging an old iPhone into a chassis is something anyone could do, even my 6 year old son. Programming an AVR from scratch is something many adults cannot do.