Slashdot Mirror

Yes it is possible (and often preferable) to learn in a pure software emulated
environment.

Start with turtle graphics. While old school this is where many have started
their programming journey.

Simulators that model real systems are critical to the design and maintenance
of all manner of real world systems. This is what many video games are...
i.e. they are simulators of real or imaginary systems.

Modern graphics invites a 3D turtle graphics environment where ants
can place blocks and build bridges to navigate turtles over.

Yes a Raspberry Pi is a wonderful learning tool.
It is possible to explore almost any programming language you can name.
And yes there is a Turtle Graphics application set.

The big value of a SBC like the Raspberry Pi is all the levels are open enough
for any level of software tinkering and they are easy to recover if your hacking
adventure steps on the OS. The logic of the Raspberry Pi is low voltage
but it is very easy to add LEDs for small change. A current limiting resistor and
an LED cost small change. The schematics of the Raspberry Pi shows how the
onboard LEDs have been interfaced.

Look at QEMU -- it is a very interesting simulator and tool kit.

Big powerful robots are expensive but the simulation
tool set is a necessary layer that any robotics project will need.
Without a good simulation expensive hardware becomes expensive junk.
https://www.willowgarage.com/p...

And if you make and document your progress there are individuals and companies that
will fund a project in areas lacking schools, funding and infrastructure.

Back in February, IEEE Spectrum reported that Willow Garage was shutting down, which led to a rebuttal from WG in which they said that they were changing, not shutting down. I guess the change wasn't profitable enough.

Sheldonbot is a Texai Remote Presence System from Willow Garage, so recreating Sheldonbot is just copying some other company's product.

A "Robotics" project sounds way to generic, A little more detail on your end goal would help you focus better. If you want a premade solutions with all interfaces I'd start here Lego Mind storms If you want to try your hand at control algorithms without spending a penny I'd start here (sharp learning curve) http://gazebosim.org/wiki/DRC/Install If you want to visually do something with your robot i'd start here, various boards and controls are included. http://www.roborealm.com/ If you want a bit more advanced hardware I'd start here http://www.ros.org/wiki/Robots For pure visual processing fun, this actually is rolled into ROS and DRC sim i believe http://opencv.willowgarage.com/wiki/

In fact is says right on the announcement page that "Beam is no robot". So no need to be 3 laws compliant.

I'm still looking for a use case for telepresence robots. It needs to be a situation where all of these things apply:

1) I need to "freely" move around where I'm not. There are lots of situations where I would want this. However the situational awareness of these things is very poor. I drove one around and ended up rolling around the Y-Combinator offices without knowing it. For a tour of a place, office, factory, photos and handheld video would be preferable. (It's very difficult to "look around" with these thing. turning is slow.)

2) I don't need to touch anything. This is sort of the breaker. If I don't need to touch anything, why not just teleconference? Yeah, teleconferences kind of stink, but they do work. (And I can screw off during the parts that don't concern me. Try doing that in person!) If I actually need to do things then I need to be there in person with my arms and hands and fingers.

3) Movement is completely unrestricted where I need to be. Doors are all automatic. No stairs. No elevators.

4) Someone has the money to spend on these. They're expensive. People tend to abuse them, which makes them a maintenance problem.

5) No one cares about the Uncanny Valley. These things are deep in it and people react not positively to them. People hit the Emergency stop button to make the telepresence go away, people sit their drinks on them. Or push them around, pick them up. Drop them. Or just ignore them. People don't react to them like they're people. And there's really no way in the near future to get them out of this valley.

The closest I get for this is a factory tour in China. (For people who live a long distance from China.) But frankly if I need to take a factory tour I have the money to do it, and it would be worth my time to fly there and do it in person.

For anything else it seems like "Skype on a stick" is more than good enough. Does anyone have a legitimate (ie: not "it's cool!") use case for telepresence bots like this?

Project 1: Do some AR projects. Can use OpenCV, http://opencv.willowgarage.com/wiki/. Lots of project ideas out there. Project 2: Put up a camera in each class and record the class to have 'OpenCourseWare' - see MIT. Of course a lot of permission to get, but would help in multiple ways. If students miss a day of class, they can catch up. Good way to evaluate teachers, or they can watch them self to improve.

Sounds like you've got it all figured out. Since no one else has thought of this, or precisely your specific "isn't that hard" implementation, I take it you are actively working on the system of which you speak?

Please let me know when I can download the beta version of your product. I would also like to subscribe to your newsletter.

I would like to see this scripting environment of yours, it sounds revolutionary... I would settle for just being able to write & run scripts myself... if only I could write the scripts faster than I can click the mouse...

For now I'll just keep typing things like:

http://google.com/search/?q=AI

http://google.com/search/?q=site:slashdot.org+AI

http://slashdot.org/index2.pl?fhfilter=ai

-- or --

Just using the search bars on those sites homepages...

Then a JS URL bookmarklet like this:

javascript:var a = document.getElementsByTagName( 'a' );var b = 0;
for ( var x in a ){ if ( a[x].href.match( /\bAI\b/ ) ) window.open( a[x].href );
if ( ++b >= 6 ) break; }
void(0);

(open the first 6 links that have AI in the URLs, probably should be a[x].innerHTML.match(\bAI\b), meh...)

But honestly, I usually just middle button click on the interesting article links (open in new tab) faster than I could write the script or explain to my robot what I want.

I think the reason no-one has this "robot" system you speak of, is that the computer gives you what you want fast enough -- hint: "[ctrl+L]googe.com[enter]AI[enter][mid-click][mid-click][mid-click][mid-click], browse away... IMHO, adding a generic all powerful robot layer would make that more complex than it needs to be.

Maybe I'm wrong -- I look forward to being proved so, having a computer system that satisfies my deepest desires without me having to use any input has only ever happened When I use XP -- It regularly reboots itself (update, crash, whatever) and my boot-loader boots the GNU/Linux partition by default (The computer somehow knew I'd rather be using Linux than MS/Windows).

Perhaps you could use Rhino to create a JS environment with the functionality of Java.awt.Robot, and OpenCV to interpret the screenshots for the AI -- now if only you had an AI to feed the data to....

You are making it a lot more complicated than it has to be. It's not hard already to do certain parts. The problem is it would take a lot of code to do it right. You don't need to use screen shots because we are talking about strings here.

It's not difficult to work with strings and regular expressions. There are a lot of capabilities already. The problem is there isn't a unified framework or backend to make it simple enough that everyone could do it.

And no I haven't decided that this will be MY project, but if nobody sees it as valuable, and I do, then at some point I will write some code and see what can and can't be done. Robots are easy to write, and so are webcrawling robots. Screen shots aren't necessary. It would have to rely on a framework from which applications run on top of.

Python scripting for example can already allow a lot of stuff. The way unix is designed, applications can communicate with each other in that the output of one application can be made into the input for another. If the framework is designed under the methodology that all applications should be able to communicate with each other, then all robots should also be able to communicate with each other. If each robot is sufficiently specialized, you can have fairly complex operations broken down into highly specialized simple tasks.

It's easy to write a bot which bro

Sounds like you've got it all figured out. Since no one else has thought of this, or precisely your specific "isn't that hard" implementation, I take it you are actively working on the system of which you speak?

Please let me know when I can download the beta version of your product. I would also like to subscribe to your newsletter.

I would like to see this scripting environment of yours, it sounds revolutionary... I would settle for just being able to write & run scripts myself... if only I could write the scripts faster than I can click the mouse...

For now I'll just keep typing things like:
http://google.com/search/?q=AI
http://google.com/search/?q=site:slashdot.org+AI
http://slashdot.org/index2.pl?fhfilter=ai

-- or --

Just using the search bars on those sites homepages...
Then a JS URL bookmarklet like this:

javascript:var a = document.getElementsByTagName( 'a' );var b = 0; for ( var x in a ){ if ( a[x].href.match( /\bAI\b/ ) ) window.open( a[x].href ); if ( ++b >= 6 ) break; } void(0);
(open the first 6 links that have AI in the URLs, probably should be a[x].innerHTML.match(\bAI\b), meh...)

But honestly, I usually just middle button click on the interesting article links (open in new tab) faster than I could write the script or explain to my robot what I want.

I think the reason no-one has this "robot" system you speak of, is that the computer gives you what you want fast enough -- hint: "[ctrl+L]googe.com[enter]AI[enter][mid-click][mid-click][mid-click][mid-click], browse away... IMHO, adding a generic all powerful robot layer would make that more complex than it needs to be.

Maybe I'm wrong -- I look forward to being proved so, having a computer system that satisfies my deepest desires without me having to use any input has only ever happened When I use XP -- It regularly reboots itself (update, crash, whatever) and my boot-loader boots the GNU/Linux partition by default (The computer somehow knew I'd rather be using Linux than MS/Windows).

Perhaps you could use Rhino to create a JS environment with the functionality of Java.awt.Robot, and OpenCV to interpret the screenshots for the AI -- now if only you had an AI to feed the data to....

They're probably doing this to compete with Wilow Garage, which has their own open-source robot operating system.

"What the Kinect Could Be, But Probably Won't" -- Been there done that -- TFA should be called, "What Kinect & LIRC hackers have realized is really lame way to control a TV or computer."

The article says a qwerty keyboard in the living room is a bad idea, without explaining why. So, why?

Thanks.

Because Dvorak is so much nicer.

On a serious note, I don't see keyboards going away any time soon (or ever). I can type almost as fast as I think and 8 times as fast as I can get my voice recognition software to recognize.

What I am seeing more of is Computers. Everywhere. In portable phone & tablet form factors, as mp3 players, as game consoles, set-top boxes and routers... Even in the dash of some cars.

Once we realize that TVs are just big computer screens, and a general purpose "desktop" computer can perform all the tasks that we currently use the set-top boxes for, it won't seem too strange to just use your keyboard in the living room. Google TV already does this... For typing in a search or composing text/emails, nothing beats a keyboard. If I'm near my computer, I use it to send text messages.

Hell, I even have a wireless USB keyboard hooked up to my XBox360 -- It's much quicker/nicer than the overpriced controller mounted keyboard.

We'll always need a pointing device -- I prefer a Wacom pen-tablet/mouse pad, but I could see a Kinect filling this role. In fact, I've used my Kinect to control the mouse pointer, but the CPU usage is ridiculous when you consider how little my Wacom uses and how much more precise it is.

As for Kinect controlling the TV -- Well, I've done that. It wasn't that hard. I've been using LIRC to control my TV with Linux for quite some time. Linking LIRC to a gesture recognizer (libFreenect + OpenCV) was a piece of cake, but not really worth it. The Kinect is far less efficient and precise than either my truly universal remote (which I use to control both the TV & computer with via LIRC), or a simple keyboard / mouse combo. Seriously though -- WAY too much CPU consumption when you consider how little an IR remote, keyboard or mouse/pen tablet consumes...

Lots of robots do use 2 eyes (cameras) for 3D vision. http://opencv.willowgarage.com/documentation/camera_calibration_and_3d_reconstruction.html

Mobile robots just aren't very good yet. But progress, after decades of frustration, is now rapid. Willow Garage is making real progress. Their mobile robot can already fold towels, starting from a pile of randomly placed towels. When it can change a bed, they'll have something useful.

My guess is that the killer app for this will be a mobile robot for hotels that can clean a room and reset it for the next occupant. Give this ten years.

Then I spent part of the afternoon, along with some others, watching the video replays of it and the unfolding tragedy in a conference room by Hans Moravec's Mobile Robot Lab, all the time hoping it was just a misunderstanding, and the astronauts were all right or something.

One of the hopes of some at the Robotics Institute was that robots could do more of the space exploration more safely, including preparing the way for humans. Was that really a quarter century ago? :-) Well, the robots are finally starting to be here:
http://www.willowgarage.com/pages/pr2/overview
http://www.hizook.com/blog/2009/08/03/high-speed-robot-hand-demonstrates-dexterity-and-skillful-manipulation
http://listcultures.org/pipermail/p2presearch_listcultures.org/2009-November/005926.html

Or in some cases, even come and gone, sadly:
http://www.ri.cmu.edu/research_center_detail.html?type=publications&center_id=7&menu_id=262
"Space Robotics Initiative (SRI)
This center is no longer active."

Always wanted to work there and make Hewey, Dewey, and Louie from Silent Running, and the space habitat biospheres they maintain. :-) But that was not exactly their focus.
http://www.pdfernhout.net/princeton-graduate-school-plans.html

That Challenger tragedy was doubly sad with a school teacher on board, considering all the school kids who had been encouraged to watch it. I can wonder if that was part of the further collapse of the US space program?

Still, as much as such tragedies are awful, I later wrote that a big problem with the US space program is that not enough people are taking risks and dying from the consequences. If you think of how many people have died in ocean voyages in the early day of sailing, an active space program seriously oriented to extending human life into the cosmos should be willing to accept hundreds or thousands of deaths a year by astronauts taking calculated and reasonable risks (as in, a 80% chance of success).

The obsession with perfection and zero risk by NASA ultimately seems to have grounded the US space program. That, and an acceptance of overly complicated designs. If astronauts are willing to accept a 20% chance of disaster so they can fly more often (or at all), I say let them. If current astronauts don't want those odds, find new astronauts.

I'm not saying take foolish risks, or 99% risks of death, or risks not worth risking death for. I'm just saying, we probably could be launching 100X as many cheaper rockets and having a lot more success, and having thousands of people going into space every year, if we accepted more causalties (on the order of 20% of launches failing like this shuttle did 25 years ago). Obviously, such a program should be voluntary and people should understand the risks as best as they can. Ideally, over time, the risks would be reduced by better engineering to that of the current risks for air travel in commercial aircraft. But it is just too early to have that expectation.

Besides, and maybe I should not say this, but TV ratings would go up for the space program if NASA did not go out of its way to make everything look so boring with astronauts who have been training for years because there are so few launches and they are so expensive. The most interesting thing I ever saw on NASA TV was when that NASA astronaut lost her bag of tools while fiddling with a grease gun. :-)
http://www.space.com/6131-astronaut-laments

I'm surprised they picked the PR2 from Willow Garage and compared with the Anybot. Willow Garage also makes the Texai robot, which has almost identical capabilities as the Anybot, and fulfills the same kind of role. PR2 and HRP are not designed for offices, but are research robots which are loaned out to universities and other institutions. Neither is designed to be a commercial robot, while Texai and Anybot are commercial products.

Disclaimer: I work for Willow Garage

It is already amazing what can be done with an optical camera.

However using depth images Andrew Johnson did some impressive work on recognising objects in 3D depth maps. And Dan Munoz recently worked on applying this kind of algorithms to Willowgarage's PR2 robot. With your Kinect driver, depth sensors are getting within reach of hobby developers.

Hope they will not be using a Haardon-Like Features face detection algorithm.

OpenCV has C interfaces and there are more that have some C code libraries. Really the coding challenge would be building the wrappers to utilize those libraries with your camera's hardware (I assume provided through CHDK APIs). My vote is for a nifty KLT implementation that allows me to take a video and extract a huge wide pan image in post processing on the camera.

That's very kind....makes my day.... There was also Ada Lovelace on the other side of the card...... Tech details... Ubuntu, python http://www.python.org/ enthought http://code.enthought.com/ opencv http://opencv.willowgarage.com/ yarp http://eris.liralab.it/yarp/ KDL http://www.orocos.org/kdl Gamera http://gamera.sourceforge.net/ Shapely http://pypi.python.org/pypi/Shapely/ few other bits and pieces Hardware: Dell xps13, Intel atom motherboards, 1gig ram, 4gig usb thumb drive, and wifi dongle, router... Arduinos, philips webcam, low cost micro servos, AX12 Bioloids servos, USB2Dynamixel, ikea lights and tables...that's about it..pen and paper...

There are free products for face detection.

There are free products for face detection.

Computation The PR2 robot has two eight-core i7 Xeon system servers on-board, each with 24 GB of RAM, a 500 GB internal hard drive, and a 1.5 TB external removable log drive. The computers and most of the sensors communicate over a 16-port gigabit Ethernet hub with a 32-gigabit backplane. The robot also has an on-board, dual-radio router that can be bridged into a WLAN, as well as a secondary, stand-alone access point for laptop or smart phone access.

Also:

The PR2 ships with sensors in the head, arms, and base. The head contains two stereo camera pairs coupled with an LED pattern projector, a 5MP camera, a tilting laser range finder, and an IMU. The forearms each contain an ethernet-based, wide-angle camera, while the grippers have three-axis accelerometers and pressure sensor arrays on the fingertips. The base has a fixed laser range finder.

That's a fair bit of grunt to throw at the OpenCV libraries, which is listed under their Supported Projects in the Software section. No surprise either, Willow Garage has taken over hosting the project from Intel.

Computation The PR2 robot has two eight-core i7 Xeon system servers on-board, each with 24 GB of RAM, a 500 GB internal hard drive, and a 1.5 TB external removable log drive. The computers and most of the sensors communicate over a 16-port gigabit Ethernet hub with a 32-gigabit backplane. The robot also has an on-board, dual-radio router that can be bridged into a WLAN, as well as a secondary, stand-alone access point for laptop or smart phone access.

Also:

The PR2 ships with sensors in the head, arms, and base. The head contains two stereo camera pairs coupled with an LED pattern projector, a 5MP camera, a tilting laser range finder, and an IMU. The forearms each contain an ethernet-based, wide-angle camera, while the grippers have three-axis accelerometers and pressure sensor arrays on the fingertips. The base has a fixed laser range finder.

That's a fair bit of grunt to throw at the OpenCV libraries, which is listed under their Supported Projects in the Software section. No surprise either, Willow Garage has taken over hosting the project from Intel.

Computation The PR2 robot has two eight-core i7 Xeon system servers on-board, each with 24 GB of RAM, a 500 GB internal hard drive, and a 1.5 TB external removable log drive. The computers and most of the sensors communicate over a 16-port gigabit Ethernet hub with a 32-gigabit backplane. The robot also has an on-board, dual-radio router that can be bridged into a WLAN, as well as a secondary, stand-alone access point for laptop or smart phone access.

Also:

The PR2 ships with sensors in the head, arms, and base. The head contains two stereo camera pairs coupled with an LED pattern projector, a 5MP camera, a tilting laser range finder, and an IMU. The forearms each contain an ethernet-based, wide-angle camera, while the grippers have three-axis accelerometers and pressure sensor arrays on the fingertips. The base has a fixed laser range finder.

That's a fair bit of grunt to throw at the OpenCV libraries, which is listed under their Supported Projects in the Software section. No surprise either, Willow Garage has taken over hosting the project from Intel.

Videos of the robot doing some pretty cool stuff. If I was into robotics, I'd definately want to be playing with one of these!

Cell phone makers are using robotic names, and even imagery, for yet another brick with a display. It's getting late for that. Robotics is moving along rapidly now, and there are humanoid "androids" and worker "droids" in small-scale production.

These aren't one-off prototypes. The Kokoro Actroid Der2 is being used by Sanrio, the company behind Hello Kitty and other cuteness. Willow Robotics' mobile utility robot is just now going into small scale production. A few years from now, robotic names for mere cell phones will sound stupid.

(A previous dumb step in that direction was "MicroPhone", a misnamed Macintosh modem program distributed before the Mac had much audio capability.)

Take a look at opencv. Proper training takes a lot of images but I like the results.

Downloaded. NOTE: the as-compiled binaries require the OpenCV libraries of the 110 variety (SourceForge holds the 200 version now). So, get older 110 binaries. From the file list

See the OpenCV Wiki on setting up and checking you OpenCV installation.

I'm still setting up, but I'll post back when I get it working...

Those are just toys, and toys at low price points. They're too weak to do much except look around and transmit video and audio.

Now look at the PR2 Personal Robot. That has real manipulation capabiilty and can be teleoperated over WiFi. Now there's a potential problem.

I agree. What I find most interesting about this is the robotic applications for touch sensitivity and dexterous manipulation.

If robots can "feel" the world, vision would not be nearly so important.

Touch sensors that are as cheap and functional as the one mentioned here will IMHO revolutionize robotic manipulation, and in turn, manufacturing and deployment of manufactured goods in unstructured settings (like doing plumbing).

Insects do really well relying a lot on touch.

While I don't follow that field very closely, it would seem to me that this invention could be game changing.

Also, one of the top priorities of a place like Willow Garage
    http://www.willowgarage.com/
is to make robots that can safely interact with humans an coexist in human space. Touch in an important part of that.

I expect to see a lot more invention along this line.