Author Topic: Robot Brains: Searching mid-high end units  (Read 3857 times)

0 Members and 1 Guest are viewing this topic.

Offline Avadon77Topic starter

  • Jr. Member
  • **
  • Posts: 15
  • Helpful? 1
Robot Brains: Searching mid-high end units
« on: November 15, 2015, 12:25:46 AM »
Hi everyone, (brand new to the group)

Been a while since I've worked on robotics and my last robot was in early/mid 2000's. Back then I was using a "Handiboard" (they still sell them) which was a pretty decent little device for the time. I scratch built it and it was a very fun project and good piece of hardware. So now I'm back after a decade and boy-oh-boy have things changed. I want to build a 1/5 scale semi-humanoid robot (Telepresence+autonomous modes). Basically a torso on a rover. But in that torso I want to put a very sophisticated and highly integrateable motherboard controller. The Handyboard was as microprocessor with servo integration right on board. That was very cool. Is there anything still like that? Or is that all a separate board?

I looked at Raspberry Pi and beagle board micro computers. Cool devices but seemed a little more for intro beginner/low cost market. I'm looking in the 400-1000$ price range. A controller that really integrates a lot of things and has a lot of room for me to grow into. Would be awesome if I can hook up servos to it or octopus controller boards from it where I can run servos from that. I'd love to know what everyone's thoughts are on cutting edge boards. I do like the ones that can run Linux, Dos, etc. Super cool and makes it easier for my IT background.

Does anyone have any experience/thoughts with any of these?

http://www.robotshop.com/en/single-board-computers.html?dir=desc&order=price

Seems like a lot of this stuff looks good but I'm at a total loss which direction looks best.

Offline Avadon77Topic starter

  • Jr. Member
  • **
  • Posts: 15
  • Helpful? 1
Re: Robot Brains: Searching mid-high end units
« Reply #1 on: November 15, 2015, 01:14:28 AM »
Okay so I just watched this
https://www.youtube.com/watch?v=pDE2qenDXKQ
Mind=Blown  haha
Seems like the answer is MCU vs FPGA and how seriously in depth do you want to go lol

Penny for your inputs

Offline mklrobo

  • Supreme Robot
  • *****
  • Posts: 558
  • Helpful? 15
  • From Dream to Design at the speed of Imagination!
Re: Robot Brains: Searching mid-high end units
« Reply #2 on: November 15, 2015, 06:17:51 PM »
 ;D Hello!
Awesome info on the FPGA on youtube! Thanks!
In reference to where you are going, from where you are, I would look into the saying;
form follows function.
What is the function that you would like to evolutionize to?
I started out with basic stamp(Parallax) to begin, then went on to Raspberry pi 2 and Axon.
I have 2 agendas to satisfy;(besides the limitation of money)
I use the Raspberry Pi and Axon for making a high level diagnostic tool to troubleshoot other devices,
to include a device simulator. The other issue is for robotics, so I can start developing "classes"
of projected robot(s) I would like to have.
Once you experience the power of the device you have, it always best to "get the most" out of your money.
A developer station is paramount, then branch out once you have acquired a direction you would like to go.
Good Luck!!   ;) ;) :) :D ;D

Offline Robo_Pi

  • Jr. Member
  • **
  • Posts: 28
  • Helpful? 4
Re: Robot Brains: Searching mid-high end units
« Reply #3 on: November 15, 2015, 06:49:43 PM »
Hi Avadon,

I agree with mklrobo that your design ideas should play a role in what hardware would be best suited to your needs.
I like to build in modules so my current robot project includes using a Notebook computer as my main development station.  On the Notebook I mainly program using C#.

The Notebook computer is then in communication with the "Robot Brains" which will consist of many different devices.  I'm using a Raspberry Pi in each robot as the main "Cerebral Cortex".   The Raspberry Pi also functions as the WiFi connection to the Notebook computer.  Under command of the Raspberry Pi I use several Arduino boards.  Each board devoted to specific functions on the robot.   I also hope to employ several FPGA chips as well.  I plan on using the FPGA chips as "Hardware Processors" for various sensory I/O.

For this reason I have chosen a very small and inexpensive FPGA board.  Although I haven't actually implemented this yet so I can't say whether I'm heading in the right direction yet or not.  I'm only just now learning  how to program FPGA configurations and I don't yet know what their power and/or limitations are.

Finally, I'm hoping to incorporate some actual neural networks in my robot project.  But that's down the road in my hopes and dreams as I don't yet know of any actual neural network boards available.  I'm thinking that I may need to design my own and I'm hoping the  FPGAs will come in useful for that endeavor.  Only time will tell. 

Real life issues tend to interfere with my robotics hopes and dreams.  ;D



Offline Avadon77Topic starter

  • Jr. Member
  • **
  • Posts: 15
  • Helpful? 1
Re: Robot Brains: Searching mid-high end units
« Reply #4 on: November 16, 2015, 05:16:00 PM »
My girlfriend and I want to basically build a rover with humanoid/android torso. Sort of like an animitronic ontop of a roving base. So first it must function as a telepresent/RC vehicle. Once I know all the servos, turret, back, arms, neck, head, etc. work well, then I'll dive into seeing if I can augment the experience, so it's not just a fancy car with a torso on it. Then still after that work on an autonomous mode. So I'm very forward thinking looking at robot brains right now, but I didn't want that to bite me in the hiney later when I know nothing about them. It's like looking at cars before you have your license, but it's good to know something about cars, even before you drive. ;)

The FPGA stuff seems unreal complicated at first look. It is manageable I'm sure once you get your feet very wet and sit down with it for a few weeks. From what I read about FPGA's is that most robotics platforms are not using them and they are indeed overkill, and not really overkill in a good way (because who doesn't like overkill) but somewhat of a diversion to a more immersive experience where one gets more into the theoreticals of design and control regarding micro processor architecture. Or at least that's what a newb like me came away from it. On another blog someone said short and sweet "If you want a robot get a MCU or embedded PC. If you want to dive deep in computer robotics processor arhitecture and study that for years then get an FPGA."   That sentiment was enough for me to start to think about other platforms rather than just throwing down a few hundred for a fancy FPGA.

I'm more of a fan of all inclusive boards but I can see that there's not too much out there in that way. The Adlink fanless computers are pretty cool, like the Adlink MX-5400, prolly way to pick and heavy though for my robot. Then I saw stuff like this TS-7800 which looks really cool  https://www.embeddedarm.com/products/board-pictures.php?product=TS-7800


All of these reccomendations above I saw from this site..

http://robotsforroboticists.com/industrial-computers/

But when it comes to a controller I'm basically looking for someting similar to what you two mention. Something I can plug a monitor,keyboard,ethernet, etc. into and operate off of. So the more I/O capacity of the board the better. And I really wanted something that runs C, Linux, or Dos and not some very proprietary programming language that I would only ever use again if I got a job in a car robotics factory in china. I'd rather learn a language that has broad and wide uses than just is specific to one niche.

Do you guys have an opinion on single board x86 and Arm systems?

Offline mklrobo

  • Supreme Robot
  • *****
  • Posts: 558
  • Helpful? 15
  • From Dream to Design at the speed of Imagination!
Re: Robot Brains: Searching mid-high end units
« Reply #5 on: November 16, 2015, 05:55:56 PM »
 :) Avadon77;
It looks like you are narrowing down what you want.
I have found, that whatever platform you use, make sure you have PLENTY of
software and community SUPPORT for it. Having a chatroom to help you work out
your problems is very helpful. Raspberry Pi has a lot of support. Typically, new items
take a while to get example code for them, and time itself tends to weed out the "Fad"
items. If they are productive, they will be here for quite awhile. You must have a way to program
the items, with IDE to work with, that is friendly. More features are great, BUT, if there is no
code examples to take advantage of these, the chip is useless. ARM architecture has been around
for awhile, and the most used. If you invest in a platform, like playing chess, choose the platform that
gives you the most learning/fun/projects/code that you can get. You can use this like a "crowbar" to
help you in the real world, if you need it. Good LUCK!!   ;) :) :D ;D

Offline Robo_Pi

  • Jr. Member
  • **
  • Posts: 28
  • Helpful? 4
Re: Robot Brains: Searching mid-high end units
« Reply #6 on: November 16, 2015, 10:33:09 PM »
My girlfriend and I want to basically build a rover with humanoid/android torso. Sort of like an animitronic ontop of a roving base.

That's basically what I'm doing too.  I started out building something like a Johnny 5 (Or "Short Circuit"), but then decided I wanted a more human-like head.   I'm already working on the track base (similar to Johnny 5), but now I'm also working on constructing a human head.

My first robot (if it ever gets built) will most likely just be the head running around atop the Johny 5 track base.  I'll deal with adding arms and torso later.   The whole project is just a journey into AI.   My main goal is to build a robot that can actually do things and have some sense of what it's being told to do, etc.   Precisely what it looks like is less important.  But after looking at how much work so many people have been putting into traditional Johnny 5 heads I decided that I may as well go full-blown human-like instead.

But like I say, the AI programming is really the focus of my project.  I want a robot that has some real smarts.  At least as much as say a dog (or even better).   After all, you can't talk to a dog in English, so the robot is already one up on a dog.  ;D

You can tell the robot not to go out on the road and it will know what you mean.  8)   Sometimes that's hard to convey to a dog.

The FPGA stuff seems unreal complicated at first look. It is manageable I'm sure once you get your feet very wet and sit down with it for a few weeks.

From what I read about FPGA's is that most robotics platforms are not using them and they are indeed overkill, and not really overkill in a good way (because who doesn't like overkill) but somewhat of a diversion to a more immersive experience where one gets more into the theoreticals of design and control regarding micro processor architecture. Or at least that's what a newb like me came away from it. On another blog someone said short and sweet "If you want a robot get a MCU or embedded PC. If you want to dive deep in computer robotics processor arhitecture and study that for years then get an FPGA."   That sentiment was enough for me to start to think about other platforms rather than just throwing down a few hundred for a fancy FPGA.

I certainly agree with this.  Most robotic projects wouldn't benefit from the use of an FPGA.   Using an FPGA when you don't really need one is a total waste of time.  There are only two situations when FPGAs are beneficial.

1.  If you are planning on mass producing a product.

In this case a single FPGA can be programmed to replace a lot of other more complicated devices and tailored specifically for the product in question. 

The only other time FPGAs are beneficial is:

2.  If you are designing something where an FPGA simply outperforms an MCU bigtime.

These types of applications often show up in video and sound production where an FPGA can simply handle a lot of data very quickly in parallel where an MCU would need to process things byte-wise thus taking a lot of time to do it.

Some examples are, pattern recognition devices, such as face recognition, object tracking, extreme navigation situations, specific sound recognition such as speech or voice recognition, etc.  The FPGAs outperform MCUs hands down in these types of situations.

Since you created as the topic heading "Robot Brains: Searching for mid-high end units", the topic of FPGAs is certainly going to come up since FPGAs are capable of doing high-end "Thinking" far faster than MCUs.  But only in specific situations.

3.  Neural Networks designed using FPGA plus Op Amp Arrays.

Another area where FPGA "may" be useful is in the design of physical Neural Networks.  This is where I am hoping to apply them.  Unfortunately I'm on my own in this department as most people who work on Neural Networks only do so via simulations using standard CPUs which is kind of self-defeating actually, unless the goal is to simply understand how neural networks work.   But to build an actual Neural Network you wouldn't use a CPU or MPU.

An FPGA itself cannot serve as a programmable Neural Network alone.  But it might be able to play a role in a larger design of one (which is what I'm hoping).   It would be great if they actually made FPNNs (Field Programmable Neural Networks), but unfortunately they don't.  Why they don't is beyond me, we certainly have the technology.  Apparently there just aren't any electronic companies aware of just how powerful such devices would be (or they haven't figured out how to make them yet).   But if they haven't figured out how to make them, then shame on them, because they certainly have the technology to do it at their fingertips.

In any case, you're right.  Using an FPGA just for the sake of using one is a big waste of time.  You need to have an application where it's going to outperform the MCU by a very large factor (as in the examples I stated above).

As far as programming the FPGA it's fairly simple depending on your previous knowledge of programming and electronics.

An FPGA can be programmed using a language called Verilog which is very similar to the structure of C.  So if you know how to program in C then you basically already know how to write Verilog code to a large degree.

However, you also need to have some understanding of electronics, and in the case of FPGAs that would be digital logic.  So if you understand digital logic electronics, and C programming language, you're basically already there.

If you want to design a Neural Network using an FPGA (as I am  hoping to do), then  you'll also need to be well-versed in analog OP AMP circuits.   There aren't any of those inside the FPGA, but the idea is to program a field of Op Amps using the FPGA as a programmable switch.  At least that's the idea I have in mind right now.  Whether or not this can even be done is yet to be seen.  Right now it's just an idea I have. 

I wish they made a Field Programmable Op-Amp array already configured as a Neural Network, (so I wouldn't need to design my own from scratch) but like I say, the electronic manufactures seem to be asleep behind the wheel on that one.   So I'm stuck with having to make do with what I can scrape together on a hobbyist's budget.

But yeah, if you want a truly "High-End" brain, programmable Neural Networks would be the way to go.  And not just computer simulations (way too slow!), but by using actual electronic neural networks (super FAST!).





« Last Edit: November 17, 2015, 04:55:07 PM by Robo_Pi »

Offline mklrobo

  • Supreme Robot
  • *****
  • Posts: 558
  • Helpful? 15
  • From Dream to Design at the speed of Imagination!
Re: Robot Brains: Searching mid-high end units
« Reply #7 on: November 17, 2015, 05:42:25 AM »
 8) Awesome info from RoboPi!!!  (+1)  8)
I feel, that, aren't we all developing our robots, a little at the time, kind of like
Young Anikin Skywalker?? ; C3PO in that movie, was just being built, :o
(I would like a light sabre, but Definitely be able to use "The Force"!)  ;D

Offline Avadon77Topic starter

  • Jr. Member
  • **
  • Posts: 15
  • Helpful? 1
Re: Robot Brains: Searching mid-high end units
« Reply #8 on: November 19, 2015, 12:10:14 AM »
That's basically what I'm doing too.  I started out building something like a Johnny 5 (Or "Short Circuit"), but then decided I wanted a more human-like head.   I'm already working on the track base (similar to Johnny 5), but now I'm also working on constructing a human head.

Wow that's funny we are working on an eerily similar project. I've been interested in and researching tracked vehiciles in my spare time since late 90's. I find it totally fascinating and it's a great base for a platform. But I set aside robotics for a good many years and now getting back to it and development, but really on a hobby level. So I'm ambitious but not trying to kill myself. lol Full scale is a challenge because things at hobby scales don't always translate well the physics of a full humanoid, which I'm sure you've already discovered. There's soo much on the market techwise compared to even 10 or 15 years ago. It's amazing looking at what is available now. Mindblowing even.

My first robot (if it ever gets built) will most likely just be the head running around atop the Johny 5 track base.  I'll deal with adding arms and torso later.   The whole project is just a journey into AI.   My main goal is to build a robot that can actually do things and have some sense of what it's being told to do, etc.   Precisely what it looks like is less important.  But after looking at how much work so many people have been putting into traditional Johnny 5 heads I decided that I may as well go full-blown human-like instead.

The head on tracks.. now i gotta see that. Might be a little frightening or humorous. Not sure which. Sounds like our main goals are extremely similar "My main goal is to build a robot that can actually do things and have some sense of what it's being told to do, etc"  There's so much sophisticated hardware out there and fairly easy to implement.

But like I say, the AI programming is really the focus of my project.  I want a robot that has some real smarts.  At least as much as say a dog (or even better).   After all, you can't talk to a dog in English, so the robot is already one up on a dog.  ;D

You can tell the robot not to go out on the road and it will know what you mean.  8)   Sometimes that's hard to convey to a dog.

Now therein lies a challenge.
[/quote]

I certainly agree with this.  Most robotic projects wouldn't benefit from the use of an FPGA.   Using an FPGA when you don't really need one is a total waste of time.  There are only two situations when FPGAs are beneficial.

1.  If you are planning on mass producing a product.

In this case a single FPGA can be programmed to replace a lot of other more complicated devices and tailored specifically for the product in question. 

The only other time FPGAs are beneficial is:

2.  If you are designing something where an FPGA simply outperforms an MCU bigtime.

These types of applications often show up in video and sound production where an FPGA can simply handle a lot of data very quickly in parallel where an MCU would need to process things byte-wise thus taking a lot of time to do it.

Some examples are, pattern recognition devices, such as face recognition, object tracking, extreme navigation situations, specific sound recognition such as speech or voice recognition, etc.  The FPGAs outperform MCUs hands down in these types of situations.

Since you created as the topic heading "Robot Brains: Searching for mid-high end units", the topic of FPGAs is certainly going to come up since FPGAs are capable of doing high-end "Thinking" far faster than MCUs.  But only in specific situations.

3.  Neural Networks designed using FPGA plus Op Amp Arrays.

Another area where FPGA "may" be useful is in the design of physical Neural Networks.  This is where I am hoping to apply them.  Unfortunately I'm on my own in this department as most people who work on Neural Networks only do so via simulations using standard CPUs which is kind of self-defeating actually, unless the goal is to simply understand how neural networks work.   But to build an actual Neural Network you wouldn't use a CPU or MPU.

An FPGA itself cannot serve as a programmable Neural Network alone.  But it might be able to play a role in a larger design of one (which is what I'm hoping).   It would be great if they actually made FPNNs (Field Programmable Neural Networks), but unfortunately they don't.  Why they don't is beyond me, we certainly have the technology.  Apparently there just aren't any electronic companies aware of just how powerful such devices would be (or they haven't figured out how to make them yet).   But if they haven't figured out how to make them, then shame on them, because they certainly have the technology to do it at their fingertips.

In any case, you're right.  Using an FPGA just for the sake of using one is a big waste of time.  You need to have an application where it's going to outperform the MCU by a very large factor (as in the examples I stated above).

As far as programming the FPGA it's fairly simple depending on your previous knowledge of programming and electronics.

An FPGA can be programmed using a language called Verilog which is very similar to the structure of C.  So if you know how to program in C then you basically already know how to write Verilog code to a large degree.

However, you also need to have some understanding of electronics, and in the case of FPGAs that would be digital logic.  So if you understand digital logic electronics, and C programming language, you're basically already there.

If you want to design a Neural Network using an FPGA (as I am  hoping to do), then  you'll also need to be well-versed in analog OP AMP circuits.   There aren't any of those inside the FPGA, but the idea is to program a field of Op Amps using the FPGA as a programmable switch.  At least that's the idea I have in mind right now.  Whether or not this can even be done is yet to be seen.  Right now it's just an idea I have. 

I wish they made a Field Programmable Op-Amp array already configured as a Neural Network, (so I wouldn't need to design my own from scratch) but like I say, the electronic manufactures seem to be asleep behind the wheel on that one.   So I'm stuck with having to make do with what I can scrape together on a hobbyist's budget.

But yeah, if you want a truly "High-End" brain, programmable Neural Networks would be the way to go.  And not just computer simulations (way too slow!), but by using actual electronic neural networks (super FAST!).
[/quote]

This is a fascinating breakdown. I also was surprised that there weren't boards more adapted to robotics. It must be that they see this as more of a hobby community and that the development cost-sales ratio just isn't there. An FPNN would be an amazing idea. To create it though you'd need a platoon of engineers to create yoru own board.

hey there is just soo much more I'd like to write but this little box and the quoting system is migraine producing. I'd much rather converse in regular email. Our projects are definitely on the same wavelength and at the very least I'm sure we could share some cool ideas, and at most maybe even be a good resource to each other. I'll pm you my email if that's cool.
« Last Edit: November 19, 2015, 02:50:15 AM by Avadon77 »

Offline mklrobo

  • Supreme Robot
  • *****
  • Posts: 558
  • Helpful? 15
  • From Dream to Design at the speed of Imagination!
Re: Robot Brains: Searching mid-high end units
« Reply #9 on: November 19, 2015, 06:44:00 AM »
 :o Holy Smoke! FPNN!  Awesome! :o
I wish they made a Field Programmable Op-Amp array already configured as a Neural Network, (so I wouldn't need to design my own from scratch) but like I say, the electronic manufactures seem to be asleep behind the wheel on that one.   So I'm stuck with having to make do with what I can scrape together on a hobbyist's budget.
That FPNN must be the next inevitable step in the robotic evolution. Yes Folks, you have seen the concept
devised right here on this forum!
Ironically, a similarity of the neural chip was also introduced on the movie, "terminator"; but the specifics
were left out, of course, because, at the time, it was pure fiction. But now...................  8)

Offline jasonhead

  • Jr. Member
  • **
  • Posts: 10
  • Helpful? 0
Re: Robot Brains: Searching mid-high end units
« Reply #10 on: November 19, 2015, 10:28:55 AM »
I'm planning out and prepping for a dextrous robotic system, with the goal of using Neural Networks over an API.
Background: Software Engineer / Developer with some hardware knowledge.

I would avoid exotic hardware like FPGAs unless you are an actual electrical engineer with a good budget :)
For neural networks, GPUs are the way to go, big companies such as Google are using them:
http://stackoverflow.com/questions/317731/cuda-vs-fpga
http://devblogs.nvidia.com/parallelforall/cuda-spotlight-gpu-accelerated-deep-neural-networks/

GPUs are in the mainstream for neural networks.

For actual software, this could turn out to be pretty useful; opensourced by Google:
http://www.tensorflow.org/

Amazon GPUs for some cloud computing could be part of a solution:
at least for the training stage:
http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using_cluster_computing.html
https://aws.amazon.com/about-aws/whats-new/2013/11/04/announcing-new-amazon-ec2-gpu-instance-type/

You may be aware of this, but for general control you could use ROS:
http://www.ros.org/

For the actual onboard hardware to operate the servos/gather sensor data I'm initially going to to with an Arduino Mega, with the heavy lifting being done on a server or my laptop - the idea is to have python programming communicate with the Arduino  - should still be pretty quick over USB.

I'm planning on a stationary system initially, but for a mobile one there are some pretty compact desktop power computers.
The potential advantage I see a computer on a board offering is response speed to commands;
I've yet to see what the limits of an Arduino would be - for me my cameras will be hooked directly to a computer, so the Arduino won't have to handle that input.

There are different ways of doing things, but using more common commodity hardware could have some advantages. I want to get my hardware up and running quickly so I can use it to test software and start that feedback loop.
« Last Edit: November 19, 2015, 10:38:11 AM by jasonhead »

Offline Avadon77Topic starter

  • Jr. Member
  • **
  • Posts: 15
  • Helpful? 1
Re: Robot Brains: Searching mid-high end units
« Reply #11 on: December 29, 2015, 08:36:46 PM »
After months of research this is what I came up with. So for what it's worth anyone asking about MCU's.... MPU's and what's new and out there might want to read my 2cents.

#1. Nvidia TX-1 (A mini ITX form factor GPU supercomputer for deep learning and robotic vision) $599 State of the art. New on the market for late 2015/early 2016. MIT/Stanford recommend. This is a robotics development platform for industry professionals and scientists. The prior version (which I believe is cheaper) is the TK-1. Nvidia is heavily invested in leading the way in automous cars, autonomous drones, and robotic vision. 

#2. Beaglebone X-15 one of the best mid range robotic platforms. Cutting edge. New for 2016 $260

#3. SolidRun: Hummingboard Edge. This is a really good mpu with everything on it. Small form factor and amazing capabilities. $252

#4. Hardkernel: Udroid-XU4 Very similar to the Hummingboard Edge with lots of capabilities in a small package $74.00 (before accesories)

#5. Intel Gallileo 2 development kit. Very nice board with intel architecture. ~$100.00

#6. Beaglebone black: Simple, very effective, daisychain-able, inexpensive ~$60.00

#7 Arduino Mega and Due  Both very good MCU robotic platforms. Inexpensive. ~$50.00

#8. Raspberry Pi2. Effective small MPU running Linux compatible with Arduino ~$50.00

#9. Creator Ci20 The MIPSĀ® Creator Ci20 board is a high-performance, fully-featured Linux and Android development platform. $65.00

Looked at Adlink and it seemed ill designed for amateur robotics and even many professional applications. It's propriety as heck, closed box, heavy, large, and very expensive. Standard laptops are rather poor for robotics because of their size and bulk and their problems with connectivity to other robotic mcu's and mpu's like adruino shields or beaglebone capes (add-on boards). I even looked at the TS-7800 which is a very old board from 2007 and at $229 it seemed archaic and outdated. I even talked to the company and they didn't even seem very enthusiastic about their old product.

Now you might argue with my placement. We all have our preference here. This is like religion, everyone has their favorite and their reasons for why they are their favorite. But these nine are more or less the best options on the market and you should look at them all before making a purchase. Most of these are very similar and can do much of the same things. When you get into the $200+ market you get a lot more for your money and you can start doing a lot more advanced things. Obviously I have my preference and that is the Nvidia TX-1 which is a supercomputer of 256 GPU cores (cuda cores) for doing robotic vision, mapping, occular odometry, deep learning and very advanced robotics projects. Yes it's expensive but you get what you pay for. They also have a Nvidia TK-1 which is a lesser expensive previous model. Hopefully this helps you all in deciding what brains to use for your robots. I believe all of these boards can hook up to MCU's for physical computing (I.e servo controllers) like Adruino shields and beagleboard capes (which are just fancy names for microcontrollers to drive motors, servos, etc.)




 


Get Your Ad Here