/robowaifu/ - DIY Robot Wives

Advancing robotics to a point where anime catgrill meidos in tiny miniskirts are a reality

Days left: 15

JulayWorld fallback document - SAVE LOCALLY

JulayWorld onion service: bhlnasxdkbaoxf4gtpbhavref7l2j3bwooes77hqcacxztkindztzrad.onion

Max message length: 32768

Drag files to upload or
click here to select them

Maximum 5 files / Maximum size: 20.00 MB

no cookies?

(used to delete files and postings)

Robowaifu Design Software Robowaifu Technician 09/18/2019 (Wed) 11:41:06 No.414 [Reply] [Last]
In this thread we discuss various CAD software for the purpose of making access easier for others to get started. I'm using Fusion360. Other good options are Blender and FreeCAD. Fusion360 is the easiest to use imho and is free unless your business makes over 100,000 dollars.

Post software, tutorials, tips and tricks you've learned, etc. Let's make designing waifus easier for each other!
23 posts and 10 images omitted.
thank you anon, it's been a long wait!
It feels like a professional platform now tbh.

Official Blender 2.8 Fundamentals playlist
Edited last time by Chobitsu on 10/05/2019 (Sat) 17:10:48.
Open file (39.18 KB 480x360 0(4).jpg)
Man that's an outdated infographic. The newer ones changed the layout so I knew it was old just at a glance.
great, thanks anon.

Open file (173.41 KB 1080x1349 Alexandra Maslak.jpg)
Roastie Fear 2: Electric Boogaloo Robowaifu Technician 10/03/2019 (Thu) 07:25:28 No.1061 [Reply] [Last]
Your project is really disgusting
Edited last time by Chobitsu on 10/05/2019 (Sat) 06:35:57.
Open file (49.86 KB 600x604 you_ok_anon.jpg)
You alright? Do you need us to call someone for you OP?
No u

Aoki Lapis model; Robot fairy Robowaifu Technician 09/16/2019 (Mon) 02:51:51 No.266 [Reply] [Last]
Height; 15cm
Type; Vocaloid

I'm making it in as natural looking way as I can, this means a bone structure and a similar layout of the electronics and components as with a living humanoid

This is a prototype version that I'm working on, I might change things later and make modifications or adjustments to the design or components
20 posts and 31 images omitted.
Teflon is explicitly designed to stay slippery. You might have better success using it over plastic insulation.

ah got it. i plan to eventually repopulate the threads here, including the electronics learning class one.

>Fab lab
cool. i've never been to one yet tbh. expensive?

>Also might be kinda tricky to 3d draw the parts properly but I might do that later actually
There's the Fusion 360 product anon, and Blender too ofc, whenever you feel inclined to design a reproducible system others can create as well.

Good luck.
Open file (7.82 MB 2138x1932 3d-printing-filament.png)
This insulation actually works surprisingly well like the video shows >>1004

Booking at a specific time probably costs, but I'm not sure how much and its kinda irrelevant for me right now anyway since I don't have money to spare
But it used to have open hours here for the public on specific days, I'm not sure if they have those open days since the schedule doesn't show it, but it still reads on the front page that they do, I'll have to check on that better later

I actually happened to get lucky the other day and got a bunch of 3d printing filament for free

Funnily when I started out making models for 3d printing I used sketchup
>I actually happened to get lucky the other day and got a bunch of 3d printing filament for free
very nice, gratz.

>Funnily when I started out making models for 3d printing I used sketchup
cool, cool. how was it?
It worked out fine, I would prefer to use something else for detail work tho
Fusion360 can get as detailed as they come. The (((Autodesk))) and (((cloud))) lock-in aren't to my tastes however. I think I'll stick with Ton & Blender.

Open file (10.89 KB 480x360 0.jpg)
TensorFlow Robowaifu Technician 09/15/2019 (Sun) 05:47:25 No.232 [Reply] [Last]
Tensorflow is currently the hot Deep Learning framework. The engine is written in C++, and the scripting API for it uses Python.






13 posts and 4 images omitted.
Heh. BTW, rather than starting out with a big investment of your own time and energy, you and your friend might consider taking advantage of some of the the prior art in this field. There are several corpora already freely available. For example:

Stanford maintains a list particularly suited to NLP:
There's a web-based TensorFlow system you can't break.
Relevant YT playlist.

TF training by an ML/AI researcher. In this I discovered Jewgle has a free, online facility to run your TF Python code on an honest TPU! or a GPU or a CPU, your choice. Did I mention it's free?


Prototypes and failures Robowaifu Technician 09/18/2019 (Wed) 11:51:30 No.418 [Reply] [Last]
Post your prototypes and failures.

I'll start with a failed mechanism, may we build upon each others ideas.
15 posts and 11 images omitted.
Maybe it would be cheaper/faster to prototype initial designs as cardboard cutouts instead?
Tried making a 10-spoke speed encoder wheel that fits the Tamiya 4WD tires, its too brittle and uneven. My plan is to have the optical speed encoder just arch over the main wheels while they spin, instead of trying to read a separate encoder disc. I will try 5 spokes and wont bother with concave shape, I'll just have the holes flat and square. The diameter with the Tamiya wheels is 28.5mm, one revolution is 9cm, so two revolutions or 10 clicks is exactly one maze block traversed.

Repost but the sumobots in action:

In the horizontal pic the bots in order of completion are 1,3,4,2,6,5.

Future bots are all waifu types, using some variation of the Fluffytail design and/or 3D doll pieces:

7. Micromouse (has its own current thread)
8. Cattymouse(larger micromouse with stepper motors)
9. Penguingirl balancebot(bluetooth control)
10. Kiwigirl balancebot(using 2.4Ghz RC radio and stepper motors)
11. Fluffy Rachnera-san(12-servo quadruped)
12. Fluffy Wresler(at least 14 metal gear servos)
13. Panzergirl waifutank (torso on tracked tank chassis)
14. Panzergirl devastator (torso on mecanum wheels). 13 and 14 will dress up like pic related.

Unfortunately neither of the future bots will stand taller than 40cm, I do not have the mechanical know-how to design for high torque, high weight applications.

I've noticed in related news, consumer home robots are closing shop: the makers of Kuri, Jibo, Cozmo, Vector. Most of the AI of those robots are in private servers so the bots will essentially die unless the AI is open-sourced.

Apparently the only successful "robots" are the Roomba, Siri, Alexa, Google Home etc. Either single-function tools or subsidized spyware. That is not good.

We must change that. Right now all I can do is work on creating a standard list of serial commands so that all my future waifubots can be controlled through a single universal arduino-compatible controller - either a physical controller that uses bluetooth or 2.4Ghz RC frequencies, or an app. Later waifus will have audio and video processing so I'm hoping I can offload that to a desktop computer.

I'm just sharing this infodump since I hope even though our paths may be plagued by a thousand PLA carcasses, there is a clear vision of what progress should look like, for me at least.
>13 and 14 will dress up like pic related.
<robowaifu qt tankgrill enforcers

As for robot shops closing, it's neither unusual nor surprising for a nascent industry to have churn. One great thing that we have on our side is that the Wright Brothers/Henry Ford style home garage operations are a real possibility [for this industry atm], as you and others are proving. We certainly don't want the botnets to be the only possibility for anons to turn to in the future though, so we have to stay focused. Fortunately we still have a few years before they manage to make it illegal to create your own waifu but have to use theirs instead. But it's definitely a race. If we make positive progress first and it becomes widely popular among hobbyists the world over, then we may be able to forestall such a move by the botnet vendors such as Jewgle & Amazon.

There are some command protocol works already out there, have you had a look at any of them?

Keep up the good work anon!

Sorry for off-topic OP
Thanks man. It is indeed the only industry I can think of that is truly still in the garage phase. Actual laws regulating humanoid robots don't exist yet outside of science fiction stories, so this is still a window of opportunity before corporate takes over.

>There are some command protocol works already out there, have you had a look at any of them?

I'm taking a look at various arduino smartcar bluetooth controller apps for Android, will try to find some common string between them so you can just download any one of them to control my bots.
>so this is still a window of opportunity before (((corporate))) takes over.
Exactly. But no one should just relax because there's a gap for now, and assume they just won't try to stop it. They will. We just have to move faster so there will simply be far too broad access to basic robowaifus before they can stonewall it.

NLP General Robowaifu Technician 09/10/2019 (Tue) 05:56:12 No.77 [Reply] [Last]
AI Natural Language Processing general thread

>"Natural language processing is a field of computer science, artificial intelligence, and computational linguistics concerned with the interactions between computers and human (natural) languages."

>Computing Machinery and Intelligence

18 posts and 5 images omitted.
>Maybe that's all we really are to each other.
Heh, maybe. But I suspect intangibles are there that go deeper than just language tbh.

Yes, I think all physical life very pattern based and dependent on learning them successfully. Fortunately the purely mathematical approach to machine learning that is now standard today does this incredibly efficiently. It's the underlying basis on both the improvements recently in audio and video recognition systems, and is the underlying approach that will soon give us self driving cars. All of this plays into bring robowaifus about too ofc.
NLP course
Open file (86.82 KB 800x250 0704234652024_25_vox.png)
Open file (9.89 KB 480x360 0.jpg)
Thanks Anon, great info!




I'm adamantly opposed to the SJW shitfest Mozilla has become now, but the project license should allow us to do whatever we want with it AFAIK, including creating cuddly cute robowaifus. I'd like any legalanons to confirm that or no please.

on-topic related crosspost (possibly a dup ITT)

Open file (3.72 KB 67x118 file.png)
, which anon pointed out here. [[5085

Discusses an NLU approach by the OpenAI guys that apparently outperforms GPT-2, and is entirely unsupervised. Use the WebText corpus.

Batteries & Power Robowaifu Technician 09/09/2019 (Mon) 06:21:14 No.23 [Reply] [Last]
Robowaifus will need power to run, and since they will be mobile this means a mobile power system too. ITT post info on batteries and other mobility capable power systems.
11 posts and 8 images omitted.
I see your standard lithium batteries being chosen, very practical. We also see some theoretical discussion, very interesting. Fuel cells may be worth looking into if you want an option that doesn't require plugging in your partner. I wonder what options will be most cost efficient, but certainly your partner if using lithium batteries can't be more costly to charge than an electric car. There is the matter of energy density however. Will there be room within your partner for sufficient energy packs to run them for a day? You may also want to make your system hot-swappable in case she starts running low on energy during a late evening out.

It may be worth it to study the power systems used for artificial hearts.
My first response for technical problems like this has always been, "how did biology solve it?"
In this case, it kind of didn't. Sleep isn't a charging cycle, as much as it is a cleaning time. Spend the less-productive night hours squirreled away, healing your body and indexing the day's memories and such. What people actually do to get their energy is eat and digest, and humans only store enough energy for about three days of actually doing things. When someone eats, they are getting all the energy from the storage of whatever they are eating. That's why you eat your own body weight in food in about a month, but it's split into a pound or two three times a day: the thing you're eating also didn't solve the energy storage problem, and you use energy almost as fast as you get it if you're an active person. It's like that all the way down to the plant level, when they use the sun's energy to turn ground-minerals into chemicals they can use.
Considering that billions of years of evolution managed to get energy storage for three days of activity, the best option probably is to have some form of on-site power generator and a steady supply of fuel, as it is much easier to generate power than store it. The reliance on fossil fuels is because they have something like 40 times the energy per kilogram of current batteries. Those biobatteries that >>797 mentioned show great potential if they can be miniaturized to fit inside of a torso.

As for artificial hearts, those work via "inductive coupling" - basically, if you have two conductors set up right, you can wirelessly transfer energy short distances, as in, a few feet at most. The transmission speed decreases very drastically with range. The heart is connected to one of the conductor (along with a rechargeable battery), while outside of the body is a much bigger battery connected to the other conductor. The rechargeable battery inside has only about half an hour of charge, almost all energy goes nonstop from the external battery to the heart. Inductive coupling can be used to transfer energy without breaking the skin, but doesn't have the flexibility to be much better than regular charging in terms of range, and takes up a lot more space, since a bigger conductor means faster power transfer at a much less efficient trade than bigger charging cables.
Any refinement possible on your '3-day energy storage based on evolution' model to account for motivated humans being able to go 40 days w/o food?

>inductive coupling
what if we made a wireless-charging futon for our robowaifus to sleep on overnights?
Could use some variety of flow battery if we're looking for in ingestion based model. Your waifu would require input of material in order to keep running however adding this new material would be quick and depending on volume could be rather convenient. As well I find the idea of my waifu having to "eat/drink" in a fashion like a live person to be appealing in some manner.
>As well I find the idea of my waifu having to "eat/drink" in a fashion like a live person to be appealing in some manner.
There's a social appeal to eating together with friends and loved ones anon. Watch the scene with K and Joi in Bladerunner 2049 and imagine instead that she was eating holographic food along with him. It becomes more charming to my thinking.

Robowaifu-OS & Robowaifu-Brain(cluster) Robowaifu Technician 09/13/2019 (Fri) 11:29:59 No.201 [Reply] [Last]
I realize it's a bit grandiose (though probably no more than the whole idea of creating a irl robowaifu in the first place) but I want to begin thinking about how to create a working robowaifu 'brain', and how to create a special operating system to run on her so she will have the best chance of remaining an open, safe & secure platform.

OS Language Choice
C is by far the single largest source of security holes in software history, so it's out more or less automatically by default. I'm sure that causes many C developers to sneer at the very thought of a non-C-based operating system, but the unavoidable cost of fixing the large numbers of bugs and security holes that are inevitable for a large C project is simply more than can be borne by a small team. There is much else to do besides writing code here, and C hooks can be generated wherever deemed necessary as well.

C++ is the best candidate for me personally, since it's the language I know best (I also know C too). It's also basically as low level as C but with far better abstractions and much better type-checking. And just like C, you can inline Assembler code wherever needed in C++. Although poorly-written C++ can be as bad as C code in terms of safety due to the necessity of it being compatible with C, it also has many facilities to not go there for the sane coder who adheres to simple, tried-and-true guidelines. There is also a good C++ project already ongoing that could be used for a clustered unikernel OS approach for speed and safety. This approach could save drastic amounts of time for many reasons, not the least of which is tightly constrained debugging. Every 'process' is literally it's own single-threaded kernel, and mountains of old-style cruft (and thinking) typical with OS development simply vanishes.

FORTRAN is a very well-established language for the sciences, but a) there aren't a lot of FORTRAN coders, and b) it's probably not the greatest at being a general-purpose language anyway. I'm sure it could be made to run robotics hardware, but would probably be a challenge to turn into an OS.

There are plenty of dujour SJW & coffee languages out there, but quite apart from the rampant faggotry & SJWtardism plainly evident in most of their communities, none of them have the kind of industrial experience and pure backbone that C, C++, or FORTRAN have.

D and Ada are special cases and possibly bear due consideration in some year's time, but for now C++ is the obvious choice to me for a Robowaifu OS foundation, Python probably being the best scripting language for it.

(1 of 2)
12 posts and 12 images omitted.
Related convo crosspost:
And, all the power, control and data wiring out to the rest of the robotics platform can more easily be home-runned and consolidated into driver banks and the RWBrain there as well. In fact, I'd guess the brain and stepper-drivers and other electronics will probably take up the majority of the internal chest cavity at a first guess.
Fantastic idea, why not build your own Linux distro though? I'd imagine that would make things easier. A custom version of Debian that was light weight, had OpenCV, PocketSphinx, and a custom program to allow easy control of SPI devices to interface with low level sensors and servo/stepper controllers. That would be fantastic. Also, I agree that Arm should be the architecture of choice. Their upcoming Cortex A76 and Mali G76 look go be very promising for AI. I'm not sure clusters are a good idea for something that needs to be inexpensive. Clusters add complexity, cost, and heat into a system. Though it is a really cool idea, perhaps a cluster based server could be an accessory that she'd wirelessly connect to for upgraded capabilities at her home? Overall, a robowaifu OS is something is very much like to see happen. Especially because I'm working on waifu hardware and would gladly test this.

From what classic anime has taught me, this is used to summon demons from CRTs.
Good points. In fact I do intend to create a lean image of Devuan (Debian sans Poettering) to place in the Node 0 ARM device. This will sort of act like a moderated gateway to my project, and could easily be used in the manner you described. And Devuan has fantastic support already for imaging working snapshots, and has an abundance of ready-made images for various ARM and other devices already in place. So, if you'd like to get ahead and prepare now for where I'm going here, then get a box set up with Devuan and start playing around with it now.

As for your point about clustering, that's kind of a separate topic and has many benefits it brings to [my] table. If you really want to understand my rationale, then I'd suggest you watch the first three videos I've already linked. That should give you a basic understanding.

Thanks for the input anon.

your file is too big onii-chan Robowaifu Technician 09/12/2019 (Thu) 02:58:03 No.152 [Reply] [Last]
it's a commendable quest to try to breath life into metal and a topic i am very interested in. but, does anyone else stop and wonder whether it is right to bring waifus to the real world?
10 posts and 1 image omitted.
>unironically objectifying human-level AI
An example of where I did/said this?
that's the only kind of AI this thread is talking about, i don't know why you've got your head up your ass about video games
Open file (23.61 KB 665x574 0705060933408_13_47b.jpg)
This shitpost belongs in my shitpost
It's obviously right, only SJW's say otherwise because they think NPCs and robots are people.
You never specified human level intelligence, at all. The thread was ambiguous at best. I also stated that I was talking about at most a fancy AI/chatbot when defining what would be considered a viable victim of a proper moral scenario. I legit don't even know the point of your thread and what you want out of it. Half of us are talking about transporting the human mind into some animu simulation whilst the other half are attempting to talk about AI in robots.

If you want to talk about a specific subset of AI, then specify that in your thread's initial post; don't just retcon it in.

AI + Brain/Computer Interface news & commentary Robowaifu Technician 09/15/2019 (Sun) 10:35:53 No.253 [Reply] [Last]
DARPA Wants Brain Implants That Record From 1 Million Neurons

7 posts and 2 images omitted.
I figure this is probably the most pertinent thread for this.
Elon Musk's startup Neuralink is about to do some kind of reveal on Tuesday evening.
Open file (7.46 KB 320x180 mqdefault.jpg)

Report/Delete/Moderation Forms

Captcha (required for reports and bans by board staff)

no cookies?