Show off your skills and solve real design problems
Robosavvy is looking for your help in gathering fresh ideas for the design of a new humanoid robot designated for entertainment and education. Our objective is to launch a new open-source humanoid robot that will look great and will be agile and smart at a reasonable cost.
Comparable education and research humanoid robots cost over $10,000 due to their plastic molded shell parts, expensive actuators and closed-source business approach. We want this to be an open-source robot that will outperform its rival platforms while keeping the cost of materials to about $1,000. We will achieve this by utilizing materials and fabrication methods available to anyone. Materials include aluminum sheet and 3D printed parts for structure, vacuum formed polycarb exterior shell, servos for actuation and low cost electronics including Raspberry Pi for brains.
We expect to have to do some modifications to the chosen design to make the robot perform the featured range of motions because we realize the designer may not have all the practical experience designing humanoid robots. We will adapt the shell if needed for polycarb vacuum forming. We will also handle the electronics involved in power management and servo control, in addition to developing the core open-source software to make the robot move, including ROS support, simple python scripts and physics simulation.
You must include IGES or STEP files and at least one rendering. Native CAD formats are optional but encouraged.
PLEASE NOTE WE HAVE UPDATED THE PRIZES, SPECIFICATIONS AND REQUIREMENTS ON AUGUST 26, 2013.
The competition is open to everyone. Multiple entries are acceptable. Team entries are welcome - prize will be transferred to team leader.
Tag your entry with “robosavvy” when uploading to GrabCAD. If you submit through the "Submit entry" button, this should be added automatically. Only entries with valid tag will participate. Only entries uploaded to GrabCAD through the "Submit entry" button on this Challenge page will be considered an entry.
By entering this competition, you 1. Warrant that the work is your original work. 2. To the best of your knowledge, it is not, and has not been in production or otherwise previously published or exhibited. 3. Neither the work nor its use infringes the intellectual property rights (whether a patent, utility model, functional design right, aesthetic design right, trade mark, copyright or any other intellectual property right) of any other person.
-----------------------------------------------
If you think an entry may infringe on existing copyrighted materials, please email challenges@grabcad.com.
Applicants are not entitled to any compensation or reimbursement for any costs. The applicant’s participation shall not constitute employment, assignment or offer of employment or assignment.
Winning designs will be chosen based on the rules and requirements. The Challenge ends on October 6, 2013 (11:59pm UTC). EXTENDED: New end date is October 20, 2013, which means the winner will be announced by November 20, 2013.
The awarded participant is solely liable for the payment of all taxes, duties and other similar measures if imposed on the reward pursuant to the legislation of the country of his/her residence, domicile, citizenship, workplace, or any other criterion of similar nature.
We will release the finalists before the final announcement to give the Community an opportunity to "vote" for their favorites in the comments. The jury will take these into consideration when picking the final winner. This new addition is based on feedback from you!
The winning design of a humanoid robot will be custom built and sent to the designer (valued in parts and labor at $4,000) + $800 cash
$450 cash and GrabCAD shirt and bottle
$250 cash and GrabCAD shirt and bottle
GrabCAD shirt and bottle
GrabCAD shirt and bottle
The jury will consist of Robosavvy and GrabCAD experts.
Hosting the most popular online community for humanoid robot builders and with a wide range of products including humanoid robots, 3D printers, DIY electronics and sensors, RoboSavvy is a leading source of materials and knowledge in the area, developing robots for the education, entertainment and research markets, selling parts needed for creative making of robots and gadgets. RoboSavvy is developing the most advanced 1m tall research humanoid robot platform and the first acrobatic 3D printable humanoid robot.
If you don't receive the email within an hour (and you've checked your Spam folder), email us as confirmation@grabcad.com.
108 comments
Tan Wei Kok over 11 years ago
Servo size, receiver, controller and some standard joints dimensions will be helpful (now too small to see). Else will be FFA ^_^ If you had the contest started last weekend. I would had "rob" the one on display in Science Centre XD
Bar Danino over 11 years ago
"The robot should be based on Robotis AX12 and AX18 servo actuators"
"The robot's chest needs to carry a 3-cell Lipo battery, miniature webcam, Raspberry-Pi or BeagleBone"
Does anyone have STP files for these parts?
Tan Wei Kok over 11 years ago
@Bar Found these:
Dynamixel AX-12/18 CAD: http://www.trossenrobotics.com/productdocs/AX12.zip
Dynamixel AX-12/18 Dwg: http://www.trossenrobotics.com/productdocs/AX-12A.pdf
The rest was looking for it until I go for my dinner 2hrs ago =)
Bar Danino over 11 years ago
Great, thanks!
Adrien Liotard over 11 years ago
@Bar I found these models on GrabCAD:
Raspberry PI - model B
BeagleBoard
http://grabcad.com/library/hexapod-robot--1 (it uses AX12 servos, maybe they can be extracted from this model)
I also wonder about what features we can implement in our designs, because the current Bioloid GP looks already super cool :)
Tan Wei Kok over 11 years ago
@Adrien Same thought here. Was wondering how to improve an already super cool robot over dinner just now...LOL I think RoboSavvy is looking at cost down...LOL Read last few words from first paragraph "Our objective is to launch a new humanoid robot that will look great and will be agile and smart at a reasonable cost." I see reasonable and reachable cost are what lacking in Bioloid GP now :p
vineeth vijayan over 11 years ago
@adrien i admit to what you say, the current bioloid gp looks super cool, and adding more design to its aluminium body can increase the price on the current model, one can cover up the body and stuff led's to give it a tron or the iron man look..one has to change or cut down the metal to meet their requirement for a reasonable price
Branko Stokuca over 11 years ago
I am just wondering about the 1st prize, What would you possibly do with a robot?
if the cost of this robot is 4000$ couldn't you just give the money prizes?
And also reward the second and third place because I think that there will be many cool entries here and it would be a shame not to reward great designs as well.
Just an idea ;)
Sara Sigel over 11 years ago
Thanks to all for the early comments, I'll get them to RoboSavvy!
Bar Danino over 11 years ago
hi sara, could you also ask for an STP file for the "bioloid GP" model?
Simonr over 11 years ago
Since the chassis has to be made from folded Aluminium, do you want the full sheet metal development of the parts (easy in Solidworks) or just 3D 'shells' and you will figure out how to fold it from sheet?
You have specified the Servos - to avoid any confusion, could you list the specifications of those servos you work to - as real world is usually different to the manufacturers spec. Mainly max speed, acceleration, resolution & holding torque.
Likewise, it would be helpful if RoboSavy could link to the models THEY use for servos etc. I've found a couple of RPi models which are different, and neither matches my RaspberryPi exactly!
Simonr over 11 years ago
Also, you have specified "will look great and will be agile and smart at a reasonable cost" as the judging criteria.
Can you give us the balance between these you are going to use to score entries.
For example, will an ugly, agile, cheap robot win over a good looking, clumsy, cheap robot?
Simonr over 11 years ago
Sorry for a third question.
In the last part of the description it says "RoboSavvy is creating the most advanced 1m tall research humanoid robot in the world, a 3D printable low cost humanoid robot that looks like Honda Asimo"
Is that what this challenge is to design? Or something along those lines - i.e. a 1m tall relation of Asimo?
sai over 11 years ago
heyi think we should team up and builds this and share money.my email is saisrinath2001@yahoo.com
Tim Titus over 11 years ago
There is no money to be shared, and you can't share a single robot (IE the prize). Nice idea though, Sai
Bar Danino over 11 years ago
Another question: do these servos have continuous rotation versions, and if so, can they be used in the design?
Bar Danino over 11 years ago
Also, can anyone tell if the servos' shafts comes out of both ends? (i.e. I can connect something to both the front and the back of the servo and have it spin.)
Bar Danino over 11 years ago
^ Found the answer, apparently only the front spins, but on the opposite side there's a hole to attach a bold to, for stability.
Gochat over 11 years ago
what options are sheet metal thickness selection and what is the maximum size of the robot?
Frederik Vollbrecht over 11 years ago
"The robot's body can all or part be covered by a very thin vacuum formed polycarbonate sheet shell"
Define "very thin". 0,375 mm or 0,5 mm?
Tim Titus over 11 years ago
I would be interested in this challenge, except that I have no use for one of these robots. I agree with others that there should be an option between a cash prize or a robot, and also various smaller prizes for 2nd, 3rd, 4th... otherwise, I predict that there won't be very many entries for something as time consuming as this will be.
Sara Sigel over 11 years ago
Thank you so much for the comments, so far. They are extremely valuable for RoboSavvy and the GrabCAD team. The awards will be updated and we'll include cash options as well. Keep the comments and feedback coming and stay tuned for updates!
Tan Wei Kok over 11 years ago
@Sara Please ask RoboSavvy to increase the number of types of servo (included dimensions & performance chart), including those they didn't carry. That will increase the variants. Restrict the overall size too. State clearly how many axis (degree of movement) and where they want them. Else we will see designs get disqualify because of some hidden requirements or last minute changes. Thanks =)
chandan over 11 years ago
should RObosavvy in assemble mode or in only one part model pls clearify it
Tan Wei Kok over 11 years ago
Guys, look what I had found: http://www.thingiverse.com/thing:5192 ^_^ Will try to assemble it later =)
Limor Schweitzer over 11 years ago
I'm from RoboSavvy and I'll try to answer your questions.
GP model: First of all, we have put together a CAD model of the GP which we have passed onto Grabcad. A link should show up on this page any time now.
servo model: You can find a detailed model of the AX12 / AX18 servo here: http://humanoids.dem.ist.utl.pt/Iden_external/overview.html
battery model: you can use any size and shape found on hobbyking weighing up to 150g and of at least 1200mah ( http://goo.gl/WP98nm )
camera: we will post the model or dimensions of a tiny usb camera found on ebay and stripped off its plastic shell.
CPU board: Rpi and BeagleBone black models can be found using google. the space should have clearance either of the two.
Julius over 11 years ago
Hi, Limor. Is it okay if we design different linkages for our humanoid robot?
Bar Danino over 11 years ago
Does anyone know what's the weight of the Bioloid GP?
Tan Wei Kok over 11 years ago
@Bar It is 1.6kg More info here: RoboSavvy Humanoid Design Intel Report =)
Limor Schweitzer over 11 years ago
Tan & vineetv: couple of points regarding your previous comments about cost, adding aluminum and shell etc.
We would like to see a nice looking humanoid robot shell, exposing as little as possible the motors and brackets. If you look at pictures of Asimo you will notice that there are no "technical" detail exposed anywhere on its body.
The cost of the robot has a lot to do with the servos and the manufacturing process of everything else. We are looking for a cool look and we may modify the design and structure slightly to reduce the cost of manufacturing.
Limor Schweitzer over 11 years ago
Brankv: You brought up the subject of a money price instead of getting your designed robot manufactured for you. We will consider the option of a money prize and also 2nd and 3rd prizes as we approach the formal start of the competition (around the 22nd Aug).
Limor Schweitzer over 11 years ago
Model of the robot: You can download the GP model by pressing the "Download Specification" button at the top of this page.
Remember that this is a good reference but as stated in the guidelines above, you are free to modify the aluminum structure to accomodate for your creative shell design.
Simonr over 11 years ago
@Limor - so you are looking for something based on your existing GP robot, with changes to the aluminium components where necessary to attach the shell. The challenge thus is to design a cool looking shell rather than to re-invent the robot itselfv
Limor Schweitzer over 11 years ago
Simonv:
"do you want the full sheet metal development of the parts (easy in Solidworks) or just 3D 'shells' and you will figure out how to fold it from sheet" - it will be nice if you can also provide the flattened version.
"...servo specification..." : the Robotis servos are not like any RC servos, they are designed for robotic applications and the specifications are pretty close to reality
"...servo models..." : i've posted a link (..ist.utl.pt) to a servo model that was precisely measured from a real AX12 servo
"...Rpi model.." : just make sure there's enough clearance somewhere for an RPi and beagle bone (either one should fit in). We can later adjust the model to ensure there are screw holes in the aluminum and/or some 3d printed base
"...1m tall robot..." : we are building a 1m tall humanoid with a nice shell design. http://goo.gl/pkbsYv . This robot design challenge is about a robot that is 40-50cm. ( fyi: To make a 1m tall robot walk you need to use servos that are 10 times more powerful and 30 times more expensive. c
Limor Schweitzer over 11 years ago
Bar: the AX12 servos can be used for continuous rotation (but the Potentiometer position reading and setting, doesn't cover the full 360deg, only 300deg.)
Limor Schweitzer over 11 years ago
Gochat: "what options are sheet metal thickness selection and what is the maximum size of the robot?" -- The aluminum can be 1.5mm-2mm. We will have to see the final design and maybe tweak to ensure it wont bend or tear, while trying to minimize weight.
Robot Height: at least 35cm and probably no more than 50cm. Depending on weight distribution he may not be able to walk if the knee/foot/hip servos need to swing too much mass X distance.
Limor Schweitzer over 11 years ago
Frederiv : the thickness of the polycarb shell can be 0.5mm. The important thing is the look of the robot. We will try to make the shell work with as thin a sheet as possible (to make it lighter) while ensuring it is not too weak at the contact points with the metal structure, to handle any stress caused by falling. The shell may need to be adjusted later to make it vacuum formable. (for example, it may need splitting into two halfs and adjusted to create nice seams due to the split..)
Limor Schweitzer over 11 years ago
Tan: "Please ask RoboSavvy to increase the number of types of servo (included dimensions & performance chart.." -- that's an interesting one. If you want to use an alternative servo to the AX12/AX18 please suggest some ideas. Please refer to hobbyking servos.
The choice of Robotis servos is because we know everything there is to know about them and they are great for robotic applications and few servos match their price/size/performance. We also know how to control them directly from Rpi/Beagle with only a small USB interface board. However, the AX12 is bulky and if you want to add a pan/tilt head to the robot, or a hand with fingers, you would have a hard time doing it aesthetically with AX12 servos.
So if you want you can use smaller size servos from hobbyking (please suggest) and we can add an additional small PWM servo controller to drive them. :-)
Limor Schweitzer over 11 years ago
chandan: "should RObosavvy in assemble mode or in only one part model pls clearify it" -- If the question is whether to send your design assembled or in parts.. the answer is assembled, but with all the parts available
Limor Schweitzer over 11 years ago
Julius : "Is it okay if we design different linkages for our humanoid robot?" - yes.
I have created one robot that uses wires instead of servos directly driving the joints. The structure was completely 3d printed in PLA which makes the robot heavier than it would be should the structure had been done in aluminum. In the page there is a picture showing another example of how a shell can potentially cover the internal structure and make the robot look better. http://goo.gl/DEHPav
Bar Danino over 11 years ago
@Limor
What is your definition of a humanoid? I've heard several definitions, where each allows or disqualifies different models than the other.
Limor Schweitzer over 11 years ago
Tan: Regarding using Bioloid plastic brackets as structure for the robot, we prefer aluminum sheet so that the robot does not depend on Bioloid kit parts. Having said that, let me know if you have a good reason to use them.
Tan Wei Kok over 11 years ago
@Limor If you are referring to RoboSavvy Humanoid Design Intel Report, it is not a Challenge entry. It's the only BIOLOID that I could found that looks close to BIOLOID GP with ready CAD files available for download and assembled for reference. It is to help the community better understand how the robot works when it is assembled, rather than guess work. It will also helps kick start more projects since we no need to work from scratch.
You don't have to worry about the plastic. Many of us here can turn it into sheet metal in a couple of days, if not hours =)
Limor Schweitzer over 11 years ago
Bar: "What is your definition of a humanoid? I've heard several definitions, where each allows or disqualifies different models than the other" -- humanoid implies human like shape and joints. ie: 2 legs, 2 arms. the more articulation the better. Can you please give examples of humanoid robots that you think may be questionable.
Limor Schweitzer over 11 years ago
Tan: "It's the only BIOLOID that I could found that looks close to BIOLOID GP with ready CAD files available for download and assembled for referencv" -- please press the "Download specification" button above. It will download the GP model with metal frame etc. The model was added only the past couple of days..
teigan over 11 years ago
i was given a bioloid for xmas. i've since donated it to a school, but i noticed it wasn't humanoid at all, apart from standing/walking bipedally. it couldn't transition into a crouch/crawl/kneel. all of these are essential to getting around the human environment. if a robot can't climb stairs, it is basically a cripple. still, that didn't keep bioloid from being a fun toy. are you satisfied with the bioloid functionality? how ambitious do you expect us to be?
Simonr over 11 years ago
@teigan - for once, I agree with you! The current generation of humanoid robots are just animated dolls. Cool, but essentially useless. It would be nice to be able to push the functionality - but I don't think that's what this challenge is about.
Tan Wei Kok over 11 years ago
Updated RoboSavvy Humanoid Design Intel Report with BIOLOID GP assy & part files. It is the same as the link above. Hope you find them useful =)
Bar Danino over 11 years ago
@Limor
Ypur definition of humanoid: "humanoid implies human like shape and joints. ie: 2 legs, 2 arms."
Poor Machamp...
http://www.pokemontradingpost.com/images/website/machamp.png
It is a good, wide definition though.
Limor Schweitzer over 11 years ago
teigan: "if a robot can't climb stairs, it is basically a cripple" -- crawling is quite a challenging feat for humanoid robots. If you can modify the GP to allow it improved mobility that's great! remember that the knees and feet need to be able to lift the upper structure during the walking gait ; and the great looking exterior shell needs to be there and not impede the movements.
Limor Schweitzer over 11 years ago
simonr: "It would be nice to be able to push the functionality - but I don't think that's what this challenge is about" -- the scope of this preliminary period before the actual competition was to get some feedback from the CAD engineering community about how far we can stretch the goals. It would be great to see some interesting structural design innovation. We are concerned about setting the bar too high. The GP is a great starting point and we'd like to see exciting shell designs that may require modifying the metal frame of the GP.
Limor Schweitzer over 11 years ago
BTW if you look at the GP, it has a battery pack and a CM510 controller pack. Neither of these are needed for the functioning of the new robot. The controller will be replaced by a RaspberryPi / Beablebone. The battery requirements were mentioned in previous post on this page.
Limor Schweitzer over 11 years ago
Regarding the webcam. It can be any miniature webcam that's easily purchased on amazon, ebay, aliexpress, dx.. We have a bunch of "Hama pocket webcam" that are really tiny when stripped off their plastic shell. They are not as easy to source any more. Here's a rough model of the stripped down camera PCB: Hema Pocket Webcam
chandan over 11 years ago
is this a assembly or part
Julius over 11 years ago
I hope you can set the bar much higher, Limor, and give us 2 to 3 months of design lead time with equivalent higher cash prizes and with more slots for cash winners. Almost all of us here are capable of complex kinematics design if you want the meatier aspect of humanoid development. As for me, aesthetics are just skin-deep and there are lots of human resources you can readily tap into that were not necessarily engineers but very skilled artisans/craftsmen and they could even share their designs for free. Also, beauty is in the eye of the beholder, such that any aesthetic design is purely subjective and cannot be fairly judged...
Limor Schweitzer over 11 years ago
Julius : "Is it okay if we design different linkages for our humanoid robot?" - yes.. here's a link to a xblog discussing how we are designing 3D printed structural elements for a humanoid robot http://robosavvy.com/forum/viewtopic.php?t=8115&postdays=0&postorder=asc&start=105
I just realized that I previously gave the wrong URL
Limor Schweitzer over 11 years ago
Julius: "beauty is in the eye of the beholder, such that any aesthetic design is purely subjective and cannot be fairly judged.." - Thank you for the comment and I tend to agree. We will set some engineering challenge criteria including designing to allow the robot to crawl, allow for pan/tilt of the head, and adding hands with grasping capabilities.
Limor Schweitzer over 11 years ago
-----------------------------------------------------------------------------------------------
Thanks to your feedback we have updated the terms of the competition. The criteria for choosing the winning design are better defined now and we have decided to opensource the platform with the objective to make humanoid robotics accessible to a wide audience.
Sesteves over 11 years ago
thanks for updating the criteria! is there an option for cash prize instead?
Sara Sigel over 11 years ago
Now, that @Limor and the RoboSavvy team have defined the specifications and we are moving into the entry phase of the Challenge. I ask that everyone keep comments specific to questions their entry, now. I have removed some comments that were not constructive. If you have any questions about using GrabCAD to upload your work, you can send those to me at sara@grabcad.com.
Julius over 11 years ago
Hi Limor. I'm glad you set the bar much higher. Yet, considering most of us here have day jobs, few of us have ample time to design a competitive entry in just 30 days. Hope you extend the deadline to allow us to thoroughly explore radical new ideas and bring them into fruition. Perhaps 2 or 3 months would be enough for everybody to participate to such intricate challenge. Any low cost yet astounding design takes months to complete, I think ;)
Joe Strout over 11 years ago
Neat contest! But why the specification for Bioloid servos? Herkulex servos are smaller, less expensive, and nearly as strong, not to mention easier to program. I think you should allow us to see what we can come up with using those, too.
Limor Schweitzer over 11 years ago
Julius : We have extended the time and as we approach the deadline, if you really need a bit more time let me know. But 3 months would be too long in this context.
Joe: We designated this robot challenge around the use the AX12 servos and not any other servos for various reasons (except as mentioned, for hands and pan/tilt head). Pls PM me for specifics about the servo you mentioned.
kaustubh about 11 years ago
please if possible extend deadline for few more days.. it might bring few more gud designs. maybe 15 days.
Sara Sigel about 11 years ago
The Challenge deadline is now October 20, 2013 at 23:59 UTC!
Pratik about 11 years ago
Thankx for extending the deadline
Pratik about 11 years ago
Working on one more entry::::::
ilyes about 11 years ago
good luck people
Sara Sigel about 11 years ago
Final push, here! Get your entries submitted before the deadline!
Che about 11 years ago
Good luck to all involved. This was fun and challenging! :-)
Luis Filipe Ferreira de Moraes about 11 years ago
How is the decision making going? There is a lot of exited folks here.
Mohomed Nusry about 11 years ago
Today it's 20th of November 2013, which means the winners will be announced today...
Am I correct Sara Sigel ?
Sara Sigel about 11 years ago
They're close! The jury is still looking through entries, as you can see, the guidelines are very complex. We will have finalists for you before the weekend and the announcement of the winners will follow a week after that. Thanks for asking @Mohomed.
Mohomed Nusry about 11 years ago
Thanks for the information Sara Sigel :-)
Sara Sigel about 11 years ago
Finalists are announced! Thanks to everyone who participated. We believe these seven entries are among the best solutions. Which one is your favorite? Did we leave anything amazing off the list? See something we should know about? Let us know here in the comments!
...
For those new to the Challenge process, we release what we believe to be the top entries. They are listed in no particular order. Over the next week, participants are able to voice their support or concerns. We encourage everyone following to leave a comment with their favorite or letting us know if we left a great entry out of the mix. Don't think of this as a vote but more of a soft announcement. We want to connect with you on these concepts and that means more than having you submit entries it also means getting your feedback, too.
Luis Filipe Ferreira de Moraes about 11 years ago
Nice selection! In my opinion the 3 best are
1st RoboSavvy Humanoid Design - unl by joshua.chung
2nd RoboSavvy Humanoid Robot V2 by mohomed.nusry
3rd Space Challenger by luis.filipe.ferreira.de.moraes-1 (ME lol)
I can't choose from the other designs, i think all of them have theyr traits and flaws..Many of them are very amazing designs but with not much visual apeal.
However, i think the choice will be simple depending on how much importance the judges will give to mobility and beauty.
Luis Filipe Ferreira de Moraes about 11 years ago
As an Add, RoboSavvy Humanoid Design - MV-03 by valerio is a very intresting design with very intresting solutions, However i do worry about the available torque of some of those servos. 67cm is a very big height for some of those servos, Am i right?
Che about 11 years ago
To all of the entries, Great work. To the finalists, I am honored to be among you. Ann Saint, I think there is more to forwarding humanoid and android reserch than simply making more human like robots.
Currently building a robot capable of playing robocup soccer will cost somewhere on the order of $12k. While I don't think the finalist's designs are capable of taking on a robot like a darwin-OP, or Nao, by combining a highly capable desktop humanoid with the computing power of the RasberryPi you are greatly reducing the cost of entry into something like a baby humanoid soccer league.
So instead of $12k a piece, and about $40k for a team of three, you are talking about $12K for a whole team. This puts autonomous humanoid robot soccer in the shooting range of regular high school classrooms. Think of the interest that would generate; the future engineers and roboticists. And that is just one pathway among a myriad of possibilities.
In anycase, Best of luck to all of you.
C- he
Sara Sigel about 11 years ago
Please keep this discussion based on selecting winners any off topic comments will be removed.
Mohomed Nusry about 11 years ago
Congratulation to the finalists! And I'm glad to have myself among you. :)
In my point of view the tops among finalist are;
Space Challenger by ttps://grabcad.com/luis.filipe.ferreira.de.moraes-1
RoboSavvy Humanoid Design - unl by Joshua Chung
RoboSavvy Humanoid Robot V2 by Mohomed Nusry (me) :p
Comparing those 3 designs, each of them have its own qualities.
When considering the appearance; UNL stands very high among those 3 designs. But still it may not stick to the proposed cost range. I think manufacturing the UNL will cost more than 10k and will take huge effort. And UNL lacks some required features compared to other two, such as hands with gripping.
As a conclusion, even the UNL stands top in the department of 'appearance', other two designs(Space challenger and V2) override UNL in the departments of 'cost effectiveness', 'advanced features' and 'ease of manufacture'.
Mohomed Nusry about 11 years ago
Please note that, above comment is just my point of view and not meant to criticize any design or designer.
Che about 11 years ago
To partially ecco Mohomed Nusry’s comments cost and buildability were important factors for me in designing the “traditional humanoid” (my design. The brackets used for the most part are sheet metal versions of standard off the shelf bioloid brackets. A prototype of my design could be up and moving around in short order. This also means that users could convert existing bioloid robots into this design inexpensively greatly increasing the user base and availability for users, and creating an environment where a maximum number of individuals are creating software, poses, walk cycles, and so on for the robot. This could be an inexpensive entry into robot soccer, and I would love to see a “baby humanoid soccer league” form which would have similar requirements as robo-cup, but use smaller robots.
I also included 2 different types of grippers in my design, 1 3D printed, which I have on 2 of my current robots, and work quite well, and the other using bent metal brackets. The use of AX-12s for the grippers was intentional because it gives the user the protection of having thermal shut down, load detection, and so on so that the gripper servos don’t overheat. Better would be a modified clutch system, but that would add costs.
I also worked to keep my design small and compact minimizing strain on ankle and knee servos.
There was some concern over using the raspberry pi as a controller and processor, There is potentially room for a small controller to handle poses and motions, and by using a camera system like the havimo you could remove a lot of the CV work from the raspberry pi also.
- Che
Valerio about 11 years ago
Hi all!
to start I must do my compliments to all the participants for the new ideas and solutions.
My praises to Robosavvy.com that promoted this challenger because I think this is a good way to diffuse the culture of humanoid robots not only in accademic reality but also in the maker's community.
Now I must promote my 3D printed humanoid robot; I think that with this solution I met the target proposed from Robosavvy: "make more economic and simple the realization of a humanoid robot".
For the economic aspect, the cost for a totally 3D printed robot consists of:
- the cost of all commercial parts like servos, controller, sensors,
- the cost of raw material for the printer; I estimated about 2 Kilos of ABS (about $ 180 on the Robosavvy store). Sure! you need a 3d printer, but on the internet you can find many that print for third with reasonable prices.
From the point of view of simplicity, realize my 3D printed humanoid is as simple as clicking on the start 3D printer button.
The model I posted is still in its infancy because I have not had much time for it, but in the philosophy of the Open-source I think may be a good starting point.
I think the strengths of my model are the joints of the legs, hands with fingers that can be managed independently, the joints of torso, the use of 3D camera for the view; for the brain I chose the BigleBone Black that I don't know but, on paper, it seems to me the more interesting.
I also placed the batteries in its feets, in this way the center of gravity of the robot is as much as possible to lower for the benefit of stability.
In my project I used Robotis AX18 servos for the legs for its high torque.
Best regards to all
Valerio
Pratik about 11 years ago
Congratulations all the finalist. Thanks to Grab Cad & Robosavvy team for bringing out this fantastic challenge. Also to the jury members for considering my design. I have enjoyed this challenge. Hoping that Robosavvy & Grab Cad will bring out similar challenges in near future. Good Luck to all the finalist.
-Pratik
Ian Turner about 11 years ago
After looking at the finalists designs, and their personal work, I've come to the conclusion that Che's design is the most viable to work and to be used. It may not be the flashiest but it is darn cute (like a short chubby Iron man). Plus the video was cool. It seems like he has talent and a lot of experience building real robots. I would buy this robot to scare my kids and cats, the grippers are perfectly designed.
Maya Kilmer about 11 years ago
I've seen many of Che's Robots, and they are all fantastic and functional. He has a lot of experience and practical knowledge, and it clearly shows in this design!
Veronica Koons about 11 years ago
Che's robots and designs are top of the line!!
Chris Farrell about 11 years ago
First of all - looking great guys! Great to see some excellent models and innovation.
There are three big things I worry about in this challenge, budget (first and foremost), weight (also a pretty big deal with AX series servos), and design complexity.
To start off - if we consider a robot using mostly AX-12s (at ~40$ each?) - a design with adequete flexibility will take about 7-800$ in servos - leaving the designer with only a few hundred dollars left for the shell and electronics. If we include battery and a rasberry pi, our budget is already nearly finished. Because of this - I would absolutely be looking at the most stripped down a simple robots (cover the metal framework where needed, but try to do so with soft foam or thermoform stuff - nothing heavy, - especially up top). If it were up to me (and the design budget was strict and you wanted it to be easily accomplished) - I would opt for having a robot with no plastic/cover pieces, just have a metal frame with additional holes and a guide for users to design their own frames (or perhaps include different frames as add-ons to the kit).
I like seeing the designs in CAD (rather than purely 3D models - show me the inner workings!), and while creative ideas (spring in the foot, gears at some joints) are cool - I'm not convinced they would work in real life (the gears remind me of the NAO, and I was never convinced of that part of their linkage). The Spring foot is interesting (One of Pratik's designs), but I fear it would throw walking calculations to the wall (ZMP is designed for flat planes the way it is - never mind stairs and such). I would love to see some videos and some exploration into this though!
Many of the fully covered bots look too heavy to me, though I like the ones that look very human in proportion (Valerio's MV-02 comes to mind here). The two things I worry about with this design are PLM in the shin (why? - is torque an issue because of all the plastic?), and of course budget (36 servos is huge!). The structure otherwise is tried and true - looks very similar to the Bioloid/Cycloid - so that's good.
I like the JISSI hollow faceplate/helmet - cool stuff!
I may be playing favourites, but I quite like Che's Traditional Humanoid design - it is simple, and I believe it would get the job done (and of the simple designs, I believe it is the most human-like - perhaps tied with Space Explorer). I would like to see the backpack hide away a little better, but you've gotta put everything somewhere!). I believe it is one of the only robots that is close to the budget (from my experience) and has low enough weight to be able to pick itself up and also move around and also seems to be near production ready!
My other picks would be Space Explorer (I would like to see hands on this model :) and the MV-02 (are those shiny pieces plastic, painted metal, anodized? thermoformed stuff?) I would say if you brought the hands down to a single DOF (with a more articulated option perhaps - and same with the head) then you would be much more in the ballpark of the budget. Love the stances.
Valerio about 11 years ago
Hi to all!
It's true! My models are expensive, but I developed the two robots following the philosophy of open-source project. In this way I think to propose a good solution for study, my robots have a big agility (24 DOF for the body), a big flexibility for the manipulation (10 DOF for the hands), an hardware for a good vision system (3D webcam to develop stereoscopic vision system like the solution of DIEGO - http://robotics.oskarpalinko.com/).
However, it's not necessary to implement all this features; it's possible to use only 1 DOF for each hand, reduce the DOF of trunk; the maker can choose the solution that prefer.
I also think that the solution more interesting is the 3D printed robot; now it's not complete because I don't have free time to develop.
A 3D printed solution is easier to make because you don't need a workshop but only a 3D printer on your desk, and if you don't have a 3D printer you can find on-line many 3D printing service.
I also think that the 3D printed solution is minus expensive.
About the weight I think that the 3D printed solution can be the more lighter. I started to draw the printed parts like filled parts, and now I'm studying the way to reduce the weight creating gaps in the pieces also useful to run the cables and put the electronic components such as gyroscopes and accelerometers.
I estimate that at the end the weight of all 3D printed parts will be minus of 2 kilos, with about $180,00 for material cost (ABS).
Lona Farrell about 11 years ago
I like the clean, smooth look of Che's design.
Seton Koons about 11 years ago
Che's design looks great! Yet another awesome piece of his work!
Michele Marshall about 11 years ago
!
Miata Martin about 11 years ago
Che-4 has an excellent design and the video on the blog is really impressive!
Cynthia Vick Allen about 11 years ago
This is such an awesome design! Che is constantly impressing me with his work!
Alida Cella about 11 years ago
I think the designs by Che look light enough to work. Over the years I have seen his robots, he never fails to impress me. His designs have become more and more sophisticated over the years. These look fantastic! Good luck Che!
Chris Novak about 11 years ago
I (of course) agree with Chris Farrell - let's make sure these robots are buildable... The Traditional Humanoid design does seem like a good pick.
Kristin about 11 years ago
Che's idea seems good because it does seem like it will be actually buildable. The other robots are too tall and heavy. Also his design would be the most cost effective.
Jen Fremd about 11 years ago
As previously stated, Che has a cost effective design; whilst regarding looks it is less sleek than others, the functionality and actuality of it seems far more likely. Particularly if it will be used in education, as an engineering student with interests in robotics myself I feel like this design is very plausible to learn from; there is a certain form of clarity on how each part will function and add to the robot as a whole. Overall it seems to fit a large sum of the requirements quite wellv
Caia Schlessinger about 11 years ago
Che's idea are innovative because they are practical yet cutting edge. They are sleek and cost effective so they are actually buildable.
Sara Sigel about 11 years ago
Such great feedback on the finalists from everyone. The Challenge was so close that these were very helpful in thinking through the winners for the jury. Winner are now announced - congratulations to all!
Pratik about 11 years ago
Congratulations everyone!!!
Che about 11 years ago
Congratulations to all involved. Can't wait to see little UNL's running around! Also, I am honored for being recognized, Thank you.
Arian Shamil about 11 years ago
Congrats to the winner and all the partecipants....i saw here a lot of amazing models! :)
Valerio about 11 years ago
Congratulations to all the participants and at the father of pleasant Unl... Joshua Chungi.
I like this challenge because I had the opportunity to share my ideas with the other designers and to find inspiration for new ideas.
Many thanks to all and and a GrabCAD!
Luis Filipe Ferreira de Moraes about 11 years ago
Congratulations to all participants!
Mohomed Nusry about 11 years ago
Congratulations to all...
Joshua Chung please share some of videos of your robot here in GC, once you received it.
Love to see it come to life. :)
Limor Schweitzer please let us know once it has been announced as a open source humonoid, so then we can contribute to the design with some developments. :)
Thnks GC and RoboSavvy for hosting another great challenge!
Please log in to add comments.
Log in