Shlomo
Canada
Asked
— Edited
is there any way to connect with someone experienced with building and remotely controlling a large robot for collaboration on a commercial project?
Related Hardware JD Humanoid
PS. I should've included my email to be contacted directly - it is [email protected]
First, welcome to the forum. there are a number of people making human sized robots who could give advice. If you can give some more info about your project goals you will probably get more input.
thx very much for your timely response. I just happened to get back to my computer so I can respond with the same timeliness (this time, at least). I don't want to publish on a public forum too much detail because of non-disclosure concerns and that is why I added my personal email address for followup conversations that could be more private. However, I can say this: the main requirements are 1) stable mobility (eg. wheelchair base), 2) instantaneous response to commands from a remote operator (eg. move forward, rear, right, left, etc.), 3) realtime audiovisual communication between remote operator and robot's environment, 4) 12 hour operation in between charges, and 5) some arm movement - to be discussed I do not have the detail orientation to build this myself. I am looking for someone to build it according to my vision (allowing for modifications based on input from the builder). I can contribute towards the expenses. The split in profits would be commensurate with the relative proportions of input from the participants and their ongoing contribution to revenue production.
OK, nothing there is outside the capabilities of the tools available here. Several ways to use ARC software remotely, some better than others. There are some license limits on commercial use that you would want to discuss with DJ Sures. If this is a small number of devices, probably not an issue. If you are prototyping something that you plan to market / mass produce, probably more of a concern.
I don't even have time to work on my own robots right now, but based on your description, I am willing to bet someone will be interested in helping. If you don't hear from someone soon you can ping me. I actually have a wheelchair base and a Sabertooth controller for it currently sitting idle waiting for me to decide on a project, so if I can find some time I might be able to help you prototype, although my build skills are not what I would call commercial quality.
Where in Canada are you? There are a few builders who are very very good up there. @Richard R I believe is in the Waterloo area, and EZ-Robot/SynthIam are based in Calgary.
Alan
Vision is great but someone needs to fund the construction of your voice activated wheel chair. There are a lot of people on here who could knock off a prototype for you pretty quick but I would also assume at minimum they would look for you to fund parts and reimburse labour costs. The build it they will come idea is great but we have seen a lot of initiatives fail in the robotics space. IF YOU HAVE FUNDS, you may want to try some smart people like Josh (mechanical engineering / Design and PTP electrical engineering / coding) https://synthiam.com/User/Profile/jstarne1-585 https://synthiam.com/User/Profile/ptp-6106
To Alan the Techguru I say: I am in the Toronto area, about an hour away from Waterloo. I wouldn't mind travelling there but you can also get a lot done via Zoom. I would love to hear from Richard R. To Nink I say: I didn't realize one could overview each member - thx. Jstarne looks good but doesn't describe broadly enough the practical applications he addresses; doesn't provide videos of finished project?! Ptp has much less completed. Still, I would be pleased to chat with either of them.
To both Alan and Nink I say: I saw that ARC sw and hw addresses my needs in a quite full way except for the fact of size. I haven't seen this product implemented on a life size robot - I wonder if the servo's (HDD) can support the weight of a larger size. The bartender robot seemed to walk very slow which was surprising if it was on an electric wheelchair base, as I was given to understand. But conceptually, it is very close to what I envision - I just don't have the detail orientation the build will take. The license issue is interesting. Right now, my idea is to build a working prototype that looks good enough to get some customers to commit to a purchase. At that time, they will describe more exactly how they want it to look and act and we will then go back and forth between price and full features that will make eveybody happy. I wouldn't be happy unless the result promised a sales volume in the 1,000's and that would include working something out with licensing, as necessary. I hadn't contemplated the remote operator using voice to activate movement, although I can see it is feasible. By "move right, move left, etc." I meant moving a joystick or sliding a finger on the screen as I have seen DJ Sures do. I think I can fund parts, perhaps labour too, but them the other party would not have any call on future profits. I was hoping to find someone who would become as enthusiastic as I about the vision (to be revealed) that they would become a sweat equity partner. Thank you both for your comments. I hope they lead to finding the match I am looking for, or even "a" match.
The EZ-B is a controller (The brain). You can directly attach and power servo's, hbridge to motors etc from the EZ-B but you can also provide power direct to the actuator. So it doesn't matter what you attach to that controller in terms of load as you are essentially just providing a signal to the actuator and an external power source.
To give you an idea of troubled robotic projects take a look at the Altair project https://synthiam.com/Robot/Introducing-The-Altair-Ez-2-Robot-563 lots of time, money and effort invested was an amazing robot but it is really hard to move these types of projects forward. I know a guy in Toronto who has spent millions building and marketing a robot (crosswing) it is a great superfast smooth omnidirectional robot but can't get passed prototype.
why can't it get past prototype? If that means nobody wants to buy it, I can understand why. The presentation does not concentrate on its benefits, what it can do for you. How well do you know this guy? Maybe I could work with him by marrying my vision with his. I once presented on Dragon's Den for an invention I called Shelevator for "shelf-elevator". It was a drawer designed for too high to reach shelves - this drawer not only went in and out but also up and down, so you could easily bring the contents down, take out what you need, and return the whole drawer to its original place. It worked well and was inexpensive but nobody liked the way it looked - I had neglected to test the market - that is why I am focusing now on getting a working prototype to test the market first.
Can't get passed prototype as like you said you need to make a certain quantity to make it cost effective "unless the result promised a sales volume in the 1,000's". So you need up front commitment to buy 1,000 robots or you need to pay for the 1,000 robots yourself and hope you can sell them. You can build one offs at $20,000 each or you can build 1,000 at $5,000 each but you then need to invest $5M so you really need to believe in your product (or get VC / Angel backing). There are a lot of smart people with a lot of money trying to make money in this industry and failed (Softbank / Honda etc ) but really the only robots that seem to be profitable are Toys (UBTECH) and Industry Robotics that replace human tasks. James Bruton Makes a living just making robots that will never move past prototypes http://www.xrobots.co.uk/
as for costs - take a look on Alibaba and you'll find several models selling for much less than 5K each at volumes of 1000 my concept includes a fundamentally different approach than everyone elses which is why I think it will be successful - still, have to test with prototype
@Nink, Good discussion and good insights. I agree 100%.
Another example of failure: https://www.therobotreport.com/rethink-robotics-closes-its-doors/
Having ideas is not difficult but implementing them is. Creating a prototype without funds is another issue basically is the chicken or the egg dilemma.
A robotic Arm is a good example, you can build a prototype with a $50 hobby servo to lift a chocolate bar. The problems will start when you want to scale up the prototype and then you have no affordable hardware (servos/actuators) options to build your prototype so you will need a 50K budget to get one robotic Arm from a hardware partner.An electric wheelchair is only a chair with motor wheels and remote control. Why are they expensive when you can create a prototype by a fraction of the price?
Jibo https://mashable.com/2017/11/08/jibo-review/#NyPo9DHiAaq5 price range was $1000, then $800 then $400 and then close the doors. Would you spend $400 when you have Alexa or Google Home for $40 or Siri on Apple device ?
@Sholomo Do you have an idea of the customer price range for your concept ?
Shlomo - It's awesome to see you have a great idea for a commercial product! Synthiam’s software platform connects robot builders with technology creators to program robot products, so you came to the right place. Robot builders get access to technologies without having to build them from ground up, and technology creators get usage of their technologies at scale. By bringing robot builders and technology creators together, the number of robot products will rapidly increase.
My role on the partners team at Synthiam is two fold. To work with technology creators (companies and individuals) so the technologies they create are added to ARC ; and to connect with robot builders who are looking to commercialize their ideas. There are many hurdles to bring a robot product to market as others have mentioned above. Synthiam aims to make that easier.
You can reach me at [email protected] to discuss the licencing concern raised above as well as where you're at and see if there are any introductions I can make that could help along your journey to a commercial product.
Hey @Alan, WOW did not know Synthiam was doing something like this. OK that is cool!
If I could give one piece of advice @Shlomo search to see if your idea already exists. If it does not exist request who ever you talk to, sign a patent or invention non-disclosure agreement before you discuss anything. You need freedom of movement to file a patent in future. Also a search maybe very eye opening as you would be amazed at what is already out there. https://www.youtube.com/watch?v=L3mPMQbJs6Q
@PTP - Want to know the common theme for those failures? They didn't use ARC
hahaha
In reality there are a two (main) reasons robotics companies fail:
The examples you provided promised the world, built massive dev teams with expertise in every aspect of the product so they could own all the IP and got lost in development, forgetting what was truly important - meeting buyer expectations. Jibo sold a ton of units in their pre-sale. That showed the demand was there and the reason they were able to raise 70+million. Once people got their hands on Jibo, they were all disappointing at the result. I would wager the majority of their funds were spent on software development. Maybe the outcome would have been very different if they integrated 3rd party tech rather than building it themselves? Maybe they wouldn't have burned through 70+million. Maybe they would have met user expectations and would still be alive today.
The industry is evolving quickly, learning from the faults of the past is a great way to progress. These historic failures shouldn't be seen as a dead end to people and companies with great ideas today, but rather an opportunity to learn from and change the way they bring their products to market tomorrow.
...a big part of a products success is also based on timing and when it comes to market.
Shlomo, don't let the naysayers kill your dream. Just because someone else failed doesn't mean you can't make it work for you. Find the right help and learn from other's mistakes. Sounds like Alan from the partners team at Synthiam has a way to help guide you and the team shares your vision.
Don't ever let them tell you it can't be done. Just find a way to make it happen. The Altair project that our good friend Nink mentions as being troubled is not dead or even that troubled. I just chatted with Tony a couple days ago. His little robot has found a now possible life as an assist robot to people that needs a hand or companionship. Will's (fxrtst) Male and Female Alan heads are alive and kicking (well maybe not kicking. They have no legs. LOL). Both these guys have the same thing in common that are making them successes, They keep trying, didn't listen to the ones that said "can't do", found help and adapted to the market. Also like Will said; "timing is everything". My two above examples above by not be "mainstream" right now but I know if the inventor builders keep with their vision their work will soon be.
Good luck and keep going.
@Dave Schulpius, I like your posts they are always positive.
There is a difference between providing examples and experience and criticizing someone.
Only to clarify I'm firmly believer you should follow your dreams and ideas, I've been following my dream: Creating, integrating, solving software problems for more than 15 years and when someones says it can't be done... I'll try my best with the available time to help and find a solution.
**** EDITED **** The previous post content reflected my thoughts. I erased the negative part to focus only on the positive side.
Thank you all for encouragement. I'll be taking this up soon with Alan re: Partners@synthiam as Alan suggested. I would also like to contact those with established projects to see if we could work together but can't find anyone current via cruising this website.
Meanwhile, a question bubbled up. If I made a life-sized robot using Synthiam's servos, would the arms be capable of lifting 10 lbs? Or, what is the maximum weight that servo can lift? (I am probably asking the wrong way because I am not conversant with the technical terms).
I'll break your question
The technical term is Torque https://en.wikipedia.org/wiki/Torque the standard unit is Newtons Meters but you can use Kg/cm, In/lbs etc. To compare torques you will need to have the same unit e.g. N/M, kg/cm, etc. Online converter tool: http://www.onlineconversion.com/torque.htm Very subjective, let's imagine your robot arm length is 80 cm and has one servo only in the arm extremity. If you add more joints/servos this servo will handle all the weight. EZ-Robot/Synthiam HDD servos torque are 19 Kg / cm source: https://www.ez-robot.com/Shop/AccessoriesDetails.aspx?prevCat=104&productNumber=1307 There are other factors like the arm weight, joints placement/rotations, but to keep the exercise simple the Torque formula:
source: https://www.grc.nasa.gov/www/k-12/airplane/torque.htmlTorque = Force * Length * cos(Angle) => Force = Torque / (Length * cos(Angle) )
What we know: Torque = 19 Kg /cm (EZ-Robot Servo) Length = 80 cm Case 1: Arm resting, Angle 0 deg. F = 19 / (80 * cos(0)) => F = 19 / (80 * 1) => F = 0.23 Kg In theory the arm can hold 230 grams
Case 2: Arm at 90 deg = 0.785 radians F = 19 / (80 * cos(0.785)) => F = 19 / (80 * 0.70 ) => F = 0.335 Kg In theory the arm can hold 335 grams
Bear in mind there are other factors, but this will help you start.
ROBOT ARM CALCULATOR: https://www.societyofrobots.com/robot_arm_calculator.shtml
Thx so much PTP. This is very kind of you to take so much time. So what kind of motor so I need to lift/support 10 lbs or 5 kg
I don't have a mechanical engineering background, there are other design factors like gearboxes, transmissions, joint placement.
for benchmark purposes a collaborative arm 80 cm 5 Kg payload (URM5) https://www.universal-robots.com/products/ur5-robot/ costs 35 K then you will need to add some addons, services, support etc.
maybe some other members can chime in and contribute their "bits"...
bear in mind we are DIY members, for a professional opinion take Synthiam/Alan's suggestion.
PTP, thx for reminding me. I convened a group from my community when the idea 1st hit me to make sure it was feasible - among that group was a young fellow in his last semester of mechanical engineering - haven't seen him in awhile because he was busy with exams, but now he will be a good source for a suggested solution I will also follow up with Alan, as you correctly suggest
Still looking for someone to help build the life sized robot -
Just want to say, thanks, great links
Awesome...this thread went off the roof since I checked it the last time!! Great to see the engagement...and the NASA Research Chart is supercool!!!
Why dont you start to develop the system on a smaller scale, to develop the software and scale the mechanical parts later?
Just an idea...great to have you on the forum!!
Mickey666Maus: thx for your encouraging participation. Your appreciation of the NASA chart nudged me to review it which then led to the following question for PTP: how could the arm at rest (Case 1) carry less weight than the arm at a 90 deg angle (Case 2)? In any case, my mech engineer friend will be the one to guide me through that. As for the last question about starting development....
Haven't you guys guessed yet? The software is mostly developed! It is waiting for final tweaking according to whatever features must be built into the mechanics. As in any marriage, you've got to start living together to work out the kinks:).
Still looking for someone to help build the life sized robot -
hi shlomo ! I think that , by life sized , you mean a walking humanoid about 1,70 meters tall. Are you aware of the problems involved ? You will need many motors with very high torque and cost, large and heavy batteries, a sophisticated control system, including accelerometers and gyroscopes , a rigid frame to hold all this, etc. i'm working on a 80 cm robot and have a lot of problems. I use Hitec servos D845WP, that seem good enough for that. To make a larger one would be very difficult !
Leonardo46: very fitting ID for inventor in Italy:). thx for commenting. Doesn't need to be walking, could be on wheels, like electric wheelchair or hoverboard; allows use of their large battery type which lowers centre of gravity. In fact, I have begun wondering if lifting 10 lbs/5kg could be done using a hoisting wire that would be powered by the same motor that turns the wheels axle and would only engage when wheels were disengaged (type of transmission??)
To move on wheels is much easier. There are many examples of such robots around. For the arms you might use cheap geared motors adding position sensors for feedback, or linear actuators (Servocity makes many types, with their own control system) or rotary servos with very high torque (I saw one with 280 kg.cm). Leonardo is my first name.
For weight distribution of carrying large loads, Boston Dynamics has demonstrated the easiest application. The trouble with humans imagining robots is that we imagine them looking like humans. The challenge with that is humans are actually quite terrible at specific tasks, and are instead merely competent at a wide variety of tasks. For example, the worst robot lawn mower you could design is a human
Here's boston dynamic's box stacking robot... Notice the counter weight to assist with it's inverted pendulum balance.
If you want a cheap hack for high torque you can always use a hacked battery drill and booker rod instead of a rotary servo for a fraction of the price. Here is Bruton's robot legs, he originally used hacked drills for (he eventually replaced them with proper motors). Sounds like you are still in the early concept and design phase. I was surprised to hear you already have the software written.
DJ: honored to have attracted your attn. Wonderful point about thinking out-of-the-(human)box, so to speak, for the "look". A business person doesn't care what the device looks like, only that it can perform efficiently. On the other hand, a consumer may yes care (or not) - I think jury may be still out. (Of course, I had seen the Boston video you linked but hadn't articulated the message you found.)
All the applications I envision need primarily an app that can control the robot device (of whatever appearance), as you have achieved with your EZ stuff. One of these applications doesn't need a humanoid body, but does need to exert more torque than 19 kg/cm. My business plan for only this one application predicts that, with only a 1% penetration into existing installs in the US, the net revenue stream would be $112M/yr., ongoing from just that penetration. As I have said before (and demonstrated) I do not have the knowhow or inclination to build the necessary hardware. But, I do have the software ready to go and need a partner to fill the missing pieces, ie. construct a working prototype (if need be with only temporary fixes such as Nink's suggestion of "hacked battery drill and booker rod") that can be shown to prospective buyers, get whatever feedback to tweak and off to the races.
Did you create the software in Synthiam's ARC? Because that would make sense in how quickly you created it!