Go Down

Topic: An affordable robotic cat runs like Boston Dynamic Dogs? (Read 1 time) previous topic - next topic

RzLi

Hi there,
I just made a post on Makezine with a demo video.
https://makershare.com/projects/opencat
A same post on Hackster:
https://www.hackster.io/petoi/opencat-845129

If you are interested in the project and want to have one in hand sooner, please like the video and share. I also love to see your comments to make it better.
Thanks for your support!

original post below:
_____________________________________________
Hi,

You may have seen Boston Dynamic Dogs and the recently released Sony Aibo. They are supper cool but are too expensive to enjoy. I'm providing some affordable alternatives that have most of their motion capabilities. Currently I have two functional prototypes:

* The mini model is a stand-alone 8-DoF Arduino motion module that holds all skills for multiple gaits and real-time adaptation. It can also be a "leg-hat" for your existing project. It's targeted at STEM education and Maker community. I'm already selling the kit to local institutes and this model will kickstart my initial campaign. Its price is similar to some robotic car kits.  

* The full version uses a Pi for more AI-enhanced perception and instructs an upgraded 16-DoF motion module. Besides Pi's wifi and bluetooth, it also carries touch, infrared, distance, voice and vision interfaces. All modules have be tested on its light weighted body. It also adopts some bionic skeleton designs to make it morphologically resembles a cat. It's targeted at consumer market with less tech backgrounds. You can imaging it as a legged iPhone or Alexa that has an app store for third party extensions. I have a regular routine for duplicating the model, but need better industrialization to reduce the labor. I expect the price to be close to a smartphone.  

* I also have an obsolete version that uses only Pi for controlling both AI and motion. The movement is not as good though.  

You can focus on coding it on either motion (Arduino with C) or AI (Pi with Python) part. And the two can communicate with string tokens. Currently it can run at about 2.6 bodyLength/sec for 75 mins. And it's pretty easy to teach it new postures and behaviors with a couple lines of codes.  

I bought my first RasPi in Jun.2016 to learn coding hardware. I made the 6 major iterations between Jul.2016 and Sep.2017. I bought an Arduino in Jul.2017 for better motion control. I really want to devote full-time for this project. However,  I'm at the bottleneck with my personal resources. Now I'm applying for several accelerators and will probably try Indiegogo. Before that campaign, I'd like to know your needs and expectations for this kind of product.

- Would you be interested in this project?  
- What features do you care the most for a quadruped robot?
- How much would you pay for your ideal robotic pet?
- Would you prefer a dog version?
- Most importantly, would you prefer it to be an open-source platform supported by our Maker community or a well-packed commercial product supported by investors? I may post the link to my demo video if appropriate.

Thanks!

vinceherman


RzLi

Hi, I have attached a photo showing its skeleton at the end of original post.

vinceherman

It is a nice drawing.  But it does nothing to indicate that you have working prototypes.
We would like to see pics and videos of the working prototypes that you have.

Graylord

So... You wish to miniaturize experimental, high end technology that took a multimillion dollar robotics company many years to develop and on top of it all somehow make it cheap?

Good luck?

Do you have any ideas of how you are going to actually achieve this?

vinceherman

#5
Feb 16, 2018, 05:09 pm Last Edit: Feb 16, 2018, 05:11 pm by vinceherman
That is why I wanted to see pics/videos of the existing prototypes.

A quadruped robot is pretty cool all by itself.
One made to look like a cat is a big step above the typical spider bots.
I am quite curious how it is arranged.  What moves the limbs?  Direct connections from hobby level servos?
That presents problems with the longer limbs reducing the force that can be generated from the torque they provide.
I want to see how the OP got it working.
Different actuators?  Hobby servos with alternate connections that allow for better geometry to provide mechanical advantage?
Does it have active balancing?
Simple quadruped robots typically drag a foot since they cannot balance on 2 legs.  BD creations actively balance and can perform a 2-leg-in-the-air gate.  Does this prototype do that?
If so, the OP has a very interesting, very marketable product.
But a drawing is not an indication of what it is.
Pics.
Videos.
Let us see.
Even if the 'mini' version is simply a cat shaped bot, it would be interesting to see it in action.

RzLi

Thanks for your insightful comments. They are actually the major challenges I had to solve during the development.

It's driven by hobby level servos (but still robust, digital & metal gear) considering price. Some elastic structures were introduced to damper the shock and protect the hardware. Making the system small actually avoids a lot engineering problems of those larger models. The simplified structure allows faster iterations and optimization, just like rats evolutes faster than elephants.

Regardless the hardware, the major control algorithm could be shared once accurate mapping of DoFs is achieved. I'm currently implementing the motion algorithm on a 32KB,16MHz Arduino, using up 99% system resources with algorithmic optimization almost everywhere. I'm going to switch to the 256KB,48MHz boards to boost the performance of active adaption, as well as allowing additional codes by future users.

I derived a motion algorithm (with a dozen of parameters) for multiple gaits. The current fastest speed is achieved with trotting(2-leg-in-air). As I constantly adding new components and changing the CoM, while the adaptive part is not fast enough, I reserved the tuning work till finalized models. New users can also experiment with their list of parameters.

I'm not saying I can reproduce the precise motions of those pioneers. I'm just breaking down the barrier from million dollars to hundreds. I believe in the power of open-source community if everyone could grab a robot and start soldering and coding.

Let this post serve as a survey on the maker market. There are more obstacles in funding&time rather than R&D during the past year. Family and friends kept asking me why I don't find a full-time job, and non-tech investors are not confirmed about its value. I'm trying to sell the mini version to validate the market and bring in some revenue. The full version is yet to be polished.

Attached is another photo of the full version on table. I'll put up a story with the demo link on Arduino Project Hub. Please reply if you are interested!

bms001

- Would you be interested in this project?
I'm very interested, but in the sense that I'd like to see and understand how it works. I don't know if it's necessarily something that I would buy.
- What features do you care the most for a quadruped robot?
Smooth motion. How it copes with different and uneven surfaces.
- How much would you pay for your ideal robotic pet?
I wouldn't buy a pet type robot. But a 'project' that can be developed further I would consider. If I did buy something I'd be unlikely to spend more than £100 but it really depends what its capabilities are.
- Would you prefer a dog version?
No, I like cats.
- Most importantly, would you prefer it to be an open-source platform supported by our Maker community or a well-packed commercial product supported by investors?
Open source.

It would be really useful to see a video to get a better idea of it.
Can you tell us any more about how the local institutes have been getting on with it?  What sort of level they are? What have they done with it so far? What are their other aims with it?

RzLi

Hi there,
I just made a post on Makezine with a demo video.

https://makershare.com/projects/opencat

If you are interested in the project and want to have one in hand sooner, please like the video and share. I also love to see your comments to make it better.
Thanks for your support!

RzLi

It's used for a robotics class I teach, approved by the department of course. I have 15 CS undergrads with diverse experience in Arduino. I'm figuring out a proper difficulty/completeness of the kit, so that future users can assemble and code it without too much challenges.

- Would you be interested in this project?
I'm very interested, but in the sense that I'd like to see and understand how it works. I don't know if it's necessarily something that I would buy.
- What features do you care the most for a quadruped robot?
Smooth motion. How it copes with different and uneven surfaces.
- How much would you pay for your ideal robotic pet?
I wouldn't buy a pet type robot. But a 'project' that can be developed further I would consider. If I did buy something I'd be unlikely to spend more than £100 but it really depends what its capabilities are.
- Would you prefer a dog version?
No, I like cats.
- Most importantly, would you prefer it to be an open-source platform supported by our Maker community or a well-packed commercial product supported by investors?
Open source.

It would be really useful to see a video to get a better idea of it.
Can you tell us any more about how the local institutes have been getting on with it?  What sort of level they are? What have they done with it so far? What are their other aims with it?

vinceherman

#10
Feb 21, 2018, 03:12 pm Last Edit: Feb 22, 2018, 02:19 pm by vinceherman
RzLi, you have done a lot of awesome work.
That video does shows how successful you have been in making your larger model move.
The mini, with its 8 DOF does walk like most 8 DOF models.  Less realism in the gait.  But much opportunity to use as an affordable educational tool.
Great work!  I wish you luck in your adventure!

RzLi

Thanks for your blessing!

I realized that the larger model is far from a consumer product. So I made a retro design (of Version0) as the current mini model. I've been challenging the system for a higher speed. It walks better when I slow it down in other gaits.

Good news is that both the two models share the same motion driver. And I could define a new model between the two to have more DOF and still easy to make.

RzLi, you have done a lot of awesome work.
That video does show how successful you have been in making your larger model move.
The mini, with its 8 DOF does walk like most 8 DOF models.  Less realism in the gait.  But much opportunity to use as an affordable educational tool.
Great work!  I wish you luck in your adventure!

Go Up