matthewwarner-bct

[matthewwarner-bct]

Trouble with servos..

So, apparently KDS (a high quality servo manufacturer) sells their servos, rated for 4.5kg/cm of torque with a female plastic splined adapter that will eventually strip (start to rotate freely) if its actually run at the 4.5kg/cm spec for any significant amount of time.  Funny they’d sell such a nice, high quality, titanium geared servo that isn’t capable of even transferring that power to anything with the inculded mounting hardware. 

For now I’ll adapt what i have and strengthen the system with a number of bolts, but this is no long term solution. The plan is to stop any stress testing on the robot until exhibition date, as the robot is already working to a satisfactory standard, working at all is better than working slightly better.  

[matthewwarner-bct]

Cheshire dog update, again. Refined the walking algorithm, and accounted for some offsets made while assembling the joints. Think I’m about done with tuning the code now, time to move on to presentation planning.  

[matthewwarner-bct]

Cheshire dog update number... I’ve lost track. Everything is up and running, wireless controller with the windows tablet and game pad is running, communicates with the robot wirelessly over serial and the robot acts on the commands its being given. Now just for a few very minor touches and the project will be all but complete.  I really think the time I took at the start of the semester to plan out exactly what I was going to do and figure out the best and most streamlined way of going about it really is to thank for not encountering any last minute issues here. Everything went as planned because a lot of time and research went into making sure all the parts would work seamlessly together and give very few issues. I’m really happy with the end result and will continue to fine tune it right up until the deadline. I think you guys will really love what I’ve done here. 

[matthewwarner-bct]

...And it caught on fire, above are the two culprits, power regulators don’t like being connected backwards!! Oops... no big deal, will replace them tomorrow and be right back on track. 

[matthewwarner-bct]

FIRST STEPS OF CHESHIRE DOG!!!! Super excited by this, my creation stumbles out its first steps, needs alot of fine tuning and work but it finally works!! 

[matthewwarner-bct]

Cheshire dog update! All legs working and all technical issues sorted out. Feeling good about things finally!

[matthewwarner-bct]

These are all the models I made for Creative audio in Maya 3D. The basic level I made was done in Unreal engine as it worked out better to have the level surfaces generated by the game as we were having some collision issues with the floor and the VR player’s z height. 

I Overall learned a lot about Maya throughout this project, and am happy at my outcome. 

[matthewwarner-bct]

Play test of our finished creative audio VR experience. 

[matthewwarner-bct]

A thought on soundboards

Me and Thomas came to the realization today that our whole project is basically a glorified sound board, and got us a bit worried about how well the project will be received, as most of the work that went into it was 3D modeling and game mechanics, none of which relate to audio. 

But then we realized that this is also true of MIDI mixers, add on devices for audio mixing, mixing consoles, DJ desks and much more are all made by product designers and electrical engineers that probably know very little about audio. What they are making is a product to enhance the audio experience and or entice people to play around with things in a better or more interactive way, which is exactly the goal of our project. 

As they say, there’s many ways to skin a cat and our method for taking on a audio project was just from a different angle than most would take, ultimately the outcome should be just as good as someone coming at it from a more conventional approach, at least I hope. 

[matthewwarner-bct]

As we got our game finished today I thought a brief panorama of the models and a basic level I contributed to the project. As this is some of my first experiences using the 3D modeling program Maya I’m fairly proud of what I’ve managed to achieve as I’ve had to grow my skill-set and re-frame my mindset for that specific side of 3D modeling. However a few weeks before the deadline I had a decent portion of our models done as I was getting more and more proficient with the software.  Thomas showed me how to code the audio samples to the objects, so it de-spawns the projectile and plays the appropriate sound and spacalised in 3D. So I imported my models and tied the appropriate samples to each of the objects. 

[matthewwarner-bct]

Short Dev log of the games progress

Uploaded a clip previewing a look into the development process for our VR Audio experience:

https://www.youtube.com/watch?v=fhbPlFkFcQ8&feature=youtu.be

In the video I’m going over each of the models I made for the game and linking them to their audio sources. Some of the models have more than 1 audio track each, so me and Thomas wrote a randomizing script to play a different audio sample each time the ball hits the object. 

Me and Thomas play tested the first prototype and were very satisfied with how it worked. Some more polish is needed but I think we’re almost there. 

[matthewwarner-bct]

Think I need to rename my blog to just; “I build stuffs, mang”

[matthewwarner-bct]

CNC’d the chassis for my The Cheshire Dog today! really stocked with how its come out.

Was a more complicated shape than we usually are able to do so had to fire up fusion360′s CAM simulation environment to generate the necessary G-code for the CNC machine, just generating the toolpaths for machining the shapes. 

[matthewwarner-bct]

VR Physics Systems

http://www.theverge.com/circuitbreaker/2016/8/24/12622554/leap-motion-unity-interaction-engine-virtual-reality-hand-tracking

It has occurred to me that there aren’t many VR physics engines that work well. Current iterations suck. Many preceding examples are very prone to glitches, and break user immersion in almost every identified precedent. The manipulation of objects in-game still remains challenging. This remains a crucial piece of the VR usability puzzle. 

As we are dealing with the interaction with objects in our project i found this article particularly interesting, hopefully some of this tech will be able to be implemented in our project to improve the user experience and immersion.  I hope to utilize the newly developed physics toolset developed by Leap Motion to improve VR project usability in the future.

[matthewwarner-bct]

Day dream VR - Googles bet on the VR market

http://www.theverge.com/2016/5/19/11713498/google-daydream-mobile-vr-virtual-reality-cardboard

I’ve always watched closely what direction google takes in new tech trends closely, as in the past they always seem to have some super human sense for exactly where a emerging market is headed (think android, and many more). Maybe they secretly invented AI long ago and put it to use analyzing trends of the technology market... Anyways i digress, Daydream VR, Google’s new bet on the direction of “mainstream” VR is cheaper, light unteathed headsets that run a new version of android developed specifically with VR in mind.

Personal I can see a lot of merit in this idea, as it makes the VR “experience” much more approachable and a friendly experience, taking the requirement for large setup times, a powerful PC and cables out of the picture will be the only way VR will work for the general market. 

Applications such as productivity, media consumption and such is where i think this platform has a leg up on the competition, and what interests me about the platform. 

[matthewwarner-bct]

Got my test print of one of the legs of the dog back from Ross, testing the fit of the servos, compliance of the leg (ability to bend and flex) and soon to test if the servos have enough torque to even lift the robot in the current configuration. Important that they can actually lift the robot right? I seem to think so, which is why I’ve got these test parts as early as possible so I can make sure my design is sound before committing to much time to it. 

Did a more complete Render of the final design aswell, overall I’m happy with how its looking.

[matthewwarner-bct]

Some more of the very much working title; Robo Doggo, project concepts. 

I’ve added the shoulder pivot points and bracing to the CAD model, just took a quick render. Everything in red is 3D printed parts and all the metal parts will be Machined by me at my place of work on our CNC machine. Time frame for the parts is still very much TBD,

I’m also not completely happy with the design, so further revisions of the design will be made. A few more days work now can save alot of trouble down the line, so I’m taking my time.

I’ve invested alot of time into ways of streamlining how the system works, reducing workload by not repeating work readily available from common electronics vendors such as Polulo and Sparkfun. While I believe I am capable of developing all the components myself, I simply don’t have the time. 

The end product will not suffer any compromises for these changes, just my pride. But I think I need to start putting things like other projects and my markable content before my own personal satisfaction.