I'd like to introduce you TechJect (a Georgia Tech spinout) that is building a robot dragonfly based on of years of academic research and $1+ Million in military funding -- aka, one hell of a toy! Early prototypes look pretty compelling (see photos and videos below). Currently, they have an ongoing IndieGoGo crowdfunding campaign that has already raised ~$200k in just a few days, where you can essentially preorder robot dragonflies for about the same price as a quadrotor platform! (Albeit with delivery sometime around September or October of 2013.) Presumably, their 4-wing ornithopter is able to independently control the pitch and amplitude of each wing to perform aggressive aerial maneuvering, which should enable the dragonfly to both fly or hover -- a unique capability compared to most other UAVs (fixed wing or helicopter). Initial product sketches suggest that the dragonflies will be ~6 inches long, weight ~25 grams, and have numerous sensors and connectivity options (ie. RC control, WiFi, IMU, cameras... the usual).
Here's the video from TechJect's IndieGoGo campaign:
If you're like me, you're probably disappointed with the ratio of robots to CAD models in that video... We'll revisit that later, but in the meantime, check out these photos of actual robot prototypes:
In the IndieGoGo comments, the TechJect folks mention an ICRA paper related to their robot: "Fixed Frequency, Variable Amplitude (FiFVA) Actuation Systems for Micro Aerial Vehicles." I just read through it.** It seems to be a survey paper exploring six (or so) different mechanisms: 4-bar linkages, direct-drive reciprocating motion, solenoid direct-drive, solenoid spring-based, "rotary-to-reciprocating" drives, etc. In short: I'm not sure which architecture is used in the actual robot dragonfly... or if there's a paper that describes the entire robot. I'd love to learn more!
** Jayant, please host an "author version" of the PDF on your website. I can't include the IEEE version here due to copyright!
I really love dragonflies (and thus this product). You see, dragonflies are one of the few (only?) carnivorous aerial insects that capture their prey mid-flight. Of all the low Reynolds Number flying insects, they are some of the most maneuverable and aggressive. In fact... my current project as a postdoc at Duke University is to (help!) develop small "sensor backpacks" that use wireless power harvesting and wireless comms to read a dragonfly's neural signals during flight.
My specific role is building the real-time backscatter communications system (a software defined radio) that can receive the high-datarate (5-100 Mbps) biotelemetry signals from our custom ASIC... but I can't talk much about it until we submit to Science or Nature. :-/ The best I can do for now is point you to an old press release.
But perhaps the most impressive ornithoper to date is AeroVironment's Hummingbird. Honestly... if I were TechJect, I'd be most worried about AeroVironment deciding to "toy-ify" their hummingbird.
TechJect has some great advisors, but I think they're making some faulty strategic decisions. Allow me to offer some armchair advice:
I'm not a fan of your IndieGoGo video. It's beautifully designed, but light on action. We all understand the concept and its potential. We love robots. We want to see more flights of the current prototypes... not CAD renderings doing all these amazing tasks. Spend your time on making videos (and photos) of real robots rather than fancy designer videos. [Though I must say... I laughed at Tucker's "I stand fully behind it" (aka, "I approve of this robot") cameo.]
Case in point... it was near-impossible to find images of your actual robot for this article. If there had been lots of awesome pictures and videos of your actual robot, I wouldn't have felt (as) compelled to fill the post with information about other ornithopters.
Delivery dates of July-October 2013? Not a chance. I'll bet you 5-to-1 that you miss your date -- severely. The prior is already against you. Only 25% of crowdfunded hardware projects deliver on time, and... you're not an iPad case or Arduino clone. Your product is a complex electromechanical system. I highly recommend reading through Orbotix's "From Concept Robot to Polycarbonate" series (parts one and two). A key quote:
First, we underestimated how expensive Sphero would be to build. We also missed on timing. No one had told us that it normally takes two years to go from a piece of paper to shipping an original consumer electronic device with complex electromechanics. We did it in 16 months.
I know that you have research prototypes, but it's still a lot of effort to go from prototype to mass-production. After that, go read Bob Christopher and Greg Appelhof's whitepaper, "The 2-year itch" about the pitfalls of bringing robots (ie. Pleo) to retail markets. There's also a good interview on Robot Launchpad.
The premise: it takes 2 years of retail sales before a consumer product company succeeds or fails. The first year is a honeymoon year, the second year is the adjustment year. If companies go out too fast the first year they'll hurt their chances of a successful second year, i.e. inventory challenges, markdowns, brand issues and retailer backlash.
Missing your deadline will really piss off users. You're pot committed now. Time to find someone with lots of overseas production experience -- someone who is willing to do dirty work, not just be an "advisor" or "consultant."
You've got three different avionics packages with different architectures, BOMs, gerbers, and (probably) code.
This is a mistake -- you're triplicating your efforts. This is fine for research-grade, low-volume R&D or prototyping... but you're going to be hard pressed to scale-up with three different boards. It's just too much overlapping effort: part sourcing, fabrication, quality control, inventory, firmware maintenance, and customer support. Your advisors should seriously discourage this... and if they encouraged it, they should be slapped!
If you're dead-set on product differentiation, then just pick the MARC-3 and don't populate certain components (eg. the camera and wifi chipsets). You can probably still alter your IndieGoGo campaign to use the MARC-3 with certain parts un-stuffed... but only if you act now!
You're going to need much more than $200-300k in preorders to finish your build-out. Orbotix (maker of Sphero) raised $5+ Million in VC funding and Ugobe (maker of Pleo) raised $26+ Million to bring their products to market. Presumably your crowdfunding wasn't (just) for the money; use the momentum to raise real money.
Personally, I would've just debuted at CES. You would've owned the show and got just as much investor interest. That's the route that Orbotix and Ugobe used to great effect. If crowdfunding was absolutely necessary, I would've picked KickStarter (as the crowdfunding market leader), though you would've had to wait until you had actual product prototypes ready (due to KickStarter's new hardware guidelines: no CAD renderings, only actual product videos).
OK, now that I got all that out... I think you're going to be a huge success! I'm really stoked to see a cool robotics company spun out of Georgia Tech (my PhD alma mater). Plus, I wish I had known about you guys while I was there... I walked passed your offices frequently on my way to work on the PR2 in the Aware Home.