I want to share my FPV setup I built on a small budget which uses some Virtual I/O iGlasses made in 1995.  I bought these in 2001 for about ~$75 and they are 600x800 in resolution and 120htz (60htz per eye).  They are capable of displaying 3D video too.  The 3D effect requires a video multiplexor, and while this has some neat applications for UAVs, I won't talk about that here.  The goggles were originally pretty expensive and universities used them for VR caves and the military used them as well.  They use a prism so you can actually see through them, I personally think this is somewhat preferable to the Oculus Rift head-in-a-box approach.  I admit on bright days you may need a dark bag anyway.  These were originally sold with an optional head-tracking unit as well.

The iGlasses use a barrel jack that is connected to an AC/DC adapter which provides 6v DC power.  I disconnected this and soldered on a USB cord I had lying around.  This way I can use an inexpensive USB power supply and make them portable.  The USB power supplies typically run about $5-10 dollars and the CR18650 lithium cells can be found for a couple dollars each or even for free when re-purposed from old laptop batteries.  If the 4500mAh rating can be believed then I have 18Ah of power for each little USB black box.  The video RX pulls more power than the iGlasses.  I'm not sure how long it will run, but far longer than my quad.  I've run an rPi for nearly 24hrs.  Another nice feature is that they can charge in a daisy chain, so I can charge several and even a GoPro at the end of the series with only one charger.    

The iGlasses are hit and miss on eBay.  Often they don't come with a cord or the head-phones are broken. Some people think they're still worth the $1,500 they paid in 1995.  Still, they frequently go for $20-80 (and by writing this I may affect the existing near zero demand).  I have flown with the headphones and they provide some information on intermittently windy days, but for the most part all you can hear is the props.  The headphones are not really necessary.  The company that makes them went out of business, but I suspect some of the video glasses for sale from China are direct descendants, derivatives of very nearly the same technology.  I have been watching them for years and some of the same selling points crop up in the new sales literature.  These iGlasses were supposed to be like watching an 80" TV from x number of feet away which occasionally shows up in the video glasses listings on eBay (the silver ones with a red-Jordi-style-visor.  There are still cheaper versions which I doubt perform as well as these do, as well as nicer ones too.    

The cord is not crucial either as the iGlasses use the same header pins most of us are already accustomed to working with for servos and flight controllers, etc.  If you look at my setup above the cords are probably the heaviest part.  I haven't started to make my own cable yet, but I'll share the pinout I stumbled on some time ago.

These goggles are pushing 20 years old.  They're not HD, but for the analogue stuff most of us are using they're great.  They're not super bright either so I typically have to fly standing in the shade when there is full sun.  I made a cardboard visor which helps and when the iGlasses were mint they came with a similar piece of plastic for the same purpose.  They work great in the evening.  I personally haven't tried any fatshark products, but a friend who has said he liked my setup better than his goggles.  The resolution is adequate for most, I believe; it allows for good compositions for photography and FPV flying.  

     The picture to the right is a small photovoltaic charge controller in the tractor (right).

It was made by lyncom around 2001 a company possibly purchased by Kyocera.  It is pretty much a toy, but the resin encapsulation is similar to higher end charge controllers such as USA made Flexcharge and should weather well (better than

 the rest of MY shoddy fabrication skills!).  The reverse flow blocking diode an LED so it lets you know if it is working.  Neither Flexcharge nor this lyncom support charging lithium ion batteries (like the Sabretooth ESC I'm using says it d

oes).  The lyncom has done a fine job boosting the battery voltage in early spring sun for even a short period of time. 
The panel is a direct from China eBay special, 10 watt polycrystalline el-cheap-o. I dremmelled out the junction box a little.

     I also made some progress with a rickety makeshift pedal switch (below).  It probably won't last long, but it works and will be covered by the plastic pedal. The tractor is a basket case right now, which is good until the end of the semester.  I'm so obsessed with robots right now that I'm questioning my life decisions.  I feel like this is exactly what was happening when I built my first computer.  That gave me valuable skills, robots might be more important than my Ph.D.  If I was a better coder I could bring some of these machine learning ideas into political science.  I'm not the next Steve Wozniak, or Jobs (since we're both ENTPs).  I'm happiest right now when I'm contributing to this corpus, or body of the mind.  

      Another small victory, I was able to VNC from VMware into the raspberri pi with X11VNC like I had hoped, but I didn't make a startup script in the init.d like I need to.  I did not get distracted by the new Kali linux, and definitely not distracted by an Rpi asterisk.  I am completely smitten with the idea of Interactive Voice Response working on the pi and having my robot talk to me.  I love the Human Computer Interaction (HCI) feature of having the Rpi speak its IP address on the asterisk iso.  I got the CMU Festival package running and played with it on the command line.  It sounded a little better than my Amiga 500 did when I was 8. I got a little distracted. 

     I remember reading that the old body style Prius (the one that looks like an Echo) had 244 D cell Ni-Cad batteries.  A friend of mine made a laptop battery from 10 Ni-Cads in battery packs taped together (10x1.2=12).  I'm guessing that the 18650's are what are in the new Tesla cars now. I have recently taken apart a few laptop batteries apart to feed my flashlights the 18650 cells I find inside.  It wasn't until reading lady ada's battery tutorial that I got a sense of exactly how volatile the unprotected cells can be. I ordered $20 bucks worth, 10 protected cells 3.6 amp/hour @ 3.7 volts from an ebay importer to tinker with (7.2amps @ 14.4 volts as a pack).  The price per amp is close to the retail I paid for the (7.5 amp/hour @12v) Sealed Lead Acid at Wall-Mart.

     The biggest problem is still figuring out how to mount the servo to the irregularly shaped chassis of the peg perego.  Everything is waiting for me to solve this problem.  I had an epiphany in bed one morning and had to draw it (below).  I haven't worked out the details, but at least I have a plan.  To compound the problem, I also ordered pneumatic tires.  I felt pretty inspired after seeing what saltydog (the url domain goes to the famous cafe which is interesting) had done at modified powerwheels.  I really admire the simplicity of his approach; put a drill in it, and move the drill's high quality switch to the pedal. 

     This provoked me to contemplate the steering linkages which may need to be replaced if the . The tie rods are already very difficult to turn and the wheel provided a lot of leverage.  The rubber tires will just make them more difficult.I may need to remove the existing plastic and replace it with some sort of ball bearing joint or even completely new spindles (a new word I've learned).  Steering is just fascinating and I have been having a very difficult time visualizing this problem. 

     I searched for Peg Perego and 3D printing and found another "powerwheel" project.  This approach felt much more steeped Mechanical Engineering than garage mechanic.  I noticed he was using Google sketchup so I downloaded google sketchup and made my first attempt at a 3D drawing.  I drew the pentagonal shaped final drive on the gear box, which if I could print and add to thingiverse, 3d warehouse or the like would solve a very common problem for folks building rovers from kids-ride-on-toys.  I may be able to use an ABS printer in the near future, so I'll need to improve my sketches considerably.    

     I hope my user generated content is of some value to other makers out there, and Mr. Anderson.  I appreciate diydrones' hospitality, I'm not sure if squatting here, or for adopting me.  I realize lots of the developers are into flying, but it's taken me a while to do my own orienteering.  The margin for error is simply too thin for me to do the kind of tinkering I like to do at those speeds.  The recent UAV safety hortatory was certainly sobering.

I recently purchased a raspberry pi from adafruit.  Without much difficulty I was able to get a demo python script for opencv running (see here).  It was able to detect my face as a face and that was kind of exciting.  It was a little choppy obviously, but still pretty fantastic.  I used an old PS2 eyetoy which you can pick up at a gamestop for about $2.  You'll get what you pay for.  One trouble with the Rpi is that they haven't released their camera yet, and it doesn't play well with a lot of USB cameras.  I was also given a GoPro Hero 3 for Christmas.  This is the silver edition, which uses different hardware from other editions.  The GoPro is an incredible little machine.  The GoPro isn't a webcam in the traditional sense, except that it is.  It charges via USB, but it doesn't stream images this way, it uses a webpage and a private Access Point (AP).  It can stream images to an iphone app (and a 3rd party Windows shareware) which begs the question as to how.  To see this work you have to connect your iPhone to the Hero3 as an access point.  My Rpi isn't going to be running Windows, so I'm not interested in the shareware unless I can port it to UNIX.  Looking at the settings on the iPhone, the Hero3 is the router @ 10.5.5.9 and the iPhone is the client at 10.5.5.119.  So I googled 10.5.5.9 & GoPro to see that someone else has already figured out what port to use; 80:80.  When I go to http://10.5.5.9:8080 on my iPhone I get a screen that looks like this:

This is really exciting because I've already seen people using Eye-Fi cards to upload their info.  I have an Eye-Fi and I hate it.  I hate the fact that it won't just let me share over samba windows networking.  That's probably unfair of me.  I'm sure it is, but I already have enough wing-ding widgets trying to run in the taskbar and I didn't want to have to leave a computer on just to transfer files.  Nor did I want to post 1,000's of pictures of my yard to facebook.  I felt really cheated of my money. 

There are 4 directories available from this little unix machine (the GoPro).  DCIM is what you'd expect; a directory for your pics and videos just like on most cards and cameras.  Live is the interesting one (and I don't really know what mjpeg and shutter are for as both were empty).  I have an axis network camera that uses mjpeg which I guess is a moving jpeg, perhaps for embedding. 

However, in the live directory are two .m3u8 files which can be used to view the "live" video feed from the GoPro.  The link above says this is an Apple filetype for streaming to iOS devices.  The files aaba.m3u8 and amba.m3u8 both stream, but the amba seems to sustain the video feed.  "Live" is in quotes because there are a few seconds of delay, just as there is with the GoPro app.  But still, Tada!  Now in theory... all I have to do is have my Rpi log in to my GoPro and I have a video feed.  I still don't know how to get Python to talk to a URL rather than a USB, but I think I can overcome this.  Everything in UNIX is treated as a file, so I doubt it will matter.  I'm not sure I can overcome the lag.  My robot is pretty fast, perhaps not the fastest, but fast enough to wreck in the lag interval.  Note to self, the next time I get it running, I should give my phone to the test driver (my daughter) and track the top speed with GPS.  I would like to stream the video without requiring two wireless cards.  My networking skills aren't awesome, and UNIX does such a great job with multiple nics (network interface connections) that I'll probably just use two cards.

One more piece of curiously juicy information. Googling "amba.m3u8" and "GoPro" returns a .pdf of patched GoPro code.  And, just to confirm, all the coolest stuff is open source: the original tarballs are available from the gopro website.  I don't really have the skills necessary to tackle a project like this.  I'm just hoping my bird-dogging helps the next person barking up the same tree.  It's encouraging to know that the API might be sufficiently accessible to make remote APM2.5 or Rpi control possible.  More likely, the people with the skills to decode an API don't need my help.

My intention is to be able to VNC into the raspberry pi to program my Arduino.  This way I could easily upload completely new sketches wirelessly. The addition of computer vihttp://www.eecs.ucf.edu/seniordesign/fa2008sp2009/g11/pictures.htmlsion is even more exciting. There are a couple of snags so far; I haven't been able to connect to the Rpi with x11VNC from my Oracle Ubuntu VM.  This is possibly because I forgot to connect to a "broadcast"ed SSID.  The Rpi will only connect to WiFi networks which broadcast their SSID.  This is sort of lame because my whole precariously cobbled together network, which is built from repurposed DSL routers as APs, etc, doesn't broadcast.  I realize it doesn't make a difference; it's just a preferance.

There really isn't anything new under the sun in the google-centric universe.  I've found a group of 4 engineering students that did the exact same project 5 years ago that I am attempting now.  They even used the exact same servo for the steering, and they struggled with the same problems in implementation (the knuckle is at an inefficient angle).  One document they posted described how many 100's of hours it took them to build which was sobering.  It made me a little taller.  I can say I'm glad I can buy an arduino for $30 and not have to build it from scratch; I'm pretty sure that saved me a lot of time.  There are probably a lot of things engineering students should have to do for their senior project that a political science student shouldn't have to do.  I have several old GPSs lying around in old iphones, a tomtom, a garmin.  I don't want to have to solder in a new one and invent the wheel.  I hope to keep my design modular enough that I'll be able to plug in an autopilot from diydrones at some point (when they go back to, or below their low of $179!) but that is still a long way off.  In the interim may Moore's law march on.  

One if by land, two if by sea.

I know everyone on here is totally stoked about air vehicles, and I admit they are way cooler. I'm not convinced that the air machines will really change society because they don't have the possibility of inexpensive persistence. Aircraft are such high energy machines; high energy, high wear parts. One major airline trip and you're made an enormous carbon footprint. A cruise isn't far off, but at least boats can float. We didn't send hexapods to Mars. Granted we did FLY our robots there on a really big rocket, once they got there, they were terrestrial. That's why solar power is so crucial for robotics. Solar power is pretty cheap as far as robot parts go, and very do-able, especially for intermittent land use. With the announcement of the 2014 Volvo with autopilot, I think it is a matter of time until autopilots will begin to be integrated into vehicles like cruise control. Then it doesn't really matter if we're even in the car. I'll just send an old golf cart to the post office for me. I can put a camera and a note on it (surely they'll be used to this by then?). Drive though food by proxy, please insert food into empty car, I'm in the garage. I know the "burrito bomber" is a fantastic idea, but I think pizza hut is going to prefer tele-presence in an empty vehicle, that doesn't want a tip to a teenage driver. Even though Chris Anderson is one of my heroes, I think "my platform," the play-ground vehicle, is going to change the world. (Ok, just kidding, I'm not changing the world). Who cares about landmines if you can send a fleet of kid's toys and powerwheels to clear the roads; they're disposable. UPS is going to want land vehicles for logistics even if they can't travel as the crow flies. My robot can carry a 50lb kid (it can probably carry me) who can then pan and tilt a hero camera. That's a pretty useful payload. I'd like it if it could just follow me to school and carry my books and crap for me.

This project has been on the back burner for most of this semester. I was weak and spent the better part of a Tuesday where I was not teaching or taking class working on the "robot". Where I had spent time agonizing over neatly soldered and heat-shrink wrapped connections I tore through the project like a jones-ing drug addict. I tore off insulation with my teeth twisted wires and slapped them into electrical tape. Before I had to pick up the kids I had debugged the code down to its most basic elements to the point where my daughter could ride it when she got home. Forwards and backwards and she did the steering. That's it,no sensors, but it finally worked! I wired in a toy-like epoxy-sealed charge controller as well. In the frenzy I lost the old foot pedal I was going to use for a brake. I scoured the house for so many different parts that I can now picture the pedal in every box of junk parts and bits I own.

My daughter had a blast, and the thing screams. It is simply too fast for my 2 year old. I hope to add a turtle/rabbit toggle switch at some point. I have a video but after getting trolled on YouTube, I don't really want to put any thing video on the internet for a while. They were like, "great video, but your voice makes me want to kill myself," to which I felt an odd and kindred empathy and wondered how I survived puberty. So for the public health; it's all stills. I don't mind posting pictures of my ghetto-spit-and-tape robot though. I got a little burned out on the fabrication side of things. I don't have the disposable income for a drill press, welder or a long list of other toys.

I finally broke down for some more parts though. The grand plan is to get the "tractor" running on par with any other normal remote control toy and upgrade to the diydrone.com autopilot. I'm still working on the steering portion. I looked on eBay for linear actuators with potentiometers but they're out of my price range or look like junk. There also isn't any place to put it that won't create a pinch hazard. I went to servocity.com (holy crap actuators were $400 bucks) and bought one of their larger servo gearboxes because folks on modifiedpowerwheels.com have had reasonable success with them. My robot is very small (1 small person vs. 2+ slightly older and larger) compared to the most easily modified vehicles (apparently there is a metal jeep that is amazing). The tractor doesn't have a trunk or much of an engine compartment. I think I may even be able to mount this servo gear box such that I can then put some plexi-glass over it out of harm's way.

I also bought a new axle which I plan on threading to make a complete disassembly easier (and repeatable without destroying the disposable hub clips). The old one is too short after adding the 2nd motor, 2 wheel drive. The next step is going to require a nearly complete tear down. At one point I was excited about the prospect of spraying it down with truck bed liner to give it a newer/rugged look. Then I went to duct-tape for structural support and aesthetics went out the window for a while.

Work exchange collective

A friend of mine and I are supposed to hang out and work together, but neither of us have been able to find the time. He has a raspberry pi while I'm working with an arduino. The prices have come down on the pi's so I've ordered one. This is in part because I only got to hold his while it was off. I don't really understand how to make the transition from the micro-controller to the RC radio. It might be easier to just let the raspberry onto my router and control it over http. The pi has USB, so I could just plug the arduino into it and not bother with the Ethernet shield... or use both. In any case, if I can get the pi to do some computer vision I'll be really excited. I want this thing to mow the yard, rake the leaves and shovel the snow, like yesterday. And tell me better jokes than siri. The dishwasher is a malapropos, it doesn't do sh!t.

Heliostat

The sun came out today so I checked the battery before and after dropping off the kids and was happy to see the charger was working. I suspect that the 10w panel might be self-sufficient in the summer. When the robot is on it should be using the arduino, the raspberry and the motor controller. The arduino may use about ~.5watts. The highest estimate I found was 35mA, so .035A x 12v = .42 watts. The raspberry should use about 3.5 watts. I don't know what the speed controller uses. The motors however use rather a lot. WV gets about 3.5 average sun hours a day provided you're on the south side of a mountain. I'm close to the top of a northern hill. Conservatively that would be about 35w/hrs. a day. The two components I can calculate are already too hungry to make it 24 hours. The summer is different though and I'm optimistic. I am hoping that I can use the panel to position the vehicle to track the sun. I think this will need either another panel or a compass. I don't want to buy a compass, gps etc., because the autopilot sold here has so many sensors for less than I could buy them individually. I have no idea how to implement the heliostat, but a 14 year old made an instructable about it and that gives me some hope. He used two photo resistors separated by a fin of cardboard. I have an old solar laptop charger that just uses monocrystaline PV's and gives a digital readout in percentages so there has to be a way. Practically speaking I'm not sure how to take the charge controller's output to the arduino.

First Robotics

I went to a first robotics group meeting and was really excited to see robots throw Frisbees. I was also a little bumbed out because it was pretty obvious that they were working on a budget vastly different from that of an adjunct professor and graduate student. I grilled every high school student I could find. I learned that they are working at a fairly abstracted level with expensive components and software. I also learned that I could attach gears to potentiometers. I had never thought of that.

The presidency.

All the while, I'm cramming all sorts of crap about the presidency into my head. I do not care about the presidency as much as I do about my robot. I love comparative policy, policy analysis, institutionalism and a lot about political science, but I don't care a whole lot about this particular n = 44 domain. In an effort to push myself to program though, I'm trying to learn some python. Maybe I can learn something interesting about executive orders and text mining.