Research

I was able to procure a replacement voltage regulator board. Instead of going with the default 2.25A onboard regulator, I purchased an 8A continuous, 15A peak regulator from Danielle's R/C, and a battery to go with it (a LiPo battery is recommended, but I'm driving it with a NiMH for the moment.) You can see the wire melee in the pic (sorry, took it with the Pre instead of my camera setup.)

I plugged in the new regulator, and the autopilot board booted right up! Unfortunately, as I was removing the old regulator, I crushed a tiny, microscopic Schottky diode out of existence. It's less than 1mm in size. I have ordered a set of 20 replacements for $1.86 from Mouser, but they will take a few days to arrive. The diode in question controls the GPS's reset line, so GPS will be offline for the moment, unfortunately. I did take the time to iron out my airframe config, and after about 8 hours of work, I finally was able to get it to compile! Even better, when I uploaded the firmware to the AP board, the servos greeted me, all working properly! (once I had reassigned a few lines.) The mode switch also is demonstrating that it works properly - on engagement of the autopilot modes, the system automatically enabled the throttle and set a course correction in motion.

Once the replacement parts come in from Mouser, I think I should be able to head to the field to do some full-manual-control testing. Once I'm satisfied that all control inputs are being executed according to my wishes, I will extensively debug my airframe config, and then give the autopilot the go! More later.

Everything seemed to be ready to go for the first test of the autopilot board. Infact, an initial test showed it worked perfectly! Servos were moving, everything seemed happy. Then I hooked up an additional, fateful wiring harness to the R/C receiver's battery input port.

Because of a wording/phrasing ambiguity in the manual for the PPM encoder board, I did the opposite of what I was supposed to do. (I seriously can only take half the responsibility here - it was terribly worded and quite ambiguous.) Unfortunately, the AP board went kablooey. More specifically, the PTH08080 linear voltage regulator died. You can see it, somewhat fuzzily, in the photo - just click on it. Sorry, I didn't put my macro lens on, the standard walkaround was acceptable enough. Fortunately, the part is easy to replace, and not expensive (about $8) and I've already ordered a replacement from DigiKey. However, this will mean more delays. At least know I know that the interfacing works properly, and we should be in the air shortly!

Hello world! It's been a while. On the research front, I thought I'd make a small posting about what's going on with the airplane. My advisor has purchased a Paparazzi autopilot module (with GPS and horizon sensing) for integration with the Alpha. I have prepared the way by upgrading the Alpha's radio system to a new Spektrum DX7 radio (7-channel, 2.4GHz DSM2) and an AR6200 6-channel receiver. The upgrades are necessary, as the autopilot module needs a dedicated channel to switch the unit into operational or standby modes. I have also changed the prop to a 2-bladed prop, permitting much faster climbing rates and faster speed. The old 3-blade prop was designed for training, and created tremendous air resistance to keep the engine from operating in its power band (and therefore higher RPM range.)

After making a few more camera mounts, a number of flights, and a few different cameras and sensory equipment (the autopilot hasn't arrived yet), I went for a number of test flights, and flew out at Cane Ridge Park, which is open on weekends and from 8-11PM. It is a bit far, but the day/time restrictions on the Warner R/C field are simply too draconian to be useful for weekend data collection. Today's flight, equipped with BOTH camera and GPS, topped out at 2427 feet off the ground! Also, at that altitude, I can see, with the Nikon camera, a remarkable 600 meters across! The autopilot should be coming in soon, so I will have much more to report and show at that time. There are many new videos, I just haven't posted them yet. More later.

All the images shown here are video stills - which explains their relatively poor quality. The first is of the forward-looking camera, mounted to the side. It's an ATC3K, on loan to me from Alex. The second is from the Nikon, belly-mounted, at 2400 feet off the ground. Lastly, there is a shot from the ATC3K, down-throw-mounted, as the aircraft was flying above the clouds.

On September 26th, I put a GPS onboard the Alpha to record altitude, path, and speed information. Here are some selected results (I did a second test run yesterday, but those results aren't included here:)

Average speed: 40-42MPH
Max speed: 60+MPH
Peak altitude (under my control): 1088 ft above the ground (in MSL, this is about 1683 feet)

More later!

At the moment, I am in Vancouver, British Columbia, Canada, staying at the downtown Holiday Inn. I am attending the 2009 Embedded and Ubiquitous Computing conference, and will be presenting a paper at the TrustCom 2009 International Symposium on Trusted Computing and Communications.

It's late, but check back later for more updates, and some PHLOG entries, if have the time to upload them. As a random fact, my laptop is picking up over 43 DIFFERENT wireless networks from my room! Thank goodness for DSSS, otherwise it'd be a nightmare - I'm betting it already is. DSSS or not, that's an awful lot of congestion on the spectrum, and something's gotta give, at least in terms of speed. This city is very lively late at night! More later.

On Monday, I took my Ph.D. Qualifying Examination. It consisted of the presentation of my project proposal, background work and preliminary results. The presentation lasted for about 1 hour and 20 minutes, with questions from the five committee members thrown in. One of my committee members was unable to be at the qualifier in person, so she came in via teleconference. She is at a secure Air Force Research Lab, so she had no access to internet (therefore no web/video conferencing), and thus came in via phone. However, I found out at last minute that the room (the only room I could schedule for such a large block that wasn't an all-out classroom) didn't have a working phone line.

Fortunately, with Skype running on my laptop, I was able to conference her in regardless, and it worked out really well. The presentation went well, with the only concerns being that I my proposal may be a bit too ambitious for a Ph.D. dissertation. However, I am optimistic, and early results are promising. Following the presentation, the committee deliberated for about 5 minutes, after which I was informed that I had passed! With this hurdle behind me, I need only to complete the research, and write and defend the dissertation. A few conference papers and journal articles thrown inbetween won't hurt either, and are part of the schedule. I'm looking forward to continuing work, after I catch a quick break and finish up some revisions to a paper and journal article that I have yet to finish.

About three weeks ago, I submitted a paper that covered a part of my proposed dissertation research to the TrustCom 2009 conference, to be held in Vancouver, Canada in late August. A few minutes ago, I found out that it was accepted, and that I am invited to speak! I'm really looking forward to it!

Many people ask what I do as a graduate student research assistant. To give a short answer, I ensure the security and availability of special-purpose computers (embedded systems) in wireless sensor networks for critical applications, such as power plants, gas and oil pipelines, nuclear facilities, and other process control applications. I design software elements and select and define optimal hardware configurations to which that software is tailored. At least, that's the work I'm involved with at the moment.

The longer answer is that I work with an infrastructure known as SCADA, or Supervisory Control and Data Acquisition. A SCADA network represents a monitoring and control systems network consisting of a hierarchy of sensor/actuator nodes, a remote in-field terminal computer, and then a central management system. This allows the network and its systems to be monitored and operated remotely; for instance, a power plant can be run completely remotely, with all critical information being sent over the internet. Of course, this presents rather obvious security concerns. What happens if a computer designed to control a valve is compromised? Does the plant blow up? The scary answer is, yes, to a degree.

SCADA networks were largely designed in the 80s and 90s and featured embedded computers that were linked by cables. With the advent of low-power wireless networks, the systems were upgraded, but without any security considerations. This means that as soon as the enterprise firewall protecting the entire network is breached, anything can happen. The work I'm currently involved with involves identifying attacks at the level of the sensors in the field to determine if an intrusion has taken place.

The work I do there is derivative from research I have done previously for my master's thesis and other publications. This is all done within the TRUST research project (Team for Research in Ubiquitous Secure Technologies) (website here), which funds my studies and living. I am a graduate student within ISIS, or the Institute for Software Integrated Systems. The TRUST project itself spans eight different schools all across the United States, with research projects in medical systems, wireless sensor networks, malware control, etc.

The SCADA research project was somewhat thrust upon me as the result of a refocusing of TRUST goals. It isn't something that I want to continue to do as my dissertation topic, so today I made the decision to choose a different but related area for my dissertation research. Because of previous work that I've done on intelligent and hybrid-stage intrusion detection mechanisms, I have chosen to combine this with new tools that I have created during my work with the SCADA project. The SCADA research project will continue, but, for the dissertation, I will likely be focusing on security aspects of a different application that I have not chosen at this time.

In any case, if you ever wondered what I do, the above text gives you a quick snapshot - though of course rather incomplete - of the research and topics with which I'm involved. So, with that said, I'm tired, and need to get to bed. I still haven't started sorting pictures yet, but I promise, they'll come soon.