Friendly Neighborhood Surveillance

I've been thinking about this for some time and decided that I really should write about it.  Its a little big brother-ish, but potentially in a distributed way.

Here's the genesis of my idea:  I live on a reasonably small road that twists and turns, but my house is on a reasonably straight portion of that road.  As a result, cars often go whizzing by at speeds far in excess of the posted 35 mph limit.  I'd like to dissuade people from doing so.

Proposed system: A simple box likely consisting of an off-the-shelf point and shoot camera, a low power laser, a tuned (i.e., focused on detecting the reflection of the laser) photo sensor and a raspberry pi or an arduino.  The system works best if it is connected to the internet full time, perhaps through a focused wifi antenna back to the house.

The laser pulses multiple times a second and determines of there is a car in the camera field of view.  If there is, the speed/direction of the car is measured from the difference in the location of the car over time.  Probably just use an average of a couple of readings when the car is in the optimal location.

When the system determines that the car is in the optimal location for a license plate reading, the camera snaps a picture, focused on the license plate area.  The picture and the speed/direction information along with the time-stamped photograph would be sent to the cloud for processing.  The key would be to pull the state and license plate number from the picture.  Better would also be to pull the make and model (more data has to be worth more money).

At first the raw pictures could be posted publicly on a website where citizens could just focus on shaming speeders or allowing parents to check in on their kids speeding habits (really anybody could check on anybody if they knew the license plate information).

My understanding is that the courts have made it clear that this is not private information - police cars routinely drive down the road with picture scanners looking for cars with outstanding parking tickets on them.  The same could be accomplished by just sitting in your front yard with a camera - the device is just making it much easier to do systematically.

One Step Beyond:   Charge for access to the processed data.  This could serve private investigators (PIs).  They could subscribe to the site for $10/month and get access to locations of a certain number of license plates.  For a premium service, they'd be able to set up a geo-fence around a location and get a text on any hits in real time (location and direction).  Eventually, home-owners could get a small payment for hosting the devices based on the revenues from the subscriptions.  Yes, I'm guessing the NSA would harvest this information with or without permission (but I'm sure that the company would rather get paid for it).

The value of subscription clearly goes up as the number of people providing data goes up.  Business owners on busy streets could post these on their signs.  Homeowners everywhere that are annoyed with speedy neighbors could provide their data to the police as justification for increased speeding patrols.  As a bonus add-on, you could create an iPhone app and let people take geo-coded pictures in parking lots to record the location of vehicles.  I don't know if it would be worth too much (and mall / shopping center owners would like become sensitive to it), but you could pay per data point.

I have a great spot near the road that has power.  If I get motivated, I may just try to hack this system together.

The State of Residential Wind Turbines

So my post yesterday about solar shingles got me thinking about roof-mounted wind turbines and what it might look like for me to install one on my house.  So, I decided to do a little bit of research on the topic today.  If you are interested in the topic, you should probably check out Bergey's Wind School first before you read more below.  If you really want to get into it, don't read anything below.  Rather go to Home Power Magazine and start your research there.  There is plenty of information on the internet to decide if you should learn more - which is to say you won't be able to conclude that you should or shouldn't buy a turbine without ultimately talking to the manufacturer, or better, talking to the installation company.  The balance of system and electrical integration costs are likely to be non-trival and need to be considered - its not just the per kW cost of the turbine!

Horizontal Turbines

These are the ones you think of when somebody says windmill or wind turbine.  The blades spin around a horizontally mounted axis, like an aircraft propeller.  Standard fare that has been around for some time and is reasonably mature (I think) in its design and engineering.  Typically, these guys use permanent magnet motors and have no gearing.  Simple is better with respect to reliability.  Some are shrouded like the TAM Energy / Honeywell turbine.

Vertical Turbines

These are not really that new, but not often used in utility scale machines.  However, from what I can tell, there does seem to be a fair bit of interest in using them in roof-mounted applications.  Unfortunately (though unsurprisingly), these appear to suffer from the same issues as the horizontal turbines with respect to average wind speed.  See below in Take-Aways.

Mounting

Roof Mounted

Two different types of roofs are dealt with - flat (common for commercial buildings) and peaked (common for residential).    Both mounting systems look quite reasonable and well-engineered, but what I can't find any information about acoustic coupling with the building.  Bergey cautions against roof mounting for this reason, but I don't know really how good or bad it can be.  TAM Energy suggests (and includes in the sale of their turbines) a "roof box" that is added on top of the original roof.  Perhaps this provides some amount of decoupling between the turbine and the roof?

Pole Mounted

I think that this is most typical of all wind turbine applications.  There are two options here: guy-wired setup and a self-standing pole.  The latter is more expensive, but doesn't require as much flexibility with your space.

Take-Aways

There is probably nothing here for me.  Bergey recommends not pursuing if you have less than 1 acre and 6+ mph average wind speeds.  I have neither.  Just shy of 1 acre in the suburbia surrounding DC.  The 30 meter (above the surface) wind speed map provided by the government suggests that I have between 4 and 4.5 mph average wind speed.  Though I do pay more than 10¢ per kWh for my electricity, its not much more.  My payback period is probably quite high and would be a riskier proposition than just continuing to purchase from the utility.  A vertical turbine manufacturer, Urwind, suggests that at 4.5 mph average wind speed, their product would produce very little output. 

Low speed performance turbine will have to improve substantially, I'm guessing, before it will make sense for me to go down this route.  This is not impossible.  One of the big achievements of utility-scale wind turbines over the last decade has been to raise capacity factors from the high 20% range to the low 40's.  And that is after choosing the best wind sites, previously.  The same could potentially change here, though the pace of change is likely to be slower given the relatively small market size for residential wind turbines.

Though the above sites suggest that measuring wind speeds at your residence for a year is likely to be overkill (expensive and can be modeled well), I think it would be interesting to measure over the course of the year.  Then, I'd cross the wind-speed duration curve with the power output curve of various wind turbines to see what opportunity to offset electricity usage that I'd have.

I think that I'd be most inclined to go with the TAM Energy / Honeywell roof-mounted turbine if I was going to do anything.  They start producing power at relatively low wind speeds - 2 mph.  I have space on my roof, so I'd spring for 2 - which would cost me $30,000.  They claim that I'd be offsetting about 2000 kWh per year for each turbine.  For me that equates to roughly (10.1¢/kWh * 4,000 kWh) = $404/year savings.  That means I'd need 75 years to pay off the investment.  The design lifetime of the device is only 20 years.  This won't really work - even if the output was double the projected output for an average site.  So much for my investment in wind power, at least for now.

To make it work, I would need to see a 7 year payback.  If the performance of the turbine can't be improved, there is really just one way to make this work: reduce the cost of the two turbines by ~90% to $2,828.  Not likely to happen any time soon.  Probably sub-scale for the balance of plant portion, too.  Which means, I think that the turbines need to improve before this all goes around.

Further Research

I have by no means exhausted the research areas for this topic.  If the turbine technology improves meaningfully, I think each of the following areas would require more thought and study:

• What are local zoning laws (if any) that need to be complied with?  Is it any different than installing an antenna on your roof?
• What is reasonable to assume the acoustic signature is for a roof-mounted turbine?
• How much should it cost to install?
• What are the additional components required to integrate into one's home electrical system?
• How would Dominion (my local utility) deal with my power costs?  How much would be truly offset?

Next up: Residential solar.  Like I have noted before, I'd really like to have a solar-driven stirling engine.  Or rather, a solar/propane CHP system.

Shingle Me Timbers

So the future is not quite now, but I think that this is a big step.  Dow Chemical is jumping into the renewable energy fray with a solar shingle, named PowerHouse.  Its a great idea.  Not cheap, but a great start.  Their marketing wisely doesn't focus on people who would install from scratch, but on those that would be replacing their shingles anyway.

Unfortunately, I live in Virginia and would not qualify for incentives, so I chose the next best place, Washington DC.  The website estimates that I could save a substantial amount of money in electricity costs (over 25 years) and raise the value of my house (presumably due to the decreased energy requirements).  Though my house is not this large, here is the view that they provide:

Unfortunately, what I took for links in the paragraph are not links.  Of course, I could talk to an expert, but I have the following complaints about the provided information:

• How is the asphalt roof of -$20k value?  Misleading comparison - costs only on the left, benefits only shown on the right.
• I'd like to dig into their assumptions and know what the pre-incentive cost is rather than just the post-incentive.
• How do they conclude my house value goes up by this amount?
• What is the $120k of energy savings based on?  What is their forecast of residential power prices?  Do they factor in DC RECs?

Perhaps the information is not shown because they are making some heroic assumptions that embarrass them.  That said, by not putting supporting information, that is pretty much the only conclusion a skeptical guy like myself can come to.

Before you dismiss this post as just a solar hater, please factor in the following about my views:

• I'm a fan of solar and innovative ways to create electricity.  If I could buy a solar stirling engine (claimed to be 1/2 the cost of PV) and put it into my front yard, I would.
• I have also investigated installing a series of nine roof-mounted wind turbines on my roof.  A relatively modest investment of $15k could largely offset my grid usage.  I'm very concerned about noise under high wind conditions.
• If organic flow batteries ever take off, the first two options will look much more interesting to me.
• However, I am a true believer in price.  We are wasting our money on subsidies.  See Marginal Revolution for a recent summary of different economists take on climate change.  My views are squarely with Bjorn Lomborg.  Stop the subsidies, ramp up the R&D.  We will soon have products that are cheaper than fossil fuels and will adopt them without market distortions.

Loss of Power

I am not an expert in the analysis of potential blackout conditions as discussed in this WSJ article.  However, I find the conclusion that by taking out only four transformers (probably actually four substations) in the eastern interconnect one could cause an 18 month blackout to be extraordinarily unrealistic.

Yes, we have some issues.  Some of them:

• It takes a long time to build high voltage transformers
• We don't have a lot of transformer building capacity in the U.S.
• We have relatively lax physical security for our substations
• We have an interdependent physical transmission network.  Losing key interconnection points will cause pain.

With respect to the last point, it should be noted that this is a good thing.  The interdependency exists because its more efficient to trade power across systems.  This is due to non-coincident electrical loads and generation outages.  This saves US electrical consumers (literally) billions of dollars a year.

What I find to be unlikely is that we can't respond to a major event like this.  Our response would be a pain, but I think that we could avoid life-altering outcomes.  We have several different ways in which we can respond (in order of escalating severity):

• Run more expensive, but more proximate electrical generation facilities.  We already have market mechanisms to deal with this and a relatively robust reserve margin to cover unexpected system conditions.
• Ask for voluntary reduction in power usage by both residential customers as well by commercial and industrial.  In the summer, this can take the form of raising the temperature in our houses another 5º.  Or not turning on the big screen TV until 7 or 8pm (after the evening peak load usage).
• Pay industrial and commercial customers to reduce their load (off of historical baselines).  There are structures (e.g., demand response) that have been effective at modifying behavior.
• Dramatically increase power rates above a certain monthly kWh usage.  This is already done to a lesser extent in the retail rate structures in many states.
• Mandate a reduction in usage.  For example, outlaw air conditioning usage, perhaps on alternate days of the week.  Impose large fines on people using air conditioning when they should not be.
• Characterize critical and non-critical buildings and usage for power.  Cut power to non-critical ones.

I could be wrong.  It wouldn't be pretty, but this would hardly bring us to our knees.

GPS Altitude Issues

I was planning on writing a rant about GPS and altitude problems.  I still will, but it will be tempered, from the start, by some quick internet searching on how to fix the problem.

My issue is that when I wear my Garmin ForeRunner 405 running or biking on an "out and back," it will often show very different elevations on the route.  Super frustrating.  I understand some of the technical side of it.  You can read here if you are interest as to why GPS is more accurate on the lat/long than the elevation, but its not necessarily an easy discussion.  GPS is reasonably complicated mathematically, but the theory is straightforward.

It turns out that there are technical solutions to this very technical problem.  Sites like Strava, for instance, do this correction in the background unless you ask them not to.  There is a fantastic post in their help area describing in detail what they do.  If you don't use Strava, all is not lost.

There are programs out there (such as GPS Visualizer) that will use the 2D representation and then map it to the "DEM (digital elevation model) database" to get the third dimension (or really, the elevation is what we care about).  Unfortunately, as described in the Strava post, the best DEMs only have 10M accuracy (which seems mind-bogglingly bad to me).

While there are solutions, it still seems as if Strava and others are not taking advantage of all of the data that could be brought to bear on the problem of determining surface elevation and thus I still feel justified in complaining that Garmin doesn't correct altitude errors when I import my data.  It would seem to me that there are portions of the surface of this earth that are well travelled by GPS-wearing folks like myself, that we ought to have accurate enough readings (from thousands of receivers) to detect a repaving of a trail or road.  Think the Washington and Old Dominion trail in northern Virginia.

Less on the fitness side and more on the house side, I'd like a 3" accuracy model of my lot.  I think that would be fantastic for modeling drainage and hardscaping/landscaping projects in the future.  Perhaps that model doesn't need to integrate with the rest of the earth, it would be kinda cool if it did.

3D Painting (with ultrasound)

This is perhaps not what you are expecting, but I'll start with a vignette about what made me think of this to begin with:

I routinely work out at the gym.  Usually, this involves a combination of lifting weights, cardio and stretching.  In my stretching, I'm usually doing some foam rolling to do that whole "myofascial release" thing.  Unfortunately, my aging body doesn't always cooperate with whatever mix of activities I've recently subjected it to.  I have one on-going issue and one new one.

The ongoing issue is in my shoulders.  They apparently loose.  Not morally, but in the sense that they are not quite firmly in the socket so there is more motion than would otherwise be ideal.  I was told by a doctor that its likely genetic and the only real solution for it is surgery (with a reasonably long recovery cycle).  And as a result of being loose, they hurt if I do too much chest or shoulder lifting.  

The new problem is my left leg.  Its not quite my knee, but behind my knee.  As in, I can barely do child's pose without it hurting badly.

As I laid there on the mat this morning, I thought to myself, "Why can't I just map my ligaments with a 3D probe and figure out if there is some tendonitis and then do the right therapy from there?"  I can't, because as far as I know, nobody has tried to put something together to do this.

So here is a proposal:

• A USB-powered wand-like device similar to that of the BodyMetrix Personal Ultrasound Body Composition tool.
• iPad or Android tablet as key interface
• Priced at USD 200
• Provides a 2D image to the user in real time
• More importantly, builds a 3D model of the body tissue while the user moves it around slowly (auditory feedback indicating ideal speed)
• Potentially use two wands, though this does get a bit tricky for the operator
• Exportable 3D model to other programs.
• Tracking/automatic measurement on various body parts:
• Body fat (like the BodyMetrix).  This is typically take based on body measurements at various points in the body (lower back, thigh)
• Muscle size
• Tendon / ligament size
• Belly
• UI will be key: this tool would need to have a fantastic approach to capturing the information and then making it meaningful to the user.

Ideally, I could have done this first when I was the model of good health. Then, when I'm in pain in some part of my body, scan it and compare the current results / sizing of everything and determine if everything is all right or I'm just being a wimp.  While the latter is likely, I'd rather not be pushing myself when I shouldn't be.  As I get older, pain seems to be a less reliable indicator of when I should stop.

Is this really that hard?  Ultrasounds are used beyond the womb.  See here and here and here and here, some of it getting into cool stuff like finding blockages in arteries or mapping blood flow using the Dopler shift from the moving fluid.  It all seems focused on 2D mapping and snapshots.  Why not go for the 3D that people can understand.  Frankly, when I went to my wife's ultrasound appointments, I only saw random black and white noisiness.  Nothing that looked like a baby.  The technician saw much more, but still had to manually measure the femur and a few other spots to accurately determine the size of the baby.  All of that seems ripe for automation and no need for a technician.  Put the woman in the room with a tub of gel and a wand and tell here to move it around until the system tells her she's done.

Fun side project.  Swallow a penny - and watch it go down through the system :)  OK, maybe this is a little gross.  But instead of a penny, use something else that isn't so gross.  And let people track it through their digestive cycle.  I don't know why that seems cool to me, but it does.  You know you've heard that story of a kid swallowing a GI Joe hand . . .

BTW: I fully understand that we have a set of people working at the FDA that can barely put their pants on in the morning.  This project would be "for entertainment only" and would be designed for pets.  Or something like that.  It seems like it would be the death of this idea if it had to be approved.

Am I missing something?  Does this already exist?  Let me know in the comments!

When Will Wearables Explode?

Not thinking in the literal sense, but in something more than the industry hype sense.  I hope that the hype has peaked and we are moving into the period of hard work where people make products attractive and usable for the masses, not just the early adopters.

Here are some ill-informed views on when I think that wearables will really take off:

• When "energy harvesting" is built-in from the start.  While I am generally not too annoyed that I have to charge my fitbit once per week, I'd rather I didn't have to.  It seems somewhat ironic, as well, that a device that is capturing my motion data, is not using that motion to its benefit.  How cool would it be if I never had to charge my fitbit (as long as I used it)?
• Data transmission is standardized.  Good example: I have a fitbit and a Garmin ForeRunner 450.  Both require proprietary USB adapters to connect to my computer.  Hopefully, devices will coalesce around using low power BlueTooth or some single suitable technology.  Better than this, it'd be great if the devices didn't use near-field technology at all, but instead connected to Wifi and take the computer out of the equation.
• More complete monitoring.  Steps are a good first step.  But not nearly enough for a compelling fitness / wellness view.  
• How about:
• 3D movement, interpreted, so that you don't have to log what activity you were doing.  Premium systems could include multiple sensors for a better read (wrist, hip & ankle).
• Heart rate (and better yet actual ECG signal)
• Blood oxygen levels
• Respiratory rate
• Perspiration levels
• Blood pressure
• GPS location, even if only through your phone
• Some of this should be real-time, but where that gets hard, at least make near real time (like Azoi could be used right after getting off the treadmill).
• Some better understanding of how people want to change themselves (or remain the same).  
• For instance, is it important to you to:
• Train for a particular event
• Just get "better"
• Remain at your current high fitness level
• Provide real feedback tailored to such.  I'm afraid that this will remain in the "paid extra" category, but I don't think that it will take off until a subscription is not required.  I'd be much more willing to buy a premium version of the product up front that has this built into the cost rather than a subscription and think that many others share my aversion.  It boggles my mind that this isn't being done already with fitbit.
• Capture "uncapturable" information and integrate it into your product.  
• AskMeEvery is a great example of how this could be done:
• Wearable user works out
• Wearable system asks user how fatigued they are after the workout (via email, if necessary)
• Wearable system determines that its pretty much the same workout as they had done three days ago
• Wearable system alerts user that they are making progress: their reported fatigue is only 8 out of 10 instead of the 10 out of 10 they reported three days ago
• To be fair, fitbit and others are doing this with food.  But it comes in the "paid extra" category and still isn't easy enough to do reliably.
• Generally decrease the uncapturable information in the system.  A few thoughts:
• Build a kitchen scale with wifi that can be used in conjunction with a phone app.  Have people weigh their food.  (Seems like this could already exist and I'm just unawares).  It has to be much more accurate that taking a picture or just estimating.  And easier.  I had 18oz of lasagna.  Done.
• [Other ideas to be added as they (hopefully) come to me.]