March 14, 2008
|
Moderately Technical Discussion |
|---|
|
Guardian Carpools & Convenience Guardian Car Technology Is Now Order of Magnitude - Great Investment Guardian & Semi-Guardian A fully developed Guardian transportation system will have no traffic accidents. No police, no bicyclists, and no pedestrians die because some fool drifts onto the road shoulder. No back injuries from inattentive rear-enders. No T-bone deaths at intersections. No inexperienced kids dying with one quick bad decision. Initially, used or classic cars, pedestrians, and
bicycles will be at least semi-Guardian.
A car or a person can be semi-Guardian with only a cell-phone (soon smaller) semi-Guardian device. The device would have location and
motion sensors to telegraph intentions, but without the automatic control of
brakes, wheel, and accelerator. A small semi-Guardian device would provide the semi-Guardian car drivers with visual or
aural warnings of a pedestrian, while also warning the pedestrian of the car. Fully-Guardian cars would avoid equipped pedestrians or bicyclists automatically. The Technology of Guardian The best cars are likely to rely on a
combination of global positioning system (GPS) and inertial navigation, perhaps
with one or two sensors on either car or road to detect a "passive"
car or obstruction. The device requires
only inertial sensors, a global positioning system, radio, and a computer all
on one silicone chip. See comments by Steve Underwood, a research scientist at the
Center for Automotive Research at A Guardian vehicle could
drive itself on a maintained road with a GPS base station or earth
"satellite" every few miles. Full-Guardian vehicles would include a drive-by-wire control system. A simple proximity
sensor might be necessary to calibrate the 2-foot follow distance. Guardian
technology is, therefore, much less expensive than individual
intelligence. It is a few thousand dollars more efficient for vehicles to
converse (communicate simultaneously) and tell each other where they are and
where they are going, rather then for each vehicle to independently track all
vehicles, in order to avoid a collision. Guardian vehicles might best be
thought of as very polite vehicles, always telling each other what they are
doing and coordinating so that everyone can proceed smoothly. Wikipedia, http://en.wikipedia.org/wiki/Global_Positioning_System has a very detailed discussion of GPS accuracy. The short
answer is that the relative position of two GPS receivers can be precise within
a few centimeters (inch). The accuracy of GPS continues
to improve and costs reduce. Also more
satellites, such as the European’s Galileo system, are launching. The higher accuracy and precision is obtained
several ways. First, the vehicles don’t
require accuracy, they require precision. Precision vs. Accuracy – Suppose you shot ten holes in a pistol
target. If you could cover ten shots
with a quarter but the quarter was 6” from the bulls-eye, you have great
precision but poor accuracy. Some
accuracy errors are caused by differing atmospheric conditions causing slight
changes in the time it takes for the radio waves to travel. Some errors are cause by clock error. Some errors persist for a few minutes after
the satellite fires a booster rocket to adjust orbit. In all of those situations, two close
receivers will calculate their distance from each other with precision of a few inches using the space satellite signals available in 2000. For survey accuracy
of less than an inch, surveyors install a base station. A base station calculates its position within
fractions of an inch by averaging out the accuracy errors over time. The base station broadcasts a radio signal
which allows every GPS receiver within about 10 miles to quickly calculate its
position to within fractions of an inch. The Trimble MS750 is a base station with a range of about 20 km (13 miles). A single base
station can be used by any number of grade control or GPS site positioning
systems at one time. Find out more about
similar units with a Google of ‘GPS base station’ There are at least two ways
for moving vehicles to update their location in tunnels or concrete canyons. One way is use a local "satellite" stuck on the
tunnel ceiling or on the wall of a building. Or use future versions of this http://www.u-blox.de/products/tim_lr.html. Another is to use inertial navigation to cover the times when satellite
signals are not available. (A more robutst system would use both.) See the April 2007 Popular Science How It Works for a description of how the Wii controller uses small and inexpensive inertial sensors. The vehicle
needs inertial sensors in any case.
Inertial sensors track linear and rotational accelerations. Acceleration for a known time generates
velocity. Velocity for a known time
equates to distance and direction.
Inertial navigation works by calculating velocity, distance, and
direction from a series of accelerations. Inertial sensors also serve
to warn other vehicles. For example, a
sudden flat tire would have an instantly recognizable “acceleration
signature.” A vehicle with a flat tire
would inform all the nearby vehicles of the flat tire quicker than the vehicle
driver could react. Also, the inertial
sensors would inform nearby cars of their performance. Suppose a Hummer is closely following a
Porsche on a dry road when the Porsche hits the brakes hard. The Hummer would inform the Porsche not to
stop at faster than the 0.7g, that the Hummer can stop at. On a snow covered road, the situation may be
reversed with the following Porsche telling the Hummer not to stop faster than
the 0.2g it can manage. (A ‘g’ is one
unit of earth gravity. Accelerating at
1g would be accelerating the same as if you were dropped off a cliff, before
air resistance slowed your acceleration.) It would be extremely
difficult for independently intelligent vehicles, such as the DARPA Challenge
vehicles, to even identify all the “threats” around them. It’s tough for sensors to distinguish a
pedestrian from a tree, let alone notice a pedestrian between parked cars or a
bicycle approaching from the left along a row of parked cars. “Robot” researchers have known for a long
time it is very difficult to make a robot that will find the milk in the
refrigerator, but it is trivial for the robot to find milk tagged with radio
frequency identification. Instead of the
milk hiding behind the eggnog, the milk is saying, “I’m in the refrigerator. Open the door. Move the eggnog, then reach to coordinates X,
Y, Z.” The Federal Highway Administration has identified a radio frequency http://www.its.dot.gov/vii for vehicle-road and vehicle-vehicle communications. Their project uses the relatively
obtuse nomencalture of Vehicle Infrastructure Integration (VII). Notice the
American site lags behind the European site in graphical presentation
http://www.car-to-car.org/index.php?id=130 The Europeans also have a better name, Car2Car Communication Consortium, for their project. Also check on the Center for Automotive Research at Guardian Weakness While the cost efficiency of
Guardian technology and the ability to protect pedestrians is its strength,
its coordinated approach is also its weakness.
As demand increases, individually intelligent vehicles will evolve on
their own. Evolving Guardian vehicles,
however, requires a deliberate and coordinated effort by both government and
private sectors. Imitating the DARPA
Grand Challenge with a Civilian Transportation Challenge, would provide that coordinated
effort.
Guardian Safety Every
fully-Guardian or semi-Guardian car, motorcycle, bicycle, or pedestrian is
protected by an invisible shield controlling the brakes of approaching
Guardian cars (and providing an early warning to semi-Guardian cars and
people). Guardian Carpools & Convenience If
we couple the accident avoidance capabilities of Guardian Cars with fuel
efficiency incentives, we can set (and enforce)
higher (and more uniform) speed limits.
Consider a 4 lane highway numbered from the left. (Cars may converse about their current
passenger miles per gallon (pmpg).) Lane
1 could be a minimum of 40 pmpg, maximum of 80 mph. Lane 2 could be 30 pmpg, maximum of 70
mph. Lane 3 could be 25 pmpg, maximum of
65 mph. Lane 4 could be 20 pmpg, maximum
of 60 mph. Except for acceleration time
during lane changes, inclement weather, or sparse traffic, all the cars in a
given lane would be traveling exactly the maximum speed. Commercial
and industrial traffic will benefit from the congestion-free and highly
predictable transit times. Liability
insurance needs would be reduced along with the reduced accidents. Guardian
Cars would make impromptu carpools easy to arrange while vehicles are
enroute. Drivers and riders would be
pre-qualified as “safe and reliable.” A
driver would phone in (using a PIN) with intended destination. The system already knows her origin and
possible routes. A rider would do the
same, and also tell the system his origin and any special needs. The rider would pay the driver with automatic
funds transfers. This system can provide
better service for elderly, disabled, emergency bicycle commuters, and even business
car pool users. The carpool could be run
by charities, business, or government. Guardian Car Technology Is Now DARPA is running an Urban Challenge for robotic vehicles racing in city streets. The first Urban Challenge race is November, 2007, http://www.darpa.mil/grandchallenge/index.asp
You are already safer with a computer controlling your brakes. 40% of 2007 model year vehicles (and all vehicles by 2010) have electronic stability control, which adjusts brake force at each wheel individually to prevent roll-over. The Federal Highway Administration has concluded that robotic cooperation between the vehicle and the roadway is necessary to achieve major safety improvements. The Federal Communication Commission has already picked a radio frequency, 5.9 GHz. http://www.its.dot.gov/vii. In September 2006, the site and the name "Vehicle Infrastructure Integration" is still too boring. The Europeans have a better name, "Car2Car," and a better web site http://www.network-on-wheels.de/vision.html (which you need to copy and paste).
Cell phones and car navigation systems are already accurate to within a few feet. The Garmin Forerunner, www.garmin.com $377 in March 2006, is a watch that uses GPS navigation to calculate calories burned, distance, pace, elevation, and heart rate. Nieman Marcus offers a dog collar that phones the owner, if the dog wanders out of a preset area. Celestron, www.celestron.com, is selling the SkyScout handheld telescope. The SkyScout uses GPS, accelerometers, magnetic sensors, and memory to guide your gaze to the celestrial object of desire. You can buy a pen and reciever from EPOS that will "track" your writing 240 times per second accuarte to half the width of a human hair.
Transportation
planners and engineers appear to be just starting to grasp the opportunities
for this technology. Although currently,
little attention is paid to having the vehicles communicate and coordinate with
each other. Also, “Guardian” is a new
(2005) term for this higher level communicating intelligence. Most of the web sites listed below are about
a year behind the current state of available technology: http://www.popsci.com/popsci/futurecar http://www.darpa.mil/grandchallenge Or
order books from the American Society of Civil Engineer www.pubs.asce.org Intelligent Vehicle Technology and Trends
by Richard Bishop or Perspectives on Intelligent Transportation Systems
by Joseph M. Sussman. As of February 2006, the above web sites reflect limited visions for future applications of
computer technology or demonstrate the kindergarten class in a candy factory
approach to investigating opportunities. Transportation Challenge (the 2005 concept) How can we most quickly save lives with Guardian car technology? The technology needs a lot of time and miles in real-life situations to debug and become idiot proof. As of 2005, researchers at the University of California PATH believed $1 to $2 billion would be needed to make the technology much more robust and reliable that human drivers (See Economic Analysis below.). And then the technology must be desired by drivers.
As of January 2008, it is evident manufacturer's have been making that investment.
The challenge has become unnecessary.
Order of Magnitude Great Investment Summary - The United States total monetary benefits are on the order of $4
trillion over ten years after an investment of $100 billion. (California, with the 8th largest global economy,could run the challenge without waiting for the Feds. Total benefits and costs would be about 1/10th. Per-person benefits and costs would be about the same.) An
investment of $300 per person, yields a return of about $6,000 per person. While every investment has some uncertainty,
this investment has one heck of a high payoff. One of our first efforts should be a rigorous economic analysis because that will provide clues for how
to reach our goals most cost efficiently.
However, the “napkin” analysis below is useful for recognizing the
challenge is a good investment. Researchers with the UC
Berkeley PATH program and past attempts to produce automated highways point out
that it won’t be easy to refine the existing technology for the robust
reliability of our goal. They estimate a
$1 to $2 billion investment over ten years. If we assume new cars will
all be “ready” for the technology with drive-by-wire features, the main expense
will be for a device, perhaps cell phone size, that plugs into motor vehicles
and is carried by bicyclists and pedestrians.
In 2005, General Motors researcher’s suggested the extra cost for their
V2V program might be $200 per vehicle.
It seems reasonable to project each device would cost less than $200 in
a production run of 300 million. A
device for every American man, woman, and child would cost us collectively
about $60 billion. The challenge may discover
roadside appurtenances are essential for robust precision of the devices or for
areas not reached by GPS satellite signals.
(See the Science Sidebar.) The
typical GPS base station covers 100 square miles with a signal making GPS
accurate within inches. $40 billion
should be plenty to cover the 3 million square miles of the Our total costs, in round
numbers, would be on the order of $100 billion (or about $300 per person). The up-front cost for the initial trials,
testing, and refining is very small. The
larger costs are 5 to 10 years from now, which virtually guarantees they will
be much less because electronic devices continue to become smaller and less
expensive, with better performance. When
considering the time-value of money, the cost of deploying the technology
happens nearly simultaneous with the benefits. One big savings would be for
auto insurance. Typical auto insurance
in A discussion of auto
insurance doesn’t begin to address the emotional benefits of fewer
accidents. What is the life of a loved
one worth to you? A second big savings would be
avoiding the cost of new roads or transit systems. This is particularly hard to peg because many
roads are already congested and it is hard to find estimates for the cost of
nearly eliminating congestion.
Californians spend on the order of 80 hours per year in congested
traffic. In January 2006, Governor Schwarzenegger
suggested a $109 billion bond to reduce congestion (using 20th Century
technology) to 65 hours per year. A third large savings would
be time. The average DARPA spent about $20 million staging its Grand Challenge. Grand Challenge contestants provided about $100 million of effort. (Mark Twain’s Tom Sawyer fence painting effect.) January 2006 news accounts of Governor Schwartzenegger's 20th Century approach to traffic congestion indicated $107 billion for no real safety improvement and 20% traffic congestion reduction (say from 100 hours per year to 80 hours per year). With Guardian technology, Californian’s spend $1 billion of public funds on the Transportation Challenge and perhaps $10 billion in private funds at $200 per car. But the Guardian technology could save thousands of lives, billions of dollars (from accidents), and drops congestion toward 10 hours per year.Per Engineering News Record, November 14, 2005, Page 61 a report by
Cambridge Systematics indicates the Highway Trust Fund annual revenue will be
$23 billion short of what’s needed to maintain current road and transit
conditions and $48 billion short for federal system improvements. View the report at www.uschamber.com/ncf/publications/default.
Mass
produced electronics (like cell phones with Global Positioning Systems) keep
decreasing in size and price. Cars are
already becoming drive-by-wire for active cruise control, anti-flip steering,
anti-lock breaks, etc. General Motors
V2V team estimated the incremental cost for Guardian technology in 2007 at
about $200 per car. This is roughly in
agreement with the projected costs for navigation equipped cell phones. www.Neimanmarcus.com
is selling a dog collar that will phone your cell phone, if the dog wanders
beyond set boundaries for $350. Computer Traffic Capacity
Calculations Guardian
technology would not allow 10,000 vehicles per hour (vph) per traffic lane
immediately. Rather the combination of
the Transportation Challenge and congestion relief incentives would evolve
higher vph as needed. Transportation
agencies should recognize the addition of three more tools for congestion
prevention – 1. a challenge; 2. Guardian technology; and 3. vehicle
demand incentives. Calculations: The
calculations are rounded off to single digits for ease of understanding and to make more clear the results are not sensitive to minor variations in the assumptions. 2005 – Non-Guardian Condition: If drivers' followed the California Drivers' Handbook, cars would be at least 3 seconds apart. Or longer, if the weather wasn't perfect, visibility unimpaired, or a host of conditions affecting human reaction time and vehicle stopping distance. Therefore, if everyone were a safe driver, cars would be passing no more than one every three seconds which gives a maximum of 1,200 vph per lane, regardless of how fast, or how slow, the vehicles are moving. Most everyone follows too close, nearly doubling the actual California freeway lane capacity to about 2,000 vph. (2,000 vph per lane is a real-world lane capacity verified with actual vehicle counts. Craming more cars per lane usually triggers a sudden speed drop from 60 - 70 mph to 0 - 30 mph.)
You may have the perception of somewhat higher capacity. At 70 mph, you cover 100 feet in a second. It seems like you have been in groups of cars with only 20 - 30 feet (0.2 - 0.3 seconds) between cars. Four perception distortions are at work. 1. You notice and remember the higher danger situations, the closest cars. 2. The foreshortened viewing angles make the cars appear closer than they are. 3. Those high density groups only persit over a fraction of a mile (less than a minute) because someone will tap their brakes. 4. You are moving, and cannot assess the average lane capacity in one spot over an hour. Now recall, how often does the situation you perceive as more than 2,000 vph (average follow distance less than 200 feet at 70 mph) suddenly transition to 0 - 30 mph? Not every time, but a lot more often than when follow distances are more than 200 feet.
We need the Transportation Challenge to establish the safe following distances for Guardian technology. That is a few years of rigorous racing to learn if car groups should be two or ten, if distances within groups should be 2 feet or twenty feet, if distances between groups should be 1 second or 4 seconds. I estimate the technology will be safe for the 2015 and 2020 situations below. 2015
– Guardian Situation, only 2005 size cars: Suppose
an average speed over 3 lanes of 70 mph (100 feet/second, fps). Perhaps the average 2015 vehicle is nearly 20 feet
long. The vehicles travel in groups of
five with 10 feet bumper-to-bumper. A
group of five vehicles is about 140 feet (or about 1.4 seconds) long. The
distance between groups is Drivers' Handbook recommended 3 seconds.
Therefore, five cars
pass every 4.4 seconds per lane or 4,000 vph per lane. (The technology will need a reaction time of less than 0.05 seconds to maintain plus or minus five feet of separation during hard braking or accelerating. Generally, simple computations and radio communications react in a few 0.000001 of a second.) 2020
– Guardian Situation, 2015 sized cars:
Suppose an average speed over 3 lanes of 70 mph (100 feet/second,
fps). In 2015, the average vehicle is 10
feet long, because many “commuter specials” are 10 feet long and 3 feet wide
traveling side-by-side in the same lane space.
The vehicles travel in groups of five with 2.5 feet
bumper-to-bumper. A group of five is 60
feet long (or about half a second). The Challenge will have improved the safe following distance between groups to about 1.5 seconds. Therefore, five cars per lane every 2 seconds sums to 9,000 vph per lane. |
|
|
|
|
No copyrights reserved. Spread the concepts with wild abandon!