Two of the biggest automotive stories of 2012 were driverless autonomous vehicles and fuel economy estimates that greatly exceeded what drivers were observing in
the real world. Turns out, there may be more of a connection between these two stories than people realize.
Over the past twenty years, as electronic control systems, sensors and actuators have become more reliable, affordable and capable, the experience of
driving has migrated from direct control to more of a simulation. While cars that we can buy aren't yet fully driving themselves, some of the technology
that will enable that capability is helping engineers hit new fuel economy targets.
Taking the driver out of the loop
Efficiency is largely a function of vehicle design, the driving environment and driver technique. Since the 1970s, regulations and market forces have put automakers under pressure to improve their vehicles in this area. At first, the only factor the engineers could control was design, so they focused on reducing
friction, improving air flow through and around the car, and maximizing combustion efficiency.
Once a vehicle is sold, engineers have no control over where it's driven and under what conditions. And until recently, they also had no control at all
over how the driver operated the vehicle.
Starting in the 1970s, electronics, such as the Hall effect sensors that replaced distributor cams and points for ignition, began to supplant mechanical systems. Those eventually gave way to even more powerful electronic controllers with individual coils on each spark plug. Similarly, carburetors were
replaced with port fuel injection, and more recently, direct injection.
The amount of power an engine produces is a direct function of the air and fuel that goes into the combustion chambers. For a long time, when the driver
pressed the accelerator pedal, the linkage would directly pull a cable connected to the throttle plate to manage the airflow into the engine.
That began to change in the late 1980s when BMW introduced the first throttle-by-wire system on the M70 V12 engine used in the 750i and 850i. For the first
time, there was no mechanical connection between the driver's right foot and the throttle. Instead, the pedal was connected to a position sensor that fed
into the engine's electronic control unit (ECU). The ECU read a number of parameters, including the pedal position and then sent a command signal to a pair
of servo motors that actually moved the throttle plates for each bank of the engine.
Over the past decade, virtually every vehicle in developed automotive markets has migrated to a similar system. As the number and resolution of sensor
signals coming into the ECU has gone up, these electronic throttle systems have increasingly taken control of the driving experience. They're used by
features like stability control, adaptive cruise control, and collision mitigation systems. Electronic throttles, variable valves and fuel injection also
lie at the heart of meeting today's emissions and fuel efficiency regulations.
Unlike their European counterparts, American drivers have long shown a preference for automatic transmissions. For a time this was a hindrance to the
engineers tasked with improving efficiency because of the transfer losses inherent in torque convertors and the additional mass of an automatic compared to
a stick shift.
Electronics again came to the rescue. Clutches were added to the torque convertors to lock them up whenever possible, thus reducing the transfer losses.
The computer brains also took on the task of deciding which gear ratio to use at any given time instead of relying on a complex hydraulic valve body.
Computerized shift management also helped to facilitate the development of transmissions with more ratios enabling the engine to run closer to its
efficiency sweet spot at all times. Without electronic controls, the eight-speed transmissions we have today and the ten-speeds that are coming soon simply
wouldn't be practical.
Just as the standardized tests your kids take in school have some flaws regarding their ability to measure student aptitude, the same goes for the tests the government uses to measure vehicle efficiency. It's not that theresults are necessarily wrong, but rather that they're limited in scope. To paraphrase a
former defense secretary, standard tests don't tell you anything about anything they don't test.
Just as there is a continuum of human intelligence, there's also a continuum of driving environments, and it's impractical to test every possible
scenario. When EPA started certifying fuel economy estimates for new cars in the 1970s, only two drive cycles were used, one that "replicated" a city
driving environment and a second that was supposed to represent highway driving in the era of the 55 mph national speed limit.
In the past decade, the EPA added three more cycles in order represent more aggressive driving as well as the use of once-rare power-sapping accessories like air conditioning. The new cycles help provide a more realistic label estimate for most cars, but there are still limits to the scope of the testing.
If everyone drove in moderate temperatures on flat ground at relatively low speeds with minimal acceleration all the time, all drivers would meet or
exceed the fuel economy numbers on the window sticker. Unfortunately, the majority of us don't drive that way most of the time.
How are engineers gaming the system?
Because schools in America are expected to reach certain goals on standardized tests, teachers generally spend a substantial portion of the school year
training students how to perform well on those specific tests.
Similarly, engineers know exactly how their vehicles will be evaluated. They know exactly how fast the car will go, and how long and how
quickly it will accelerate or decelerate. When engineers program the control logic, they can monitor parameters that correspond to the test cycles, such as speed, acceleration and pedal
position, tand select the gear ratios, throttle positions and air-fuel ratios that will deliver the minimum possible fuel
Electronics let engineers take control away from the direct influence of the driver. Drivers can request acceleration,
braking or directional changes through the gas, brake and steering wheel and the vehicle will mostly comply with those requests. What the driver may not realize is that his inputs
are filtered to help avoid a loss of control and minimize fuel consumption. Nevertheless, there will always be times when drivers don't want to drive in the most
AnyIf you've driven a modern car with an automatic transmission, you may notice that the markings near the shifter have changed in recent years. Traditionally, the shift gate was labeled with "P R N D L" for park, reverse, neutral, drive and low. Drive was the usual all-encompassing automatic mode for forward drive.
For those instances when a driver was going down a long grade or had some other need for more engine braking, the low position would prevent the
transmission from shifting into the top gear or two.
Most of today's automatics have swapped the low position for sport instead, generally denoted by an S. The reason for this? Standardized test cycles. The maximum
acceleration on the city and highway drive cycles is a mere 0.15g or 4.8 ft/s2 and the average speeds are 19.5 and 48.2 mph respectively.
In many instances where mild acceleration is called for, such as those EPA drive cycles, modern automatics will actually launch from second gear instead of
first and then short shift, sometimes as low as 2,500 rpm. This is particularly helpful in getting inflated mpg numbers, even with the higher-powered
engines often used today.
When drivers buy mainstream sedans with 200 horsepower or more, this can lead to unexpectedly lazy acceleration when passing or merging onto a freeway.
Enter the sport mode. In sport, the transmission always starts in first and usually holds a gear until closer to redline before shifting. Hitting the
brakes will often trigger a quick downshift that may never happen in drive. Thanks to increased integration of the control systems, selecting a different
transmission mode will often affect the responsiveness of the throttle, steering, and sometimes even the damping as well.
It's also possible to use filters that look for acceleration close to the test cycle and automatically minimize the movement of the throttle even if the
driver's foot is fluctuating on the pedal. If the driver presses harder on the pedal, indicating that they're probably not doing the test cycle, then the
controls can switch to a different mode to provide the performance the driver appears to be asking for.
Enter the hybrid
Now that hybrids have joined the mix, engineers have even more flexibility in how they can maximize the results they get on those standard tests. Besides
controlling the engine and transmission, they now have a motor and battery to blend into the equation.
Honda and Hyundai were getting most of the unwanted attention earlier in 2012 with a raft of complaints and lawsuits about cars falling well short of EPA
label estimates. Toward the end of the year, the spotlight shifted to Ford and its newest hybrids. Ford called a lot of attention to itself in late
summer when it announced that the C-MAX and Fusion hybrids would both be rated at 47 mpg on the city, highway and combined estimates. Unfortunately, many
reviewers and customers found their vehicles struggled to get into the high 30s.
When Hyundai announced in early November that its fuel efficiency estimates were being reduced by three percent, the manufacturer acknowledged procedural
errors in its testing. When reports began surfacing about problems with the new Fords, representatives from Dearborn insisted their numbers were correct and
that they followed all the rules.
It's too early to tell at this point what's really going on, but what we're most likely seeing is semi-autonomous control and hybrid components
that are set up to maximize fuel economy label numbers. That's great when it comes time to market the cars, but it's also leaving some drivers disappointed when the real-world numbers don't live up to the hype.
It's probably nocoincidence that several hybrids introduced in the past two years, including those from Hyundai, Kia, Porsche and Ford, are all
capable of running on electricity alone at speeds up to 62 mph (100 km/h). The maximum speeds on the EPA city and highway drive cycles used for calculating
corporate average fuel economy numbers are 56.7 and 59.9 mph respectively, meaning that these vehicles can potentially operate engine off everywhere in the
Engineering is always a matter of balancing opposing requirements. The oft-uttered phrase among engineers is, "fast, cheap and good; you can pick any
two." If you want something fast and good, it'll cost you. Fast and cheap will be poor quality, and so on. When designing a car, the goals are to get the
best performance/efficiency at an affordable price and deliver it on time.
The added cost and weight of hybrid components has always imposed constraints on what engineers can implement. As battery cost and capacity have
improved and engineers have developed more efficient motors, the maximum electric drive speed has crept up from 24 mph on the first Toyota hybrids, to 47
mph on the 2010 Ford Fusion hybrid, to 62 mph on the latest models.
The lithium-ion batteries that replaced nickel-metal-hydride units on the new Fords can absorb energy from regenerative braking faster as well. Ford
claims the batteries can recover up to 90 percent of the car's braking energy, although they don't go into specifics of when and how this happens. In all
likelihood, this number corresponds to the deceleration required during the EPA driving cycles.
The result is that during the course of the drive cycle, the hybrid powertrain probably has an almost fully charged battery much of the time when
acceleration is needed, an electric motor that is sufficiently powerful and capable enough to handle the work that would normally be done by the engine.
That yields big numbers for the window sticker.
Unfortunately for drivers in the real world, that brings us back to the whole standardized test problem. If we still lived in a world of a 55 mph national speed
limit, drivers would probably meeting or exceed the EPA numbers on a regular basis. Most American highways have speed limits of 65-75 mph,
however, and most drivers regularly go faster than that. The effects of aerodynamic drag increase with the square of the speed, so going a little bit
faster requires a lot more power, putting the 47 mpg rating on the latest Ford hybrids even further out of reach.
For now at least, Toyota drivers see real-world results that are closer to the label than drivers of the latest Fords. The will likely change when the next
update of the Hybrid Synergy Drive system arrives. Odds are, it'll boost the maximum electric speed from the mid-40 mph range to 62+ mph and adopt a
lithium-ion battery. The window sticker values will probably be boosted as engineers optimize the performance for the tests and the real-world gap
observed will probably grow.
It's not all bad
Window stickers and advertisements have always carried the "your mileage may vary" disclaimer. It greatly depends on where you live and how you drive.
Standardized tests mean those numbers are really for comparison only, and any divergence you see will be relatively consistent for almost any car. If you
fall short with a Hyundai or a Ford, you'll probably do likewise with a Toyota or a Chevy. Unless the EPA adjusts its test
procedures again, we'll probably see this happen more often as we continue to electrify vehicles.
Despite that, vehicles are getting more efficient. The cars you buy today use less energy than those from 5 to 10 years ago. Those we buy five
years from now will use less energy than those we buy today. We just need to shop carefully and be aware that we live in the real world, not a dynamometer lab. A
decade from now, when we're trundling around in fully autonomous vehicles, many of the problems we just discussed will simply go away as the driver is taken out of the loop