|Electronic Transmission Control
by Frank Owen
(last revised 2 August 2008)
Gear shifting is yet another automotive subsystem that has been affected heavily by automotive mechatronics. Prior to the explosion of micro-processor-controlled systems in cars, there was, of course, the automatic transmission. As we shall see, both the manual transmission and the automatic transmission are still in use, as well as yet another transmission that allows the car to change gears without removing the load.
A manual transmission consists of two shafts: an input shaft from the engine and an output shaft that delivers power to the drive wheels (see Figure 1). These two shafts can be coupled together in a number of different ways by gear pairs. Each gear pair represents a different speed reduction ratio. Usually a car has five forward speeds and one reverse speed. When shifting gears, the load on the gear pair must first be removed, so that the transmission is not transmitting any load while the gear change is taking place. This is done with the manual clutch. (See http://auto.howstuffworks.com/transmission.htm for a more detailed explanation of how a manual transmission works.) From a stop the driver applies gas. When the first gear has “wound out”, he or she pushes in the clutch while removing gas, then shifts into second gear. Thus the power transmission is interrupted while the gear changing is taking place.
Figure 1 – Simplified model of a manual transmission
An automatic transmission, by contrast, is a set of planetary gears or multiple sets really. By keeping some of these gears stationary and allow others to move, various speed reductions corresponding to the various forward gears can be achieved. There is also a torque converter in the transmission that represents a hydraulic coupling between the engine and the drive wheels. A hydraulic linkage decouples the engine and the drive wheels, so that shifting can take place without damaging mechanical parts. (For a more complete description of the function of a hydraulic automatic transmission, see http://auto.howstuffworks.com/automatic-transmission.htm.) To a driver the gear changing is automatic, that is takes place without his or her direct intervention. But the driver can perceive the discreet gears by feeling the transition and watching the drop in engine speed on the tachometer with each specific gear speed.
Both manual and automatic transmissions are now controlled by automatic shifting systems. With a manual transmission, the action of pushing in a clutch and then moving the gearshift lever are performed by electro-mechanical actuators that make these motions for the driver. Thus the distinction between a manual transmission and an automatic transmission has been blurred, as far as the driver is concerned. He or she just steps on the gas. With electronic control, the direct link between the input (driver) and the output (gear) has been severed. The microprocessor mediates the input and decides what the output should be. The gear shift lever becomes merely an input device for the driver’s wishes. The microprocessor senses the car’s environment and makes an intelligent decision on what the appropriate gear should be.
In the simplest case, consider a straightforward acceleration from a stop with an automated transmission. By “automated transmission” I mean a transmission under control by a central processor. The microprocessor senses the speed of the car and the speed of the engine. When, for a certain speed, the engine RPM reaches a certain value, the microprocessor sends out commands to actuators to shift the transmission into the next gear. If the car has an automated manual transmission (AMT), i.e. a transmission with an automated clutch and gear shift, the microprocessor sends out commands to actuators to
actuate the clutch
reduce the throttle to idle
let out the clutch
Of course the commands would be slightly different with an automatic transmission because the transmission itself is different. But the shifting strategy, at what engine RPMs to shift gears, would be the same because a given engine has a given performance curve.
Regarding the AMT, the shifting sequence can be optimized so that the unloading, shifting, reloading is made as smooth as possible. Thus programming and the control by an automated system would allow someone who doesn’t know how to drive a stick shift to drive one like an expert. A slightly more complicated shifting strategy allows the shifting points to change. For example, when you are driving a manual transmission and you want to accelerate quickly, you usually stay in any given gear longer before shifting up to the next highest gear. A microprocessor can be programmed to sense this wish to accelerate quickly and then to delay the normal shifting points until the car reaches respectively higher speeds. A wish for a leisurely acceleration could also be sensed and early up-shifting implemented.
Besides these operation advantages, the mediation of the microprocessor also allows for the correction of driver errors, for example accidentally shifting into reverse or shifting the transmission into an inappropriately low gear when the car has too high a speed. Thus, again, the insertion of a microprocessor into the system allows for convenience, flexibility, and safety not found in a non-automated transmission.
In a fully automated vehicle, the transmission control can be integrated with the electronic engine control and indeed with the overall electronic management of a fully mechatronic automobile. Transmissions must work closely, for example, with the ABS system and the ESP (electronic stability control) systems in modern mechatronic cars.
To bring home the point that the transmission has become a programmable unit, consider Figure 2. This shows a gearshift. What is obviously missing is the transmission. This is possible because the gearshift lever is simply a wish input device in today’s mechatronic auto. Note that activating the gearshift lever simply generates electronic signals that are then interpreted by the car’s central control system.
Figure 2 - Today’s mechatronic gearshift
Note also that this device is manufactured not by BMW, Audi, Mercedes, but rather by a third-party manufacturer. These third party developers and manufactures work very closely with the large auto manufacturers to produce almost everything in the car except the body and the engine.
Additionally this company, Kostal, produces a steering column module, a roof module, a network control unit, a door control unit, a seat control unit, and control panels and switches. Note that more than one manufacturer may use Kostal products in their cars.
My wife recently purchased a Smart Passion. This is a little two-seat, economical car, popular in Europe and now being imported into the U.S. (see Figure 3). The car has an automated manual transmission. It has a manual stickshift but no clutch. It has two shifting modes—manual and automatic. In manual mode, the driver decides when to shift and inputs this wish by deflecting the gearshift lever forward (up-shift) or backward (downshift). In automatic mode, the driver puts the transmission into “D” and simply drives, as if the transmission were an automatic, instead of an automated manual transmission.
Figure 3 – Smart car
At first the shifting was a little hard to get used to. Actually the transmission control did a pretty good job of shifting into the next gear. But, being used to an automatic transmission, the shifting seemed slow. But when the driver thought of what it would be like to operate the shifting manually, the sluggish shifting actually wasn’t sluggish but was comparable with how the shifting would be if done manually.
So at first both my wife and I used the manual shifting, because of the perceived clumsiness and slowness of the automated shifting. After further reading of the owner’s manual, however, my wife saw that the automated shifting was optimized for economy. So she switched to “D” and got used to the automated shifting.
One other problem we had was that in “D”, when the car was shifting, we each naturally let off on the gas pedal. She read further and informed me that we are not supposed to do that. This can be confusing to the micro-controller. It could be interpreted as a desire to stop accelerating. So the proper procedure is simply to express the desire to accelerate as a steady position of the accelerator pedal (i.e. don’t let off on the pedal while the car is shifting). With these adaptations, we are becoming acclimated to an automated car.
Different Types of Transmissions
Manual transmissions are still very common in Europe. Today they have five, now often six, and sometimes even seven forward gears. Manual transmissions have the highest drive efficiency because the engine is mechanically linked to the drive wheels. There is no inefficiency due to the hydraulic link in a hydraulic automatic transmission. Manual transmissions are not as comfortable to the passengers in the car. Ride comfort is a weakness.
The automated manual transmission (AMT), manual transmissions under the control of a microprocessor have been introduced in the last few years into such cars as the Opel Corsa, the Ford Fiesta, and the Smart. They are also used in some sporty models such as the BMW M3 & M5 and the Ferrari. The automatic control of the manual transmission compensates somewhat for ride comfort. The controller also prevents shifting mistakes. In stop and go traffic, with automatic control the transmission will be in the correct gear more often than it would be under manual control. Thus the automatic control also leads to better fuel efficiency and thus reduced emissions output.
Large trucks often have 12 to 16 gears. It is very complicated to shift these trucks manually. So automated shifting has been applied in such vehicles for some time now.
At first the AMT was built as an add-on. A normal manual transmission was simply fitted with actuators to actuate the clutch and the gear shift. Since then manual transmissions have been redesigned with integral actuators. This led to room and weight savings.
Automatic transmissions are oil driven. Early on, however, these transmission were outfitted with governing electronics. For example electrical components were employed to manage the bypass clutch, a device that locks the transmission mechanically once the car has reached highway speeds. Thus for long-distance trips at highway speeds, the internal hydraulic losses of the transmission are avoided.
The double clutch transmission is a relatively new development (see Figure 4). This transmission was developed by Volkswagen in 2003. It has a single input shaft and two output shafts. The output shafts have respectively gears 1, 3, and 5 and gears 2, 4, and 6. This enables the load to be transferred from one output shaft to the other gradually. Thus the gears may be shifted without removing the load. So, for example, in shifting from first into second gear, the load is gradually removed from first gear while it is being gradually applied to the second gear.
Figure 4 – Double clutch transmission system from Volkswagen
Besides just gear shifting, other systems—specifically all-wheel-drive and hybrid drives—have benefited from electronic control. In the latter, the wheels are driven either by electric motors or by the engine. It would be hard to imagine the manual management of a hybrid drive.
Basic Shifting Strategy
Figure 5 shows a typical shifting characteristic curve. Note that this curve is plotted on a scale of accelerator pedal position vs. car speed. As is normal in a microprocessor-mediated system, the gas pedal is simply a wish-input device. So the accelerator pedal simply is telling the central microprocessor in the car how fast the driver wants to accelerate.
Figure 5 – Shifting characteristic curve
Note that the characteristic consists of two curves, one for shifting up and one for shifting down. The “hysteresis” is built into the design of the shifting strategy to prevent constant shifting up and down if a driver happens to be driving near a shifting point. Rückschaltung means downshift and Hochschaltung means up-shift. The best way to understand these two curves is to consider simple examples first. So let’s see what happens from the AP, the Arbeitspunkt or just simply a typical operating point.
As a first case, consider the situation where the driver does not change the accelerator pedal position. So the shifting state stays on a horizontal line. If he/she encounters a hill, the speed of the car will drop off. Eventually point A will be reached on the Rückschaltung curve, so the car will downshift. If, on the other hand, the car starts to go downhill, point C will be reached and the transmission will shift into a higher gear.
Now consider a vertical movement on this shifting characteristic. This would be a sudden change in the driver’s wish without a change in speed. For example, if the driver wanted to pass a vehicle suddenly, he/she would step on the gas pedal and reach point B. The car would downshift to achieve quick acceleration. On the other hand, if the driver suddenly lets up on the gas pedal, point D will be reached, and the car will up-shift to reduce engine RPMs for less engine braking and more economical operation.
Of course there is such a pair of curves for each gear change. And since the shifting strategy is electronically controlled, these shifting characteristics can be postponed (pushed to the right on the speed axis) for sportier operation or shifted to the left for more economical/environmentally friendly operation.
The Big Picture
As previously mentioned, in a fully mechatronic car the shifting control does not happen in isolation. The electronic control of the transmission needs to be aware of other electronically controlled systems and interact with them. At the same time, the shifting strategy of a manufacturer is unique and is separate from the actual hardware that constitutes the transmission. The strategy for managing this complexity is to adopt a layered approach. This scheme is illustrated in Figure 6.
Figure 6 - Overview of transmission control
Driving Strategy – The most abstract function is the driving strategy. This incorporates the shifting characteristic curve previously discussed. Note that there are 2(n-1) such curves, n being the number of gears, since one cannot up-shift from the highest gear nor downshift from first. Thus the normal five-speed transmission has eight such curves.
These curves can be modified for
Eco-friendly driving – early up-shifting, later downshifting
Sport – the opposite of eco-friendly, late up-shift, early downshift
Normal – between 1 and 2
Loaded shifting – going uphill or carrying a trailer, late up-shifting, late downshifting (to avoid flip-flopping gears)
Unloaded shifting – going downhill, late up-shift, early downshift (to use the engine as a brake)
Special – winter (cold engine, heat catalytic converter)
Note that since these driving strategies are programmed, they are selectable. I.e. the same car can behave differently simply by choosing a different driving strategy from a menu.
So, in general the scheme for selecting the gear can be characterized as shown in Figure 7
Figure 7 - Gear selection strategy
But this simplified scheme is not optimal for all driving situations. For example, if the drive wheels slip, it looks to the sensor as if the car is going faster. You wouldn’t want to up-shift in this case. Also in a curve it may not be wise to up-shift or downshift. Also there is the “fast-off” response, where a driver quickly moves his/her foot off the gas pedal because of a perceived obstacle ahead. A delayed shifting response to such an immediate return of the gas pedal to the null position can be implemented.
Driving strategies vary from manufacturer to manufacturer. Each manufacturer does lots of tests under all kinds of conditions and situations and then incorporates the developed strategies into their shifting control. In fact this can be considered proprietary information, part of what makes a BMW and BMW. The developed software embodies a driving strategy worked out painstakingly and with the outlay of a great deal of money and effort by a manufacturer.
It is worth noting also that Fuzzy Logic control is now being implemented for assessing driving situations and producing responses thereto. It is also worth noting that since driving strategies eventually wind up expresses as software, it is theoretically possible that 1) third-party developers could come up with their own strategies for improving driving strategies and then implement them on various manufacturers’ automobiles and 2) improved driving strategies could be developed which could be up-loaded into a car’s control system after its date of manufacturer. So as an owner, you could be faced with being able to actually improve the performance of your car with as little as a software upgrade. The mechatronic car is becoming more and more like a computer.
Driving Functions – One big purpose of this layer between Driving Strategy and Transmission Control is to separate them, so that a manufacturer’s driving strategy can be implemented on more than one type of car and transmission. For example, a manufacturer could have both an automatic transmission and an automated manual transmission, yet the driving strategy is the same for this manufacturer, regardless of the type of transmission.
At this second level, there are subfunctions that are important. For example, starting, stopping, how to implement shifting, dealing with the slip of couplings, gathering the overall state of the transmission (not low-level stuff but a layer above this), how to deal with the engine torque during shifting, evaluating gas and brake pedal positions, implementing characteristic engine performance. Also at this stage one finds the management of the lower level, a function call set to interface with all the servo-valves and solenoid actuators in the transmission. With the double-clutch transmission, it is at this level that the transfer of torque steady from one output shaft to the other takes place.
Lower Level Functions – This level interfaces with the actual actuators and sensors in the transmission. For example, to transfer into a certain gear, it may mean applying an increasing and controlled pressure to a coupling. This pressure is supplied by a hydraulic servo valve. The pressure would be governed by a current to a valve. So the control loop would appear as shown in Fugure 8.
Figure 8 - Servo valve controlling pressure in an automated transmission
There may be an adaptation to this because of hysteresis in the servo valve, wear in the system, aging, change in operation of the system as fluid warms up, change due to altering fluid properties over time (fluid gets dirty, burnt), etc.