Emerging engine technologies require a tremendous amount of intelligence, that is, embedded hardware and software. Here's a look at what's involved and where the challenges and opportunities lie.
Engineers in Detroit and around the world are hard at work designing hybrid and fuel cell vehicles, both of which will require numerous embedded systems. What are the challenges these systems present for their designers? While some of these systems run on simple microcontrollers and communicate with other systems over a rudimentary network, much more complex microcontrollers with larger memories are being introduced, too. Though many of the control algorithms and the subsystems they control are not new, putting them all together in a single product and ensuring they meet the many stringent requirements of an automobile is a challenge. In this article, we'll discuss the challenges inherent in designing embedded systems for hybrid and fuel cell vehicles.
As their name implies, hybrid propulsion systems combine at least two different means of propulsion. In automotive applications, the typical hybrid system consists of an internal combustion engine (either gasoline or diesel) and one or several battery-powered electric motors. One advantage of such a system is that the internal combustion engine can be designed to always operate at peak efficiency with the electric motors used to provide primary or additional propulsion during high-demand periods.
There are principally two types of hybrid architectures: parallel and series. The distinction is mechanical. In a parallel hybrid system, the internal combustion engine provides mechanical power to the wheels in parallel with the motors providing electrical power. In a series hybrid system, the internal combustion engine must first generate electricity, which is then converted into wheel power through the electric motors.
Within each hybrid architecture, specific designs are numbered in the dozens. For example, a Society of Automotive Engineers (SAE) technical paper on the upcoming GM Parallel Hybrid Truck indicates the designers considered seven different parallel architectures during design.
Figure 1: Parallel hybrid system
Figure 2: Series hybrid system
Figure 1 shows a schematic of one type of parallel hybrid system. Figure 2 shows a schematic of a series hybrid system. Both figures show the electrical system complete with the various embedded control systems.
In Figure 1, which shows essentially the layout of the Honda Civic Hybrid available today, the purpose of the electric motor is to assist the internal combustion engine during acceleration (enabling use of a smaller, more efficient engine), recovering energy during deceleration and braking, and starting the engine.
The Civic Hybrid attains decreased emissions and fuel consumption in three ways. First, energy normally lost during vehicle deceleration and braking is used to recharge the batteries. Second, fuel is saved by turning the internal combustion engine off when the vehicle is idling. The electric motor is large enough to both restart the engine and begin propelling the vehicle simultaneously. Finally, a smaller displacement engine is employed in this car than the standard Civic, since the electric motor is able to assist during high-demand periods (such as rapid acceleration).
The series hybrid system shown in Figure 2 employs a motor at each wheel and is essentially the same concept described by F. H. Moeller. The series design can decrease emissions and fuel consumption in the same ways as the parallel design, but has several additional advantages. First, traction and energy recovery can be switched by wheel, thereby allowing rapid traction control and antilock braking via the motors. This series architecture also provides all wheel drive capabilities without expensive transfer cases and differential gearing. A second advantage of this architecture is its natural expandability to any number of wheels, which is particularly important in military applications. Finally, the decoupling of the engine from the wheels may be more efficient if the engine/generator is used to supply external power (no residual loads from the transmission).
As you can see from both figures, numerous embedded controllers are required to make a hybrid vehicle work properly. These controllers are all interconnected, usually via a high-speed controller area network (CAN) bus (see Pfeiffer and Murphy in this issue). The firmware in these controllers must control their part of the system and coordinate with the other controllers in real time. This requires extensive design and development of new control algorithms and software, even for mature control systems such as engines, transmissions, and antilock braking systems.
Figure 3: Fuel cell powered vehicle
Figure 4: Fuel cell stack and controls
Fuel cell vehicles are electric vehicles powered by batteries and a fuel cell. The fuel cell converts hydrogen and oxygen to electricity, with water and heat as the only by-products. The basic fuel cell configuration used in vehicles is shown in Figure 3. This configuration is similar to the series hybrid architecture, except that the internal combustion engine is replaced by the fuel cell stack.
Although fuel cell technology has been around for a long time, it's new in the uncontrolled environment of automobiles. Significant durability, reliability, and cost hurdles must be overcome before mass-market fuel cell vehicles can be realized.
The fuel cell, shown in Figure 4, is itself a complex system. It requires precision control of the temperature and humidity of the fuel (hydrogen) and air entering the cell, along with the temperature of the cell itself. For example, the temperature of the air entering the fuel cell must be within 2° C of the temperature of the membranes in the stack, and the relative humidity must be in the range of 70 to 90%. If these requirements aren't met, damage may occur to the stack. The science of fuel cells is best left to the chemists and scientists. These individuals, however, are the source of many requirements that must be met by the embedded systems that control fuel cell stacks.
Without the intelligence provided by embedded systems, hybrid and fuel cell vehicles won't be possible. Many of the components of these embedded systems exist today in isolation. Internal combustion engines (both gas and diesel) and vehicle transmissions are all now controlled by embedded systems. Electric motor controls and battery-charge management are also well established in other industries. And fuel cells have been used in space exploration for decades. Combining these various embedded systems into a single vehicle is key to the success of hybrid and fuel cell vehicles.
Many assumptions made in the design of existing electronic vehicle systems don't apply to the hybrid/fuel cell architecture. For example, many vehicle systems today assume that the internal combustion engine is running before anything occurs. If the engine isn't running, you don't have power, right? Wrong. Series hybrid systems can run just fine with the engine shut offfor as long as the battery has sufficient charge. Electronics (both hardware and software) may account for less than 25% of the cost of the final propulsion system, but the amount of engineering effort required to develop the algorithms and software for these components may well exceed 75% of all development costs.
Whenever a new electronic component is added to an automobile or other moving vehicle, many challenges must be overcome. The primary challenges are in the areas of cost, power consumption, operating environment, reliability and durability, safety and security, and regulation. These well-known constraints exist in any embedded systems project, but they take on a different aspect in hybrid and fuel cell vehicles. I discuss these challenges in the following sections and offer some suggestions for meeting them.
Automobiles are cost-sensitive products, and this price pressure is felt right down to the design of the individual components. Since hybrid vehicles include a traditional powertrain plus an electric vehicle system (motor[s] and batteries), the buyer of the vehicle is essentially paying for two powertrains. Although the cost savings in reduced fuel consumption will somewhat offset this greater sticker price for the consumer, the market will still demand the initial cost be kept down.
Power management has two important challenges. The traditional automotive challenge remains unchanged: reduce power dissipation at each controller and get excess heat out of the box. Driver circuits are notoriously high power because they drive external actuators, and this power must be dissipated to avoid damage to the electronics. Furthermore, adding electric motors to the vehicle means even more power must be dissipated.
The new challenge stems from the vehicle's purposefuel efficiency. The embedded systems in the vehicle must be very efficient and require minimum power to operate. It's useless to save energy by shutting off the engine if the energy saved is then wasted in keeping controllers or actuators running (especially if actuators are not needed).
Many battery-powered embedded devices use algorithms to increase battery lifetime by going into a low-power mode when not actively performing functions. It's imperative that even this small amount of power savings be applied in hybrid and fuel cell vehicles, especially given the total number of processors. By going to a low-power state when the software is waiting to take action, the embedded control system contributes to overall improved energy efficiency.
The electronics that control automobiles must operate in a harsh environment, including both temperature and vibration extremes. Suppliers of traditional automobile electronics are familiar with this environment. For example, the temperature of underhood electronics (including microcontroller-based systems) is specified as -40 to 125° C. This is many times mistaken as the ambient temperature underhood, but is actually the junction temperature at which the electronic components are rated. Tests performed under very harsh conditions have shown that the actual maximum underhood temperature is only 110° C, which is still hot enough to boil water.
Vibration is also a key concern. (Just think how you feel when you hit a pothole even though you're sitting on a cushioned seat.) Hybrid and fuel cell powered vehicles will bring suppliers of electric motors, batteries, and fuel cells into the automotive industry for the first time. These new suppliers will have to design to vibration specifications that in many cases even rival requirements in the aerospace industry.
Reliability and durability
Automotive electronics are required to reliably operate under all conditions for 10 years/100,000 miles for automobiles and 15 years/150,000 miles for trucks (these numbers vary slightly between OEMs, but are typical). This means that the electronics must do the same thing every time without breaking or being reset.
Some of these controllers actually never shut down; they continue to operate at a reduced-power level, even when the vehicle is turned off. Imagine turning on your computer in 1993 or your workstation in 1988, and still using it today without ever restarting it!
Safety and security
Many plaintiffs and lawyershoping to sue the engineer's employer for damagesare looking over engineers' shoulders to see if a poor design decision has caused an accident. This scrutiny and the risk of legal action have produced many sleepless nights for automotive engineers. Tens of thousands of engineering hours must thus be spent in Failure Mode and Effects Analysis (FMEA) and Fault Tree Analysis (FTA) to ensure that designs are safe under all conditions, even if a component fails.
Many OEMs require that for any single failure (one component failing completely) the vehicle must be able to continue to a service location safely; fortunately, a reduced capability such as lower power may be acceptable. Included in this safety loop is the firmware, which in many cases will have to take the safest action if it detects a failure.
Security is defined as the vehicle not executing any action the driver does not intend to make. For example, unintended acceleration became a major case a few years back, where people were driving their vehicles through the back of their garages without knowing why. The actual cause was never determined, but the potential for this to occur has produced many more sleepless nights. Security has become a greater concern with the advent of “by-wire” systems, where the operator of the vehicle is detached from the device. Hybrid and fuel cell systems are inherently drive-by-wire, brake-by-wire, and steer-by-wire systems. In many cases, this will require the design of fault-tolerant systems and redundancy.
The automotive industry is one of the most highly regulated industries in the world. Key regulatory bodies in the U.S. include the California Air Resources Board, the National Highway and Traffic Safety Administration, the Environmental Protection Agency, and the Department of Transportation. In particular, the California Air Resources Board has a set of requirements called OBD-II (On-Board Diagnostics v2), which are complex and require extensive study to understand. In a nutshell, OBD-II says that if a component can fail and that failure may negatively affect either the vehicle's emissions or fuel consumption, that component's failure must be detected and the emissions light turned on.
By using modeling tools such as UML, Matlab/Simulink, and AscetSD, you can develop a much more bulletproof design. In particular, using products that enable the system to be simulated offline on a workstation will be a major benefit. With modeling, we'll be able to analyze reliability, safety, and security early in the program to ensure that a design works under all conditions, not waiting until the final product is tested to identify a concern.
Many of the early conceptual vehicle designs that were made public have used Matlab/Simulink and automatic code generation on a rapid prototyping system to test in an actual vehicle. The major hurdle will be to move all of this intellectual property into production-ready code that meets the challenges previously discussed.
Software-in-the-loop and hardware-in-the-loop simulation provides the greatest opportunity to meet the design goals of hybrid and fuel cell systems. Such simulation can assist in rapidly performing many tests on multiple designs prior to actual hardware availability. It's also key in the development of safe, secure, reliable, and durable systems.
By simulating the system first at the design level and then as individual components become available, we can test the function of the system under multiple environmental and failure scenarios. By the time the software reaches the end vehicle, we'll have performed extensive testing and development, and can then use the vehicle to validate the model, not to develop the control system. This doesn't take away the need for vehicle-level validation, but will reduce the risk and improve the quality of the end system.
Simulation and modeling work well together, since the simulation will be run directly in the modeling environment many times. The biggest challenge here is to develop the simulation models so that they truly represent the simulated device and are efficient enough that the simulations can complete in a reasonable time.
It remains to be seen, of course, if all of the challenges in hybrid or fuel cell vehicle design can be met in products that consumers will want to own and drive. But if these architectures do succeed, it's certain to be a big success for the embedded systems designers involved.
Joe Lemieux is the chief technologist for the Engineering Solutions group of EDS Engineering & Manufacturing Services. He is the author of the book Programming in the OSEK/VDX Environment. Joe holds a BSEE from General Motors Institute (now Kettering University) and an MSEE and MBA from the University of Michigan. His e-mail address is .
- Evans, David G. et. al., “Powertrain Architecture and Controls Integration for GM's Hybrid Full-Size Pickup Truck,” SAE World Congress, 2003.
- Ogawa, Hiroshi, et. al., “Development of a Power Train for the Hybrid Automobilethe Honda Civic,” SAE World Congress, 2003.
- Moeller, F. H., “Prime Movers for Series Hybrid Vehicles,” SAE World Congress, 1997.
- Gurski, Stephen, et. al., “Design of a Zero Emission Sport Utility Vehicle for Future Truck 2002,” SAE World Congress, 2003.