Jagannathan Sarangapani, professor of electrical and computer engineering, and Jim Drallmeier, professor of mechanical and aerospace engineering, believe the sophisticated controller shows the most promise with exhaust-gas recirculation (EGR), a technique used to reduce nitrogen oxide emissions.
"A spark-ignition engine needs both fuel and air to operate," Drallmeier says. "If, however, I can give it less fuel for the same amount of air or dilute the mixture with inert gases, as from EGR, the engine will behave differently," he explains. "And that's what we're doing here."
The two researchers and their students created a neural-network controller that is implemented in software. Artificial neural networks are adaptive systems that learn by keeping successful connections between neurons or nodes.
"The neural network observer-part of the controller will assess the total air and fuel in a given cylinder in a given time," Sarangapani says. "It will send that estimate to another neural network, which generates fuel commands for the engine in each cycle." Speed is a critical factor. "This controller observes what an engine cycle is doing, makes measurements in that period of time, reduces that data, and decides how you need to push the engine in the next cycle," explains Drallmeier. "It does all that before the next cycle starts. We're talking about a matter of milliseconds."
Significant theoretical challenges encountered during design must be overcome before the controller can be implemented on the hardware. The primary problem is too little information coming from the engine, says Sarangapani.
Although more EGR can reduce nitrogen-oxide emissions, it can cause a "cycle-to-cycle variability in engine output," Drallmeier says. "A good example is when you're sitting in your car at a stop light and feel the car shaking. The more EGR you add, the lower your nitrogen-oxide emissions. The question is how far you push it and stay in a reasonable range."
The National Science Foundation and the Environmental Protection Agency are jointly funding the three-year, $515,000 project.