Cyber-Physical-Human Systems: Making the whole greater than the sum of Its parts
“I’m not the only one who’s gotten off an interstate and had the car shoot up the ramp when they thought they had disengaged the cruise control twenty miles earlier by a tap of the break—when it was actually dormant,” said CS Professor Alex Kirlik.
Automation has become an integral part of people’s lives, and it is often taken for granted. However, there are instances when automation needs to have a human step in, such as a car speeding up an exit ramp.
Led by CS Professor Alex Kirlik, Illinois researchers have received a $950,000 National Science Foundation (NSF) grant to study cyber-physical-human systems to further investigate how humans need to be added to the equation of automation for many aspects of our lives. In addition to Kirlik, researchers on this project will include CSL professors Lui Sha (Computer Science), Carolyn Beck (Industrial and Enterprise Systems Engineering, who will lead the anesthesiology portion of the study) and Naira Hovakimyan (Mechanical Science and Engineering), along with University of South Carolina Assistant Professor Xiaofeng Wang.
This project brings together researchers from computing and engineering disciplines to study ways to create, validate, and demonstrate the value of new techniques for ensuring that systems composed of combinations of hardware, software, and humans can operate in a synergistic and safe fashion.
The researchers will look at ways to improve the interactions of the cyber and physical systems. Cyber-physical systems are those in which there is hardware with embedded software for sensing, monitoring and systems control. An example would the inclusion of sensors and computing systems into the structure of bridges to provide a continuing monitoring of the bridges strength and safety. Increasingly, these types of systems are used within areas where humans find themselves interacting.
“Our goal is to design systems that outperform what may be possible by either automation or by unaided humans alone,” Kirlik said. “You make automation and humans team players that can then perform better than either of them operating individually.”
For this study, the researchers are going to be looking at aviation and anesthesiology as areas where today the cyber-physical systems also interact closely with humans. “We will develop an integrated framework for the design and evaluation of what we call cyber-physical-human systems, or CPH systems,” Kirlik said.
The team will focus on the areas of flight simulation and anesthesiology.
In aviation, automation of flight control has been increasing as airplanes and flight technology have become increasingly complex. The calculations and adjustments made by these machines are at the level that could not be made by humans alone.
Nevertheless, said Kirlik, “No engineer could foresee a double bird strike after taking off from LaGuardia Airport that would require landing the airplane in the Hudson River. We don’t have—not only do we not have the adaptive control automation, but the artificial intelligence yet to be able to make the types of real-time seat of the pants judgment that that pilot was able to make.”
In the example Kirlik gave of Captain Sullenberger who landed the US Airways plane on the Hudson River, it was essential that a human step in. It is these instances that Kirlik and his team will investigate. “We are interested in this class of systems where there’s hardware, software, yet still a human in the loop,” he said.
Kirlik and his team want to investigate the cyber-physical-human aspects of both flight flight and anesthesiology.
Like flight, anesthesiology is a highly automated field, but there are times when it is essential that the human step in to maintain patient safety. “In anesthesiology, we will have a similar need to design interfaces to these drug delivery systems that give the anesthesiologist a very good picture as to what these automated systems are doing,” Kirlik said.
The ultimate goal of the entire project is to develop CPH systems where safety and performance exceed what could be achieved with either skilled humans or with automation alone.