Light as data and power source

Self-driving cars are equipped with sophisticated cameras and optical sensors that use powerful computer processors to analyze their environment. What if the sensors and cameras could be powered by taking advantage of the properties of light?

Mark Lawrence, assistant professor of electrical & systems engineering in the Preston M. Green Department of Electrical & Systems Engineering, plans to develop such sensors and cameras with a five-year, nearly $560,000 CAREER Award from the National Science Foundation. CAREER awards support junior faculty who model the role of teacher-scholar through outstanding research, excellence in education and the integration of education and research within the context of the mission of their organization. At least one-third of current McKelvey Engineering faculty have received the award.

Under Lawrence’s plan, these sensors could be used to perform medical diagnostics, surveillance, or safety and security monitoring without needing a camera, computer chip or battery.

“Our idea is to design and build physical systems that analyze a scene by coaxing light waves scattered from objects to flow entirely passively through complicated optical networks, performing important machine vision or smart imaging tasks, like object recognition and segmentation, without needing a computer processor. Light is both the data and the power source,” Lawrence said.

The ambitious concept is based on bringing nonlinearity into optical image processing systems, which is notoriously challenging.

“Although most discussions on the role of hardware in AI revolve around how quickly and efficiently large matrices can be multiplied together, neither the human brain nor digital neural network algorithms display intelligence without thresholding filters,” Lawrence said. “To perform universal computation, at intermediate steps it is necessary to ignore signals below a certain value, while transmitting or amplifying signals above that value. That's something we can't do with light very easily, but we're trying to with our new devices we're developing through this work. We’re going to bring that crucial gradient of nonlinearity into all optical image processing systems.”

Lawrence said he’s not trying to replace GPUs and CPUs with optical accelerators but instead hopes to introduce advanced computing capabilities where light exists, but computation doesn’t, using established physics to access nonlinearity at unprecedentedly low levels of light.

“I want to take cheap optical sensors and make them powerful, as opposed to replacing existing powerful computational tools,” he said.

In addition to the research, Lawrence is planning to establish an optical devices apprenticeship for individuals with limited exposure to, or education in, photonics and nanoscience. The bootcamp would prepare them to work as an intern or entry-level employee in an integrated photonics company.

“There is a blossoming industry in integrated photonics,” he said. “People are using integrated photonic circuits and nanoscale photonic structures to do all sorts of things, from quantum computers and fiber optic interconnects to virtual reality goggles, so we hope to expand the workforce by not requiring people to enter those fields with a PhD.”

In addition, he plans to offer outreach modules in optical nonlinearity for the public in collaboration with the St. Louis Science Center.

Click on the topics below for more stories in those areas

• Research
• Electrical & Systems Engineering

Back to News