Qiangfei Xia, professor of electrical and computer engineering, was recently awarded a grant of approximately $1.1 million grant from the Defense Advanced Research Projects Agency (DARPA) under a program to develop a camera that can track fast events more quickly and efficiently by only transmitting data from pixels that have detected changes.
Xia’s award is part of the multimillion dollar effort led by BAE Systems under the Fast Event-based Neuromorphic Camera and Electronics (FENCE) program launched by DARPA. Neuromorphic cameras are an emerging class of sensors that only measure the motion of what they are recording, such as changes in brightness. Each pixel operates independently, producing significantly less data yet operating with much lower latency and at lower power.
The objective of FENCE, according to DARPA, is to develop and demonstrate the camera and learning algorithms that use combined spatial and temporal (location in space and time) information to enable intelligent sensors for tactical applications ranging from autonomous vehicles and robotics to IR search and tracking.
“The goal is to develop a ‘smart’ sensor that can intelligently reduce the amount of information that is transmitted from the camera, narrowing down the data for consideration to only the most relevant pixels,” says Whitney Mason, FENCE program manager.
According to Xia, the UMass team will provide design, development, and fabrication expertise in spatio-temporal algorithms, integrated circuit design, and neural processing. Xia will share the expertise from his Nanodevices and Integrated Systems Lab with BAE for the program.
The BAE team will develop an asynchronous read-out integrated circuit with low latency. The read-out and processing layers will enable an integrated FENCE sensor to operate on less than 1.5 watts of power.
This article was originally published by the UMass Amherst Office of News and Media Relations.