Matters of Perception
by A.J. Kleber
Professor Vivek Goyal’s research areas can sound a bit like magic, when put into plain terms: he’s known for studying how to see (and “take pictures”) around corners, or using “invisible” light. This apparent wizardry, of course, is a matter of science–and mathematics. Goyal’s work, while varied, has the common theme of using physics, algorithms, and related methods to expand perception past what the human sensory suite is biologically capable of. Two new projects, recently funded by the NSF, are no exception.
Improved 3D imaging, for safety you can afford
As driverless taxis hit Boston’s streets for the first time, safety concerns about autonomous vehicles (AVs) have never been more relevant. In order to “see” hazards on the roadway, AVs rely on sophisticated realtime 3D mapping and modeling. These technologies are important to a variety of fields and applications, from augmented reality (AR) systems to surgical medicine. Advanced as current computational imaging methods are, the critical importance of many of these applications make continued improvements imperative.
With the support of a $240K NSF grant, Professor Goyal is taking an already very fast and accurate method, termed frequency-modulated continuous-wave (FMCW) lidar, and deconstructing it down to first principles, to see if he can eliminate its biggest disadvantage: cost.
Underlying FMCW lidar are elegant principles that govern the observation of sine waves, including the Doppler shifting effects that cause the sound of a siren to depend on the direction of travel of a police car. The mathematically simple basis of FMCW lidar depends on having carefully calibrated and highly complex hardware to ameliorate physical effects that the calculation method does not account for. This trade-off of simple math for complex hardware makes for a very expensive system, which is less than ideal for use in consumer products like AVs. To address this, Goyal plans to break the methodology down to first principles, developing new algorithms and other strategies that can provide accurate distance and velocity measurements with less stringent hardware requirements. Preliminary results have been promising.
Beyond night vision goggles & wildlife cams
Long-wave infrared (LWIR) light, which is beyond the visible spectrum, is useful for applications such as search-and-rescue operations, law enforcement, and wildlife cameras used to study the behavior of nocturnal species (or just find out what critters are visiting your back yard at night). Per Professor Goyal, however, there is considerable underutilized potential in this “light we can’t see.” A $600K NSF grant will support his explorations of some of this potential.
Some of the specific technologies Goyal is interested in pursuing using LWIR include 3D imaging and remote sensing to measure air temperature and composition, with applications for navigation, monitoring pollution, and much more. He plans to use both learning and physics-based approaches to formulate and solve inverse problems, working backwards from observed data to determine causal factors, in order to define parameters and methodology for the target applications.
Professor Vivek Goyal joined the ECE faculty at Boston University in 2014. He is a Fellow of the AAAS, IEEE and Optica, the recipient of a 2025 BU College of Engineering Dean’s Catalyst Award, a 2024 John Simon Guggenheim Memorial Foundation Fellowship, several best paper awards, and a 2023 Frontiers of Science Award in Computational Optics, among other accolades. His research interests revolve around computational imaging, information representation and signal processing.