September 27, 2019
Takashi Tanaka, an assistant professor in the Department of Aerospace Engineering and Engineering Mechanics at The University of Texas at Austin, is a recipient of the DARPA (Defense Advanced Research Projects Agency) Young Faculty Award. The award program’s objective is to identify young faculty early in their careers by supporting them with funding, mentoring, and industry DoD contacts, with the “long-term goal of developing the next generation of academic scientists, engineers, and mathematicians who will focus a portion of their career on DoD and National Security issues,” according to the DARPA website.
The award will support Tanaka’s work on developing a new theoretic framework for predictive vision modeling, which could revolutionize vision-based control planning and technologies such as visual navigation of ground robots. Tanaka and his team will develop a new top-down, (task-dependent) mathematical model of visual attention and demonstrate its similarities to a human’s visual cortex functions, which are known to be task-dependent, allowing for more efficient decision-making in dynamic environments.
Although technology has been developed to incorporate artificial neural networks that resemble the human visual cortex, the current technology is largely restricted because these systems rely on feature extraction and classification of static images. However, daily tasks where this new vision-based technology might be applied such as driving a car, flying an aircraft even or navigating a robot, require a more dynamic, decision-based vision-based system. According to Tanaka’s research proposal, there is currently no compelling theoretical framework in place for this type of model.
To meet the challenge of reproducing a human-like visual attention mechanism, Tanaka’s research proposes to introduce a top-down visual attention model based on the information-theoretic concept known as directed information. Once the framework is complete, Tanaka’s team will demonstrate the new system by developing autonomous ground robots and equipping them with an on-board GPU-based artificial visual cortex.
The work will be conducted over a two-year period, with an optional additional 12-month period, and includes a postdoctoral researcher, graduate students, and undergraduate students, who will participate through the Oden Institute of Computational Engineering and Sciences.
Tanaka was also selected to receive an Air Force Office of Scientific Research Young Investigator Award for his work on path planning.
Tanaka joined the department as an assistant professor in 2017. His research interests include systems and control theory, networked and multi-agent systems and information theory control. He specializes in stochastic and non-stochastic optimal control, robust control, distributed and networked control, optimization, and game theory. He is an affiliated member of the Oden Institute. Learn more about Tanaka’s work on his research website.
