2017 Fellow

Dhanvin Mehta's research goal was to bring robots out of confined and controlled environments and into the real world by developing planning algorithms for autonomous navigation among pedestrians. His research group has developed Multi-Policy Decision Making (MPDM), a framework for navigation in dynamic and uncertain environments where higher-level behavioral decision-making is decoupled from low-level control. The behaviors or policies of dynamic agents (cars and pedestrians) in the environment are estimated from observed motion. Based on these estimates, MPDM dynamically switches from amongst a set of complementary behaviors that allow the robot to adapt to different situations in a risk-aware fashion [1,2]. The team is currently expanding the MPDM framework to incorporate more expressive policies, as well as finding other application domains.

Dhanvin Mehta is a Ph.D. graduate from the Department of Computer Science and Engineering. He worked with Professor Edwin Olson in the April Lab.