Home research People General Info Seminars Resources Intranet
Time - Wednesday, December 14, 2016, 4:00 pm
Location - 302 HRBB

Motion Planning for Active Data Association and Localization in Non-Gaussian Belief Spaces

Saurav Agarwal
Department of Aerospace Engineering
Texas A&M University


We have developed a method for motion planning under uncertainty to resolve situations where ambiguous data associations result in a multimodal hypothesis on the robot state. Simultaneous localization and planning for a lost (or kidnapped) robot requires that given little to no a priori pose information, a planner should generate actions such that future observations allow the localization algorithm to recover the correct pose of a mobile robot with respect to a global reference frame. We present a Receding Horizon approach, to plan actions that sequentially disambiguate a multimodal belief to achieve tight localization on the correct pose in finite time. Experimental results are presented for a kidnapped physical ground robot operating in an artificial maze-like environment.


Saurav Agarwal is currently a Ph.D candidate in the Dept. of Aerospace Engineering. His primary research is focused on motion planning under uncertainty for both known and unknown environments and he is also developing methods for accurate long-term SLAM. He recently concluded a research internship at Qualcomm Research, where he applied motion planning and computer vision techniques for autonomous aerial robotics.