PhD Seminar Notice: "Informative Sampling & Path Planning in Random Fields" by Shamak Dutta

Monday, September 12, 2022 10:00 am - 10:00 am EDT (GMT -04:00)

Candidate: Shamak Dutta
Title: Informative Sampling & Path Planning in Random Fields
Date: September 12, 2022
Time: 10:00 AM
Place: REMOTE PARTICIPATION
Supervisor: Smith, Stephen L.

Abstract:
In this talk, I will discuss some recent results in sampling and path planning for robots in random fields. First, I will discuss the subset selection problem in a spatial field where we seek to find a set of 'k' locations whose observations provide the best estimate of the field value over a finite set of prediction locations. We propose a greedy algorithm that searches only over the prediction locations and the centroids of cliques formed by the prediction locations. We demonstrate the effectiveness of this approach, in terms of solution quality and runtime, over a greedy algorithm that searches over a grid discretization, which is a popular approach to solve the problem.

Next, we present a new mixed integer formulation for the discrete informative path planning problem in random fields. The objective is to compute a budget-constrained path while collecting measurements whose linear estimate results in minimum error over a finite set of prediction locations. Our approach expands the search space so that we optimize not only over the measurement subset but also over the class of all linear estimators. This allows us to formulate the problem as a mixed-integer quadratic program whose objective is convex in the continuous variables. In simulations, we demonstrate the benefit of our approach over previous branch and bound algorithms.

Finally, I will conclude with the work I am currently focusing on relating to adaptive sampling in random fields, where the covariance structure is unknown.