Informative Path Planning for Active Tracking of Agile Targets

This paper proposes a method to generate informative trajectories for a mobile sensor that tracks agile targets. The goal is to generate a sensor trajectory that maximizes the tracking performance, captured by a measure of the covariance matrix of the target state estimate. The considered problem is a combination of estimation and control, and is often referred to as informative path planning (IPP). When using nonlinear sensors, the tracking performance depends on the actual measurements, which are naturally unavailable in the planning stage. The planning problem hence becomes a stochastic optimization problem, where the expected tracking performance is used in the objective function. The main contribution of this work is an approximation of the problem based on deterministic sampling of the predicted target distribution. This is in contrast to prior work, where only the most likely target trajectory is considered. It is shown that the proposed method greatly improves the ability to track agile targets, compared to a baseline approach.