An Optimal Control Approach for the Data Harvesting Problem

TL;DR

This paper introduces a trajectory optimization method for data harvesting by mobile agents, using parameterized paths and real-time optimization techniques to handle stochastic data sources efficiently.

Contribution

It presents a novel trajectory parameterization approach combined with Infinitesimal Perturbation Analysis for real-time optimization in data harvesting tasks.

Findings

Trajectory parameterization with elliptical and Fourier series is effective.

The method is robust to stochastic data generation.

Scalability is demonstrated with larger event sets.

Abstract

We propose a new method for trajectory planning to solve the data harvesting problem. In a two-dimensional mission space, $N$ mobile agents are tasked with the collection of data generated at $M$ stationary sources and delivery to a base aiming at minimizing expected delays. An optimal control formulation of this problem provides some initial insights regarding its solution, but it is computationally intractable, especially in the case where the data generating processes are stochastic. We propose an agent trajectory parameterization in terms of general function families which can be subsequently optimized on line through the use of Infinitesimal Perturbation Analysis (IPA). Explicit results are provided for the case of elliptical and Fourier series trajectories and some properties of the solution are identified, including robustness with respect to the data generation processes and scalability in the size of an event set characterizing the underlying hybrid dynamic system.